Show all abstracts
View Session
- Front Matter: Volume 7513
- 3D Shape Measurement and Profilometry
- Optical Methods for Industrial Applications
- Remote Sensing, Lidar, and Environmental Applications
- Imaging, Processing Algorithm, and Analysis
- Poster Session
Front Matter: Volume 7513
Front Matter: Volume 7513
Show abstract
This PDF file contains the front matter associated with SPIE Proceedings Volume 7513, including the Title Page, Copyright information, Table of Contents, Introduction, Symposium Committee, and Conference Committee listing.
3D Shape Measurement and Profilometry
All-fiber optical dynamic Fourier transform profilometry
Show abstract
The Young's dual-pinhole interference fringe generated by the structure of Mach-Zehnder interferometer was
projected to the object. High-speed signal acquisition card based on FPGA and multiple DSP is developed for dynamic
use. High-precision and broadband LD controller is also developed. CMOS camera is used in the image acquisition.
And the whole optic path of the system uses fiber optic, which helps to realize the fiber optical integration and
miniaturization of the system.
Stripes extraction technique of projection pattern in frequency domain for 3D shape measurement based pattern projection technique
Ke Sun,
Cunwei Lu
Show abstract
Because the 3-D shape measurement technique using optimal intensity-modulation pattern projection (OIMP)
method can detect a lot of stripes numbers by one projection, the application is expected by the spread of the digital
camera and the pattern projector. However, there is a problem that the accuracy of stripes recognition decreases in the
case of the intensity range of the observation image of the measurement object is insufficient. Or since two images are
needed to convert the observation image by the image correction method, saying, OIMP method, the measurement time
is long. In order to solve above problem, we propose an image analysis method based on Fourier transform technique to
correct the intensity of the observation image. By this method, the stripes can be extracted only using a single
observation image, and then, a fast 3-D shape measurement is realizable.
Compact camera system for 3D profile measurement
Show abstract
Pattern projection method based on phase-shifting technique in fringe analysis has been one of popular principles for
three-dimensional shape measurement and profilometry. However, a few drawbacks are pointed out such as "large in
size" and "heavy in weight" for practical usage in industrial applications. To improve these difficulties inherent to the
conventional three-dimensional system based on pattern projection or structure light method, we try to incorporate recent
digital devices such as a MEMS scanner. Due to this revision in projection optics, first of all, a compact measurement
system is easily attainable, and, when we adjust the scanner to produce the reference pattern with non-equal periodical
structure, the projected pattern is so formed as to be equal in period on the reference plane. In addition, the pattern
becomes sharp over the whole field of measurement when the Scheimpflug condition is satisfied in optical arrangement.
This brings easier analysis of the captured pattern and attains the three-dimensional profilometry system with deeper
range of focus, wider field of measurement and higher result in accuracy of measurement.
Spatial phase-shifting moiré deflectometry
Show abstract
A novel spatial phase-shifting method is presented from the moire phenomena. With the analysis to moire patterns from
the scalar diffraction theory, there is stable phase shift between different order diffraction interferograms of moire
deflectometry. This phase shift just depends on the distance between two gratings and the grating period. In this spatial
phase-shifting method, no beam splitter, polarizer or wave plate is used and the disadvantage of uncertain phase shifts
due to the methods using polarizer or wave plate is removed.
Measurement of inner and/or outer profiles of pipes using ring beam devices
Show abstract
Inner profile measurement is an important matter in such fields as medicine, dentistry and anthropology as well as
mechanical engineering and industry. Here we propose a measurement method for inner diameter of pipes and/or
holes. The key device in this technique is a ring beam device which consists of a conical mirror and a laser diode.
And the fundamental principle is based on optical sectioning without any contact probe. The optically sectioned
profile of an inner wall of a pipe-like object is analyzed to give the inner profile in addition to the inner diameter.
This optical instrument with a simple and small configuration is now under development for practical uses. In the
hitherto-tried experimental works, the availability of this instrument has been highly evaluated and usability for
practical applications is expected, especially, for measurement and inspection of mechanical components and
elements besides pipes. This ring beam device consisting of a conical mirror and a LD is assembled to form a disklike
light beam sheet. We show measurement result of pipes and holes, and, at the same time, report a compact inner
profile measuring instrument. Both the ring beam device and a miniaturized CCD camera are fabricated in a glass
tube. Availability of this instrument is shown by measuring the inner profiles of various pipes. In response to this
trial, there appeared a strong request that not only the internal but external profiles should be measured
simultaneously. Therefore we propose an improved method for measuring the external profile in addition to the
internal profile. In our arrangement, one pair of concaved conical mirrors is used for the external profile
measurement. In combination with the inner profile measurement technique, simultaneous measurement of the inner
and outer profiles becomes attainable. A measurement result on a bevel gear shows availability of newly proposed
principle. Now we are aiming to realize simultaneous measurement of the internal and external profiles of various
types of pipes and similar objects.
Study on measurement accuracy of stereo visual sensor based on experimental design
Show abstract
The impact of stereo visual sensor parameters on the measurement accuracy of the sensor is studied based on
experimental design. Using a mathematical model of stereo vision, errors caused by stereo visual sensor parameters are
analyzed. Based on this, main parameters are identified and used in experimental design. In order to find the optimized
combination of the parameters, optimum regression design method is used. Eleven primary experiments with different
combinations of parameters are designed and conducted. According to the experiment results, a regressive equation is
established. By resolving it, the optimized combination of the parameters is got.
Study on removal of phase lines in welding pool surface shape sensing
Show abstract
In recent years, arc welding pool surface shape sensing becomes a hot spot in the field of welding automation. In order to
restore the pool surface shape, we first need to photograph the pool surface, and then extract useful information from the
acquired images. In arc welding surface shape sensing system based on structured light projection, the raster images
obtained by charge-coupled device (CCD) are seriously affected by strong arc and spatter, etc. resulting in errors of
phase unwrapping, and thus seriously affecting the surface shape recovery. To address phase lines of unwrapping errors,
this paper presents a two-neighborhood method. First we analyzed the characteristics of phase lines in the phase diagram,
then by comparison of phase diagrams or phase difference diagrams processed before and after, the effectiveness of
two-neighborhood method was confirmed, finally this method was applied to the actual pool phase diagram processing,
experimental results also confirmed this two-neighborhood method is feasible in removal of phase lines.
Exposing digital image forgeries by 3D reconstruction technology
Yongqiang Wang,
Xiaojing Xu,
Zhihui Li,
et al.
Show abstract
Digital images are easy to tamper and edit due to availability of powerful image processing and editing software.
Especially, forged images by taking from a picture of scene, because of no manipulation was made after taking, usual
methods, such as digital watermarks, statistical correlation technology, can hardly detect the traces of image tampering.
According to image forgery characteristics, a method, based on 3D reconstruction technology, which detect the forgeries
by discriminating the dimensional relationship of each object appeared on image, is presented in this paper. This
detection method includes three steps. In the first step, all the parameters of images were calibrated and each crucial
object on image was chosen and matched. In the second step, the 3D coordinates of each object were calculated by
bundle adjustment. In final step, the dimensional relationship of each object was analyzed. Experiments were designed to
test this detection method; the 3D reconstruction and the forged image 3D reconstruction were computed independently.
Test results show that the fabricating character in digital forgeries can be identified intuitively by this method.
Face detection and face authentication using 3D shape measurement system
Hiroya Kamitomo,
Cunwei Lu
Show abstract
In the security field, a technique which can recognize human faces under various transitions of the scene situation is
requested. The recognition technique that uses 2-D image of the face might fail under various transitions of the scene
situation. We propose to use 3-D image measurement based on pattern projection methods. In this paper, first of all, the
principle and the achievement method of the proposal technique are explained. are explained. Next, the application to the
direction detection of 3-D faces is described. Finally, the 3-D face authentication for security or an access control is
described.
Extraction of tire size code using local phase
Show abstract
Tire size code (TSC) specifies the dimensions of the tire, and some of its key limitations, such as load-bearing ability and
maximum speed. The key challenge of vision based TSC identification is the low contrast of TSC to tire background. In
order to extract TSC from low-contrast tire images, this paper proposed to consider the contours of TSC as the symmetry
axes of the neighbor pixels along the contours. The symmetry axes were detected using local phase information which is
robust to image brightness and contrast. Firstly, the ring shaped tire region was segmented and then expanded to a
rectangular region, where all characters were upright. Due to the intensity values of the neighbor pixels along the
contours of TSC were symmetric, the contours in low-contrast tire image were enhanced using local phase information,
which is computed by Log-Gabor wavelets. Finally, the thin contours were extracted by non-maxima suppression and
thresholding with hysteresis. Experiments on real tire images showed that, in comparison with intensity based
algorithms, the proposed algorithm were robust to uneven illumination and low-contrast.
Optical Methods for Industrial Applications
Motion blurred image restoration algorithm for IC wire bonder
Show abstract
Chip packaging is an important process during IC manufacturing. Wire bonders are key equipments of chip packaging.
The manually type of wire bonders can be updated to automatic with a pattern recognition system (PRS). However, the
resolution of PRS is quite sensitive to the mechanism's trembling. Additional waiting time has to be taken before the
image sensor can take a shoot, which limits the wire bonder's productivity. In this paper, method for motion blurred
image restoration is presented. The method uses point spread function (PSF) learning to estimate the degradation
function and power spectrum analyze to estimate the image's signal to noise rate (SNR). Experimental results are given
to illustrate the proposed image restoration algorithm.
Study on visual image information detection of external angle weld based on arc welding robot
Xiaorui Liu,
Nansheng Liu,
Wei Sheng,
et al.
Show abstract
Nowadays, the chief development trend in modern welding technology is welding automation and welding intelligence. External angle weld has a certain proportion in mechanical manufacture industries. In the real-time welding process, due to hot deformation and the fixture of workpieces used frequently, torch will detach welding orbit causes deviation, which will affect welding quality. Therefore, elimination weld deviation is the key to the weld automatic tracking system. In this paper, the authors use the self-developed structured light vision sensor system which has significant advantage compared with arc sensors to capture real-time weld images. In the project of VC++6.0 real-time weld image processing, after binaryzation with threshold value seventy, 3*1 median filter, thinning, obtain weld main stripe. Then, using the
extraction algorithm this paper proposed to obtain weld feature points, and compute position of weld. Experiment result verified that the extraction algorithm can locate feature points rapidly and compute the weld deviation accurately.
Fault diagnosis system for the inspection robot in power transmission lines maintenance
Li Zheng,
Ruan Yi
Show abstract
Presently electricity utilities make use of a number of inspection tools to survey their transmission lines and electrical
distribution equipment. Mobile robot capable of crossing obstacles is developed on overhead ground wires. A
teleoperated robot with cameras realizes inspection and maintenance tasks on power transmission line. This paper
examines the possibility of remote detection of defective devices on power transmission lines. Infrared, visible and
ultraviolet cameras are the latest visual diagnostic tools available to utilities to identify potential failures on electrical
equipment. Infrared cameras are used to collect images related to temperature differences that arise on components such
as compression splices and insulators. The clever analysis of thermograms will then reveal possible defects. Corona
effects and arcing are symptomatic of certain types of defects on insulators, line spacers, and conductors, etc. These
possible defects can be detected by the ultraviolet. As new imaging technologies were developed, line inspectors began
employing infrared and ultraviolet imaging to locate hot spots and electrical discharges. This paper highlights the design
and development a inspection robot with cameras which combine ultraviolet, infrared and visual imaging techniques for
advanced diagnostic inspections. A novel inspection robot was designed to inspect the 500KV power transmission lines.
Results of experiments demonstrate that the robot can be applied to execute the navigation and inspection tasks.
Study on fruit quality measurement and evaluation based on color identification
Show abstract
A non-destructive measuring and evaluating method for fruits is proposed based on color identification. The color
images of fruits are taken firstly. Then, images' RGB histograms are calculated and used as quality parameters for fruits.
