The readily-available, vendor state of the art in real-time color digital image display systems is summarized. The display systems considered are those having multiple refresh memories, multiple graphic overlays, interactive capability and full color operation using at least a 512x512 element image array. Also presented is a comparison matrix of a host of technical issues for ten of the currently available, state-of-the-art display systems. The technical issues considered include display system architectures, refresh memories, graphic overlays and cursors, lookup tables, processing capabilities, and display presentation.
More information can be displayed for human visual interpretation with a full color presentation than with a monochrome (black and white) presentation. However, image analysis systems have not been able to utilize color to the extent that should be possible. This paper presents reasons and application examples showing why the traditional RGB additive color approach has been ineffective. A system which overcomes the limitations of RGB, and which utilizes the fundamental human perception parameters of intensity, hue and saturation is described.
Image capture with a combination of video camera, digital-to-analog converter and digital video processor is described. Techniques for facilitating camera calibration are addressed. Followed by discussion of reference and sample data capture. Physical registration methods using real-time difference images are considered. Capabilities for functional manipulation, reference correction and multiple frame integration are elaborated with examples. Representative post-acquisition processing operations are outlined.
There have been many unsuccessful efforts to construct a practical three dimensional display. A display that could produce an up-dateable three dimensional image would have many applications ranging from air traffic control to the display of the higher order structure of proteins. A prototype that demonstrates the fundamental feasibility of a new technique has been constructed; and a more sophisticated, computer-driven display is under development. The new method is loosely based on older schemes involving the deformation or vibration of a lens or mirror, but it avoids the primary problems of these older displays. The technique consists of rotating a lens in front of a stationary, two dimensional light source in which the appropriate pixels are illuminated at calculated lens rotation angles. A real, luminous image is produced in front of the lens. A display based on this method is mechanically simple, produces high quality, large size images, is easily up-dateable, and can be in color.
The display and processing of large images (on the order of thousands of pixels per side) is one of the most difficult problems faced by digital softcopy display systems. Even simple operations on these large images, such as scrolling and magnification, are performed slowly and with great difficulty on currently available systems. In this paper, a new architecture is proposed based on high-rate video channels, parallel transfer disks, asynchronous interleaved memories, and special-purpose processing elements. This design offers orders of magnitude increased performance in the display and processing of large images at relatively low cost, and makes use almost exclusively of commonly available hardware.
This document describes the architecture and design philosophy of the COMTAL Vision One/20 digital image processing system. The Vision One/20 is a dual ported RAM refresh memory system which affords multiple user access to a common expandable data base (4096x4096x8 bit pixel memory images are available) with dynamic partitioning to afford a multiplicity of different applications. Real time roaming with a window size of up to 1024x1024 pixels through the data base is possible with zooming and 3x3 convolution all implementable in 1/30 second. With graphics overlay memories, annotation, labeling, outlining and arbitrarily shaped multiple small area monochrome or color correction are possible. Due to the dynamic allocation of the data base memory, digital loop movies in real-time are possible as is left-right, right-left, up-down or down-up scrolling of new imagery into the refresh memory and viewing window. These and other features are expanded upon herein. Pipeline processors, freeze frame iterative image array feedback, firmware-burned control and instruction com-mands are descriptive of modern image processing architectures used in the Vision One/20.
A few years ago the requirement for true three-dimensional displays in a variety of applications seemed to be answered by the development of holography. Holograms allow stereoscopic group viewing without glasses. As the observer moves his head he sees the scene with parallax from different horizontal and vertical points of view. However, the limitations of holography soon became apparent. These include: the inability to reproduce continuous motion; lack of color; difficulty of recording under any but very special conditions; limited screen size; and the requirement for a laser in both pick-up and display. Intensive research has provided some solutions to these limitations and has raised further problems. Additional improvements will certainly be made during the coming decades. Autostereoscopy of the parallax panoramagram type utilizes optical techniques similar to holography. Recent developments make possible three dimensional film and television displays, without glasses. These displays have the features of: full motion; full color; excellent resolution; horizontal parallax; conventional illumination sources; and are recorded with more conventional filming or television techniques. The applications include: aerial and satellite reconnaisance pictures in depth; computerized axial tomography stereoscopic displays; remote manipulation of dangerous materials and remote driving of vehicles by 3D-TV viewing; stereoscopic radar presentations; spacial simulation displays; and education and entertainment pictures in depth.
