Two kinds of electron bombardment CCD (EB-CCD) camera are newly developed, employing an EB-CCD sensor made by Hamamatsu Photonics. The slow scan cooled CCD camera installs the full frame transfer type EB-CCD sensor with 512 X 512 pixel format and standard video rate camera installs the frame transfer type EB-CCD sensor with 658 X 490 pixel format. For slow scan camera, EB-CCD sensor is cooled down to -25 degree C to realize high sensitivity utilizing long exposure time and very low noise performance. And instead of mechanical shutter, high speed electrical gated function is installed in this camera. Standard video rate camera has a function of recursive filter and auto background subtraction as a standard in order to get high quality image. Furthermore, camera can select both inter- laced scan and progressive scan readout mode. These cameras have been compared with the other high sensitive camera like SIT camera, photon counting camera and I-CCD camera which results in proving that these EB-CCD camera are suitable and ideal.

Adaptive optics applications require a camera and detector that are capable of very high frame rates, high sensitivity and low noise. The PixelVision Adapt III camera achieves over 85 percent quantum efficiency and 12 electrons read noise at 1500 fps using the Adapt III, back illuminated CCD, designed by PixelVision, Inc. and manufactured by Scientific Imaging Technologies, Inc. The CCD's imaging area consists of 80 X 80 36 micrometers square pixels and utilizes full frame transfer architecture with an additional 80 rows in the storage region. To achieve low noise at high speeds, every two columns are mixed together into one output amplifier. There are 40 output amplifiers on the CCD and 40 analog channels. Digital data is mixed to a series stream and is transmitted over fiber optic cables to a PCI bus data acquisition board. Windows95 software drivers allow continuous acquisition of data into system memory. A water- cooled housing eliminates turbulence created by forced air cooling.

A color line scan camera family which is available with either 6000, 8000 or 10000 pixels/color channel, utilizes off-the-shelf lenses, interfaces with currently available frame grabbers, includes on-board pixel by pixel offset correction, and is configurable and controllable via RS232 serial port for computer controlled or stand alone operation is described in this paper. This line scan camera is based on an available 8000 element monochrome line scan camera designed by AOA for OEM use. The new color version includes improvements such as better packaging and additional user features which make the camera easier to use. The heart of the camera is a tri-linear CCD sensor with on-chip color balancing for maximum accuracy and pinned photodiodes for low lag response. Each color channel is digitized to 12 bits and all three channels are multiplexed together so that the resulting camera output video is either a 12 or 8 bit data stream at a rate of up to 24Megpixels/sec. Conversion from 12 to 8 bit, or user-defined gamma, is accomplished by on board user-defined video look up tables. The camera has two user-selectable operating modes; lows speed, high sensitivity mode or high speed, reduced sensitivity mode. The intended uses of the camera include industrial inspection, digital archiving, document scanning, and graphic arts applications.

To provide real-time imaging, x-ray diagnostic imaging relies entirely on the combination of the x-ray image intensifier and the high-performance television camera. Although these devices have been pushed to remarkable degrees of performance, they remain complex electro-optical assemblies with significant built-in errors, instabilities and degradation mechanisms. We describe a replacement for these systems utilizing as a sensor a large array of amorphous silicon photodiodes and thin-film switching transistors. Specially, the equipment described is a replacement for a 9-inch dual-mode x-ray image intensifier with a high-performance 2000-line digital tv camera capable of operating in both real-time video and high-performance spotfilm modes.

The advantages of digital photography are well documented, and digital photography is seeing increased use in demanding photography applications. Of the many implementations of digital cameras, the three-CCD camera provides the optimal resolution, temporal sampling, and color reproduction, which when examined together form the information content of the sensor - a physical measure of the detector's imaging performance. So that uncertainty between the object of interest and the reproduced image is reduced, high quality, precise, color photographic records require sensors with the highest possible information handling capability. Furthermore, high information content images, due to the increases information about the scene, can be compressed to smaller file sizes than can lower fidelity images -thus, allowing reduced transmission data rates. Whereas traditional film is hypersensitive in the blue, conventional CCD imagers have reduced blue response as compared to their red response. A class of CCD called the back-illuminated CCD has uniform spectral response throughout the visible, UV, and NIR spectral regions. By integrating the uniform spectral response of the back-illuminated CCD with ultra-low noise amplifiers, high dynamic range pixels, a high pixel density, a large area detector, and a 3-CCD color prism architecture, a nearly ideal digital camera can be realized. This paper discusses a development effort at PixelVision Inc. to realize a nearly ideal color digital camera. So that a system for evaluating solid state photographic imagers can be established, a methodology of determining the information content of an imager is introduced and the information content of a back-illuminated CCD camera is compared to conventional film and to currently available studio digital cameras.

