Show all abstracts
View Session
- Front Matter: Volume 8002
- Multispectral Image Acquisition
- Multispectral Image Processing and Analysis
Front Matter: Volume 8002
Front Matter: Volume 8002
Show abstract
This PDF file contains the front matter associated with SPIE Proceedings Volume 8002, including the Title Page, Copyright information, Table of Contents, and Conference Committee listing.
Multispectral Image Acquisition
Building outline detection based on height and intensity information of airborne laser scanning data
Show abstract
Urban environment is extremely complex due to a multitude of features with different heights and structures. Traditional
methods available to extract information regarding the buildings by using optical remote sensing images are highly
labor-intensive and time-consuming. This paper developed a new method to detect building outlines based on height and
intensity information of Airborne LiDAR data. Texture, relative height and intensity characteristics were first extracted
from the LiDAR point cloud. Then, Support Vector Data Description was used to detect buildings with training
knowledge. Finally, building outlines were obtained after data post process including small region removal, raster to
vector conversion and so on. Experiments show the method proposed in this study is reliable and could be widely used in
other urban areas.
Design of wireless power driving and controlling system for electrically driving liquid crystal micro-lens array
Show abstract
A novel wireless power driving and controlling system is proposed, which is based on the electrical driving and
controlling characteristic of liquid crystal micro-lens array with tunable focal length. The system consists of two isolated
circular copper stacked coils with circular cross section. Both of coils are air-cored. The big one, used to generate
electromagnetic field, has the wire diameter of ~1mm, and the coil diameter of ~38mm and the height of 10mm. By
contrast, the small one, used to produce electromotive force by induction effect, is ~1mm, ~13mm and ~18mm,
respectively. The varying electromotive force is generated in the small coil by the variable intensity of magnetic field
resulting from changing the time of current through the big coil. The agreement of experimental results with the
theoretical analysis proves that the electromotive force can be continuously changed from 0 to 30Vrms (root mean
square), when the duration of current through the big coil is reasonablely selected. It is very important that the clear
image can be got and the image quality of liquid crystal micro-lens array can be comparable to that of conventional
optical micro-lens array.
Analysis of effect of HgCdTe passivant on the performance of long-wavelength infrared(LWIR) detectors
Peng-xiao Xu,
Ke-feng Zhang,
Wei Wang,
et al.
Show abstract
The performance of HgCdTe infrared photoconductors is strongly dependent on the semiconductor surface conditions. In
this paper, the effect of fixed charge density(Qox) due to passivation on the responsivity of HgCdTe photoconductive
detectors is analysed both theoretically and experimentally. A profile responsivity model is used here for calculation,
which mainly includes the contribution of minority carrier lifetime and the shunt resistance resulting from the
accumulation layer at the surface. In this model, the profiles of surface majority carrier concentration and surface
mobility are taken into consideration. A gate-controlled photoconductor structure is designed and fabricated to investigate
surface effects on HgCdTe infrared photoconductive detectors. And it is used to evaluate and optimize surface passivation
layers. Minority carrier lifetime, resistance and responsivity of the device have been measured as a function of the gate
potential in this structure. The measured variations have shown a reasonable agreement with our model. It is predicted
that the optimization of surface fixed charges at the MCT-passivant interface can bring a great improvement in the
responsivity of photoconductive detectors.
Research on one manual zoom liquid lens
Show abstract
In liquid zoom lens, the accurate fixed-focus location needs the accurate calculation of focal length. That is to say, the
accurate calculation of surface curvature must be obtained. Therefore, the research of their relation in liquid zoom lens
becomes extremely important which has a directly theoretical instruction for the design of liquid zoom lenses. In our
paper, the relation is studied. A manual liquid lens is reported which has upper and nether components with specificaton
Φ12×5mm. The zoom lens is obtained by the upper component circumvolving down into the nether components.
External swing method for forward motion compensation of airborne digital camera
Show abstract
The mechanism of image motion of Airborne Digital Camera was analyzed and a method of external swing forward
motion compensation was proposed in this paper. The forward motion compensation for external swing was achieved
through the design of a compensation device to control the rotation of the camera system. In order to simulate the actual
motion effect at the most under the laboratory condition, the Stepper Motor Controller (SC) stepper motor controller was
designed as the simulation installment to control the guide rail movement. Using this device, compensation speeds and
compensation could be controlled, and hence the evaluation of compensation effect could be conducted by image
sharpness. The experimental results of the forward motion compensation, show that the compensated remain image
motion was less than 2μm (1/3 pixel), illustrating the feasibility and effectiveness of the method proposed in this
paper.
Range gated ladar imaging data simulation
Show abstract
Range gating is a kind of effective method which can eliminate backscattered light and increase the SNR of the
laser range gated (LRG) imaging system. In the paper, we describe a computer model that was developed to simulate the
imaging data of the LRG system. It is used to help design LRG system, and also acquire LRG imaging data used for
further research on surveillance, city topography, combat identification, and other applications. The experiments are
made in the condition of varied pulse energy, range gate width, atmospheric transmition, then the simulation result is
analyzed and the related conclusion is obtained. It demonstrates that the simulating image data is suit for further study on
many areas.
Look-up table construction based on 6S model
Rui Zhang,
Ting Jiang
Show abstract
Look-up table can accelerate the atmospheric radiation transmission simulation in remote sensing imaging simulation
effectively. Because of easily using and convenient for re-development of 6S, the construction of look-up table based on
6S was discussed in this paper. After the introduction of 6S, the parameters of 6S which should be including in the
look-up table and have the mainly relativity in the processing of atmospheric radiation transmission calculation were
analyzed. Then the construction scheme was designed, and look-up table was generated based on 6S executable program.
And the validity of atmospheric radiation transmission calculation using look-up table instead of 6S model was verified.
The experiment shows that the luminance of target in every geometric conditions which interpolated by look-up table has
nice precision, and the look-up table construction scheme discussed in this paper can meet the needs of anticipation.
Simulation of uneven blurred images impacted by the aero-optical effect of high-speed flow field
Zhi-guo Cao,
Shuai Wang
Show abstract
In this paper, a new approach is proposed to simulate images impacted by the Aero-Optical Effect
(AOE) of High-Speed Flow Field. We first study the mechanism of the Aero-Optical Effect of
High-Speed Flow Field, and its influence on the imaging system of the aircrafts. After reviewing
existing methods of simulating blurred images impacted by the AOE, we show a novel method of
simulating uneven blurred images which makes a great contrast to the traditional algorithms.
Experimental results indicate that the blurred images generated by our methods are much closer to
authentic situations than the previous results.
A comparison of three retrieval methods with single Doppler radar data
Shenghui Zhou,
Ming Wei,
Li Gao,
et al.
Show abstract
In single Doppler radar wind retrieval technique there are three methods proposed in recently years, which is VAP, VPP
and SVVP method. Three methods were employed to retrieve wind field with the purpose to find out which one is more
suitable in practice according their accuracy by using synthetic data and real data.
