The characteristics of a computer controlled intensified CCD camera system are quantitatively studied. The light response of the computer controlled camera system is found to be non- linearly dependent upon the gains of the intensifier, the CCD amplifier, and the imaging board. Controlling the intensifier gain is considered the most efficient method of controlling the system's sensitivity and is the chosen working mode. A 3D analytical model of the camera's light response is generated. An inverse transformation was designed. By employing a look-up table (LUT) to map the inverse transformation using hardware in the imaging board, the non-linearity of the camera's light response at any given operating point can be corrected in real time.

We propose a two-stage linearly working integration method for low light level imaging systems. The influence of integration and background substraction on the signal to noise ratio is discussed in the paper. The introduced linear approach avoids thresholding and creates images with enhanced spatial resolution. After 5 minutes integration time the system reaches a sensitivity of 0.6 Photons/(Pixel X second). The linearity of the system can be checked with a suitable calibration device. We propose an electronic light standard with a pulsed LED, that allows us to cover a wide range of brightness by variation of the pulse frequency. Very low light intensities of luminescence can be produced in a stable and reproducible manner. The system has powerful software functions for bioluminescence applications like bacterial cloning as well as measurement functions for chemiluminescence applications within the DNA or protein blotting. Utilities for documentation, image printing, and data export to spreadsheet or graphics programs are also provided.

In contrast to existing methodologies employed for multispectral imaging based on filter wheels, prisms or gratings, the acousto-optic tunable filter (AOTF) provides significant advantages. The AOTF is digitally accessible providing wavelength selection and rapid optical switching in a compact instrument. In addition, AOTFs can be employed as multiplexed polychromators in which several wavelengths are tuned at the same time by injecting multiple rf signals into the AOTF simultaneously. We apply spectral multiplexing for Raman microscopy in which the integrated image on the CCD detector is the linear combination of scattered radiation corresponding to several vibrational modes. Hadamard transform mathematics govern the multiplexing sequence allowing efficient recovery of the spectrum. In this manner, the optical throughput of the AOTF is increased and spectral images can be collected rapidly.

This paper addresses the color medical imaging system (CMIS) program, which entails the development of a prototype system to evaluate spatial correlation techniques to convert microscopic images into full color digital electronic files. Program objectives were directed toward the creation of high resolution 2D images using spatial template matching. Full color image segments were captured using NTSC CCD array cameras. These lower resolution segments were captured in an overlapping coverage and combined at their borders as complete seamless high resolution files. The use of segmented capture overcomes the resolution limitation of the capture system, and expands the field of view of the microscope for a fixed magnification. CMIS was used to capture image segments from medical glass slide specimens using a light microscope, and to convert these segments into full color electronic image files. This paper describes the CMIS system and its image capture and conversion process.

Although high contrast between signal and the dark background is often claimed as a major advantage of fluorescence staining in cytology and cytogenetics, in practice this is not always the case and in some circumstances the inter-cellular or, in the case of metaphase preparations, the inter-chromosome background can be both brightly fluorescent and vary substantially across the slide or even across a single metaphase. Bright background results in low image contrast, making automatic detection of metaphase cells more difficult. The background correction strategy employed in automatic search must both cope with variable background and be computationally efficient. The method employed in a fluorescence metaphase finder is presented, and the compromises involved are discussed. A different set of problems arise when the analysis is aimed at accurate quantification of the fluorescence signal. Some insight into the nature of the background in the case of comparative genomic hybridization is obtained by image analysis of data obtained from experiments using cell lines with known abnormal copy numbers of particular chromosome types.

In the emerging field of biotechnology, new technologies that enable early detection of genetic aberration or cancer are becoming a reality. These methods label specific parts of the DNA of the chromosomes with fluorescence fluors. The development of the techniques, as well as the correlation of the genetic aberrations to cancer, require a scientific imaging system coupled with an advanced image analysis method. In this paper we outline the problems associated with both the imaging and image analysis. We present a coherent approach to the design of the system. The system parameters are considered simultaneously so that the acquisition process is fast, yet the image quality is good so that its analysis is reliable. A fast new technique for scanning/focusing using extrapolation and interpolation is described. Moreover, a novel approach for a robust analysis through a combination of morphological operators and statistical optimization is outlined. Experiments and results that evaluate the system performance are presented.