A BP neural network with three layers is established. Its input and output are the RGB histograms and evaluating results,
respectively. After training, the qualities of fruits are identified by the BP network according to the histogram of RGB of
fruits' images. For verifying the proposed method, the qualities of bananas are measured and evaluated. Experiment
results show the reliability and feasibility of proposed method.
Method of detecting underwater bubbles film based on distilling the characteristics of laser signal
Show abstract
Based on the characteristic of the scattering light by the underwater bubble film that the declining edge has a obvious hypo-peak region, the arithmetic of distilling the characteristic
parameters of the region is designed and applied to the experimental signals. These parameters serve as the input of the perceptron neural network and the network's output judges whether the
bubble film exists or not. The result indicates that this method has effective detecting capability against the different underwater bubble films.
The election-pass filtering and image fusion algorithm based on the wavelet transform for CW THz image processing
Show abstract
The method of election-pass filtering and image fusion based on the wavelet transform is presented for eliminating
the fringes in the continuous wave (CW) Terahertz (THz) Images. The CW THz images have the appearance of
sinusoidal waves covering the image with standing wave patterns. The patterns result from reflected inside of the
imaging system interfering with the images coming from the sample. The pattern interferes with contrast
enhancement techniques, and can make the analysis of small areas difficult. An analysis is taken on the traditional
method dealing with this problem by Frequency Domain Filtering and notch filtering, and the disadvantages of the
traditional methods are shown. The Wavelet Transform is chosen to deal with the THz images and eliminate
interference fringes. In the method, when the size of object is larger than one interference fringe, the character of
the object can also be reserved. And the image fusion based on the Wavelet Transform is used to combine the
background and the target with pixel-level fusion, which is proposed to improve the speed of identification and the
quality of the fused image. Comparing the image restored via wavelet transform with other images, it can be found
that the scope of continuous wave THz imaging's application can be extended and clearer images can be acquired.
Ascertaining the relation between underwater bubbles' parameters and valve's outlet pressure with the application of a high-speed camera
Show abstract
To fix the strict relation between the valve pressure and bubble film's parameter the
high-speed camera is used in this paper to record the situation of underwater air bubble films in
the flume. Then the single picture is intercepted from the record to calculating the size distribution
and numeric density of bubbles through partitioning the picture into some special districts
combined with element analysis. Moreover the moving speed of bubbles could be also fixed by
drawing the track of bubbles when several consecutive intercepted pictures were compared
together. These results can provide some reference of the target character for laboratorial
underwater laser detection of wake bubbles.
Process monitor in thermal denaturation of albumin using dynamic speckle method based on wavelet entropy
Show abstract
The process of thermal denaturation of the albumin was investigated using dynamic speckle method
based on wavelet entropy and analyzed by light scattering theory. In experiments, the dynamic
speckle patterns sequences generated by albumin colloid during denaturing were acquired using a
CCD camera. By analyzing the variations of wavelet entropy values of the THSPs (the time history
of speckle patterns), the thermal denaturation process of albumin could be divided into two stages.
At former heating process, the values of wavelet entropy were bigger; correspondingly, the protein
particles were aggregated and flocculated quickly. Conversely, at latter heating process, the wavelet
entropy values decreased drastically, which meant there was slow aggregation. According to those,
the movement properties of the protein molecule ensemble were analyzed during thermal
denaturation of the albumin. The results show that this method is effective to analyze the process of
movement and aggregation of protein molecules quantitatively. The experiment proved that this
method is an useful tool to investigate the particles motion in solution.
Image enhancement for liver CT images
Yanmei Liang,
Li Yang,
Hailun Fan
Show abstract
Image enhancement methods for liver computed tomography (CT) images are studied in this paper. The liver region is
first segmented from the whole CT image by simply using the characteristics of the gray distribution and the size of the
liver; some algorithms are then used to enhance the contrast of the segmented liver image. The method of selective
histogram equalization is proposed to enhance the segmented liver CT image. Some existing enhancement algorithms are
explored and compared on the effect of the enhancement. It is proven from the experimental results that by this two-step
method, namely, the segmentation-enhancement method, the visual effect of the segmented image can be effectively
improved and the focus is obviously highlighted.
Remote Sensing, Lidar, and Environmental Applications
Application of satellite remote sensing images in urbanization and surface energy balance of Beijing
C. L. Meng,
L. P. Lu
Show abstract
This paper first uses data from USGS (Retrieved from NOAA AVHRR) in 1993 and MODIS in 2003 to study the
urbanization in Beijing, and then uses both data as the input to the urbanized CoLM land surface model to discuss its
effect on surface energy balance etc. The result indicates that Beijing was experiencing a rapid urbanization process in
the research period. As the result of the urbanization, the vegetation cover decreases and the impervious surface area
increases. The radioactive temperature is also increased according to the heat island effect and the surface wetness is
decreased. For the effects of surface energy balance, as the result of decreases of evapotranspiration, the latent heat flux
decreases and surface sensible heat flux increases and the Bowen ratio increases in the urban areas too, but the total
sensible heat flux is not increased.
Efficient visualization of LIDAR datasets
D. Mongus,
S. Pečnik,
B. Žalik
Show abstract
The purpose of presented point-based rendering method is to visualize data retrieved from 3D scanners, especially from
Light Detection and Ranging (LIDAR) scanners. LIDAR data scanners capture highly accurate and dense data and are
used to gather vast spatial information about geographical areas. As a result, a lot of points are obtained and optimization
techniques are needed for their real-time visualization. Fully interactive system with high frames-per-second (FPS) ratio
is presented in this paper. The efficiency is achieved by reduction of a graphics workload, based on space subdivision,
using a quad-tree. The points maintained in the leaves of the quad-tree are aligned evenly through nine or more depth-levels,
depending on the number of points in the leaf. Such subdivision allows effective control of level-of-details (LOD)
as it minimizes the number of calculations, needed for the visualization in real-time. The changes in the organization of
data during the run-time are very rare and therefore, the LOD hierarchy can be established in the preprocessing. A
method for achieving a detailed look of the scene, even at low LOD degree, is implemented, too.
Image quality degradation analysis induced by satellite platform harmonic vibration
Show abstract
The paper discusses how different frequencies vibrations influence the image quality and limit resolution. The main module of image quality degradation is based on the physical optical imaging process. This model builds up the accurate relationship between image space and object space, and reproduces entire imaging process from time-space joint relation. In this experiment, fixed amplitude harmonic vibrations with frequencies between 1Hz to 2000Hz applied to the 2-Dimesion unit impulse function image and standard knife edge image. Then those degenerated images are analyzed to
measure the PSF (Point Spread Function) and MTF (Modulation Transfer Function). As the result, with the increase of the ratio of Te T0 , the PSF and MTF become constringent. When the ratio of Te to T0 is greater than 10, the maximum of relative error of the PSFs is less than 5%. This experiment reveals that the very high frequency vibrations have almost the same effects to degenerate the image as the relative lower frequency vibrations, which means that the
frequency is not as important as the vibration amplitude.
Imaging, Processing Algorithm, and Analysis
A color image analysis approach using fusion of Markov random fields in different color spaces
Show abstract
This paper presents a new color image analysis approach which fuses several processing maps by Graph-Cuts algorithm
for Markov Random Field (MRF) in different color spaces. Recently, graph-based image analysis methods, such as
Graph-Cuts, have been achieved exciting results for approximate inference in Markov Random Field. One color image
can be represented in many color spaces, such as RGB, LAB and HSV, but existing Graph-Cuts approaches often
compute global energy function only in one color spaces and ignore that each color space has an interesting property for
certain applications respectively. This paper processes images in MRF and represents them in one MRF model firstly.
Then Graph-Cuts algorithms are used to process images in each color space and generate one map. Several processing
maps can be acquired from some color spaces. These maps are fused to get more reliable and accurate results. We select
stereo matching which can get depth maps from multi-view images to evaluate our image analysis approach. The
experiments herein reported in this paper illustrate the potential of this approach compared to existing Graph-Cuts
methods from processing results.
Level 0 and Level 1 data processing for a type of hyperspectral imager
Xiaohui Li,
Changxiang Yan
Show abstract
Hyper-spectral imaging (HSI) is a kind of optical remote sensor that can simultaneously obtain spatial and spectral
information of ground targets. We are now designing a data processing system for a type of space-borne push-broom HSI,
which has 128 spectral channels covering the spectral range from 400nm to 2500nm. With its large amount of spectral
channels, the HSI collects large volume of spectral imaging data need to be efficiently and accurately processed and
calibrated. In this paper, the detailed Level 0 and Level 1 data processing steps for the HSI were presented. The Level 0
processing refers to a set of tasks performed on the data downlinked from the spacecraft, including decoding to extract
science data, separating the science data into files corresponding to different tasks (e.g. ground imaging, dark imaging,
and onboard calibration), checking data integrity and instrument settings, data format conversion, and Level 0 files
creation. The Level 1 processing performs several steps on Level 0 data. Firstly, it corrects the image artifacts (mostly the
SWIR smear effect), subtracts the dark background, and performs the bad pixel replacement according to the prelaunch
measurement; then it performs radiometric and spectral calibration based on the ground calibration results and onboard
calibration collection. The detailed algorithms for bad pixel replacement, radiometric and spectral calibration were
presented. After processing, the digital numbers downlinked from the spacecraft can be converted into at-sensor absolute
spectral radiance of ground targets, thus providing accurate quantified spectral imaging data for various applications.
Optical image hiding based on interference
Show abstract
Optical image processing has been paid a lot of attentions recently due to its large capacitance and fast speed. Many
image encryption and hiding technologies have been proposed based on the optical technology. In conventional image
encryption methods, the random phase masks are usually used as encryption keys to encode the images into random
white noise distribution. However, this kind of methods requires interference technology such as holography to record
complex amplitude. Furthermore, it is vulnerable to attack techniques. The image hiding methods employ the phase
retrieve algorithm to encode the images into two or more phase masks. The hiding process is carried out within a
computer and the images are reconstructed optically. But the iterative algorithms need a lot of time to hide the image into
the masks. All methods mentioned above are based on the optical diffraction of the phase masks. In this presentation, we
will propose a new optical image hiding method based on interference. The coherence lights pass through two phase
masks and are combined by a beam splitter. Two beams interfere with each other and the desired image appears at the
pre-designed plane. Two phase distribution masks are designed analytically; therefore, the hiding speed can be obviously
improved. Simulation results are carried out to demonstrate the validity of the new proposed methods.
Light source target recognition based on horizontal projection and differential algorithm
Show abstract
In the image-based bridge deflection measurement system, changes in weather and interference of background cause light source target hard to accurately recognize. Through analyzing the principle of
image-based deflection measurement, a conclusion was drawn that small error in light source target in image recognition cause larger error in deflection because of amplification of optical measurement
system. In order to solve the problem, analyzed gray level characteristic of target image, and proposed an image enhancing method. In addition, a new method based on horizontal projection and differential algorithm is adopted for recognition of light source target. Finally, we use gray weighted centroid algorithm to realize sub-pixel location of light source target. Moreover, experiments were done for target image with poor quality. The experimental results show that the method is adaptable to the changing weather.
Motion clustering and object detection via modulated integral imaging
Show abstract
A moving object detection method based on the dense velocity field detection followed by an unsupervised
spatiotemporal clustering by means of the mean shift algorithm is proposed. The method firstly makes use of a novel
imaging system and an algebraic solution to detect the extremely dense velocity vector distribution with a pixel-wise
spatial resolution and a frame-wise temporal resolution. It is based on the complex sinusoidal-modulated imaging with a
three-phase correlation image sensor (3PCIS) and an exact algebraic inversion method based on an equivalent weighted
integral form of the optical flow partial differential equation. Since the inversion method is free from time derivatives,
any limitations on the object velocity and inaccuracies due to approximated time derivatives are thoroughly avoided.