When tests applied with uniform data, the orders of magnitude for relative error of radial velocity is 10-13 for SVVP,
much lower than that of VAP and VPP which both are 10-4. Furthermore, the results under condition of wind direction
unchanged and wind velocity varied linearity with altitude are 10-8, 10-3, 10-3 for SVVP, VPP and VAP, respectively. In
real wind field of typhoon "Pearl", the authenticity of surrounding wind and wind shear retrieved by SVVP and VPP is
better than VAP, the deviation of simulation would be increased at some points missing data. Relative errors of radial
velocity achieved by VAP, VPP and SVVP are 7.02, 5.78 and 3.1 respectively. It indicated that SVVP method's
performance is better than that of other two and suitable in practical applications.
Probability density function and the speckle number on receive aperture
Show abstract
We develop an analytical expression for probability density function of speckle intensity on receiving aperture and
express integral speckle intensity variance when receive aperture area is s. Analytical and numerical results of speckle
number are given for rectangular and circular diffuser apertures. Laser radar intensity image are simulated in different
speckle numbers on receive aperture. The results shows that the image is becoming more and more clearly as speckle
numbers increase by the increasing of receiving aperture and decrease of the light beam waist.
Infrared radiant spectrum calculation of the gas in the atmosphere based on physical model
Show abstract
This paper proposes a theoretical method based on physical model to calculate infrared (IR) radiation spectrum. IR
physics establishes that emission abilities of objects are measured by emissivity, while the emissivity of objects is based
on the emissivity of a blackbody. Planck formula shows how to calculate the spectral energy density of a blackbody, and
IR radiation spectrum of a sample can be easily obtained if its emissivity is known. Kirchhoff's law shows the emissivity
of an object is equal to its absorptance at the same conditions. By calculating the absorption coefficient using line-by-line
model, the absorptance can be achieved through the relationship between them. The performance data of the technique is
from public databases.
Parameters selection study of high-operating temperature MCT photoconductor detectors
Wei Wang,
Jingtong Xu,
Peilu Jiang,
et al.
Show abstract
This paper concerned HgCdTe (MCT) infrared photoconductor detectors with high operating
temperature. The near room temperature operations of detectors had advantages of light weight, less cost
and convenient usage. The selection of material could greatly reduce cost caused by low performance
detectors, with usage of unsuitable wafers. Many characteristic parameters were tested before the
production period, material were chosen to produce the high yield detectors. Using these parameters for
model construction and hypothesis testing, we could efficiently select the material most suitable for high
performance detectors. Using the methods of statistical inferences, the product's fabrication management
and maintenance abilities could also be greatly improved. In this paper, data of detectors were collected
from detectors used in railway observation. The MCT detector was photoconductive with an 800μm×
800μm active area, which responses to radiation from 3μm to 5μm, the detector material was grown by
the Bridgeman technique. The hypothesis test was a measure of how close the performance of selected
fiber was to the empirical distribution, the Kolmogrov-Smirnov test, Anderson-Darling test and
Chi-square goodness-of-fit test were used in this paper. The selection of material had a statistical
credibility if the null hypothesis was not rejected. The reliability of detectors could also be improved with
the selection of detector parameters.
Controlling system for smart hyper-spectral imaging array based on liquid-crystal Fabry-Perot device
Show abstract
A research for developing a kind of smart spectral imaging detection technique based on the electrically tunable liquidcrystal
(LC) FP structure is launched. It has some advantages of low cost, highly compact integration, perfuming
wavelength selection without moving any micro-mirror of FP device, and the higher reliability and stability. The
controlling system for hyper-spectral imaging array based on LC-FP device includes mainly a MSP430F5438 as its core.
Considering the characteristics of LC-FP device, the controlling system can provide a driving signal of 1-10 kHz and 0-
30Vrms for the device in a static driving mode. This paper introduces the hardware designing of the control system in
detail. It presents an overall hardware solutions including: (1) the MSP430 controlling circuit, and (2) the operational
amplifier circuit, and (3) the power supply circuit, and (4) the AD conversion circuit. The techniques for the realization
of special high speed digital circuits, which is necessary for the PCB employed, is also discussed.
Design and fabrication of sub-wavelength metal polarization gratings used in polarization imaging
Show abstract
This paper designs a sub-wavelength metal polarization gratings array composing two orthogonal micro polarization
gratings as one unit. Effects of different metal materials and grating profile on the TM and TE polarization transmittance
as well as the extinction ratio are analyzed by the Finite Difference Time Domain (FDTD) method. Based on the
requirement of the visible light polarization imaging and the resolution of the holographic lithography, we obtain the best
parameters for the grating: period is 250nm, Al thickness is 260nm, and duty cycle is 0.4, the transmittance is higher than
45% and extinction ratio is higher than 100. Then, the sub-wavelength metal polarization gratings have been fabricated
by two methods: the holographic - ion beam etching - Al oblique deposition or the holographic - reactive ion etch (RIE) -
ion beam etching. Preliminary results indicate the polarization information has been obtained. A prototype metal
polarization gratings array will be fabricated in late 2011.
Three dimensional modeling for ship image based on MultiGen
Show abstract
To build a suit of integrated ship model database, a new method of constructing ship three-dimensional models is
proposed which partition ship into basic structure and special structure of ship body. According to the distributing of ship
structure, segmenting the whole ship area, and then marking off the basic hierarchies of practice ship, finally
disassembling to basic unit structure, thus setting up corresponding tree hierarchies of ships. The modeling procedures of
build ship 3D model are presented, Simulation results of three-dimensional ship models have gained vivid solid effect.
Multispectral Image Processing and Analysis
Super resolution of remote sensing image based on structure similarity in CS frame
Show abstract
In this paper, a novel super resolution (SR) method for remote sensing images based on compressive sensing (CS),
structure similarity and dictionary learning is proposed. The basic idea is to find a dictionary which can represent the
high resolution (HR) image patches in a sparse way. The extra information coming from the similar structures which
often exist in remote sensing images can be learned into the dictionary, so we can get the reconstructed HR image
through the dictionary in the CS frame due to the redundance in the image which has a sparse form in the dictionary. We
use K-SVD algorithm to find the dictionary and OMP method to reveal the sparse coding coefficient's location and value.
The difference between our method and the previous sample-based SR method is that we only use low-resolution image
and the interpolation image from itself rather than other HR images. Experiments on both optical and laser remote
sensing images show that our method is better than the original CS-based method in terms of not only the effect but also
the running time.