This paper describes continuing work with both 2-D and 3-D line-scan vision systems using rotation of both the camera and of an object to produce 2-D images. A rotating line-scan camera arrangement has been used to produce images of an area surrounding the camera stage. Due to the nature of the movement parameter these images can be arranged to cover a variable field of view in the movement axis, i.e.: from as small as 5 degree(s) up to 360 degree(s). This same arrangement has been used to produce images of a rotating object allowing a single 2-D image to contain information from the front, back and both sides of the object of interest. This 2-D system is to be extended into a full stereoscopic line-scan vision system. This will enable an investigation into extracting 3-D coordinate information from both the rotating camera and the rotating object.

Recent advances in charge-coupled device (CCD) coatings and manufacturing have resulting in CCDs sensitive to short UV wavelengths. These UV-enhanced CCDs can be used to image in the UV without the negative attributes of the intensified systems. This paper compares the characteristics of two systems: (1) a non-intensified UV-enhanced CCD sensor system, and (2) an intensified sensor system using a photocathode/micro-channel plate (MCP)/fiber optic bundle with a high quantum efficiency CCD. Both sensor systems have advantages and disadvantages in terms of noise performance; and the better choice is very much dependent on mission requirements. The results of the analysis illustrate parametrically the performance for both intensified and non-intensified sensor systems as a function of light level and range to target. Additionally, other sensor system noise parameters are discussed such as photo-cathode and CCD dark current, charge transfer efficiency, readout rate, and ADC quantization. For the analysis, a detailed computer model was constructed to account for all relevant intensifier, CCD, and electronics noise. The computer model is described, and the details of accounting for intensifier gain, photocathode noise, and CCD performance are discussed. Finally, the sensor system model is used along with atmospheric transmission predictions to estimate the NET of the UV sensor system (both intensified and non-intensified) operating through the atmosphere where UV attenuation is quite high. The predictions are valid against a wide range of remote sensing requirements, and the computer model constructed is generic with respect to remote sensing missions.

Demand for high frame rate operations using precision slow scan CCD cameras increases. In an overview, different commonly used readout techniques are described. CCD timing is then analyzed in order to find the most limiting factors for high frame rates. Different approaches to clocking and their impact on frame timing are discussed and compared. Parameters such as sub-array dimensions, parallel shift times, and pixel read rate are compared and the tradeoffs are discussed. Finally some approaches are shown for reading sub-arrays with frame rates higher than 300 frames per second.

The purpose of this paper is to explain the applications of automatic exposure adjusting apparatus for a digital still camera. Suppose one image has 512 X 512 pixels and the gray levels of each pixel range from 0 to 255, it will result in 256^{512 X 512} kinds of images. The next question is: how can you differentiate the proper luminance from these images? The fuzzy c-means algorithm of the fuzzy pattern recognition is used. When the degree of the membership function, representing each element in the eight orthogonal vector spaces calculated by variance criterion, is equal to 0.5, it represents neither light images nor dark images and is the proper luminance. The equation `the degree of the membership function equal to 0.5' is derived. Applying the third form of the fuzzy control involves using this equation to differentiate proper luminance of images from all images. The experimental data demonstrate satisfactory results.

This paper describes the requirements, design, and results of a modular data acquisition system with a resolution of 12 bits at up to 20 MHz sampling frequency. The modularity enables the analog-digital conversion to be separated from the digital processing/storage. This allows the latest, best performing ADC (analog-digital converter) to be easily integrated into the system by a re-design of only the AD board with the rest of the system unchanged. The converter employed operates at frequencies up to 20 MHz. The complete system produces measured quantization noise figures of -75 dB and integral non-linearity of -72 dB. The unit can sample video or non-video waveforms. For video applications, an active clamping system is used to ensure that the black level is accurately maintained. The framestore is connected externally using a high-speed digital data bus. This facilitates the inclusion of real-time digital processing units. The framestore used is doubly buffered to permit simultaneous acquisition and readout. The store is 8 Mbytes to accommodate HDTV images and has an input data rate of 40 Mbytes per second.

In an effort to realize a compact HDTV camera with high performance, we have developed a prototype equipped with four 2/3-inch CCDs. A smaller image format is preferable for downsizing TV cameras. However, this causes shrinkage of the unit pixel size and inevitably makes it more difficult to produce an image-pickup device with required HDTV qualities, especially sensitivity, and dynamic range. We have overcome this problem by using CCD imagers with high performance but with a relatively small number of pixels and by increasing the number of CCD chips used in a camera to secure the necessary spatial sampling points for HDTV. In the newly developed color-separating system of the camera, two of the four CCDs are assigned for the green (G) light component and one each for red (R) and blue (B). We succeeded in improving the resolution by introducing spatial pixel offset imaging. This new method has two major advantages: it prevents resolution degradation caused by chromatic aberration and improves the resolution of colored signals over a wide range.