Secondly, in order to segregate dominant velocity regions from noisy background and identify the exact moving object,
the mean shift clustering method is employed to the spatially and temporarily dense velocity vector distribution. To
avoid too many cluster centers or over clustering, the traditional mean shift method is improved by setting an emerging
condition. An experimental system was constructed with a 320×256 pixel 3PCIS device and a standard PC for inversion
operations and display. Several experimental results are shown including an application to dense motion capture of face
and gesture and traffic scene, which including several independent moving objects.
SoPC implementation of combined real-time non-uniformity correction of IRFPA
Show abstract
With the increase of pixel density and scale of IRFPA (Infrared Focal Plane Array), NUC (Non-Uniformity Correction)
technology becomes more and more important to access high quality infrared images. However, traditional one-point or
two-point NUC methods based on calibration technology can't achieve ideal performance because they can't overcome
the non-linearity and drift of the detector response parameters in both spatial and temporal regions effectively. A novel
combined real-time non-uniformity correction method is proposed based on FPGA (Field Programmable Gate Array)
technology, which adopts SoPC (System-on-a-Programmable Chip) architecture based on Nios II processor core to
implement the total NUC processing functions inside only one chip. The NUC processing chooses the reference-based
binomial fitting algorithm to remove the main non-uniformity of the detector, and the remained non-uniformity is
compensated by using the improved scene-based temporal high-pass filter algorithm. The experiment results show that
the combined method based on SoPC architecture can access the ideal efforts with IRFPA size of 320×240×14bit @ 25
frames per second. The block diagram of hardware circuit and the processing flow are described in details.
Poster Session
Study on the particle-filter-based motion filtering algorithm for digital image stabilization systems
Show abstract
Towards to the existent problems in current motion filtering algorithms such as the low-pass filter, the Kalman filter and the
extended Kalman filter. We proposed a novel motion filtering algorithm base on particle filter to separate random jitter from
the global motion vectors. In this method, the variation of global motion parameters are regarded as the state-variable of
system, the uniform motion model of camera is used, and the motion filtering is carried out to dynamic image sequences
according to the features of digital image stabilization systems. Experimental results prove that this particle-filter-based
motion filtering algorithm can achieve real-time filtering effect, and the filtering effect can be affected by the number of the
particles only and almost irrespectively with other factors. This method can be used agilely and is very suitable for digital
image stabilization systems application. We realized the programs based on the TI TMS320C6416DSP processing chip and
got very perfect experiment results.
Fuzzy synthesis evaluation for image target recognition performance
Show abstract
With the rapid development of optoelectronic tracking and measurement technology, the testing and evaluation for the
tracking system and its inner algorithms are urgently demanded. Automobile target recognition(ATR) technology for the
image is a key part of the tracking system based on the image, and develops advanced and fast, which makes the
performance evaluation difficult and complex. There is not a reliable and effective evaluation method adaptable to the
developing technology. Therefore, a fuzzy synthesis evaluation method for ATR system or its detection and recognition
algorithms group was proposed. The evaluation indexes were selected and designed, which weights were calculated by
the direct method, W-road method and the change weight method. The simulation testing conditions, the size and the
hypothesis test methods of the statistic swatch were discussed. The mean, the covariance dependency and the distribution
indexes of the probability of detection(Pd) were effectively tested. The statistic ranges corresponding to evaluation ranks
of these indexes were established. Finally, the simple model and the division model of the fuzzy synthesis evaluation
algorithm were discussed. Tests show, this method is valuable for getting the occurrence probabilities of the different
performance racks of the system or the algorithms group corresponding to vary environment levels.
Ultraweak bioluminescence of maize under NaCl stress
Show abstract
In this paper, the delayed ultra-weak luminescence (UWL), leaf dry/fresh weight and reactive oxygen species (ROS) of
the maize (Jundan 20) were investigated under various NaCl concentration. The results showed that following with the
increase of NaCl, ROS contents increased obviously, but the UWL intensity decreased, and the decrease speed increased
following with the increase of NaCl concentration, the decrease extent of delayed luminescence ranged from 10.4% to
27.2%. It was also found that both dry and fresh weight of seedling leaves decreased, but the ratio of dry/fresh weight
increased. According to these results, we speculated that the decrease of UWL was mainly closely associated with the
destruction of seedling normal physiological activities and metabolic disorder which were caused by NaCl stress, rather
than ROS only. This study revealed that the spectral analysis of UWL is a useful tool for studying plant response to salt
stress.
Free space optical communication system based on wavefront sensorless adaptive optics
Liqiang Han,
Katsunori Shida,
Qi Wang,
et al.
Show abstract
Free space optical communication technology approach has a number of advantages; unfortunately, it can be severely
affected by atmospheric turbulence. Adaptive optics is the technology for correcting random optical wavefront
distortions in real time. This paper is the research of a wavefront sensorless adaptive optics system which, in contrast
with conventional AO systems, would offer reduced complexity, size and cost and would facilitate widespread
application. Resulting aberration correction is demonstrated in laser communication system.
Interline transfer area CCD imaging system with FPGA for real-time processing
Show abstract
The paper presents the hardware and software design of an area CCD imaging system with FPGA for real-time
processing, which can be used in image capture system. The interline transfer area CCD uses ICX424AQ from Sony
Company. The paper makes a detailed analysis on the structure parameters of ICX424AQ and the image capture
principle. On this basis, the design realizes the CCD control sequence and the timing logic of the whole capture system.
The area CCD image sensor is covered by Bayer color filter array (CFA). Each pixel has only one component of three
primary colors. In order to obtain full chromaticity at every pixel, the paper adopts an improved bilinear interpolation
algorithm. The CCD is working under progressive scan mode. All pixel signals are read out continuously at the exposing
time of 0.32ms. The whole system is controlled by FPGA, the pixel data transmits by transmitting chip SiI1162. At last,
the design realizes the real-time display on TFT-LCD.
Determining the optimal imaging position in tomographic interference microscopy
Show abstract
Tomographic interference microscopy is a method which can obtain the three-dimensional refractive index of live cells
and tissues. In this paper, the cone-shaped light with the transverse scanning is adopted, which offers non-contact, highresolution
and real-time cell magnifying imaging facilities relatively to the traditional parallel beam projecting on the
tissue sample. However, the index-induced focal shift is the common disadvantage for this tomographic interference
microscopy, which leads to the decline in image quality and impacts later reconstruction. Then the image sharpness
metric based upon the histogram of image to determine the index-induced focal shift is introduced. The experimental
results show that the variances of histograms are compatible with a Gaussian function. The peak value of the Gaussian
curve corresponds to the optimal imaging position's histogram variance.
Meaningful region extraction based on three-stage unsupervised segmentation algorithm
Show abstract
From a theoretical standpoint, meaningful region segmentation based only on gray level or color usually presents over
segmentation or non-continuous regions. In view of this, we adopt a number of classical powerful algorithms (mean shift
clustering, edge detection and region growing) to extract the meaningful regions adds spatial information. These
algorithms are subjectively connected together and impact the results each other. The experiments indicate that the
proposed method can avoid over-segmentation phenomenon and the results can be easily accepted by human eyes.
Experimental results are superior to that of kmeans clustering method in both real-time performance and image
segmentation performance. Finally, we achieved a new procedure to extract meaningful regions by clicking some place
of a color image. It possesses a good application prospect and owns an effective real-time performance.
The design and experiments of the super-resolving annular binary filters based on genetic algorithm
Show abstract
Annular binary filters that can fulfill the superresolution are designed and produced. The transverse and axial resolution
gains of the filters are defined. The transverse gain is optimized by genetic algorithm to obtain the optimal structure
parameters. The designed filter has a transverse gain of 1.37. The distributions of the transverse and axial intensity of the
designed filters and the general circular pupil is illustrated, from which we can see that the spot size of the former is
reduced by 30% and the depth of focus is increased compared with the latter. The optical system for testing the
performances of the binary annular filters was build. The experimental results show the binary annular filter has fulfilled
the superresolution.
3D reconstruction based on wavelet pyramid image fusion algorithm for digital microscope
Show abstract
Optical microscopy is an important technique in petrological and biological. However, because of the limited
focus depth of a traditional optical microscope, 3D objects can't be observed with all its information in one
scene. The commonly used method to solve this problem is to use a digital microscope to collect a sequence
of multi-focus images with a constant change of the focal length and then get the final sharp image by fusing
the images. Wavelet pyramid algorithm is one of the commonly used fusion algorithms and has a good
performance. In this paper, a novel 3D reconstruction algorithm is presented based on the wavelet pyramid
image fusion algorithm. First, a sequence of multi-focus images is collected with a digital microscope and
then the images are fused with the wavelet pyramid fusion algorithm. After that, the deviation between the
fused image and each of the original multi-focus images are calculated pixel by pixel. For convenience, the
absolute value of the deviation is calculated. The smaller the deviation, the closer the corresponding pixel to
the focal plane and vice versa. Thus a 3D deviation matrix is obtained. For each pixel position (x, y), the z
index of the smallest deviation value in the third direction of the deviation matrix is considered the pixel's
height index. A complete height index map is obtained by selecting the indices of these smallest deviation
values. In order to eliminate the noise effect, a median filter is applied to the height index map. The height
index map times the height factor (the unit step length along the optical axis when collecting the multi-focus
images) is the object's actual height map. Finally, the surface of the 3D object is reconstructed with the
object's height map. In order to test the reconstruction algorithm, a dedicated high resolution Complementary
Metal Oxide Semiconductor (CMOS) digital microscope is designed and fabricated to collect the multi-focus
images. Experimental results show that the proposed algorithm produces a nice looking surface of a 3D
object.
Two-wavelength moiré deflection tomography in studying the dispersive characteristic for flame flow fields
Yun-yun Chen,
Yang Song,
An-zhi He,
et al.
Show abstract
In this paper, the dispersive characteristic of flame flow fields is studied. Two-wavelength moire; deflection tomography is introduced to verify the dispersive characteristic for flame flow field, while 532nm and 808nm are chosen as the probe wavelengths. The combustion of propane in air is chosen as a practical example for theoretical and experimental investigation. The theories manifest that the dispersive capability of flame is determined by temperature, pressure, components as well as the probe
wavelength, and the contribution of an electron which is one of the components of flame to the refractive index is nearly 102 or 103 times that of a neutral particle in the given wavelength region. In short, the theoretical analysis and discussion match well with the experimental results.
Study on preparation and electrical properties of Mo-doped vanadium oxide thin films by organic sol-gel
Show abstract
An organic sol-gel method of preparing Mo-doped vanadium oxide thin films was described. Vanadium oxide powder and molybdenum trioxide powder were dissolved into a mixed solution of benzyl alcohol and isobutyl alcohol to obtain a vanadium solution, which was then spin-coated on glass substrates followed by an anneal at 500°C. Square resistance was measured by the four-point probe, and film thickness was measured by the effects of molar ratio. The effects of
doping concentration and film thickness on the electronic properties of vanadium oxide thin film, and film uniformity as well as stability are studied. Results show that the TCR (30°C)
of VO2 films was increased greatly, and the square
resistance was reduced by Mo-doped.
Investigations on titanium oxide film as an alternative heat sensitive material
Show abstract
In this article, substoichiometric titanium oxide thin films (TiOX) are prepared with the reactive DC sputtering in an
oxygen and argon atmosphere and then annealed in an oxygen atmosphere. Under a set of optimum deposition
parameters, TiOX film with sheet resistance value of 167.9 KΩ/(box) and TCR value of -3.30%/K obtained. Transmittance of
the optimum film is obtained in the 300≤λ≤1100nm wavelength range at room temperature. Thickness, near-infrared
absorptance, and other properties of the optimum film are also investigated in this paper.