Rapid assessment of large scale vegetation change based on multi-temporal phenological analysis
Show abstract
Detecting vegetation change is critical for earth system and sustainability science. The existing methods, however, show
several limitations, including inevitable selection of imagery acquisition dates, affection from vegetation related noise on
temporal trajectory analysis, and assumptions due to vegetation classification model. This paper presents a multitemporal
phenological frequency analysis over a relatively short period (MTPFA-SP) methodology to detect vegetation
changes. This MTPFA-SP methodology bases on the amplitude components of fast Fourier transforming (FFT) and is
implemented with two steps. First, NDVI time series over two periods are transformed with FFT into frequency domain,
separately. Second, amplitude components with phenological information from Step 1 are selected for further change
comparison. In this methodology, component selection shows physical meanings of natural vegetation process in
frequency domain. Comparisons among those selected components help enhance the ability to rapidly detect vegetation
changes. To validate this MTPFA-SP methodology, we detect changes between two periods (2001-2005 and 2006-2010)
in the eastern Tibet Plateau area and make two kinds of assessments. The first is for a larger scale, including statistic
analysis of altitudinal zonality and latitudinal zonality. The second assessment is for rapid detection of vegetation change
location. Landsat TM image were employed to validate the result.
Design and fabrication of diffraction imaging elements for common Gaussian laser beam in terahertz frequency
Show abstract
A special software is constructed effectively for reconstructing the fine phase distribution of the diffracted Gaussian laser
beams in the terahertz frequency range, according to common diffraction theory. The fine surface microrelief patterns of
the elements, which originate from the simple patterns in photomask and further etched onto the surface of {100}-
oriented silicon wafer by a low cost and rapid method, are created by the software above according to the phase
distribution designed. Being different with the traditional silicon diffractive lenses fabricated by multiple level processes,
the elements produced by the method introduced by us can transfer common Gaussian beams into desired images
through created fine patterns over the surface of the elements. Two typical type of diffractive elements, which are used to
transform common Gaussian laser beams in terahertz frequency into highly focused spot or so-called common focus, and
the desired figure of the "umber one", are designed and fabricated. For testing the element, the LASER SIEIR 50 of
Coherent Company is used to generate common Gaussian laser beams (the diameter of the beams is ~10mm), and the
PYROCAM THERE of Spiricon Company is also used to display the images acquired. Experimental results show that
the elements can be used to form needed light fields and expected images, respectively.
Compressed image transmission based on fountain codes
Show abstract
In this paper, we propose a joint source-channel coding (JSCC) scheme for image transmission over wireless channel. In
the scheme, fountain codes are integrated into bit-plane coding for channel coding. Compared to traditional erasure codes
for error correcting, such as Reed-Solomon codes, fountain codes are rateless and can generate sufficient symbols on the
fly. Two schemes, the EEP (Equal Error Protection) scheme and the UEP (Unequal Error Protection) scheme are
described in the paper. Furthermore, the UEP scheme performs better than the EEP scheme. The proposed scheme not
only can adaptively adjust the length of fountain codes according to channel loss rate but also reconstruct image even on
bad channel.
Evaluation of hyperspectral classification methods based on FISS data
Show abstract
With the deterioration of ecological environment, rare plants on the earth are decreasing rapidly, so there is an urgent
need for the study on sophisticated vegetation classification. Hyperspectral data have great potential in sophisticated
classification. FISS(Field Imaging Spectrometer System) is a newly developed system, and pixels of FISS could be
considered as pure pixels with high spatial and spectral resolution, which makes FISS a perfect option on the study of
methodology. This study aims to evaluate different methods based on FISS data and find out the best one of
sophisticated vegetation classification. The methods are as follows: Maximum Likelihood (ML), Spectral Angle
Mapping (SAM), Artificial Neural Net (ANN), Support Vector Machine (SVM) and Composite Kernel Support Vector
Machine (C-SVM). Firstly, segmented principal components transformation is adopted for spectral dimensionality
reduction, and all bands are divided into 2 subsets according to the correlation matrix. Secondly, 16 principal
components are kept. After that, 5 methods mentioned above are tested. The Overall Accuracy and Kappa coefficient of
C-SVM, SVM and ANN are higher than 90%, and C-SVM obtains the highest accuracy, which is consistent with visual
interpretation. The result shows that C-SVM, SVM and ANN are more suitable for sophisticated vegetation classification
of hyperspectral data, and C-SVM is the best option.
Detail-preserving image recovery based on respective restoration on phase and amplitude
Show abstract
This paper firstly improves D.L.Plillips's representation about image restoration and then points out
that image restoration is just a 'partial ill-posed' problem rather than a 'total ill-posed'
problem---amplitude restoration is ill-posed but phase restoration is well-posed. Basing on the
viewpoints, the paper proposes a restoration method, which cuts down phase-pollution caused by
traditional regularization methods, that amplitude restoration is realized by regularization and phase
restoration is achieved by algebraic method. Experimental results indicate that the proposed method
performs well. It can efficiently restore image phase and elaborately preserve image details.
A structured sub-pixel target detection method base on manifold learning method
Show abstract
The manifold learning theory is firstly used to transform the hyperspectral images into a low-dimension feature spaces.
The reconstruction error is computed to get discriminative information. Then a structured matched subspace detector is
developed. This method can effectively avoid the contamination by targets and spectral anomalies to backgrounds
subspace and detect sub-pixel targets with better performance than traditional methods.
A novel non-uniformity correction method based on ROIC
Xiaoming Chen,
Yujue Li,
Chao Di,
et al.
Show abstract
Infrared focal plane arrays (IRFPA) suffer from inherent low frequency and fixed patter noised (FPN). They are thus
limited by their inability to calibrate out individual detector variations including detector dark current (offset) and
responsivity (gain). To achieve high quality infrared image by mitigating the FPN of IRFPAs, we have developed a
novel non-uniformity correction (NUC) method based on read-out integrated circuit (ROIC). The offset and gain
correction coefficients can be calculated by function fitting for the linear relationship between the detector's output and a
reference voltage in ROIC.
We tested the purposed method using an infrared imaging system using the ULIS 03 19 1 detector with real nonuniformity.
A set of 384*288 infrared images with 12 bits was collected to evaluate the performance. With the
experiments, the non-uniformity was greatly eliminated. We also used the universe non-uniformity (NU) parameter to
estimate the performance. The calculated NU parameters with the two-point calibration (TPC) and the purposed method
imply that the purposed method has almost as good performance as TPC.
One SAR image filtering method based on the heterogeneity measurement
Shao-bo Chen,
Tao Zhang
Show abstract
The deviation coefficient (Cv) and the arithmetic-geometric mean ratio (A/G) are two kinds of different measurement
methods of SAR image heterogeneity. They are widely used for SAR image processing. This paper analyzes the Kuan
filtering algorithm based on the deviation coefficient and puts forward the Kuan filtering algorithm based on the
arithmetic-geometric mean ratio. And meanwhile, it analyzes and compares the two types of adaptive speckle noise
filters through the experiment.
Researches on ultrasonic image revise for phased array ultrasonic test of train wheel
Yu Zhang,
Li Wang,
Xiaorong Gao,
et al.