The expanding use of CCD sensor based, digital cameras necessitates the development of a reliable method of determining an equivalent ISO speed for such cameras. This paper will briefly describe the physical differences between image capture, and the signal and noise properties of film and CCD based imaging systems. The model first addresses the ISO speed of a monochrome CCD sensor. The results of this model indicate that an equivalent ISO speed can be found when the exposure, E_m, in Lux-seconds at a signal-to- noise ration of 30 is used to calculate the speed from the expression developed for film based systems, ISO speed equals 0.8/E_m. The model shows that calibration exposures of 0.01 seconds or less must be used to eliminate the effect of dark current on determining the 'correct' ISO speed. The model is then sued to calculate the effective ISO speed for a series of CCD sensors for different pixel areas, nose characteristics and quantum efficiencies. The same model is used for color cameras by introducing a color filter. The model is also used to determine what set of color filters will provide a balanced exposure for a given light source and what electronic processing must be done to compensate for different equivalent ISO speeds between the color filter elements. The effect on ISO speed and color balance as a function of illuminant is considered. The model is also used to map the signal-to-noise characteristic of the CCD camera under usage equivalent to a film based camera.

The sensitometric dynamic range of a silicon-based CCD imager is generally based on the ratio of the array element's well capacity to store electrons and the readout noise level of the imager. The readout noise level is analogous with the base+fog level of film, and exceeding the capacity of a pixel's well generates excess electrons resulting in overexposure of the image. Therefore, CCD imaging cameras exhibit S-shaped H and D curves with range limiting toes and shoulders similar to silver halide based film and paper. It is generally accepted that to capture all the tomes of a typical subject image, the capturing imager should have a minimum dynamic range of 7 f-stops. Current professional film and paper support this 7 f-stop requirement, with some films capable of 8-9 f-stops of tonal dynamic range. This paper describes a recent research project conducted at the Institute to characterize the sensitometric response of several DCS-class professional CCD digital imaging cameras. Cameras tested included to Eastman Kodak DCS 420, DCS 460, and the Canon EOS-DCS 3.

This paper investigates the complexities associated with the development of next generation digital cameras due to requirements in connectivity and interoperability. Each successive generation of digital camera improves drastically in cost, performance, resolution, image quality and interoperability features. This is being accomplished by advancements in a number of areas: research, silicon, standards, etc. As the capabilities of these cameras increase, so do the requirements for both hardware and software. Today, there are two single chip camera solutions in the market including the Motorola MPC 823 and LSI DCAM- 101. Real time constraints for a digital camera may be defined by the maximum time allowable between capture of images. Constraints in the design of an embedded digital camera include processor architecture, memory, processing speed and the real-time operating systems. This paper will present the LSI DCAM-101, a single-chip digital camera solution. It will present an overview of the architecture and the challenges in hardware and software for supporting streaming video in such a complex device. Issues presented include the development of the data flow software architecture, testing and integration on this complex silicon device. The strategy for optimizing performance on the architecture will also be presented.

Digital cameras as a class of digital imaging devices are gaining in popularity. Many digital imaging companies now have products in the market ranging from the low end consumer devices to high end professional digital cameras. As with any other product, different classes of the product brings a variety of challenges from a tecimological standpoint. We will lay the foundation with a generic image processing architecture for digital cameras and build on that to identify distinctions in the various classes of cameras. One point of interest is that most manufacturers have treated digital cameras as embedded devices as opposed imaging devices. This offers image scientists the ability to implement a variety of processing steps in photo capture that were not possible before. For example, histogram equalization can be performed in camera as opposed to a PC with the advent of cheap, low-cost computing horsepower embedded in the device. At the highest level, a digital camera captures an image, processes it, and either compresses it for storage or displays it. Today's digital cameras implement little more that basic processing and compression. Some of this is because of the time pressures on the development teams and some because most cameras use either VGA (640x480) or 1/2 VGA (320x240) CCD sensors today (at the low end) where image quality is an issue even from the standpoint of raw images. Remember that these images are often captures with a mosaic array and the effective color resolution is lesser than the raw ccd pixel resolution.

In this paper, we propose a method of region based histogram specification resulting dynamic range expansion effect. Main goal of this paper is enhancement of the image captured in back light condition, in which the captured image contains tow extremely contrasted regions simultaneously. The proposed method has two steps. The first is region segmentation in the spatial domain by using threshold values derived from the histogram domain. The second step is execution of independent histogram modification for the segmented regions. The histogram modification for each region is similar to typical histogram equalization except the interval to be histogram equalized. The resulting regionally histogram specified image can be obtained by combining the results of the previous steps. The proposed segmentation algorithm has an advantage in computational cost in that it can be performed with simple neighborhood operation by using the thresholding intensity level derived fast from histogram domina. The proposed algorithm shows better result than the typical histogram equalization by referring to specific region of interest to be enhanced.