Study on aero-optical effect of a hypersonic missile infrared image guide
Show abstract
When a hypersonic missile with a side mounted IR seeker is launched in the atmosphere, a serious aero-optical effect is formed and affects the quality of the detector's imaging. And in the course of the missile flight time,aero-optical effects changes over time, which makes real-time, accurate correction of optical distortion becomes very difficult. Therefore, it is necessary to study the relationship between the optical distortion and time.In this paper, Fluent,a computational fluid dynamics(CFD) internet applications is used to make researches on effect of optical transmission of the flow field outside the IR window. And a thermal finite element analysis (FEA) of an IR window is used to study the aerodynamic
heat effect.
TOCM digital color photography
Show abstract
In this paper, total optical color modulator (TOCM) digital color photography is presented. TOCM has the character of
multi-wave superposed in spatial domain and separated in frequency domain. If TOCM is close-contacted with the image
plane of a black-and-white (B&W) CCD, the encoding B&W CCD is formed. Image from the encoding B&W CCD are
digital encoded by the TOCM. The decoded color image can be obtained by computer program. The program includes
four main steps. The first step is Fourier transforming of the encoded image. The second step is filtering the spectra of
the first and zero order in frequency domain. The third is inverse Fourier transforming of the filtered spectra. The last is
melting the image with zero order. Then the digital color image will be shown on the display of the computer. The
experiment proves that this technique is feasible. The principle of encoding color information in B&W image can be
applied to color-blind sensors to get digital color image. Furthermore, it can be applied to digital multi-spectra color
photography.
Detection technique of targets for missile defense system
Hua-ling Guo,
Jia-hao Deng,
Ke-rong Cai
Show abstract
Ballistic missile defense system (BMDS) is a weapon system for intercepting enemy ballistic missiles. It includes
ballistic-missile warning system, target discrimination system, anti-ballistic-missile guidance systems, and
command-control communication system. Infrared imaging detection and laser imaging detection are widely used in
BMDS for surveillance, target detection, target tracking, and target discrimination. Based on a comprehensive review of
the application of target-detection techniques in the missile defense system, including infrared focal plane arrays
(IRFPA), ground-based radar detection technology, 3-dimensional imaging laser radar with a photon counting avalanche
photodiode (APD) arrays and microchip laser, this paper focuses on the infrared and laser imaging detection techniques
in missile defense system, as well as the trends for their future development.
Research on application of LADAR in ground vehicle recognition
Show abstract
For the requirement of many practical applications in the field of military, the research of 3D target recognition is active.
The representation that captures the salient attributes of a 3D target independent of the viewing angle will be especially
useful to the automatic 3D target recognition system. This paper presents a new approach of image generation based on
Laser Detection and Ranging (LADAR) data. Range image of target is obtained by transformation of point cloud. In
order to extract features of different ground vehicle targets and to recognize targets, zernike moment properties of typical
ground vehicle targets are researched in this paper. A technique of support vector machine is applied to the classification
and recognition of target. The new method of image generation and feature representation has been applied to the
outdoor experiments. Through outdoor experiments, it can be proven that the method of image generation is stability, the
moments are effective to be used as features for recognition, and the LADAR can be applied to the field of 3D target
recognition.
Image sharpness function based on edge feature
Show abstract
Autofocus technique has been widely used in optical tracking and measure system, but it has problem that when the
autofocus device should to work. So, no-reference image sharpness assessment has become an important issue. A new
Sharpness Function that can estimate current frame image be in focus or not is proposed in this paper. According to
current image whether in focus or not and choose the time of auto focus automatism. The algorithm measures object
typical edge and edge direction, and then get image local kurtosis information to determine the degree of image
sharpness. It firstly select several grads points cross the edge line, secondly calculates edge sharpness value and get the
cure of the kurtosis, according the measure precision of optical-equipment, a threshold value will be set beforehand. If
edge kurtosis value is more than threshold, it can conclude current frame image is in focus. Otherwise, it is out of focus.
If image is out of focus, optics system then takes autofocus program. This algorithm test several thousands of digital
images captured from optical tracking and measure system. The results show high correlation with subjective sharpness
assessment for s images of sky object.
Image restoration for sparse aperture systems based on wavelet-Wiener algorithm
Show abstract
The sparse aperture system, which utilizes several sub-apertures in place of monolithic surface, has attracted more
interests because it has the advantages of lower cost and lighter weight, and keeps the same aperture size to reach the
demanded angular resolution. However the image quality degrades because of smaller effective aperture areas. When the
diffraction-limited sparse aperture optical system is imaging ideally, since the optical transfer function is known in the
diffraction-limited sparse aperture system, wiener filter is thought to be the best tool to restore the images. In the actual
imaging process, an image must been disturbed by varieties of noises so that the ability of Wiener filtering image
restoration degrades obviously, the restoration effect of the images with noises by using Wiener filtering is not to be
efficient. This paper proposes an improved de-noising algorithm by analyzing traditional wavelet threshold de-noising
method. For the images created by using the simulated sparse aperture optical system, first, we can remove the noises in
the images using the improved wavelet threshold method and enhance the signal-to-noise ratio, and then obtain the more
ideal image formation in the greatest degree, and finally perform restoration of the preprocessed images based on the
improved Wiener filtering method. The simulated experiments are fulfilled with sparse aperture system Golay6 with
different fill factors designed with the aid of the optical designing software system ZEMAX. The simulated results
demonstrate that the algorithm proposed in this paper is superior to normal Wiener filtering or the improved
Wavelet-Wiener filtering method.
Infrared small target detection based on Danger Theory
Show abstract
To solve the problem that traditional method can't detect the small objects whose local SNR is less than 2 in IR images,
a Danger Theory-based model to detect infrared small target is presented in this paper. First, on the analog with
immunology, the definition is given, in this paper, to such terms as dangerous signal, antigens, APC, antibodies. Besides,
matching rule between antigen and antibody is improved. Prior to training the detection model and detecting the targets,
the IR images are processed utilizing adaptive smooth filter to decrease the stochastic noise. Then at the training process,
deleting rule, generating rule, crossover rule and the mutation rule are established after a large number of experiments in
order to realize immediate convergence and obtain good antibodies. The Danger Theory-based model is built after the
training process, and this model can detect the target whose local SNR is only 1.5.
Generation and optoelectronic reconstruction of binary CGH based on detour phase encoding
Su-juan Huang,
Xiao-jing Liu,
Shuo-zhong Wang,
et al.
Show abstract
A binary amplitude Fourier computer-generated hologram is produced, from which both numerical and optoelectronic
reconstruction can be implemented. The hologram is formed on the computer screen based on improved detour phase
encoding, and saved as a digital image at suitable resolution without printing and photographical reduction. Numerical
reconstruction can be obtained directly from inverse Fourier transform of the hologram, resulting in only the original and
conjugate images. Optoelectronic reconstruction can be realized by using an electronic addressing reflective
liquid-crystal display (LCD) spatial light modulator. A binary digital hologram from the computer is loaded on a LCD
panel. A clear image can be observed in the distance by illuminating a reference light from a laser source to the
liquid-crystal panel. The digitally stored hologram can be used for dynamic electro-holographic display, unlike the
existing methods based on detour phase encoding which can reconstruct hologram either numerically or optically. The
proposed method significantly reduces noise in the reconstructed images. Experimental results show effectiveness of the
method.
Three-dimensional scene reconstruction by using lenslet array model in computational integral imaging
Show abstract
Integral imaging is a promising technique for both 3-D scene capturing and reconstruction. Recently, computational
simulation has been used to generate the free view of reconstructed scenes without optical devices, which can easily
overcome image quality degradation due to the physical limitations of optical devices. In the reconstruction process of
integral imaging, current researches focus on the pinhole array model which regards lenslet array as pinhole array for
simplicity. But in fact, the optical characteristics of the lenslet such as the focal length, the aperture size of the lenslet,
and so on, have significant impact on the reconstructed 3-D scene. In this paper, we proposed a lenslet array model in
computational integral imaging. The elemental images were picked up by using a well developed computer graphics
programming library OpenGL. And then 3-D scene was reconstructed by an ideal diffraction-limited integral imaging
model which taken into account of the effect of the focal length and the aperture size. We presented some simulations
and evaluated the image quality by the peak-to-peak signal-to-noise ratio (PSNR). Experimental results show that the
proposed lenslet array model increase the depth of field.
Color sorting algorithm based on K-means clustering algorithm
BaoFeng Zhang,
Qian Huang
Show abstract
In the process of raisin production, there were a variety of color impurities, which needs be removed effectively. A
new kind of efficient raisin color-sorting algorithm was presented here. First, the technology of image processing basing
on the threshold was applied for the image pre-processing, and then the gray-scale distribution characteristic of the raisin
image was found. In order to get the chromatic aberration image and reduce some disturbance, we made the flame image
subtraction that the target image data minus the background image data. Second, Haar wavelet filter was used to get the
smooth image of raisins. According to the different colors and mildew, spots and other external features, the calculation
was made to identify the characteristics of their images, to enable them to fully reflect the quality differences between
the raisins of different types. After the processing above, the image were analyzed by K-means clustering analysis
method, which can achieve the adaptive extraction of the statistic features, in accordance with which, the image data
were divided into different categories, thereby the categories of abnormal colors were distinct. By the use of this
algorithm, the raisins of abnormal colors and ones with mottles were eliminated. The sorting rate was up to 98.6%, and
the ratio of normal raisins to sorted grains was less than one eighth.
Nondestructive quality inspection of fruits based on multispectral imaging
BaoFeng Zhang,
Xiaohui Li
Show abstract
A new method of non-destructive quality inspection of fruits was investigated based on multi-spectral imaging technology, and the system was also developed, which consists of lighting chamber, multi-spectral light source, CCD camera, frame grabber and a computer. According to the basic color theory[1]; multi-spectral images were abstained by this system, and analyzed using multi-spectral images fusion methods. It was found that the multi-spectral images had the highest recognition rate when the spectral bands of multi-spectral fusion images are red spectral band, blue spectral
band, near infrared spectral band. After the feature extraction of the multi-spectral images, the fruits were classified according to size, shape, defects and other parameters. It was shown by the experiment that the accuracy of the fruits quality inspection can be achieved 90.25% using multi-spectral detection technology on non-destructive.
Tracking device of multiple targets based on IR fish-eye system
Show abstract
A tracking device of multiple target is proposed, whose field range reaches hemisphere airspace(sr) because of the using
of IR Fish-eye system. The infrared ray is collected by Fish-eye lens, by this way can get the thermal image about the
whole space, whose range depends on the lens used. Thanks to the Fish-eye system and Ptsi staring focal plane array, the
tracking device can get a big range, that means can get multiple target at one time. Depends on the algorithm adopted to
process the image has been gotten, the tracking device specify some points with intensive ray which means elevated
temperature, and that will be the targets. At the next step after targets detection, the algorithm will compute out the speed
and orientation of every target, then figure out a quotient-for-danger about every target. With this quotient, the algorithm
can judge which one is the most dangerous from those flying to us, and it should be treated firstly, then the less
dangerous ones, etc. To ensure the whole system run at real-time, all of the processing has to be finished at a moment. A
TM320 DSP is adopted to run the algorithm. The result will be displayed at a LCD with every target labeled by their
quotient-for-danger. The reaction speed and precision of this tracking device has been proved on our multiple target
experiment platform.
An experimental device to track multiple targets on IR fish-eye lens system
Show abstract
Large field of view IR fish-eye system can survey real-time the target in hemisphere airspace. For the single target, it can
find the cooperative target and experimentalism easily. But multitasking for the multiple detecting, tracking and judging
the level of threat it could find the cooperative target and prove the studied algorithms and hardware system very difficultly[1]. The paper brings forward an experiment system which is used for identifying multiple target. Through the system the simulation of the parameters in fairway, distance, direction and velocity are achieved. The system is composed of many controlled fountain of heat, which are made up of symmetrically heated flat and temperature
controlled equipment. Each fountain of heat is fixed on the carrier moving free in three dimension space, in front of which diaphragms are fixed for changing the size of diaphragms and distributing of temperature. The fountain of heat moves on the different track at different speed and direction when it's working. The experiment system is proved to be feasible and precise through simulating single target and tracking the plane and other targets. Based on this system, the algorithms of multi-target tracking are validated.