Show abstract
Ultrasonic image formed by Phased Array Ultrasonic probes placed on wheel tread is deformed. Especially for ultrasonic
image around area of wheel disk hole, because that area has a complex construction and a long ultrasonic path.
According to probes layout and beam angle, correction algorithms are designed to revise deformed ultrasonic images.
Tests on CRH reference wheels with manmade defects and tests on CRH wheels on train shows that the correction
algorithms proposed can correct all types of CRH wheel's ultrasonic images, and the calibration can meet the
requirement of field application.
A spectrum fractal feature classification algorithm for agriculture crops with hyper spectrum image
Junying Su
Show abstract
A fractal dimension feature analysis method in spectrum domain for hyper spectrum image is proposed for agriculture
crops classification. Firstly, a fractal dimension calculation algorithm in spectrum domain is presented together with the
fast fractal dimension value calculation algorithm using the step measurement method. Secondly, the hyper spectrum
image classification algorithm and flowchart is presented based on fractal dimension feature analysis in spectrum
domain. Finally, the experiment result of the agricultural crops classification with FCL1 hyper spectrum image set with
the proposed method and SAM (spectral angle mapper). The experiment results show it can obtain better classification
result than the traditional SAM feature analysis which can fulfill use the spectrum information of hyper spectrum image
to realize precision agricultural crops classification.
PolSAR image speckle reduction algorithm based on TV-PDE
He Li,
Zhiyuan Qin
Show abstract
Good polarization SAR speckle reduction effect is very helpful to follow-up treatment for such as Polarization SAR
image classification and interpretation. Based on analysis of the existing polarization SAR image filtering algorithm, an
algorithm of PolSAR speckle reduction was proposed to maintain the relative phase information with SAR image
speckle reduction of the total variation of partial differential equations (TV-PDE). First, three polarization combinations
which could be represented by three color channels were got from Polarization SAR data with Pauli decomposition.
Then, the color channels were operated with TV-PDE speckle reduction algorithm. The experiment results showed that
the algorithm is effective and of good performance in the filtering and edge preservation. Comparison was made with
other speckle filtering algorithms taking speckle index as the evaluation criteria.
Watermarking spectral images with three-dimensional discrete wavelet transform and singular value decomposition under various illumination conditions
Show abstract
Kaarna et al. [pro. Scand. Cof. Image Analysis, SCIA 2003, pages 320-327] proposed a watermarking method based on
the three dimensional wavelet transform for spectral images. kaarna et al [J. Imaging SCI. Technol. 52, pages 30502-1 -
30502-18, 2008] reported that the robustness of the watermarking method to different illumination conditions. The
spectral image database provider stores the reflectance or radiance spectra of the images. Depending on the client's
requirements, the effects from illumination can be added to the spectra, i.e., the viewing conditions change the perceived
color of the spectrum. External illumination can be compensated through convoluting the spectra of the image with the
spectrum of the illuminant. In this paper, a hybrid watermarking method based on the three-dimensional wavelet
transform and singular value decomposition is proposed. The proposed method is compared with the 3D-DWT method
of kaarna et al in the cases both with and without effect of different illumination conditions. Experiments were
performed on a spectral image of natural scenes. Inlab2 was selected. The color reproduction is done using CIE XYZ
basis function with D65 light model. Inlab2 image have the following dimensions: 256x256 pixels, and 31 spectral
components per each pixel. Images were captured by a CCD (charge coupled device) camera in a 400-700 nm
wavelength range at 10 nm intervals. The image selected was taken indoor (in a controlled environment, i.e. dark-lab or
glass-house). The performance of the proposed technique is compared with the work of kaarna et al against different
illumination conditions and attacks including median and mean filtering, lossy compression. The experiments indicate,
the proposed method outperforms the work of kaarna et al.
Diagnosis method of cucumber downy mildew with NIR hyperspectral imaging
Youwen Tian,
Tianlai Li,
Lin Zhang,
et al.
Show abstract
This study was carried out to develop a hyperspectral imaging system in the near infrared (NIR) region (900-1700 nm)
to diagnose cucumber downy mildew. Hyperspectral images were acquired from each diseased cucumber leaf samples
with downy mildew and then their spectral data were extracted. Spectral data were analyzed using principal component
analysis (PCA) to reduce the high dimensionality of the data and for selecting some important wavelengths. Out of 256
wavelengths, only two wavelengths (1426 and 1626nm) of first PC were selected as the optimum wavelengths for the
diagnosis of cucumber downy mildew. The data analysis showed that it is possible to diagnose cucumber downy mildew
with few numbers of wavelengths on the basis of their statistical image features and histogram features. The results
revealed the potentiality of NIR hyperspectral imaging as an objective and non-destructive method for the authentication
and diagnosis of cucumber downy mildew.
Infrared image denoising and enhancing algorithm using adaptive threshold shrinkage in a new contourlet transform
Fei Wang,
Xiaogeng Liang,
Yankai Cui,
et al.
Show abstract
Edge and detail of infrared image are blurry or loss after denoising with the threshold shrinkage arithmetic. A new
adaptive denoising and enhancing algorithm with detail enhancement based on a new Contourlet Transform with Sharp
Frequency Localization(CT-SFL) is proposed to preserve the edge better. CT-SFL has the characteristic of well-localized
in the frequency domain compared with the original contourlet. Firstly, CT-SFL, instead of the original contourlet, is
employed as the multiscale decompositon to decompose the infrared image into subbands. Secondly, the hierarchical
adaptive denoising threshold of new Contourlet coefficient is estimated respectively by each location from different scale
and directional subband, the noisy image is denoising with soft threshold related to the transform scale and direction,
then the denosing image is enhanced by taking decomposable scale and directional energy into account with
intrasubband and interscale dependencies. Thirdly, inverse CT-SFL is used to reconstruct the denoising and enhancing
image. Finally, in order to reduce significant amount of aliasing components which are located far away from the desired
support because of the new Contourlet Transform, cycle spinning is accomplished to the whole denoising and
enhancement process to overcome the lack of translation invariance property and suppress pseddo-Gibbs phenomena
around singularities of denoising image. Numerical experiments on infrared noisy image show that the proposed novel
algorithm can significantly outperform some arithmetics based on contourlet like 3 sigma, VisuShrink and Bayes
Shrinkage in all kinds of noise spectral density both in terms of PSNR(by several dB) and in visual quality, which can
enhance image's detail and stretch its contrast with nearly similar computational complexity.
An improved blind restoration algorithm for multiframe turbulence-degraded images
Show abstract
This paper proposes an improved blind deconvolution algorithm, which adopts maximum likelihood method to find the
most similar estimation of the PSF and object with Poisson-based probability model. The algorithm integrates Cauchy
probability distribution model into the estimation of the PSF under the condition of low SNR, uses the characteristic of
short-exposure image sequence that the adjacent images have similar PSF to get restored image with frames as few as
possible. The experimental results show that this method is robust with high ability of resisting noise in the restoration of
turbulence-degraded images.
