Portable electronic still cameras have been available for some years although, in general, the image quality has fallen short of the 35 mm film quality benchmark. To obtain improved image capture and reproduction, higher resolution CCD imagers with wider dynamic ranges must be employed in these cameras. A camera system composed of a 35 mm Nikon F3 camera body, a camera back containing a CCD-imager, and a portable hard drive for storing digitized images was constructed and employed to acquire images distributed over photospace. This paper describes the camera's hardware and capabilities, the software post-processing, the camera's characteristics, and a method for evaluating the camera's performance.
This paper describes two electronic camera systems which use a common CCD sensor but in different modes. The first system is a black and white Still/Video System and the second system is a color Still-Still/Video System. Although both systems are distinctly different in their modes of operation, they use many of the same components to achieve their goals.
The maturing camera industry has been employing a RS 170 formatted Image Intensified CCD Camera. This left some users with limited image resolutions. In a typical RS 170 camera the CCD sensor's resolving pixel density is in the order of 500 X 400 squared pixels. To surmount this limitation for a particular user, a high resolution image intensified CCD camera was developed. This new high resolution camera's active image plane is 1096 X 961 squared pixels. It was designed to operate from overcast starlight to daylight. Other requirements imposed upon the design were the form factor and the camera size. These requirements have lead to a small and uniquely shaped camera. Although the camera was designed to meet a specific customer's needs, rather than the RS 343 specifications, the camera is compatible with the RS 343 format.
Electronic cameras provide near real time image evaluation with the benefits of digital storage methods for rapid transmission or computer processing and enhancement of images. But how does the image quality of their images compare to that of conventional film? A standard Nikon F-3TM 35 mm SLR camera was transformed into an electro-optical camera by replacing the film back with Kodak's KAF-1400V (or KAF-1300L) megapixel CCD array detector back and a processing accessory. Images taken with these Kodak electronic cameras were compared to those using conventional films and to several still video cameras. Quantitative and qualitative methods were used to compare images from these camera systems. Images captured on conventional video analog systems provide a maximum of 450 - 500 TV lines of resolution depending upon the camera resolution, storage method, and viewing system resolution. The Kodak Professional Digital Camera SystemTM exceeded this resolution and more closely approached that of film.
Typical one-chip color cameras use analog video processing circuits. An improved digital camera architecture has been developed using a dual-slope A/D conversion technique, and two full custom CMOS digital video processing ICs, the 'CFA processor' and the 'RGB post- processor.' The system uses a 768 X 484 active element interline transfer CCD with a new 'field-staggered 3G' color filter pattern and a 'lenslet' overlay, which doubles the sensitivity of the camera. The digital camera design offers improved image quality, reliability, and manufacturability, while meeting aggressive size, power, and cost constraints. The CFA processor digital VLSI chip includes color filter interpolation processing, an optical black clamp, defect correction, white balance, and gain control. The RGB post-processor digital IC includes a color correction matrix, gamma correction, two-dimensional edge-enhancement, and circuits to control the black balance, lens aperture, and focus.
A greater number of factors determine lens suitability for electronic cameras than for conventional photography. Included in this paper are a description of characteristics to be evaluated, standard techniques for measuring these characteristics, and a comparison of the suitability of three lenses for a sample system. The characteristics to be evaluated include general lens performance criteria, specific system performance criteria, and criteria derived from features of the image sensor.
Multichannel detectors offer potential advantages over single-channel detectors in analytical plasma emission spectroscopy because of the wealth of information available in the spectrum. Simultaneous detection over a wide wavelength range provides information which can be used to improve accuracy, precision, and reliability. The very large dynamic range of plasma sources exceeds the dynamic range of common array detectors however, so means for changing the integration time between faint and intense spectral lines are required. To be successful, this approach requires that saturated areas of the detector do not spill photogenerated charge into adjacent areas (bloom). The application of a recently introduced, large-format, anti-blooming CCD to atomic emission spectroscopy is described. The results of the characterization of the performance parameters of the camera system relevant to analytical emission spectroscopy are presented. The camera system is used with a custom-built spectrometer based on an echelle grating and a CaF2 prism. The performance of the spectrometer, as well as design trade-offs, is described. Examples of emission spectra recorded with the instrument and a discussion of qualitative and quantitative analysis are presented.
A camera system suitable for microholography has been constructed, tested, and applied to the imaging of biological materials. The design of this instrument is compatible with operation over a very wide spectral range spanning from visible to x-ray wavelengths. In order to evaluate its properties, visible light Fourier transform microholograms of biological samples and other test targets have been recorded and digitally reconstructed using a glycerol microdrop as a reference wave scatterer. Current results give a resolution of approximately 4 (lambda) with (lambda) equals 514.5 nm.