Multiple target tracking algorithm research on IR fish-eye lens system
Show abstract
The paper presents the algorithms of data process on multi-target tracking and threat level sequence in large visual area
IR (infrared ray) fish-eye system. The targeting information is associated and filtered with the polar coordinates. Firstly,The characteristic point of every target is picked up and tracked with special restriction for each of them. Secondly,the parameters of fairway, which are gained by aberration revising and researching the characters of temperature area, are picked up from serial images. The threat level of attacking objects is sorted by the speed, acceleration and direction of all targets. Finding out the most prior target which is in the certain distance in order to make
sure that effective warning information is emitted in time. The algorithms are proved to be quick and effective through
simulation and the muiti-target experiment system developed by the writers.
Double-layer PMN bimorph deformable mirror system optimization design and fabrication technology
Show abstract
In this paper we present a novel bimorph deformable mirror composed of double-layer lead magnesium niobate (PMN) slices. Compared with lead zirconate titanate (PZT), PMN has better physical features served as bimorph mirror. The optimization design of the mirror electrodes pattern and mirror size are investigated. The type of conductive adhesive and coating process are proposed. An improved mounting apparatus using O-type rubber rings for flexible mounting the mirror is innovated. Also the control circuit model and the interface unit with computer are investigated and designed. A mature
design proposal and technology of bimorph deformable mirror system is presented.
Front-based image edge detection
Show abstract
Fronts are elementary structures that separating domains of different steady states in reaction diffusion systems. The local dynamics of front structures have been investigated by a
non-gradient reaction diffusion model of the FitzHugh-Nagumo type. It is found that stable and stationary fronts can be extracted under certain conditions. Based on the local dynamics of front structures, a novel method of image edge detection has been proposed in this paper when fronts are considered as edges of images.
Modeling and analyzing stripe patterns in fish skin
Show abstract
The formation mechanism of stripe patterns in the skin of tropical fishes has been investigated by a coupled two variable reaction diffusion model. Two types of spatial inhomogeneities have been introduced into a homogenous system. Several Turing modes pumped by the Turing instability give rise to a simple stripe pattern. It is found that the Turing mechanism can only determine the wavelength of stripe pattern. The orientation of stripe pattern is determined by the spatial inhomogeneity. Our numerical results suggest that it may be the most possible mechanism for the forming process of
fish skin patterns.
A study of moment invariants based on pattern recognition
Show abstract
Using moment invariants algorithm extract the eigenvector for the characteristics of the human ear. The algorithm is not subject to restrict on translation, rotation and the ratio of changes.
Canny algorithm used for edge detection of human ear base on moment invariants. Experiment show that Canny algorithm can not miss the real edge and put the non-edge points as the edge point detection.
It detects the edge points as possible in the actual center of the edge and Canny algorithm also inhibit the false edge response.
A study of optical system and image processing in Vickers hardness photoelectric detection system
Show abstract
Although the New Vickers Hardness Tester has integrated with today's high and new technologies, and has been digitalized, automated, intellectualized, and gradually developed into a video-based photoelectric detector, one of the key technologies still be the advanced Optical Imaging Technology. Putting this technology into use, an Optical Image Acquisition System with fine image quality can be designed and created. The key technologies of Vickers Hardness
Tester are discussed in this paper with an example of Anchor Clamp Hardness Photoelectric Detection System including the Special Infinite Image Distance Optical System and Image Processing Software. A Special Infinite Image Distance Optical System composed of high amplification Semi-apochromatic Microscope Objective lens with Flat-field and long working distance and Photographic Objective lens is designed, and taking the advantage of optical device based on the Special Infinite Image Distance Optical System, the image of the shape of impression of a diamond cone can easily be
acquired when it is pressed into Anchor Clamp. Indention image processing are carried out through the self-designed Image Processing Software. At present, this system has been successfully applied to Anchor Clamp Hardness On-line Lossless Automatic Detection Apparatus.
The research of underwater imaging based on the wavefront coding system
Qin-xiao Liu,
Ting-yu Zhao,
Yan-ping Chen,
et al.
Show abstract
Current research on the underwater imaging often needs particular techniques. Since the specificity of the underwater
imaging system, it must overcome the impact of the turbulent and turbid of the water. In this paper we present a new
application of the wavefront coding (WFC) system in the underwater imaging. Wavefront coding system is a two-step
imaging system, by using a cubic phase mask with optimized parameters; we get the defocus-insensitive intermediate
image in the imaging plane. Then the final sharp image can be obtained by deconvoluted the intermediate image with the
point spread function (PSF). Initially the wavefront coding system is introduced to extent the depth of the field of the
optical system and the allowance of the system can be great improved by using the optimized cubic phase mask. Because
of the defocus-insensitive property, this system can eliminate the aberration brought by the turbulent of the water. The
deconvolution can improve the imaging quality by deblurring intermediate image and reduce the impact brought by the
turbid of the water. In order to get excellent result, we propose to use the global generalized minimal residual method as
the deconvolution algorithm. Antireflective boundary conditions (BCs) are considered to eliminate ringing effect, and
Kronecker product approximation is introduced to reduce the computation consumption.
This method can give excellent deconvolution result. There is little ringing effect on the image border and little vibration
on the edge. Theory analysis and the experiment show that this is an effective system in the underwater imaging. It is an
economical system.
The effect of sampling on FFT-based direct integration method in digital holography
Show abstract
The Rayleigh-Sommerfeld formula(RS) has proved accurate for evaluating diffraction of the optical field from a planar
aperture. Thus the FFT based direct integration method for the RS(FFT-DIRS) can provide a more exact reconstructed
image from sampling points of the diffraction field of the object than the numerical method for the Fresnel formula(FR)
that is an approximation of the RS. Although the FFT-DIRS has been proposed and studied in some literatures, an
important problem remains to be solved, that is the effect of sampling on it. Sampling of the object diffracted field leads
to a periodic or quasi periodic shifting of the reconstructed image. If these spatial replicas overlap, the desired image can
not be recovered without the aliasing noise. So the overlapping period plays an important role in employing the FFTDIRS
for the practical applications. In this paper, a formula of this overlapping period is obtained through the
relationship between the RS and the FR. Then the validity of this formula at different distances is investigated by the
experimental results.
Multiple spiral patterns in a cardiac tissue
Zhanguo Bai,
Xia Li
Show abstract
Ventricular fibrillation (VF) is the major cause of sudden cardiac death, the leading cause of death in the
industrialized world. However, the mechanisms for its onset are still not well understood. Recent experiments indicate
that VF is induced by transitions of cardiac electric propagationg waves from a single spiral wave to multiple waves. To
further understand the underlying mechanism of VF, we investigated the interaction between two waves in a
two-dimensional excitable media. Three types of multiple spirals including multi-arm spirals have been found depending
on the rotation direction and the distance among spiral waves.
Research on scene-based Narcissus correction algorithm in infrared focal plane arrays
Show abstract
Cooled infrared focal plane array (Cooled IRFPA) has been applied widely in military and civil fields. In the cooled
focal plane array detector infrared imaging systems, narcissus as a unique defect has a degenerate effect on the image
quality of cooled infrared imaging systems, such as blocky, annular or "corner heat" fixed pattern noise in infrared
images. Therefore, the correction of narcissus effect using image processing method is important to infrared imaging
systems. Through analyzing manifestation of narcissus effect by statistical approach, a cooled infrared imaging system
model considering narcissus effect is established, by which a self-adaptive narcissus correction algorithm based on scene
statistic and radial compute is presented. From the simulation and the experiment on an actual cooled IRFPA module,
this algorithm is proved very effective on correcting narcissus effect and improving image quality. Moreover, the peak
signal to noise ratio (PSNR) also increases obviously.
Research and development on embedded omnidirectional vision tracker for vision navigation
Show abstract
Omni-directional vision appears the definite significance since its advantage of acquiring all vision information
simultaneously. In this paper, an integrated omni-directional vision tracker based on the configuration with CMOS,
FPGA and DSP has been implemented. Fisheye lens is one of the most efficient ways to establish omni-directional
vision systems, however, it appear with an unavoidable inherent distortion. An imaging system model which consists of
fisheye-lens and the embedded tracker has been proposed. A novel beacon owning the feature of particular topology
shape can be identified by the appropriative recognition processes. Particle filter has been programmed as an
intersectional arithmetic structure which processes the same step of several particle filters simultaneously. We called it as
Multiple Intersection Particle Filter. MIPF makes multi-targets tracking efficiently and successfully on embedded
platform. A rectification technique based on equidistant projection theorem is used for correction some distorted image
point. The localization method just employs the picture position of two objects to estimate the space position and
orientation for AGV. After target recognition, vision tracking, rectification, and object positioning functions actualized
on the embedded omni-directional vision tracker, autonomous navigation has been demonstrated on experimental AGV.
Target recognition and tracking method for one-off aerial imaging system
Shu-le Ge,
Ting-fa Xu,
Guo-qiang Ni,
et al.
Show abstract
One-off aerial imaging system could work in rugged environments too hazardous for piloted or autonomous aircraft to
cruise through, thus it has abroad potential utility in disaster assistance, exploration and some other fields. In this paper
we develop a method for target recognition and tracking in image sequences from this kind of system. Target recognition
is performed with local features of salient points, which could cope with partially obscured situation. We choose SIFT
descriptor that is able to perform effective recognition with incoherent frames of one-off imaging system. Matching
vector in high dimension is accelerated by our categorization scheme with a decision tree. In our method, no target
detection operation is needed as targets are recognized directly by matching salient points. When initialization for
tracking is finished, fast tracking methods start to record their position in following frames. The recognition algorithm is
performed periodically to cope with entering and leaving problem and provide context for tracking method. Test with
simulated image sequence demonstrates that our method can recognize and track predefined objects effectively in image
sequences from one-off imaging system.
A novel compact parallel optical correlator
Show abstract
We present a compact parallel optical correlator (CPOC) based on joint transform correlation manipulation that can be used for various applications. Two innovative approaches are adopted to make
the whole system more compact with a volume of 17×5×3cm3. Two demo systems, i.e., guidance of guided missile and on-orbit rendezvous and docking have been established. The experimental result showed that CPOC system currently operates at approximately 300 Hz and can accurately determine the positioning information when a 1024×768 pixel spatial light modulator (SLM) is employed.
Fast target recognition and tracking in optical correlator using genetic algorithm
Show abstract
Optical correlation manipulation presents great potential in future machine vision systems, which
can be used in a large variety of application fields. Nevertheless, how to extract the tracking signal
effectively and fast for the feedback system is still an open question. In the present paper, we present
novel target recognition and tracking approach in optical correlator system. The tracking signal is
extracted by using genetic algorithm. In this way, a photoelectric detector with short response time
can be employed in the system instead of digital camera, the tracking signal can be extracted by
iteratively evaluating the signal detected by the photoelectric detector until converge. Numerical
simulations were performed to validate the validity of the approach presented in the manuscript.
Online measurement for geometrical parameters of locomotive wheel set
Show abstract
Locomotive is the most important parts of a train. Wheel set is the major running components of a locomotive. Wheel set
tread is the contacting part with the rail and tread will be worn down gradually. The wearing degree of the wheel set
tread is one of the main factors that influence the safety and stability of running train. The measurement of wheel set
wear is usually static and by handwork, which limits the accuracy and reliability. An automatic measurement method for
geometrical parameters of locomotive wheel set based on optoelectronic technique was proposed. Geometrical
parameters include flange thickness, flange height and rim inside distance. Linear structured laser light was projected on
the wheel tread surface. The geometrical parameters can be deduced from the profile image. An online image acquisition
system was designed based on asynchronous reset of CCD. Precision hardware time-delay and asynchronous reset pulse
generation circuits were designed. The entire time sequence of asynchronous reset was researched. Images were acquired
only when wheel sets moved into the designed position. The image acquisition was fulfilled by hardware interrupt mode.