Variational adaptive image denoising model based on human visual system
Wenjun Li,
Chanjuan Liu,
Hailin Zou
Show abstract
A variational image adaptive denoising model based on human visual system is proposed by introducing control
parameter p which can determine the diffusion intensity to Total Variation (TV) model. The model can adaptively
select the value of parameter p according to human visual system noise visibility value of each pixel which makes
diffusion intensity close to edges smaller than those far away from edges. For this method is more consistent with human
perception, human eyes can perceive the improvement of image quality intuitively. Numerical experiments show that the
proposed method can overcome staircase effect, remove the noise while preserving significant image details and better
performance has been achieved.
The study on disaster assessment of snow in pastoral areas based on remote sensing data
Juan Nie,
Wenbo Li,
Tong Tang,
et al.
Show abstract
In this paper, snow disaster assessment model based on analytical hierarchy process, snow range extraction model and
snow depth inversion model have been established. The established disaster assessment models and MODIS data have
been employed to evaluate snow situation in north Xinjiang from November 2005 to January 2006. The results show that
the proposed model can effectively assess snow disaster situation and be used in disaster situation management.
Multispectral and hyperspectral image processing based on the waveform characteristics of spectral remote sensing classification method of large area
Show abstract
Determination of land cover spatial distribution and change scientifically and accurately have a great significance for
monitoring the regional ecological environment and the impact of human activities. This paper proposed a method basing
on the waveform characteristics of spectral for mapping large area using TM data simply, quickly and accurately. In the
experiments, we segment the image to object and order the object value of the TM5 1~7 except the thermal infrared band
and each order result given the only value. In the next, using manual methods assigned the order to get the typical surface
feature such as water, soil, vegetation. Classification using spectral waveform characteristics can effectively avoid the TM
image classification error, due to spectral differences by accessing time and spatial, spectrum differences on same object
and spectral mixing .This method can map the large area of typical surface is simply, fast and accurate, and support on
theoretical basis.
Soil moisture evaluation in the Three Gorges Reservoir area using ENVISAT ASAR data
Show abstract
This paper presents a method to estimate the soil moisture using ENVISAT ASAR data in the Three Gorges Reservoir
area in China. Firstly, this study introduces a semi-empirical model for bare surface scattering and a water-cloud model
for the elimination of the impact of the vegetation cover. Secondly, a new combined roughness parameter is introduced to
describe the roughness of soil surface. Thirdly, we analyze the relationship between soil roughness and radar
backscattering coefficient and that between soil moisture content and backscattering coefficient respectively. Then the
soil moisture inversion algorithm is achieved through programming. Finally, an evident logarithmic relationship between
radar backscattering coefficient and soil moisture is presented. In conclusion, the experiments prove that the method is
fast, efficient and widely applicable.
A novel method to suppress noise in marine radar images based on pulse-pulse correlation
Show abstract
As the X-band marine radar often suffers from interference of electromagnetic waves of the same frequency transmitted
by radars in its vicinity, the acquired images frequently contain co-channel interference noise. The noise degrades the
quality of the marine radar images and is unfavorable to the processing and interpretation of the marine radar images.
To suppress the noise in marine radar images, a novel method based on pulse-pulse correlation is proposed. This method
includes three steps: threshold segmentation, noise extraction and noise fixing. In the threshold segmentation step, the
threshold T is calculated based on the K distribution sea clutter model. In the noise extraction step, a 3×3 window is
applied. By using the window, the pixels of noise can be extracted, and at the same time the pixels of non-noise can be
discarded. In the noise fixing step, the strategy of piecewise interpolation is applied. At the region near to the image
center, the triangulation with linear interpolation algorithm is applied; at the region far from the image center, the nearest
neighbor algorithm is applied.
The real X band marine radar image was used to test the performance of the proposed method. The obtained results show
that the proposed method is able to reduce the co-channel interference noise from the marine radar images significantly
and keep the information of objects in the images such as ships and islands. Besides, the proposed method can be fast in
speed of operation.
Multispectral remote sensing image cross simulation based on nonlinear spectral fitting model
Jinxiang Shen,
Liao Yang,
Xi Chen
Show abstract
The remote sensing image recorded the ground object spectral responses with a special spectral, temporal, and spatial
resolution. There are some complex relationship may exist between the remote sensing images with different spectral,
spatial, and temporal scale. In this study, we try to use a nonlinear regression model - Cubist regression tree model to
mining the spectral relationship between the image bands. The Landsat5 TM image was used as reference image to
collect samples to train Cubist model, and then the target image - SPOT5 image was used to predict its lacked TM-liked
band1 and band7 with the TM-trained Cubist model. The experiments shows that the Cubist nonlinear regression model
could simulate TM band1 and band7 with a high accuracy and the TM-trained Cubist model also could be used to predict
SPOT5 lacked TM-liked band1 and band7.
Hyperspectral image segmentation using spectral-spatial constrained conditional random field
Show abstract
In this paper, we propose a hyperspectral image segmentation algorithm which combines classification and segmentation
into Conditional Random Field(CRF) framework. The classification step is implemented using Gaussian process which
gives the exact class probabilities of a pixel. The classification result is treated as the single-pixel model in CRF
framework, by which classification and segmentation are combined together. Through the CRF, the spatial and spectral
constraints on pixel classification are exploited. As a result, experimental results on real hyperspectral image show that
the segmentation precision has been much improved.
Scene-based nonuniformity correction using sparse prior
Show abstract
The performance of infrared focal plane array (IRFPA) is known to be affected by the presence of spatial fixed pattern
noise (FPN) that is superimposed on the true image. Scene-based nonuniformity correction (NUC) algorithms are widely
concerned since they only need the readout infrared data captured by the imaging system during its normal operation. A
novel adaptive NUC algorithm is proposed using the sparse prior that when derivative filters are applied to infrared
images, the filter outputs tends to be sparse. A change detection module based on results of derivative filters is
introduced to avoid stationary object being learned into the background, so the ghosting artifact is eliminated effectively.
The performance of the new algorithm is evaluated with both real and simulated imagery.
Multisensor image registration using modified Hausdorff distance matrix metrics
Qing Li,
Guang-zhou Qu,
Xin Zhao,
et al.
Show abstract
As to multi-sensor image registration, a novel algorithm for registration synthetic aperture radar (SAR) image to Optical
image based on Weight Hausdorff Distance Matrix (WHDM) Metric is proposed in the paper. In the proposed method,
edge feature is used for common feature, which can solve the key problem brought by different senor. WHDM is
designed as a similarity measure. Experimental results using synthetic and real images demonstrate that the proposed
WHDM metric outperforms most existed Hausdorff Distance based (HD-based) methods, especially in higher level noise
and that the proposed algorithm is robust, and can achieve high registration accuracy.