A digital camera has been designed based on a 1317 by 1035 pixel, full frame, CCD which offers the large area, high density sampling but with the options of programmable sub-area scanning and multiple independent x and y binning and programmable gain and offset. The sub-area scanning and binning can be used to increase the effective frame rate of the device for such applications as fast focusing and object location. The binning also provides greater sensitivity at the cost of spatial resolution, but is ideal in low light level applications such as fluorescence microscopy. The programmability of this camera permits the switching between the various clocking modes within a single frame time. Hence, the image can be auto-focused in sub-area scanning mode and then within a single frame time the full frame image can be acquired. The programmable clocking modes make this device ideal for quantitative imaging applications requiring a high throughput and the programmable gain and offset allow the user to fine tune the device for his specific needs. A description of the programmable microimager design and some initial measurements of noise, linearity, frequency response, and stability are presented.

This paper describes the configuration of a super high definition (SHD) motion picture film digitizing system and performance evaluation results. A still/motion picture film digitizing system is developed using a CCD line image digitizer, which can capture images at a maximum resolution of 4096 X 4096 pixels, 12 bits/pixel. Motion picture sequences are obtained on a frame by frame basis from a 70 mm motion picture film with the aperture size of 57 X 57 mm. Several versions of SHD motion picture test sequences were so digitized and the frame to frame positioning error was within +/- 1 pixel/2048 pixels. An SHD image processing software environment was newly developed to provide all of the image processing functions needed in the digitizing system and a comfortable user interface. A color matching method, which is based on direct color measurement/approximation between the original image and the reproduced image, was proposed for the digitizing system. A color matching experiment was performed between a film projector and a CRT display. The improvement ratio of mean squared color distance was 1.30. Although the improvement ratio is relatively small, a noticeable improvement in color reproduction accuracy was observed.

With the use of polycrystalline cadmium telluride (CdTe) film as the photoconductive target a 1' x-ray imaging camera tube (vidicon) was developed. This vidicon yields an x ray signal current of 200 nA cm^-2 for 2.58 mc kg^-1 min^-1 radiation. This is a higher value when compared with that of conventional x ray vidicon (45 nA cm^-2) which uses PbO as the photoconductive target. At the target voltage of 35 V a dark current of 5 nA cm^-2 is generated in the developed vidicon at room temperature. When this vidicon is used with a diode mode electron gun and a mixed field type deflection electrode, a resolution of more than 15 micrometers can be obtained. Rf sputtering method produces large and uniform films, therefore, 1.5' films could be easily fabricated. The vidicon developed with 1.5' CdTe film yields better shading characteristics, a wide field of vision which is twice as large, and a higher signal current compared to those of 1' type vidicon.

A zoom lens system using gradient-index lenses was studied to realize a compact video camera. This study investigates the relationship between the total index change and the amount of aberration correction in order to effectively apply existing gradient-index lenses to the zoom lenses. In addition, it examines how the effective dispersion of the gradient affects the chromatic aberrations while zooming. A zoom lens system consisting of a reduced number of lens elements using axial and radial gradient-index lenses is presented that can match the optical performance of a homogeneous lens system.

The quarter megabuck cost of VGA-resolution infrared focal plane array cameras is sufficient grounds to look at alternative designs for high-speed infrared imaging systems. In addition, one has to deal with flawed sensors. Using linear models and linear thinking, the state of the technology in optical design has apparently crested. The existence of 6.5-to-1 lossless image compression for infrared is sufficient to establish that sufficient redundancy exists to reconstruct the flawed data. We are convinced that most real-time imaging systems could be improved with real-time morphological coprocessing. By building in additional redundancy using a 768 X 3 linear infrared sensor, we have designed an infrared camera system with a single rotating mirror which produces better pictures than existing focal plane array systems with comparable data rates to the existing focal-plane array systems, and one tenth the electronics of existing focal-plane array systems. Flaws are removed by morphological processing, converting redundant linear scans of the 768 X 3 sensor to a 768 X 1 linear array. As the mirror turns, a full 2-D image is recovered, producing a super-VGA (1024 X 768) resolution image.

