MonoColor CMOS sensor
A new breed of CMOS color sensor called MonoColor sensor is developed for a barcode reading application in AIDC industry. The RGBW color filter array (CFA) in a MonoColor sensor is arranged in a 8 x 8 pixels CFA with only 4 pixels of them are color (RGB) pixels and the rest of 60 pixels are transparent or monochrome. Since the majority of pixels are monochrome, MonoColor sensor maintains 98% barcode decode performance compared with a pure monochrome CMOS sensor. With the help of monochrome and color pixel fusion technique, the resulting color pictures have similar color quality in terms of Color Semantic Error (CSE) compared with a Bayer pattern (RGB) CMOS color camera. Since monochrome pixels are more sensitive than color pixels, a MonoColor sensor produces in general about 2X brighter color picture and higher luminance pixel resolution.
The implementation and spectrum response analysis of multi-finger
photogate APS pixels
Photogate APS pixels use a MOS capacitor created potential well to capture photocarriers. However, optical absorption of the poly-silicon gate reduces photon transmission. We investigate multi-fingered photogates with opening in the gate to increase photon collection. 0.18 µm CMOS standard and multi-fingered photogates were implemented where the enclosed detection area is divided by 1, 3 and 5 poly fingers. Preliminary spectral response comparison with standard photogates suggested the sensitivity of 1-finger pixels dropped ~15% implying open areas collected 70% of the photocarriers. The sensitivity of 3 and 5 finger pixels increased ~15 – 30% over standard, with open area collection ~130 – 220% of the photocarriers due to fringing field created potential wells. These results indicated at least 55% of the incident light is absorbed by the poly-silicon gate. In spectral response multi-fingered pixels showed an increase in sensitivity in the red (631 nm) – yellow (587 nm) - green (571 nm) wavelengths but a relative decline in the blue (470 nm) possibly due to more absorption in the SiN insulator layers. Some Silicon Nitride (SixNy) compositions have higher absorption coefficients in the Blue than poly-silicon and thus may dominate the absorption in these photogates structures.
Nanoplasmonic filters for image sensors
As pixels shrink in CMOS detectors, scaling effects could lead to potential issues with existing coloured filters because of their thickness (approx. 1µm).
In this paper, we propose to investigate a new generation of filters that are potentially thinner by approximately a decade.
Several years ago , nanometric metallic gratings have been found to have very unusual transmission properties. Especially, unexpected high transmission has been measured at specific wavelengths. Since these first experiments, studies have shown that very different optical processes can be responsible for these resonant transmissions, depending on the geometry of the grating.
With our designs, we show that such components are suitable for RGB colour filters.
We first discuss the theoretical performances and the integration of these components through modelling with rigorous electromagnetical techniques (RCWA and FDTD) in 2D and in 3D. As an example we evaluate the impact of the pixelization, the technological errors (over etching, presence of slopes on the gratings) and the illumination conditions on the filter performances. Thanks to an algorithm that can optimize colour correction matrix, we show that we have a satisfactory colour rendering (dE<5).
In a second part we realize these samples. We sputter Al layers on glass substrates and structure them with Focused Ion Beam technology that enables approximately 30 nm resolution. These structures are compatible in size with pixel dimensions (1.5 µm X 1.5 µ) and are tested with a dedicated micro-spectrometer.
 Ebbesen T.W. et al, Nature 391, 667-69 (1998)
Thin Color and Stop-InfraRed Metal-dielectric Filters for CMOS Image Sensors
Camera system for multispectral imaging of documents
A multispectral imaging system for documents has been constructed and used in the Library of Congress Preservation Research and Testing Division to characterize the condition, the components, and to provide baseline information for the 1507 world map by Martin Waldseemüller, the first printed map to reference “America.” It was also used to image sections of the Carta Marina 1516 map by the same cartographer for comparative purposes. The system uses a monochrome sensor with custom-built LED illumination for reflectance and transmission imaging from the near ultraviolet to the near infrared. By applying techniques used in astronomical and medical imaging, this system revealed printing techniques, preservation challenges, and obscured text and drawings to provide new insights into the cartographic, mathematical and scientific concepts and methodologies.
The integrated MegaVision Monochrome E6 camera utilizes a 39 megapixel Kodak CCD sensor array with 7216 ´ 5412 pixels of 6.8 micron pixel dimension. The images exhibit a dynamic range of 12 bits per channel. The Equipoise Imaging EurekaLight™ LED illuminators and SideLong™ raking light illuminators were used to generate images in twelve distinct spectral bands or any combination thereof. Raking illumination was provided in two spectral bands from either side of the object. A team of imaging and preservation scientists used this imaging system to produce 300dpi 8-bit TIFF images of each of 12 map sheets for all illuminations with associated metadata for each unprocessed and processed image. Higher resolution 600dpi images were also collected and stitched to create large-format images of each map sheet. In addition, 1200dpi images were collected of key areas of specific interest, e.g., paper watermarks, in both reflected and transmitted light. Infrared imaging also revealed previously hidden text. The scientific team performed multispectral image processing, including pseudocolor renderings, of images in areas of key interest on the map. This imagery highlighted the different spectral responses observed with wavelength analysis from paper, parchment, inks and colorants. In a novel imaging application, this also revealed original woodblock printing techniques.
This paper will address the development and optimization of this and similar systems for digital imaging of fragile and unique cultural objects over a range of spectral bands for the purpose of capturing important digital records to make the information accessible to researchers, professionals, and the public. Building on prior experience in the imaging of the Archimedes Palimpsest, this paper will discuss the advantages of monochrome imaging for preservation science studies, including during the imaging and post-processing. It will cite some of the challenges faced with color vs. monochrome sensors, and the enhanced information acquired with this system without the Bayer array. It will demonstrate the potential of this combination of sensor and illumination for not only preservation science, but also for other scientific studies of paper, inks, and colorants.