Gradients-predict filter of multiple-scale template
Show abstract
Noise filtering is an essential part of any image processor, whether the final image is utilized for visual
interpretation or for automatic analysis. Arithmetic Mean Filter (AMF) and Standard Median Filter (SM) tend to work
well for fixed-valued impulses but poorly for random-valued impulse noise. The authors review the current trend of
image filter development. Some papers present new ways to identify corrupted pixels, while others strongly emphasize
on suppressing the noise ratio. And some summarized the both and combined to present new complicated scheme. As
widely known, Vector Median Filter (VMF) has the disadvantage of replacing too many uncorrupted image pixels. New
method form R. H. Chan, Chung-Wa Ho, and M. Nikolova in 2005 is capable of restoring images corrupted by
salt-and-pepper noise with extremely high noise ratio, but the calculation is so complex that it is only can be treated as a
post-processing image enhancement procedure. In the fact, the modern imaging equipment is good enough and will not
produce more than three percent of salt-and-pepper noise. But as a pretreatment of Image Analysis, filter should focus on
preservation the original details and simple or fast for the engineering. Our purpose in this paper is to present a simple
scheme to preserve uncorrupted, original pixels, but still enables to remove corrupted ones with a good balance the
algorithm complexity and the efficiency.
Object matching task-oriented image quality assessment
Zhiguo Cao,
Xiyao Duan
Show abstract
As the development of Automatic Target Recognition (ATR) technology, the performance evaluation for it becomes
more and more important. Image Quality is a very important part of ATR performance evaluation. Whether the image
quality is good or bad is tightly related with algorithm performance. In this paper, we construct a system of image quality
assessment related to object matching task. In our framework, image quality measures published and newly proposed are
analyzed first, and then a set of effective measures are selected from above. At last, we construct a model of relationship
between selected measures and matching. Given an image, the system will indicate whether an image is suitable for a
specific matching task. Gray level correlation matching and histograms of oriented gradients -matching are demonstrated
and their experimental result shows that the image quality assessment system works effectively.
Adaptive stereo image joint compression based on characteristics classification
Show abstract
On the basis of analyzing the characteristics of the stereo image pair obtained by the remote sensing mapping,
considering the continual increment of data obtained by stereo mapping satellite on orbit and the trend of the hardware
development, a joint compression algorithm for the stereo image pairs based on the classified characters is proposed.
First, stereo image pairs are matched through SIFT feature matching; and then the images are classified adaptively by the
features according to the matching results. For the region where the characteristics are flat, the predicted residual-error
image is coded based on the coarse match and the fine match using the adaptive block in order to improve the accuracy
of matching with the linear prediction for the radiation compensation. For the feature-rich region, according to the
correlation of the images, individual compression or joint compression is selected. When the joint compression algorithm
is selected, the match alignment of the left and right image pair is done by the affine model estimation and the linear
prediction is adapted to the radiation compensation. Because different compression scheme is adopted for different
characteristics regions, the results show that the PSNR and the geometric precision of the reconstructed images are
improved effectively, especially when the images contain many characteristics of elevation information.
The estimation of sub-pixel NDVI time series based on down-scaling technique
Show abstract
In this paper, based on the theory of down-scaling, we propose the methods for linear mixed model disaggregate mixed
pixels in coarse resolution images. Exploiting information about within mixed pixel each component fractional cover
derived from high spatial resolution classification map, Sub-pixel reflectance for the different land-cover classes are
calculated by solving a linear system of equations for each pixel of a coarse resolution image and producing the subpixel
level NDVI time series curve of different component. Results showed that application of the algorithm provided
good estimates of sub-pixel NDVI time series even for poorly represented land-cover classes. The main advantage of the
proposed technique is that could analysis the land-use and vegetation biomass change better.
Selection of matching area in SAR scene-matching-aided navigation based on manifold learning
Bin Li,
Junbin Gong,
Jinwen Tian
Show abstract
Selection of suitable matching area is one of the key issues for image-matching-aided navigation system,but it is also a
very challenging mission, especially with the multi-source image matching tasks. In this paper, a novel method to
analyze the matching suitability of the satellite optical photograph to the realtime SAR in candidate flying regions is put
forward. At first, several typical low-level image features are extracted. Then manifold learning is used to reduce the
dimension of the sampled features, so as to generate new high-level image features with better discrimination ability.
Finally, with the new features generated by manifold learning, we used support vector machines (SVM) to divide the
candidate regions into two classes for suitable or unsuitable for matching. The experimental result shown that the
proposed method is valid and effective.
Horizontal tilt correction for license plate image
Show abstract
License plate location is the basis of LPR, and it is the most important part of it. Tilt vehicle license plate has an adverse
effect on its character segmentation and recognition. In this paper, tilt models of a plate are analyzed and a novel
approach for number plate tilt correction is presented. Horizontal incline angle is obtained by combining horizontal or
vertical edge of a plate with its rotation projection. Bilinear interpolation rotation correction is performed to the plate
which is tilt to the horizontal line. Experimental results show that the method can be implemented easily and offers
robustness when dealing with dirty number plates and license plates in variant lighting conditions.
Defocus blurred image restoration by minimizing second-order central moment
Show abstract
A new blur identification and restoration method is presented. We observe that blurring increases the second-order
central moment (SOCM) of image and introduce a new parametric blur identification method by minimizing SOCM. The
method applied to finite support images, in which the scene consists of a finite extent object against a uniformly black,
grey or white background. The method has been validated by direct comparisons with other methods on simulated
images. Our experiments show that the SOCM minimization measurements match well with methods than maximize
PSNR.
Combining shape and texture features for infrared pedestrian detection
Show abstract
This paper presents a robust pedestrian detection algorithm that works on infrared imageries. Our algorithm is
applicable to images captured from surveillance infrastructure as well as moving platforms. Firstly, we introduce a
local binary pattern (LBP) texture feature for infrared pedestrian representation. Secondly, motivated by the recent
success of multiple cues pedestrian detection in visual imagery, we combine both shape and binary pattern texture
features for effective infrared pedestrian description, providing a level of robustness to variations in pedestrian
shape and appearance in infrared images. Finally, a support vector machine (SVM) classifier is utilized to classify
sub-windows into pedestrians or background. Experimental results demonstrate the robustness and effectiveness of
our method.
Analysis and research on precipitable water vapor in autumn of Chengdu region
Show abstract
The rainfall process in autumn of Chengdu region has significant regional characteristics. Since GPS detection
technology has characteristics such as all-weather, high accuracy, high spatial and temporal resolution and low cost, its
tracking and monitoring technique on water vapor has achieved rapid developments in recent years. In this paper, it
makes use of GPS-PWV data from 6 foundation GPS observation stations of GPS observation network in Chengdu
region from September 2007 to November 2007 and from September 2008 to November 2008 these two falls which have
30min intervals. Fast Fourier transform was used to obtain the variation principle that in autumn the rainfalls change in
the time period and there also has one quarter season within the fall, which is around 22 days or so. After the analysis, it
finds that PWV drastically decreases at late October which is closely correlated to the local climate changing cycle. After
we conduct composite analysis on diurnal cycle by integrating PWV with other meteorological elements we can find:
There is a negative correlation between PWV and temperature; there is more obvious positive correlation between
humidity and PWV. GPS precipitable water vapor may increase or maintain the characteristics within high value area
from midnight to early morning which can bitterly correspond with the actual rainfall process of this region.