A high-performance image processing system has been developed to provide 2-D image acquisition for a variety of optical applications. The system is built around two high- performance visible-light charged coupled device (CCD) arrays. The arrays and associated circuitry were designed to operate at high frame rates, with an optical power dynamic range of over 30 dB. The 16-port, 512 X 512 pixel array is designed to operate at over 800 frames per second; the 32-port, 1024 X 1024 pixel array, at over 300 frames per second. The system has two main units: a stand-alone high-performance camera array module (HPCAM) and a VME-based image capture and display system (ICADS). HPCAM contains a CCD array and associated drive and control electronics. ICADS consists of an analog multiplexer, digitizer, digital frame buffer-integrator, main central processing unit board with user and Ethernet interfaces, graphics driver, and high-resolution monitor. The CCD arrays used in this work are designed for use in optical processing systems. This system provides a method for capturing and displaying CCD array output data, giving the researcher a method to analyze and develop CCD array and optical processing systems.

The commercial confocal laser scanning microscope (CLSM) has made 3D imaging with submicron resolution broadly available. CLSM combines focused illumination and spatially filtered detection to reject out-of-focus light, yielding images of thin optical sections within thick specimens. Images taken at different focal planes are combined to form a 3D reconstruction. Using this technique, fluorescent biological indicators are now quantified within volumes of approximately 0.1 micrometers ³. Live cell fluorescence imaging is limited by photodynamic properties of the fluorophores. Finite fluorophore excited state lifetimes limit imaging speed, and multiple spots in the specimen must be illuminated to increase temporal resolution. However, multiple illuminated spots decrease 3D image quality. The trade-off between imaging speed and quality is presented for several microscope designs. Fluorophore photobleaching limits the total signal available for 3D imaging, and photobleaching in out-of- focus planes limits the quality of 3D reconstructions. Femtosecond laser pulses focused to a diffraction limited spot can excite fluorophores by a 2-photon absorption that occurs only in the focal spot. Two-photon excitation limits photobleaching to the plane of focus and allows extended viewing of thick samples. Two-photon absorption can also locally release biologically active `caged' molecules and indicators.

The electro-optical performance characteristics of commercial camera products are specified using a variety of different terms, metrics, and measurement methods. These terms and metrics are often ambiguously defined, presented along with unclear or incomplete descriptions of the testing and measurement methods, and with little or no commonality among camera vendors with respect to the meaning and usage of technical terms. In addition, many important imaging characteristics of cameras are omitted entirely from vendor specifications. As a result, many camera product specifications are barely useful to the inquiring engineer, making comparative camera shopping difficult. While the basic questions of correlation between objective image quality measures and subjective judgments have not been completely answered, the value and utility of camera specifications can be enhanced significantly through establishment of an accepted set of metrics, test methods, and technical terms.

In this paper we present methods for characterizing CCD cameras. Interesting properties are linearity of photometric response, signal-to-noise ratio, sensitivity, dark current, and spatial frequency response. The techniques to characterize CCD cameras are carefully designed to assist one in selecting a camera to solve a certain problem. The methods described were applied to a variety of cameras: an Astromed TE3/A with P86000 chip, a Photometrics CC200 series with Thompson chip TH7882, a Photometrics CC200 series with Kodak chip KAF1400, a Xillix' Micro Imager 1400 with Kodak chip KAF1400, an HCS MXR CCD with a Philips chip and a Sony XC-77RRCE.

This paper describes preliminary work on a real-time neutron radiography imaging system. Initially, the system provided neutron radiographs on photographic film using a gadolinium foil converter. The photos were digitized by a flat scanner. The system implements digital image processing functions for: counting and sizing, measurement, analysis, enhancement, geometric operations and scripting. After the development of these functions, experiments were done on the real-time imaging system and the functions were applied on images captured by a CCD camera. This is pioneer work related to neutron radiography in Brazil. This project is particularly original because our nuclear reactor has unique features: its core lies in a 6 m pool and the neutrons must be driven to the surface by a pipe. Another challenge of this project is the elimination of any light intensifier between the converter screen and the CCD camera. Neutron converter deficiencies have been compensated by applying digital image techniques to the neutron images. A low cost system has been achieved by exploring the benefits of digital image processing. Initial results provided by this system have encouraged us to continue the research and to enlarge its capability.

This paper describes the design of a 10 bit, 23,000 frame per second camera with digital control. The camera is based around a 64-tap 256 X 128 element interline transfer CCD with a current mode output. Readout is at 12 MHz per tap for an aggregate data rate of 768 megapixels per second. Sixteen MHz operation (30,000 frames per second) is possible with reduced performance. Digitization is to 10 bits with 8.5 bits rms effective. The signal processing chain allows for digital control of the analog gain and features a feedback loop to maintain offset stability. Linearity of the processing chain is 0.2%. Additional features included digital control of the integration time and readout rate, and digital compensation of light level. Digitized data is transmitted over high speed serial links to a remote rack that emulates the target processing system. An automated analysis system is able to exercise the system and measure performance characteristics.