The measuring system was installed along the straight railway section. When the locomotive was running in a limited
speed, the devices placed alone railway line can measure the geometrical parameters automatically.
Micro-structure imaging and measurement based on digital holographic system
Show abstract
Based on the phase reconstruction method which is combined digital holography (DH) with digital lateral shearing
interferometry (LSI) and the idea of the automatic aberration compensation (AAC), a method to eliminate the linear term
of the sheared phase map is proposed. The procedures for reconstruction phase image using the LSI-DH method have
been analyzed in detail, which indicates that one-dimensional phase unwrapping must be applied to the original phase
map before lateral shearing. Moreover, the condition to obtain correct phase image is that the phase difference between
the two adjacent pixels of the shearogram map is less than 2π. One of the prominent advantages of this method is that
when the small amount of shearing is applied the two-dimensional phase unwrapping can be avoided. The computer
simulations, which are based on the lensless Fourier transform holographic system, demonstrate the validity of the
proposed method. It also indicates that the method has large tolerance to noise.
Autofocus method for amplitude and pure phase objects reconstruction in digital holography
Show abstract
Four autofocusing evaluation functions, variance, Fourier spectrum, Gradient Square and the modified standard deviation
correlation, are analyzed and demonstrated based on Fresnel transform algorithm for amplitude and pure phase objects.
The results show that for amplitude objects variance function, weighted Fourier spectral function and standard deviation
correlation function, which get the maximum at the focus plane, are the optimal autofocusing evaluation ones in digital
micro-holography. Moreover, Fourier spectral evaluation function is the preferred function. For pure phase objects the
weighted Fourier spectral function, the gradient square function and the modified standard deviation correlation function,
which get the minimum value at the focus plane, are available in digital holographic autofocusing. The focusing
calculating time with the gradient square function is more longer than others.
Study on the depth of focus of lensless Fourier transform digital holographic imaging system
Show abstract
Based on the principle of digital holography and the criterion of the depth of focus of coherent lens imaging system, the
depth of focus (DOF) of off-axis lensless Fourier transform digital holographic system is analyzed. The DOF expressions
of digital holographic systems with symmetrical and unsymmetrical offset of reference point sources are deduced and
presented. By using a minute vertical bar object the corresponding simulations are performed. The recorded holograms
are reconstructed digitally at the focal plane and its vicinity. The results verified the correctness of the analysis and show
that the DOF of digital holographic imaging system is not only dependent on the optical wavelength and the recording
numerical aperture but also strongly on the optical setup, which indicates that the DOF of digital holographic image
system isn't the same as that of the coherent lens imaging system.
Digital holographic imaging technology
Show abstract
The recording and numerical reconstruction of digital lensless Fourier hologram is analyzed. Based on the quadratic
phase mask of the reconstructed image field the two common methods for phase reconstruction are discussed and
compared by computer simulations. The results show that the automatic phase compensation approach has stronger
toleration for noise than that of the digital holography combined with the digital lateral shearing interferometry approach.
But the latter method can avoid the irksome two-dimensional phase unwrapping process. In addition, the standard onedimensional
phase unwrapping procedure must be applied before shearing procedure in order to eliminate the phase
jump of the original phase map. An effective method to eliminate the linear term of the shearogram automatically have
been proposed and demonstrated.
Shape measurement of microscopic object by using digital lateral shearing holography
Show abstract
A phase reconstruction method which is combined digital holography with digital lateral shearing interferometry has
been analyzed and verified by computer simulation. The results show that one-dimensional phase unwrapping procedure
must be applied to the original reconstructed phase field before lateral shearing. An effective method to eliminate the
linear term of the shearogram automatically have been proposed and demonstrated. Moreover, the simulation results also
show that the maximum of the difference between two adjacent pixels of the shearogram must be less than 2π in order to
reconstruct phase map correctly. If the value is more than 2π, an adoptable method is padding zeros to the recorded
hologram appropriately.
Lifting wavelet method of target detection
Show abstract
Image target recognition plays a very important role in the areas of scientific exploration, aeronautics and
space-to-ground observation, photography and topographic mapping. Complex environment of the image noise, fuzzy,
all kinds of interference has always been to affect the stability of recognition algorithm. In this paper, the existence of
target detection in real-time, accuracy problems, as well as anti-interference ability, using lifting wavelet image target
detection methods. First of all, the use of histogram equalization, the goal difference method to obtain the region, on the
basis of adaptive threshold and mathematical morphology operations to deal with the elimination of the background error.
Secondly, the use of multi-channel wavelet filter wavelet transform of the original image de-noising and enhancement, to
overcome the general algorithm of the noise caused by the sensitive issue of reducing the rate of miscarriage of justice
will be the multi-resolution characteristics of wavelet and promotion of the framework can be designed directly in the
benefits of space-time region used in target detection, feature extraction of targets. The experimental results show that
the design of lifting wavelet has solved the movement of the target due to the complexity of the context of the difficulties
caused by testing, which can effectively suppress noise, and improve the efficiency and speed of detection.
Image denoising method based on 3D transform and local noise analysis
ZhiHui Li,
XiaoJing Xu,
HaiZhen Liu,
et al.
Show abstract
This paper proposes a novel image denoising method based on filtering in 3D transform domain. At first, we apply
block-DCT to the image point-by-point, then, all the transform data block are collected to build a 3D array. All the
coefficients from the same band in this 3D array can be an 1D-array. We apply global 1D-DCT to those 1D-arrays, and
get local de-noise threshold by local noise analysis and local structure analysis. After threshold process, we can get result
by inverse 1D-DCT and inversing block-DCT corresponding to forward transform. The experimental results show that
the proposed technique deliver state-of-art denoising performance, both in terms of objective criteria and visual quality.
Estimation method of CCD and CMOS response functions based on a single image
Su Qiu,
Weiqi Jin
Show abstract
Response function is one of the important characteristics of CCD and CMOS. Traditional measure method of response function need to fulfill strict measurement conditions, and has the shortcoming of
low measurement efficiency and high measurement cost. This paper use PCA method to compute 5 eigenvectors corresponding to 5 maximum eigenvalues of the response function space base on 201
existing response functions. Express the existing response functions by the linear combination of the 5 eigenvectors. Combine all the linear coefficient in a probability distribution in the form of Gaussian Mixture Model (GMM). Estimate the parameter of this Gaussian Mixture Model. Base on this GMM, process the color edge pixels clusters in the image, estimate the RGB response functions. This method has high precision and wide applying prospect.
Design of a novel micro-detection system based on image dynamic range enhancement
Show abstract
In the microscopic image detection, the image quality is the key factors for accuracy of measurement. Owing to the
different material and uneven surface of work pieces, different regions of the work piece have a different response to
light intensity in imaging. In order to obtain complete information of detection of objects, it is necessary to expand the
dynamic range of microscopic images. A new design of micro-detection system based on high dynamic range imaging
technology is proposed. The hardware of our system mainly includes computer system, light intensity control system and
micro-imaging system equipped with a digital CCD image sensor. The effective synthesis algorithm for high dynamic
range image based the quantity of information of image is also proposed. In our algorithm, the images are divided into to
many layers levels by the laplacian pyramid algorithm and the weight of each layer is determined according to the grads
and entropy of image, then the laplacian pyramid inverse algorithm is applied to obtain a high dynamic range image. The
detail features in both areas with high and low illumination of the view can be increased. The experiments have shown
that the entropy and definition of composite image is better than each images of different exposure. Therefore, the
method we propose can effectively improve the resolution of the microscopic images and the accuracy of detection.
MCP gated x-ray framing camera
Houzhi Cai,
Jinyuan Liu,
Lihong Niu,
et al.
Show abstract
A four-frame gated microchannel plate (MCP) camera is described in this article. Each frame photocathode coated with gold on the MCP is part of a transmission line with open circuit end driven by the gating electrical pulse. The gating pulse is 230 ps in width and 2.5 kV in amplitude. The camera is tested by illuminating its photocathode with ultraviolet laser pulses, 266 nm in wavelength, which shows exposure time as short as 80 ps.
Variable-step constant statistics algorithm for removing residual fixed pattern noise of infrared images as second non-uniformity correction
Wei Zhang,
Hong-Bin Nie,
Qing-Yu Hou,
et al.
Show abstract
Regarding the appearance of fixed pattern noise (FPN) in the IR images of an IR observation system introduced by errors
in assembly, environment, etc. Non-Uniformity Correction (NUC) is an important technique for IRFPA. Because the real
radiation response of pixels in the given dynamic range is nonlinear and the existence of 1/f noise, especially the high
temperature scaling point changes the thermal balance of the IR observation system, using the traditional linear
approximate method (temperature scaling method) is hard to obtain the perfect corrective images. On the other hand,
because of Scene-Based Non-Uniformity Correction (SBNUC) does not rely on specialized hardware, SBNUC is very
attractive alternative to radiometric calibration for infrared sensors, thereinto, Constant Statistics (CS) is the best known
approach, but it lies on the scene content and has intimate correlation with the sample quantity. So, in this paper, we
present a novel approach which inherits the rapidity of temperature scaling method and also consider the astringency of
CS, using variable-step constant statistics (VSCS) as second non-uniformity correction in the spatial and time domains of
infrared images to eliminate the residual fixed pattern noise which resulted from the theoretical and methodological
errors of temperature scaling method. The experimental result for the real infrared images data is a solution which
effectively eliminates the residual fixed pattern noise, and at the same time, it proved the effectiveness of this algorithm.
Study on laser backscattering of cone-shaped objects
Chunping Yang,
Meilin Kang,
Chunyan Zhen,
et al.
Show abstract
With the physical optics method, the laser backscattering property of a cone-shaped object is introduced in this paper.
First, the far field scattering amplitude of a rough surface illuminated by a plane light wave is deduced with the
Kirchhoff approximation. The stationary phase method is employed to deal with some complicated integral factor.
Finally, analytical form of a laser scattering intensity is derived for a cone-shaped object. To verify this model, the far
backscattering intensity distribution is calculated and analyzed with its conic angle and surface roughness. These results
show that the far backscattering intensity of cone-shaped objects is related with conic angle, its surface roughness and
light wavelength.
A wavelet and least square filter based spatial-spectral denoising approach of hyperspectral imagery
Show abstract
Noise reduction is a crucial step in hyperspectral imagery pre-processing. Based on sensor characteristics, the noise of
hyperspectral imagery represents in both spatial and spectral domain. However, most prevailing denosing techniques
process the imagery in only one specific domain, which have not utilized multi-domain nature of hyperspectral imagery.
In this paper, a new spatial-spectral noise reduction algorithm is proposed, which is based on wavelet analysis and least
squares filtering techniques. First, in the spatial domain, a new stationary wavelet shrinking algorithm with improved
threshold function is utilized to adjust the noise level band-by-band. This new algorithm uses BayesShrink for threshold
estimation, and amends the traditional soft-threshold function by adding shape tuning parameters. Comparing with soft
or hard threshold function, the improved one, which is first-order derivable and has a smooth transitional region between
noise and signal, could save more details of image edge and weaken Pseudo-Gibbs. Then, in the spectral domain, cubic
Savitzky-Golay filter based on least squares method is used to remove spectral noise and artificial noise that may have
been introduced in during the spatial denoising. Appropriately selecting the filter window width according to prior
knowledge, this algorithm has effective performance in smoothing the spectral curve.
The performance of the new algorithm is experimented on a set of Hyperion imageries acquired in 2007. The result
shows that the new spatial-spectral denoising algorithm provides more significant signal-to-noise-ratio improvement than
traditional spatial or spectral method, while saves the local spectral absorption features better.