Color image fusion method using the multi-scale retinex and directional support value transform
Yaocheng Xie,
Sheng Zheng,
Cuimei Guo,
et al.
Show abstract
Completely different from many exiting multi-source image fusion methods, a novel color image fusion method for a
single original color image, combining the multi-scale Retinex (MSR) with directional support value transform (DSVT),
is presented in this paper. The applied MSR is an (centre/surround-based) Retinex algorithm and the directional support
value transform, an anisotropic and multi-scale transform, is deduced under the weighted mapping least squares support
vector machine (LS-SVM) framework. Using the MSR to an original color image, an enhanced color image, producing
more detail information hidden in shadow areas of original color image but easily lapsing into color distortion, is
obtained. In the HSV color space, the good color information of original color image and more detail information of
enhanced color image are integrated into a fused color image based on the DSVT. Series of color images under different
environment are chosen for color image fusion experiments and the performance of DSVT is compared with other
methods also used in color image fusion, including Laplacian pyramid, discrete wavelet transform and support value
transform. The experimental results demonstrate that: the proposed color image fusion approach is effectively making
the fused color images not only present more clearly detail information but also maintain color fidelity; DSVT is superior
to other three methods used in color image fusion.
Joint spatial and spectral analysis for remote sensing image classification
Linlin Shen,
Sen Jia
Show abstract
With the development of sensors, the spatial and spectral resolutions of remote sensing data are getting much higher, which presents
new possibilities and challenges for pixel based material classification. When most of the methods available in literature extract
features in spectrum domain for land material classification, the rich information contained in hyperspectral data is not fully used. As a
result, the classification accuracies reported in literature are not satisfying. In this work, we aim to use joint spatial and spectral
analysis technique to extract information about signal variances in space, spectrum and joint space-spectrum domain. The feature thus
extracted can better represent the signal variances and can thus improve overall classification accuracy.
Locust habitats monitoring based on multi-temporal CCD data of HJ-1 satellite
Jingjing Li,
Jian Chen,
Shijie Sheng
Show abstract
The classification of locust habitats is one of the most important tasks in terms of the monitoring and controlling of the
locusts damages. Taking the Bohai development zone in Hebei province as the study area, this paper generated firstly the
14 scenes of temporal NDVI data from January 31 to December 9, 2010 based on image preprocessing and NDVI
calculation from HJ-1 satellites. And then the classification system was developed according to the characteristics of the
study area and locust habitats, including dense reeds, sparse reeds, folder wasteland, grassland, crop1, crops2, alfalfa,
orchard, soil, salt water, building and road. At last, this paper classified the locust habitats using decision tree method
based on the multi-temporal NDVI data. Verification data collected through ground surveys and 14 scenes of temporal
CCD data were used for assessing the accuracy of the classified results. The results showed that some surface features
prone to misclassification in a single phase for they have similar spectral characteristics were easy to classify from multitemporal
data because their different temporal signature.
The electrified insulator paramater measurement for flashover based on photogrammetric method
Jianwu Jiang,
Ling Zhao,
Jianwei Wang,
et al.
Show abstract
It is an important work to measure insulator geometric parameters for preventing pollution flashover on power
transmission lines. This paper presents an effective method to measure insulator parameters based on non-touch
photogrammetric method. In this paper, the insulator reconstruction takes full advantage of the rotating body's geometry
structure about the symmetry axis. The spindle of insulator can be calculated from its profile extracted from multiple
insulator images acquired by calibrated cameras and radius is obtained by the homologous image points of insulator
profiles. Results from real data are presented, demonstrating the efficiency of the proposed methods.
Information extraction from laser speckle patterns using wavelet entropy techniques
Show abstract
A novel speckle patterns processing method is presented using multi-scale wavelet techniques. Laser speckle patterns
generated from the sample contained abundant information. In this paper, we propose a method using wavelet entropy
techniques to analyze the speckle patterns and exact the information on the sample surface. In our case, we used this
approach to test the solar silicon cell surface profiles based on the sym8 orthogonal wavelet family. According different
wavelet entropy values, the micro-structure of different solar silicon cell surfaces were comparative analyzed.
Furthermore, we studied the AFM and reflective spectra of the wafer. Results show that the wavelet entropy speckle
processing method is effective and accurate. And the experiment proved that this method is a useful tool to investigate
the surface profile quality.
Parametric blind deconvolution for passive millimeter wave images basing on image decomposition
Show abstract
This paper deals with deconvolution problem for passive millimeter wave images with poor resolution and low SNR. A
passive millimeter wave images super-resolution algorithm based on semi-blind deconvolution is put forward. The
proposed method is based on two characteristic of imaging system. First, the PSF of imaging system is certain, can be
modeled by parametric function. Second, the noise imposes different influence degree on the low frequency and high
frequency parts of the pass-band of the image, the low frequency and high frequency part have high and low SNR,
respectively.The image is decomposed using the bilateral filtering into low frequency base layer and high frequency
detail layer. The base layer contains the large-scale structures and nearly frees with noise thus has the higher SNR,
whereas the detail layer includes both small-scale details and noise and has the lower SNR. The base layer is restored by
semi-blind deconvolution. The system PSF is modeled as a parametric Gaussian form. Edge structures information of the
image is extracted basing on Mumford-Shah model and used to adjust the regularization term adaptively, and iterative
method is used to estimate image and blurred kernel parameters. The detail layer is adaptively denoised by combining
the joint bilateral filtering method and with edge preservation in the guidance of the base layer. Finally, the high
resolution image is obtained by combining the base and detail layers. Comparative experimental results show that the
proposed method can effectively suppress noise, reduce artifacts, and improve the spatial resolution.
A hybrid interpolation method edge-preserving for passive millimeter wave images
Show abstract
This paper deals with edge-preserving interpolation for passive millimeter wave images with poor resolution and low
SNR. A hybrid interpolation method basing on image decomposition via bilateral filtering is proposed. The low
resolution and noisy image is first decomposed using the bilateral filter into the base and detail layers which represent
large and small scale features, respectively. The base layer contains the large-scale structures and nearly frees with noise
thus has the higher SNR, whereas the detail layer includes both small-scale details and noise and has the lower SNR. The
detail layer is adaptively denoised by the joint bilateral filtering method and subsequently interpolated with edge
preservation in the guidance of the base layer. The base layer is interpolated with edge-preserving method directly.