Higher performance and more sophisticated functions have been required in the digital still camera market growing at the rapid rate. On the other hand, the hardware cost has been increasing to meet these requirements. In light of this situation, a new image processing technique is presented to achieve the digital still cameras with the higher performance at lower cost. The present technique focuses on the real time continuous recording of single CCD images using the fixed block truncation coding (FBTC). FBTC is an embedded compression scheme we proposed to handle the image data with high quality in the hardcopy peripherals. In the digital still camera using the present technique, the continuous recording rate is up to 30 frames per second for video graphic array size full color image. Moreover, FBTC coding scheme can reduce the memory size to capture the raw image from CCD by less than a half, and enable to store over twice continuous recording frames than the conventional technique.

The digital camera market is growing at an explosive rate. At the same time, the quality of photographs printed on ink- jet printers continues to improve. Most of the consumer cameras are designed with the monitor as the target output device and ont the printer. When a user is printing his images from a camera, he/she needs to optimize the camera and printer combination in order to maximize image quality. We describe the details of one such method for improving image quality using a AGFA digital camera and an ink jet printer combination. Using Adobe PhotoShop, we generated optimum red, green and blue transfer curves that match the scene content to the printers output capabilities. Application of these curves to the original digital image resulted in a print with more shadow detail, no loss of highlight detail, a smoother tone scale, and more saturated colors. The image also exhibited an improved tonal scale and visually more pleasing images than those captured and printed without any 'correction'. While we report the results for one camera-printer combination we tested this technique on numbers digital cameras and printer combinations and in each case produced a better looking image. We also discuss the problems we encountered in implementing this technique.

The authors have newly developed measuring methods to assess and characterize color reproduction of digital still cameras and digital video cameras which capture color images and output corresponding color information in red, green, blue digital image data, the proposed new methods incorporate spectral responsivity characteristics measurements and other characteristics measurements. They are made possible by using newly developed test charts with a dark box and a spectral light source. The proposed set up for the measurements is specially effective to eliminate the influence of automatic functions commonly equipped in digital cameras for low-end consumer use. In this paper, arrangement of measurement equipment, definition of test charts, and raw data handling are described together with worked examples.

We have developed a color reproduction software for a digital still camera. The image taken by the camera was colorimetrically reproduced on the monitor after characterizing the camera and the monitor, and color matching between two devices. The reproduction was performed at three levels; level processing, gamma correction, and color transformation. The image contrast was increased after the level processing adjusting the level of dark and bright portions of the image. The relationship between the level processed digital values and the measured luminance values of test gray samples was calculated, and the gamma of the camera was obtained. The method for getting the unknown monitor gamma was proposed. As a result, the level processed values were adjusted by the look-up table created by the camera and the monitor gamma correction. For a color transformation matrix for the camera, 3 by 3 or 3 by 4 matrix was used, which was calculated by the regression between the gamma corrected values and the measured tristimulus values of each test color samples the various reproduced images were displayed on the dialogue box implemented in our software, which were generated according to four illuminations for the camera and three color temperatures for the monitor. An user can easily choose he best reproduced image comparing each others.

As an imaging dye absorbs a single photon, there exists a non-zero probability that the dye will bleach as the result of a photochemically induced reaction. Two possible mechanisms, which can cause this photochemical event, are radical formation and/or radiative energy transfer. A photophysical investigation was completed in order to identify a dominate mechanism governing the lightfast properties of imaging inks. Time and spatial domain spectroscopic studies were conducted on an azo class dye, Cibacron Brilliant Red 3B-A, and Rhodamine 6G bound to humic acid - polymer phenol with known radiative and binding characteristics to polyaromatics. Stern-Volmer analysis illustrated that Rhodamine 6G binds strongly to humic acid with a binding constant of approximately 3000 L/g - C, as opposed to, Brilliant Red which binds about 100 times less strongly to the polymer additive. Both dyes were found to bind statically to humic acid. Accelerated photo degradation studies revealed that dye bound to humic acid bleached at a rate proportional to the binding association of each dye. Time resolved fluorescence decay results indicated this bleaching is not the result of polaron or long distance radiative energy transfer. Radical quenching studies suggest that the primary mechanism involves a ground state reaction with a radical intermediate.