The acquisition of high resolution, large area, microscope images is a crucial aspect of image cytometry. A unique quantitative visible light microscope was developed for this purpose, by positioning a high density, charge-coupled device (CCD) in the primary image plane of a chromatic aberration free, flat field objective lens. The CCD clocking and output circuits of this device were designed to minimize dark currents, to allow 10 bit digitization, 500 kHz to 8 MHz readout rates and variable integration time. A circuit was also implemented to compress the CCD pixel array on the chip by a factor of four. This pixel additive circuit results in an approximate fourfold increase in signal strength, increased frame rates and smaller amounts of data while maintaining the large field of view. This is particularly useful in: (1) detecting images at the low light levels encountered during fluorescence imaging; and (2) searching for objects while screening a microscope slide. The geometric and spectral properties of the bright-field optics were also optimized for quantitative CCD imaging by providing homogeneous illumination over the entire field of view and minimizing chromatic and geometric aberrations. With this device, bright-field images of a 0.23 micron diatom striation pattern and fluorescence images of 7 X 103 molecules of equivalent soluble fluorochrome (MESF) fluorescein-labeled intensity beads have been acquired and digitized.
The use of a cooled CCD camera for confocal light microscopy provides several advantages. For low-light level applications such as reflected light microscopy from semi-transparent biological samples slow scan cooled CCD imaging devices provides an image with high dynamic range, linear response, and geometric stability of the pixels. These characteristics are required for accurate quantitative microscopy. The only cases where these detectors are unsuitable is when the temporal response of the image exceeds the readout time of the camera. For single shot events, even this limitation can be overcome with a partially masked CCD. The required spatial resolution of the object in the image plane is related to the spatial resolution of the camera detector. The detector should have twice the highest spatial resolution of the camera detector. The detector should have twice the highest spatial resolution of the object. In order to demonstrate the utility and application of a cooled CCD camera in microscopy, we have coupled a Photometrics cooled CCD camera containing a Tektronix 512, thinned, back illuminated CCD to a Technical Instruments, K2Bio confocal microscope. The high quality of the confocal images demonstrates the properties of the cooled CCD camera as a suitable detector for quantitative confocal microscopic imaging of weakly reflecting biological samples.
Image processing in biomedical applications such as the analysis of electrophoresis gels, digital microscopic imaging or the computer-assisted quantitative analysis of angiographic images recorded on 35 mm cinefilm quantitative coronary arteriography (QCA) can be improved substantially if the conventional TV-based image digitizers are replaced by devices offering a higher geometric resolution and an increased dynamic range. Before high resolution two-dimensional CCD Sensors were introduced a few years ago, these improvements could only be implemented cost-effectively by using scanning devices such as the digital camera developed earlier at the Institute for Biomedical Engineering in Zurich, Switzerland. This camera, based on a 2048-element CCD line array, offers a geometrical resolution of up to 2048 X 3000 pixel and a dynamic range of up to 12 bit. It is being used in the QCA system developed jointly by the Institute for Biomedical Engineering in Zurich, Switzerland and the University of Texas Health Science Center in Houston, Texas, U.S.A. Clinical evaluation of this system as well as an analysis of the technical properties of the X-Ray systems involved indicate that coronary arteriograms can be digitized with a resolution of about 1024 X 1024 pixel without sacrificing measurement accuracy. This fact reduces computational effort and suggests the use of two-dimensional CCD sensors. Therefore, a new digitizing camera based on a KAF1400 (Kodak) full frame sensor suitable for the QCA system has been developed. Its design concept and performance are discussed in the paper.
A lung imaging fluorescence endoscope has been developed which can be used for detection and localization of early lung cancer. We exploited tissue autofluorescence alone or in combination with fluorescent tumor localizing drugs to create pseudo images which can clearly delineate the diseased sites from the surrounding normal tissues. With this technique it is possible to detect early lung cancer as well as pre-cancerous lesions of one to two millimeters in diameter and only a few cell layers thick.