Finally, the high resolution image is obtained by combining the base and detail layers. Experimental results show that the
proposed method outperforms the conventional methods while preserving edges and suppressing jagging thus is suited
for PMMW images interpolation to enhance resolution.
Livewire based single still image segmentation
Show abstract
In the application of the video contactless measurement, the quality of the image taken from underwater is not very well. It
is well known that automatic image segmental method cannot provide acceptable segmentation result with low quality
single still image. Snake algorithm can provide better result in this case with the aiding of human. However, sometimes the
segmental result of Snake may far from the initial segmental contour drawn by user. Livewire algorithm can keep the
location of the seed points that user selected nailed from the beginning to the end. But the contour may have burrs when
the image's noise is quite high and the contrast is low. In this paper, we modified the cost function of Livewire algorithm
and proposed a new segmentation method that can be used for single still image segmentation with high noise and low
contrast.
Nonlocal means SAR image despeckling using Principle Neighborhood Dictionaries
Show abstract
The Principle Neighborhood Dictionary (PND) filter projects the image patches onto a lower dimensional subspace using
Principle Component analysis (PCA), based on which the similarity measure of image patch can be computed with a
higher accuracy for the nonlocal means (NLM) algorithm. In this paper, a new PND filter for synthetic aperture radar
(SAR) image despeckling is presented, in which a new distance that adapts to the multiplicative speckle noise is derived.
Compared with the commonly used Euclidean distance in NLM, the new distance measure improves the accuracy of the
similarity measure of speckled patches in SAR images. The proposed method is validated on simulated and real SAR
images through comparisons with other classical despeckling methods.
Hybrid digital fingerprint based on multi-type characters for collusion-resistance
Zuxi Wang,
Wenzong Wang,
Hanping Hu,
et al.
Show abstract
Collusion tolerance is an essential requirement for digital fingerprinting, and existing collusion-resistant encoding and
tracking methods are generally based on combinatorics. However, those methods have some deficiencies such as low
tracking efficiency and large fingerprint database for matching. Owing to those weaknesses, a hybrid fingerprint
encoding method based on residual characters is proposed in this paper. The fingerprint is generated by encoding the
user information into a binary image composed of letter, numeral and Chinese character. As a result of the complex
structure of Chinese character, more residue characters can be reserved when the fingerprint is damaged. Thus, the
fingerprint has a good performance in anti-tampering. Meanwhile, it has the capacity for large users because of the
diversity of Chinese character. In addition, the paper proposes a tracking method based on grid feature, which greatly
reduces the computational complexity and improves the tracking efficiency. The robustness and collusion tolerance of
the digital fingerprinting scheme are confirmed by theoretical analysis and experimental results.
Signal-dependent noise filtering of SAR image
Show abstract
The multiplicative noise model is often used to describe the speckle noise in the SAR image. This speckling process
defines the conditional probability of the SAR image intensity I from RCS ó. The MMSE algorithm of the SAR image
reconstruction is basis on the multiplicative noise model. However, the MMSE algorithm does not define how
calculating the look number of the SAR image. This paper advises an algorithm calculating the look number of the SAR
image. This algorithm is basis on the local statistical information. Therefore, the MMSE algorithm can adapt all kinds of
the SAR images. Furthermore, when the iteration MMSE is used, the MMSE algorithm can adapt the change of SAR
image statistical information.
SAR and panchromatic image fusion based on contourlet hidden markov tree model
Huihui Li,
Kun Liu
Show abstract
Contourlet-domain hidden markov tree(HMT) model can reflect the coefficients' dependencies of different scales and
directions, then a SAR and optical image fusion algorithm based on it is proposed. Firstly source images are decomposed
by Contourlet transform, low frequency and high frequency subband coefficients are obtained; secondly, the high
frequency subband coefficients are modeled using hidden markov tree and the model is trained using
Expectation-Maximization(EM) algorithm to get the posterior probability of the coefficients; Thirdly, using the posterior
probability to guide the fusion rules' design of high frequency subband coefficients in order to preserve the salient
features of original images well and obtain good noise suppression; Finally, the fusion coefficients are inversed to get the
final fusion result. SAR and Panchromatic images were taken to fuse, and the results were evaluated by Difference
Coefficient, Correlation Coefficient and Signal-Noise Ration. The experimental result shows that the proposed algorithm
is better than traditional fusion algorithms based on Contourlet and wavelet-domain hidden markov tree model.
Infrared image denoising algorithm based on adaptive dictionary learning
Deqin Shi,
Wei Yang,
Junshan Li
Show abstract
A novel infrared image denosing algorithm is proposed based on adaptive dictionary learning over sparse and redundant
representations. The dictionary which can yield sparse representations is learned from the corrupted infrared image itself,
instead of using the prechosen set of basis functions such as curvelet or contourlet. Meanwhile, the over-completed
dictionary is updated adaptively in the online learning procedure other than batch learning method to improve the
learning performance. And the learning and denoising procedure are fused together into one iterated process naturally
and properly. Experimental results demonstrate the effectiveness of the denosing algorithm for infrared images.
Improved NSCT based shrinking threshold denoising algorithm for infrared image
Show abstract
An improved NSCT based shrinking threshold denoising algorithm for infrared image is proposed. The improved NSCT
is constructed based on the Nonseparable Wavelet Transform via non-linear lifting method with redundancy structure,
which can produce better image processing performance for its better detail capturing capability, shift invariance, multiresolution
and multi-direction. After analyzing the current threshold functions and threshold selecting methods, an novel
threshold function suitable for the improved NSCT is established with high-level continuous derivative to improve the
denosing performance for infrared image. Experimental results show that the proposed algorithm has better denoising
performance and detail-preserving capability.
Compound tetrolet sparsity and total variation regularization for image restoration
Liqian Wang,
Liang Xiao,
Zhihui Wei
Show abstract
Image restoration is one of the most classical problems in image processing. The main issue of image restoration is
deblurring as well as preserving the fine details. In order to restore the high quality image, we propose a compound
regularization method which combines the tetrolet-based sparsity and a new weighted adaptive total variation (ATV).
Tetrolet transform is a geometric adaptive Haar-type wavelet transform. It finds the optimal partition to fit the local
image structures and the tetrolet coefficients can capture the textures and details information in different image scales.
ATV adds two directional gradient operators into the original anisotropic TV. It not only seeks the intensity continuity
horizontally and vertically, but also seeks the intensity continuity diagonally. Combining the tetrolet-based sparsity and
ATV together, our model can restore the local structures and details by the tetrolet-based sparsity regularization while
suppress the noise and recover piecewise smooth images with sharp edges along four directions by the ATV
regularization. For solving the minimizing problem, we propose an algorithm which consists of the variable splitting
method and the Dual Douglas-Rachford splitting method. The Experimental results demonstrate the efficiency of our
image restoration method for preserving the structure details and the sharp edges of image.