An intensified/shuttered cooled PC-based CCD camera system was designed and successfully fielded on proton radiography experiments at the Los Alamos National Laboratory ALNSCE facility using 800-MeV protons. The four camera detector system used front-illuminated full-frame CCD arrays fiber optically coupled to either 25-mm diameter planar diode or microchannel plate image intensifiers which provided optical shuttering for time resolved imaging of shock propagation in high explosives. The intensifiers also provided wavelength shifting and optical gain. Typical sequences consisting of four images corresponding to consecutive exposures of about 500 ns duration for 40-ns proton burst images separated by approximately 1 microsecond were taken during the radiography experiments. Camera design goals and measured performance characteristics including resolution, dynamic range, responsivity, system detection quantum efficiency, and signal-to-noise will be discussed.

The design and application of a digital high sped image data capturing system with a following image processing system applied to the Bremer Hochschul Hyperschallkanal BHHK is the content of this presentation. It is also the result of the cooperation between the departments aerodynamic and image processing at the ZARM-institute at the Drop Tower of Brennen. Similar systems are used by the combustion working group at ZARM and other external project partners. The BHHK, camera- and image storage system as well as the personal computer based image processing software are described next. Some examples of images taken at the BHHK are shown to illustrate the application. The new and very user-friendly Windows 32-bit system is capable to capture all camera data with a maximum pixel clock of 43 MHz and to process complete sequences of images in one step by using only one comfortable program.

The paper presents the achievements in research and development of the compact optical CCD based guiding system. The ultimate goal of the project is a modest size, low mass and rugged systems to be applied for sub-arc second optical ground-space guiding and tracking. The system includes the optics, the CCD sensor with the readout and an image- processing algorithm. The optics consist of a diffraction limited objective, four-element lens system with eh effective input aperture 90 millimeters. The objective focal length 300 mm is extended by the additional relay optics. The resulting effective focal length is 3000 millimeters, the focal spot size is 65 micrometers Airy disc diameter. The combination of the diffraction limited objective design, focal extender and mechanical construction permitted to keep the overall length bellow 600 millimeters and the total mass bellow 5 kilograms while maintaining high ruggedness at one arc-second level. A sensor, the Texas Instrument CCD chip 192 X 164 pixels, 15 micrometers size is used. the custom designed readout and data processing hardware has been developed. Parallel communication maintains image download time 0.6 second with 12 bits amplitude resolution. The data acquisition and image processing software package running under MS Windows 95 or NT provide all functions for the camera control, data acquisition and image processing for precise target position evaluation. The position is evaluated as the center of mass of square neighborhood of the brightest CCD pixel. Indoor test of the ultimate position resolution using different diffraction limited images and sizes are described. The image position resolution +/- 0.03 pixel has been achieved. It corresponds to 0.03 arc seconds of angular resolution of the entire guiding sensor.

Imaging spectrometry has mainly been a research tool, employing laboratory spectrographs and scientific cameras. This paper describes an add-on imaging spectrography that provides a unique combination of high quality image in a small, rugged, industrial, easy-to-use component. The spectrograph is based on a prism/grating/prism dispersing element which provides straight optical axis, astigmatism free image and polarization independent throughput. A volume holographic transmission grating is used for high efficiency. The tubular optomechanical construction of the spectrography is stable and small, D30 X L110 mm with F/2.8 numerical aperture and 2/3 inch image size. Equipped with C-mounts, the spectrography plugs between lens and area camera, converting the camera to a spectral line imaging system. The spectrograph allows the utilization of rapidly developing monochrome camera techniques, like high speed digital cameras, smart cameras and CMOS sensors, in color and spectral analytical applications. It is the first component available for upgrading existing industrial monochrome vision systems with color/spectral capability without the need to change the basic platform hardware and software. The spectrograph brings the accuracy of spectral colorimetry to industrial vision and overcomes the complex calibration that is needed when an RGB color camera is applied to colorimetric applications. Other applications include NIR imaging, spectral microscopy, multichannel fiberoptics spectroscopy and remote sensing.

A solar blind UV bandpass filter technology has been developed which is appropriate for use with all types of UV imaging devices under full daylight conditions. This technology is based on dye doped polymer films that provide the critical solar UV blocking and the required sharp rejection slope,due to their intense absorption bands in specific portions of the UV. These films are combined with other components that block visible and near IR light to form selective and very efficient sun blocking UV bandpass filters. This technology is very flexible and permits tailoring of the filter transmission and its blocking characteristics. Solar blind bandpass filters based on this technology have been produced for use with MCP imagers with CsTe and RbTe photocathodes. Filter characteristics include 10-25 percent peak transmission, bandwidths of 16-22 nm and out of band blocking levels in excess of 12 OD. Using these filters, under direct midday summer sun, background signal levels less than 10 photons per second were obtained, when integrate dover the entire field of view of the imager. Under development is a filter for use with back illuminated CCDs. Target specifications include 15 percent peak transmission, 20 nm bandwidth and substantial blocking of solar radiation from 290 to 1100 nm. Feasibility for CCD imaging in the solar blind range was proven with a BCCD camera.