The Advanced Imaging System is a slow scan, high precision CCD camera system designed specifically for low noise image acquisition and precise, highly flexible CCD testing and characterization. In addition, the system is designed to allow CCD mosaics to be supported with separate, programmable clock voltages and output amplifier operating points for each device. A high speed digital signal processor acts as the timing generator and allows all clock voltages and timing states to be adjusted through a set of downloadable parameters. Virtually any CCD can be operated with modest changes to the system hardware and software. In addition, the entire control program may be downloaded from the host computer at any time to facilitate radically different camera operation. The initial version of the camera system supports up to eight separate simultaneous readout channels with readout rates as high as 100 kHz at 16 bit precision and noise levels below 5 e- at 50 kHz. Control and data signals are connected to the host computer through a fiber optic interface for maximum distance and isolation. The system is a highly flexible CCD camera designed to operate individual CCDs and mosaics in slow scan systems where low noise is of primary importance.
The requirements of a charge coupled device (CCD) autoguider camera and the specifications of a camera that we propose to build to meet those requirements will be discussed. The design goals of both the package and the electronics will be considered.
Time delay and integration (TDI) imaging is an ideal technique for digitizing images of continuous webs and other large, flat objects in motion. It offers vastly improved sensitivity and set-up ease compared to line scan techniques, while retaining the line-scanner's variable- rate capabilities. The light sensitivity is comparable to that of area arrays. TDI cameras require an external horizontal drive signal in synchronization with the motion of the object being scanned. Vertical resolution is critically dependent upon synchronization of the motion of the moving object with the motion of the lines of charge as they move across the CCD sensor. This has been achieved by obtaining a synchronization signal directly from the moving object, utilizing an electro-optical resolver slaved to the object. Operation has been achieved with line rates of less than 500 Hz up to RS170 line rates -- 15,750 Hz. The integration phenomenon is readily seen: Halving the scanning rate will double the video signal for a given light intensity. The full horizontal resolution of the CCD sensor is achieved by careful alignment of the CCD sensor; in our case, in the order of 600 pixels
A high performance page width linear image sensor with a new multiplex structure called 'meander lines' has been developed. The sensor has an A3 page width (310 mm) with resolution of 400 dot per inch (DPI) and is operated at 4 MHz. The meander lines have been adopted to eliminate the crosstalk coupling and also to get small line capacitance closely related to high photoresponsivity. The sensor achieved a photoresponsivity of 20 V/lx(DOT)s and more than 64 gray scale reading capability. A compact image scanning unit was made of size 420 mm (W) X 130 mm (L) X 30 mm (H). It was tested and the reproduced image was good quality.
An A4 page width and 300 dot/inch hydrogenated amorphous silicon thin film transistor (a- Si:H TFT) driven contact image sensor which can read more than 128 gray levels has been developed. Crosstalk due to the coupling between data lines in the multiplex circuit has prevented high gray scale reading. In order to eliminate crosstalk, a sensor with a new multiplex structure has been developed with a ground mesh shield layer inserted at the crossover points between each data line. The ground mesh shield pattern was designed to optimize the gray scale reproduction ratio R. With this sensor, R is more than 0.992 for a single bit, thus achieving 128 levels of gray. This design was compared to the performance of two other sensors, one without a ground mesh shield, the other using a data line meander pattern. This technology is also applicable to higher performance image sensors with greater than 400 dot/inch resolution.
Colorimetric calibration techniques were studied for scanners that do not adhere to colorimetric requirements (the Luther condition) from the aspects of accuracy and signal noise. A newly developed calibration technique was compared with conventional polynomial regression models and was found to be more accurate. Computer simulations were performed to find the relationship between scanner responsivities and resultant chroma noise. The scanner matching the Luther condition exhibits larger chroma noise than a scanner having three separated responsivities due to the overlap among responsivities of each channel.
This paper will describe the development of a method for measuring the point-to-point global geometric accuracy of a very accurate, large, drum scanner that has been recently developed for the cartographic market. The process described is used to verify the accuracy of a scanner whose media handling capability is 44 in. X 64 in. and whose accuracy specification is to be verified as better than .003 in. circular error throughout its entire scanning area. The technique developed is based on proper measurement fundamentals and is applicable to various types and sizes of scanning and plotting instruments.
This paper describes image reading technique using multi area sensors for building a high- speed full color still image reader. An experimental system has been created using four CCD imagers which are located symmetrically on the optical axis and optically coupled to the same lens. To produce four images at each surface of the imagers, a pyramidal mirror is placed between the lens and its focal plane. Each CCD sensor reads a quarter section of the original image, and the four images are combined into one image like the original by signal processing. Image discontinuity, which is caused by the differences in characteristics among the CCD imagers, can be decreased below the perception of the human eye by signal correction processing and the read image can be output as a seamless image.