Research on microscopy automatic focusing based on gradient function
Show abstract
In the digital imaging system, how to achieve automatic focusing rapidly and efficiently is one of key researches on
imaging system. The stepping motor subdivision technique is introduced into automatic focusing system in this paper,
and quick large range focusing strategy with combination of coarse tuning and trimming is presented, applying focusing
system to high power objective lens of more minimal depth of field, further enlarging the application range of automatic
focusing. Finally after experimental verification, the requirements of large range, high precision, quick focusing are met
in the system which has wide application.
A universal denoising algorithm with trilateral filter and impulse detector
Ying-hui Liu,
Kun Gao,
Guo-qiang Ni,
et al.
Show abstract
In this paper, we introduce a new edge-preserving nonlinear filter for removing the mix of Gaussian and impulse noise.
Built from Prasun Choudhury and Jack Tumblin's trilateral filter, the new algorithm incorporate a local gradient statistic
for detecting corrupted pixels in images with impulse noise and a switching mechanism for smoothing the gradients of
impulse noise samples and the gradients of impulse noise-free samples with different parameters. If the central pixel is
impulse-like and has a high statistical value in gradient domain, the impulsive component of the weight in gradient
bilateral filter is more heavily to suppress large impulses. It smoothes image toward a sharply-bounded, gradient
piecewise-linear approximation which provides stronger noise reduction and better edge-limited smoothing behavior.
Compared to most other spatial domain nonlinear filters, the proposed algorithm consistently yields good results in
removing the mix of Gaussian and impulse noise and more notable edge-limited smoothing behavior. Like the trilateral
filter, the proposed algorithm easily extends to N-dimensional signals.
A new algorithm for image denoising
Show abstract
The differences between the wavelet denoising method and traditional denoising method have been analyzed in this
paper. Wavelet coefficients of the neighborhood characteristics have been analyzed on the basis of original global
threshold value. A new algorithm based on the characteristics of the neighborhood has been proposed by establishing
windows for each coefficient. It is found that the new algorithm is more effective than traditional algorithms at the aspect
of improving the quality of the denoised images.
Image denoising algorithm based on the relevance of wavelet coefficients
Show abstract
In this paper the wavelet transform and the wavelet transform theory have been elaborated. A wavelet coefficients'
model in which individual shrinkage thresholds are selected for each coefficient has been established. The relationships
between different wavelet coefficients are analyzed. According to this new model, a new algorithm of image denosing
has been proposed. It is shown in the experiments that the new algorithm is more effective than traditional algorithms.
High-resolution linear CCD application in the recognition of cuttings' lithology category
Bing Mu,
Wen-dong Li,
Kai Cheng,
et al.
Show abstract
It is difficult in the description of rock cuttings' lithology category with a traditional manual method since PDC bits have
been widely used in the process of drilling. To recognize the tiny cuttings' lithology category effectively, a digital
compact system based on high-resolution color linear CCD and microphotography is established and introduced in this
paper. The system is mainly constructed by high resolution color linear CCD, micro lens, white LED light sources and
conveyor. The Color linear CCD with up to 4080 pixels per line combined with micro lens leads to a resultant
reproduction ratio of 1:1. It ensures that tiny cuttings' images can be captured clearly, and hence the detailed information
of cuttings' color and texture can be obtained. With the linear scanning, we can get images with both high-resolution
and larger field of view than array CCD. Optimal parameter configuration is determined by the experiments of hardware
control program and optical parameter adjustment. The software to recognize cuttings' lithology category is developed
by extracting multi-character values from the image information and by using the LIBSVM classifier. With the optimal
parameter configuration and the developed recognition software, the cuttings' lithology category has been well described
with a average recognition rate about 95% for mudstones and about 90% for sandstones, which proves that the cuttings'
images captured by the high-resolution linear CCD meet the requirement of the subsequent recognition algorithm. The
obtained results show that the digital compact system is greatly potential to be used in oil field logging.
Performance estimation of optical system simulation based on MTF
Show abstract
Optical imaging system performance estimation is crucial in most related practical engineering applications.
The most prevailing tool for performance estimation is MTF, and knife-edge technique is the most popular method to
compute MTF. However, traditional knife-edge technique, which was designed for film cameras and other devices that
form continuous images, does not consider about the problems of sampling. Focusing on digital image application, this
paper proposes a simple method to compute MTF for the purpose of performance estimation of optic simulation system
base on traditional knife-edge technique. The method extends traditional knife-edge technique to discuss the problems of
sampling, coordinate change and increasing frequency density in digital simulation system. The experiment shows that,
in the noise controllable simulation system, digital knife-edge technique could compute MTF rapidly and correctly, and
estimate the system performance quantitatively and efficiently.
Research and application of edge detection technology in costume pattern cutting
Show abstract
Research and application of edge detection technology in costume pattern cutting. I made a software with Visual C++6.0
to realize contour extraction of costume pattern. It also contains gray balance, Laplace edge detection, image's erosion
and so on. Besides, I deeply compared applicability of each algorithm and all combinations of them. In this experiment, I
succeed in extracting contour of complex costume patterns. Furthermore, controlling laser-cutting machine cut real
costume patterns by software. In conclude that I not only have realized automatic extracting contour and cutting by this
software in several minutes but also improved the precision of contour extraction.
Degraded parameter estimation using quantum neural network
Show abstract
In this paper, an approach based on the quantum neural network is investigated to guide the process of selecting an
optimal estimation of Gaussian degraded parameter. In fact, we first formulate the nonlinear problem by maximum
likelihood estimation. Then we modify and apply the quantum neural network algorithm, which combines the advantages
of both quantum computing and neural computing, to solve the optimal estimation problem. The new algorithm does not
suffer from the morass of selecting good initial values and being stuck into local optimum as usually accompanied with
the conventional techniques. The simulation results indicate the soundness of the new method.
Reconstructing photorealistic 3D models from image sequence using domain decomposition method
Show abstract
In the fields of industrial design, artistic design and heritage conservation, physical objects are usually digitalized by
reverse engineering through some 3D scanning methods. Structured light and photogrammetry are two main methods
to acquire 3D information, and both are expensive. Even if these expensive instruments are used, photorealistic 3D
models are seldom available. In this paper, a new method to reconstruction photorealistic 3D models using a single
camera is proposed. A square plate glued with coded marks is used to place the objects, and a sequence of about 20
images is taken. From the coded marks, the images are calibrated, and a snake algorithm is used to segment object from
the background. A rough 3d model is obtained using shape from silhouettes algorithm. The silhouettes are decomposed
into a combination of convex curves, which are used to partition the rough 3d model into some convex mesh patches.
For each patch, the multi-view photo consistency constraints and smooth regulations are expressed as a finite element
formulation, which can be resolved locally, and the information can be exchanged along the patches boundaries. The
rough model is deformed into a fine 3d model through such a domain decomposition finite element method. The textures
are assigned to each element mesh, and a photorealistic 3D model is got finally. A toy pig is used to verify the
algorithm, and the result is exciting.
A semi-automatic 3D laser scan system design
Show abstract
Digital 3D models are now used everywhere, from traditional fields of industrial design, artistic design, to heritage
conservation. Although laser scan is very useful to get densely samples of the objects, nowadays, such an instrument
is expensive and always need to be connected to a computer with stable power supply, which prevent it from usage for
fieldworks. In this paper, a new semi-automatic 3D laser scan method is proposed using two line laser sources. The
planes projected from the laser sources are orthogonal, one of which is fixed relative to the camera, and the other can be
rotated along a settled axis. Before scanning, the system must be calibrated, from which the parameters of the camera,
the position of the fixed laser plane and the settled axis are introduced. In scanning process, the fixed laser plane and the
camera form a conventional structured light system, and the 3d positions of the intersection curves of the fixed laser
plane with the object can be computed. The other laser plane is rotated manually or mechanically, and its position can be
determined from the cross point intersecting with the fixed laser plane on the object, so the coordinates of sweeping
points can be obtained. The new system can be used without a computer (The data can be processed later), which make it
suitable for fieldworks. A scanning case is given in the end.
Simulation on effects of atmospheric transmission on remote sensing images
Show abstract
In order to study the effects of atmospheric transmission on remote sensing images, a simulated system on effects of
atmospheric absorption and scattering on remote sensing images is proposed in space field. The atmosphere model is
build based on MODTRAN. Due to the large effects from neighbor pixels in high resolution images, atmospheric cross
radiation is considered in the mode, and a FFT-based computation method for the adjacent effect is proposed, which is
practical in digital image process. The simulation experiment results show that the quantization measurement is made
for atmospheric influence, and images from the ground to the entrance pupil of satellite camera are rebuilt effectively.
Imaging system design and image interpolation based on CMOS image sensor
Show abstract
An image acquisition system is introduced, which consists of a color CMOS image sensor (OV9620), SRAM
(CY62148), CPLD (EPM7128AE) and DSP (TMS320VC5509A). The CPLD implements the logic and timing control to
the system. SRAM stores the image data, and DSP controls the image acquisition system through the SCCB (Omni
Vision Serial Camera Control Bus). The timing sequence of the CMOS image sensor OV9620 is analyzed. The imaging
part and the high speed image data memory unit are designed. The hardware and software design of the image
acquisition and processing system is given. CMOS digital cameras use color filter arrays to sample different spectral
components, such as red, green, and blue. At the location of each pixel only one color sample is taken, and the other
colors must be interpolated from neighboring samples. We use the edge-oriented adaptive interpolation algorithm for the
edge pixels and bilinear interpolation algorithm for the non-edge pixels to improve the visual quality of the interpolated
images. This method can get high processing speed, decrease the computational complexity, and effectively preserve the
image edges.
Image denoising based on wavelet cone of influence analysis
Wei Pang,
Yufeng Li
Show abstract
Donoho et al have proposed a method for denoising by thresholding based on wavelet transform, and indeed, the
application of their method to image denoising has been extremely successful. But this method is based on the
assumption that the type of noise is only additive Gaussian white noise, which is not efficient to impulse noise. In this
paper, a new image denoising algorithm based on wavelet cone of influence (COI) analyzing is proposed, and which can
effectively remove the impulse noise and preserve the image edges via undecimated discrete wavelet transform
(UDWT). Furthermore, combining with the traditional wavelet thresholding denoising method, it can be also used to
restrain more widely type of noise such as Gaussian noise, impulse noise, poisson noise and other mixed noise.
Experiment results illustrate the advantages of this method.
A method to obtain distance resolution information based on streak-tube camera
Longxiao Ma,
Shaokun Han,
Chenning Ma
Show abstract
Three-dimensional imaging based on Streak Tube Camera is a new Laser imaging method. Distance resolution is the
important basis of three-dimensional imaging. In this paper, a method of measuring the distance resolution is proposed
based on the structure principle and internal parameter, and the target is imaged using streak tube camera. In this
foundation, we verify the feasibility of obtaining distance resolution, and supply major foundation to three-dimensional
imaging.
Application of digital pulse delay device in range-gated control for range-gated imaging lidar
Chaokai Yuan,
Siying Chen,
Yinchao Zhang,
et al.
Show abstract
Digital pulse delay device is one of the key techniques of range-gated imaging lidar. At present, Digital
method and analog method are the two main implementations of pulse delay device. Digital method is mainly achieved
by counter or FIFO memory. With the development of Complex Programmable Logic Device (CPLD), the digital delay
device can be achieved with a single chip of CPLD. With this method, the digital delay device enjoys the advantages of
high integration, high reliability and strong ability of anti-electromagnetic interference. However, since the maximum
clock frequency of CPLD is limited, the improvement of temporal resolution is restricted. Analog method is mainly
realized by the delay-line, which is one of the dedicated integrated circuit. Using this method, a higher time resolution
can be arrived. In this paper, the timing characteristics of the delay signal are analyzed. Three design options are
presented and the advantages and deficiencies are discussed. Based on the theoretical analysis and numerical simulation,
the digital delay device combined with the delay chip AD9501 and Field Programmable Gate Array (FPGA) is chosen
because of its large dynamic range and high accuracy. Besides, the output pulse width can be adjusted conveniently. The
digital delay device is simulated and the result shows that the delay control for range-gated imaging lidar is feasible.