Motion picture film, as a major source of program material for television, is transferred to the video domain using a telecine. A high performance CCD telecine, operating at real time frame rates, has been developed for high definition television. The optical system for this telecine utilizes a feedback stabilized xenon arc lamp, pre-gate optical and spectral signal balancing, a diffuse illumination system, an apochromatic projection lens, a full resolution CCD luminance sensor, and a half resolution RGB trilinear CCD sensor.
Integrating cavities are being used for illumination in scanners. An integrating cavity is hollow, with entrance and exit holes, and a diffusely reflecting white interior. Light enters, undergoes multiple reflections to distribute the light uniformly, and exits toward the film or document. An arbitrarily shaped exit hole provides a uniform, diffuse light source of any desired shape, most commonly a line source. Integrating cavities also are being used as light collection systems, such as in laser scanners, where typically light enters a slit and exits a round hole, continuing to a photodetector. In this paper, integrating cavities are theoretically modelled by Mont Carlo ray tracing. This approach is very versatile, capable of modelling various types of cavity surfaces (diffuse, specular, refractive), arbitrary geometries, and arbitrary input light distributions. Results are presented for cavity efficiency, spatial uniformity and angular distribution of the exiting light. Complex cavity geometries from scanners in use are modelled. Theoretical results are compared with laboratory measurements.
There are a number of illuminating system for film copy based on the well known dichroic mirrors and electromagnetic slits -- valves. Light of a lamp illuminator is reflected separately on two mirrors (red and blue) and thus the third green channel is subtracted. Intensity of light in each channel is controlled by electromagnetic valve and then the final color is formed in the exit slit behind a tappered lightguide by mixing of the three principal colors. The new design of an illuminator working in a pulse mode being introduced in this paper is maintained with no mechanically moving parts. Principal part of the new illuminator is a cube-formed electro- optic modulator. The new illuminator operates at higher frequency of color modulation, with higher energy transmission (by 50% over the classical illuminator), and with the setting of greater number of color modes (73 intensity levels of each principal color). Moreover there is more compact design of the illuminating module over the one with mirrors and the new module can be easily connected with a computer control.
CCD signal-processing schemes attempt to reduce the effect of KTC, 1/f, and broadband noise on the output
signal. A number of schemes have been reported over the years. These schemes employ time delay and
subtraction to eliminate KTC noise and attenuate 1/f noise. They also include a low-pass function to reduce
the effect of broadband noise. Signal processing schemes include dual-slope integration, double-correlated
sampling, a variation of double-correlated sampling referred to in this paper as "switched exponential filtering,"
and transverse filters.
Transfer functions for signal and noise are presented for each of these schemes. Performance comparisons
are given with emphasis on their applicability to relatively high-speed CCD readout applications (readout rates
of 1 megapixel per second and faster).
Recent technology has produced CCDs with high-output gains in terms of volts per electron. Such devices
feature very small output gate capacitance and hence, reduced KTC noise. Signal processing that does not
use delay and subtraction to eliminate KTC noise is discussed. Again, the emphasis is on applicability to
relatively high-speed readout.
Video motion analysis has entered a new era with the introduction of the Kodak Ektapro EM Motion Analyzer, a high-speed solid-state video recording system. No longer confined to the limitations of film or tape for high-speed image acquisition, the electronic memory solid-state recorder offers new recording schemes never before possible. The Ektapro EM Motion Analyzer can be manually controlled by way of the keypad, or set into a trigger dependent mode to allow the actual application to control the recording. Trigger parameters include switch closure, sound, temperature, change in illumination, or a change in voltage. By writing over old frames with new, it is possible to have the high-speed video system in a continuous record mode and ready to capture an intermittent or uncontrolled event without the constraints of recorders that are dependent on consumable media. This paper will focus on innovative recording techniques (Record, Record-Stop, Record-Trigger, and Record-On-Command) and how the world of high-speed electronic image recording vastly differs from conventional high- speed tape and film methods.
In certain test, measurement, and research applications of CCD TV systems, the greater accuracy than usual 8-bit frame-grabbers can provide is demanded without the system being too expensive. The paper presents the concept and features of the high-precision low-cost digital CCD TV system intended for obtaining 12-bit monochrome images of immobile or relatively slow moving objects. The increase in accuracy is achieved by the specific digitization procedure -- one column per frame, which combines the benefits of a slow A/D converter with real-time TV imaging compatibility. To reduce speed restrictions on sample- and-hold circuits, a zoomed pixel read out cycle, corresponding to the pixel to be digitized, is proposed. The system provides great flexibility in choice of integration times and readout rates by means of a programmable readout sequencer, and is easily adaptable to various user demands and CCDs types.