Research on width detection of building crack algorithm in embedded system based on image
Show abstract
The width of the crack is an important indicator in evaluation of the quality of the buildings. Hand-held embedded device used for measuring the width of the cracks can reduces the labor intensity of staff, improve the efficiency and reliability of measurement. In width measuring based on embedded image processing, some problems are existed, such as: "isolated island" noise, the edge of burr which reduce the accuracy of calculation result. The algorithm in the paper can solve these problems. By means of sub-block to increase the speed of image processing of the cracks, then searched
the edge of the crack in the blocks which contains crack. this way can reduce image processing time by at least one order of magnitude while getting the preliminary ruling edge of the crack. Additionally, presents a method that use the basic morphological operations and chain to smooth the edge, then obtains the correct edge of the crack. Finally, get the width of the cracks through "Fast hierarchical vector". Experiments confirmed that the correct width of the crack can be got with this algorithm, Accuracy is up to 50μm.
Research on obtaining multiband laser source using spectral beam combining
Yunxia Li,
Qiaoli Li,
Shengbao Zhan,
et al.
Show abstract
In order to obtain the multiband laser source of laser multi-spectrum detection, the method using spectral beam combining in an external cavity to obtain the multi-spectral laser was proposed. The experimental results of spectral beam combining for two fiber lasers using large-mode-area Er3+/Yb3+ co-doped double-cladding fiber as the gain medium in an external cavity are reported. 3dB linewidth of 29.3nm and 11.52nm respectively between two spectra was obtained. Experimental results show that spectral beam combining in an external cavity is a feasible approach for
obtaining the multi-spectral laser sources.
An evaluation of stereo feature matching with enhanced Hough transform
Tiecheng Fan,
Shuanghe Yu
Show abstract
This paper presents an evaluation result of stereo feature matching algorithm using a kind of enhanced Hough transform.
A planar circle pattern is utilized to collect template data by a binocular stereo camera. In the feature extraction process,
the grayscale and binarization of the left and right images are firstly performed, and then the canny algorithm is adopted
to detect the edges of the target. Finally, the feature matching process is performed by extracting the image coordinates
of target points in the left and right images by using enhanced Hough transform to look up similar data in the template.
The matching results of planar circle pattern are experimentally validated with epipolar constraints.
Augmented reality based surgery navigation system
Show abstract
A surgery navigation system based on augmented reality is presented. The system is based on 3D visualization and 3D
registration techniques with an infrared tracking device and a 3D scanner. After reconstructing the 3D model of the
patient's organs and scanning the surface of the patient's face, the system uses Iterative Closest Point (ICP) algorithm to
calculate the transformation between the 3D model of patient and the three-dimensional scanner. During the surgery
navigation, 3D model can be overlaid onto the image of the real patient. The proposed system doesn't require the
attachment of markers because of the adoption of 3D scanner. Experimental result shows that the tracking accuracy of
the system is appropriate for the requirements of actual surgery and can bring down the risk of endoscopic surgery.
Novel interactive virtual showcase based on 3D multitouch technology
Show abstract
A new interactive virtual showcase is proposed in this paper. With the help of virtual reality technology, the user of the
proposed system can watch the virtual objects floating in the air from all four sides and interact with the virtual objects
by touching the four surfaces of the virtual showcase. Unlike traditional multitouch system, this system cannot only
realize multi-touch on a plane to implement 2D translation, 2D scaling, and 2D rotation of the objects; it can also realize
the 3D interaction of the virtual objects by recognizing and analyzing the multi-touch that can be simultaneously
captured from the four planes. Experimental results show the potential of the proposed system to be applied in the
exhibition of historical relics and other precious goods.
Infrared marker-based tracking in an indoor unknown environment for augmented reality applications
Yetao Huang,
Dongdong Weng,
Yue Liu,
et al.
Show abstract
Marker based tracking requires complicated preparation work that impedes its use in augmented reality applications.
This paper presents a novel tracking scheme to be used in an indoor unknown scene by adapting simultaneous
localization and mapping (SLAM) algorithms. An infrared (IR) marker system is specifically designed to simplify the
feature recognition and tracking in SLAM process. With one initial IR marker, the other markers can be projected
randomly onto a large-area environment. The pose of camera can be estimated with a monocular IR camera in real time.
Experimental result demonstrates that the proposed system meets the requirements of accuracy for large-area tracking. A
prototype system is built to show its feasibility in unknown environment and potential use in applications.
Camera calibration method with a coplanar target and three-dimensional reconstruction
Kai Nie,
Wenyao Liu,
Jinjiang Wang
Show abstract
This essay proposes a method for camera calibration. This method calibrates camera's all intrinsic parameters and
extrinsic parameters step-by-step under the condition of considering the first order radial distortion. Compared with
Tsai's two-stage method, this method only needs to use a coplanar target and can calibrate the image center and the
scale factor. It also introduces a new distortion model and then can completely calibrate camera linearly, avoiding the
trouble and the instability of non-linear optimization. It's simpler. Then, this essay introduces a computing model of
three-dimensional reconstruction of the two cameras' system, and proves that this calibration method has high accuracy
after the three-dimensional reconstruction.
A new method for the extraction of region-of-interest based on visual attention
Xiaoguang Shao,
Kun Gao,
Guoqiang Ni
Show abstract
Region of interest (ROI) plays an important role in image analysis, yet traditional ROI extraction result may be different
from human's subject feeling. In this paper, a new approach based on visual attention is proposed considering that
primates have remarkable abilities to interpret complex scenes real-time. This approach uses visual attention model to
locate salient points. As a result, the most salient region is selected and marked as ROI. The proposed work is based on
bottom-up influences of attention and, therefore, purely unsupervised. Furthermore, experiment results of the ROI
extraction mentioned in the paper are given and show perfect performance when comparing with other traditional
approaches.
Common hyperspectral image database design
Show abstract
This paper is to introduce Common hyperspectral image database with a demand-oriented Database design method
(CHIDB), which comprehensively set ground-based spectra, standardized hyperspectral cube, spectral analysis together
to meet some applications. The paper presents an integrated approach to retrieving spectral and spatial patterns from
remotely sensed imagery using state-of-the-art data mining and advanced database technologies, some data mining ideas
and functions were associated into CHIDB to make it more suitable to serve in agriculture, geological and environmental
areas. A broad range of data from multiple regions of the electromagnetic spectrum is supported, including ultraviolet,
visible, near-infrared, thermal infrared, and fluorescence. CHIDB is based on dotnet framework and designed by MVC
architecture including five main functional modules: Data importer/exporter, Image/spectrum Viewer, Data Processor,
Parameter Extractor, and On-line Analyzer. The original data were all stored in SQL server2008 for efficient search,
query and update, and some advance Spectral image data Processing technology are used such as Parallel processing in
C#; Finally an application case is presented in agricultural disease detecting area.
System analysis and simulation of airborne scannerless 3D imaging lidar
Pan Guo,
Qiwei Hao,
Siying Chen
Show abstract
Airborne non-scanned 3D imaging lidar is a recently developed method for remote sensing. The design
method and flow of the system parameters round with the spatial resolution are established and explained in
detail with examples. An evaluation indicator of data coverage is proposed to optimize the imaging control
method. Pixel aliasing in all directions are analyzed, the possible factors cause the aliasing are stated,
including the time control error, atmospheric disturbance and platform shake. At last, a parallel data output
format is proposed to eliminate the timing mismatch of image data and POS parameters.
Modeling and tracing of errors caused by coordinate conversions with Dopplerlet transform analysis in airborne lidar system
Shengzhe Chen,
Tian Lan,
Yinchao Zhang,
et al.
Show abstract
The systematic errors caused by coordinate conversions in airborne lidar system under moving conditions belong to one
of the main error sources to measurement accuracy. In order to reduce systematic errors and to attain high observation
accuracy for three-dimensional imaging airborne lidar, the tracing of the systematic errors caused by coordinate
conversions is realized. Besides, there has been substantial focus on both error analysis and development of error model
for the airborne lidar system in this paper. On the basis of review of the coordinate conversion methods adopted in most
airborne lidar systems, the Dopplerlet Transform analysis is applied to the systematic error mainly caused by coordinate
conversions in airborne lidar for correction. And the effects on the observed results are evaluated. Finally, both the
validity of the developed error model and the precision of the analytical method are verified.
Analyzing and modeling on coordinate conversion errors of airborne lidar detection data
Lifang Jiang,
Tian Lan,
Yinchao Zhang,
et al.
Show abstract
Lidar is an important technology for capturing data of three-dimensional surface information in high-resolution airborne
earth observation system. To realize the positioning of airborne lidar targets, the coordinates of the laser footprints have
to be converted from laser scan coordinate system to WGS-84 coordinate system. The existence of coordinate conversion
errors is sure to affect the earth observation accuracy. This paper realized the coordinate conversion from laser scanning
coordinates to WGS-84 coordinates and analyzed the influence of coordinate conversion parameter errors on the
accuracy of the targets' three-dimensional coordinates. These errors include the range error, scanning angle errors,
attitude angle errors, mounting errors, the GPS dynamic positioning errors and so on. Based on the analysis, the
coordinate conversion error model was developed. Monte Carlo Method was applied to simulate the influence of several
parameter errors respectively on positioning accuracy. Finally, the relationship between positioning errors and the
coordinate conversion parameters was simulated when the values of coordinate conversion parameter errors were
constants. These simulation results have reference value on the error compensation of detection data.
Simulation of atmospheric brightness distributions from a visible earth sensor using MODTRAN4
Show abstract
Due to the scattering and reflection of the terrestrial atmosphere, especially when there are clouds or any different
geographical circumstances, the visible earth sensor on the satellite will receive changing atmospheric radiation.
Therefore, the images received by the detector of visible earth sensor will have fuzzy boundaries, thus affecting the
measurement accuracy of geocentric vector. In order to eliminate the impact of the existence of the atmosphere in the
Sun-Earth-Sensor path on earth sensing imaging, the emphasis of our study lies on the issue of atmospheric brightness
distributions. Firstly, we briefly discuss the influence factors of atmospheric brightness at the visible earth sensor, using
the atmospheric radiation transfer model MODTRAN4. Then, by running MODTRAN4 under a specific physical model,
we successfully acquire both graphs and numerical results of the atmospheric brightness distributions in the full field of
view of the visible earth sensor under different circumstances. Based on the research of this paper, given the specific
structure and parameters of the detector of earth sensor, calculating the illumination intensity on the image plane is
possible.
A method of automatic recognition of airport in complex environment from remote sensing image
Show abstract
In this paper, a new method is proposed for airport recognition in complex environments. The algorithm takes all
advantages of essential characteristics of the airport target. Structural characteristics of the airport are used to establish
assumption process. Improved Hough transformation (HT) is used to check out those right straight-lines which stand for
actual position and direction of runways. Morphological processing is used to remove road segments and isolated points.
Finally, we combine these segments carefully to describe the whole airport area, and then our automatic recognition of
airport target is realized.
Fringe image processing based on structured light series
Show abstract
The code analysis of the fringe image is playing a vital role in the data acquisition of structured light systems, which
affects precision, computational speed and reliability of the measurement processing. According to the self-normalizing
characteristic, a fringe image processing method based on structured light is proposed. In this method, a series of
projective patterns is used when detecting the fringe order of the image pixels. The structured light system geometry is
presented, which consist of a white light projector and a digital camera, the former projects sinusoidal fringe patterns
upon the object, and the latter acquires the fringe patterns that are deformed by the object's shape. Then the binary
images with distinct white and black strips can be obtained and the ability to resist image noise is improved greatly. The
proposed method can be implemented easily and applied for profile measurement based on special binary code in a wide
field.