The paper discusses some of the findings of a technology survey performed via a one month visit to Japan in April/May 1979. The purpose of the study was to visit companies performing advanced development on both CRT and new display technology programs and to establish status and availability of display devices. It was found that the companies visited were actively pursuing a wide variety of display technology programs. These included: advancements in color CRT technology, development of high resolution (1125 line) direct view and large screen projection displays, investigation of color flat panel displays for the large screen consumer market, development of extremely compact calculator size portable television receivers, and numerous projects directed at the alphanumeric and graphic terminal markets. Although considerable research was also being performed on the development of improved digital and small alphanumeric displays, this article was limited to discussion of the more complex larger number of picture element type displays. Differences in funding strategies between U.S. and Japanese companies are also discussed.
A multimode liquid crystal large screen display system under development promises improved display imaging for fleet command, shipboard information centers, and related tactical data systems. Multimode technology combines monochromatic television with multi-color graphics for visual effect enhancement. The system employs a liquid crystal light valve (LCLV)--an optical birefringent and reflective liquid crystal device--which rotates polarized light during image generation. The LCLV voltage is CRT controlled to induce sufficient birefringence effect to disperse polarized white light. The result is the generation of different colors, so that a multicolor display is created simultaneously with a monochrome television image.
The development of the current high contrast, sunlight legible, Thin Film Electroluminescent display technology will be reviewed and device structures and operating characteristics will be described. Drive and addressing considerations will be presented in relation to device types, power dissipation, multiplexing and viewability. An overview of the industry status will be given including device availability and problems remaining to be solved. Applications of the technology will be reviewed covering both user requirements and device capabilities. A survey of future prospects will cover improvements in efficiency, multicolor displays and various device configurations.
Many parameters must be considered to obtain a display that is maximally useful to the human operator. Direct observation of a candidate display by the designer and survival in the marketplace provide some guidance, but whenever possible design decisions should be based on relevant test data. This paper summarizes the more critical design parameters, describes some of the available test data, and illustrates some of the characteristics of the test situation that determine whether a set of test results are relevant to a particular design problem. This discussion of test characteristics applies both to the evaluation of available test results and to the design of new test programs aimed at providing data relevant to display component subsystem or system design.
A two-phase study was conducted which related the confusions among dot matrix characters to the two-dimensional spatial frequency similarity of these characters. During the first phase of the study, subjects were shown single alphanumeric characters from four different dot matrix fonts and five matrix size/character subtense combinations. Data from this phase of the research were analyzed in terms of both correctness and character confusion frequencies. The second phase of the study consisted of digitizing and analyzing all characters from two of the fonts used in the first phase. The fonts chosen represent the most and least confusable of the four, based on the performance data obtained. These characters were scanned photometrically using a computer-controlled X-Y stage and subjected to a 512 X 512 point fast Fourier transform (FFT). The Fourier coefficients were correlated for all possible character pairs within each font-matrix/character size cell. These correlations provided an objective similarity measure among characters based upon their 2-D spatial frequency spectra. In addition to the spatial frequency similarity measure, a simple digital Phi coefficient was calculated for each character pair. The final analysis performed in this study was the correlation of observed performance (confusions) with objective similarity measures (2-D spectra and Phi coefficients).
The emphasis in this paper is on frequency domain measures of image quality. The development of a quality measure is illustrated with a series of studies using the optical power spectrum. Relationships with information extraction performance are used to evaluate the effectiveness of the resulting quality measures. The studies cover several image types: conventional photography, CRT display and digitized imagery. The status of physical quality measures, as tools for display development and evaluation, is assessed.
A model describing the decomposition of imagery in the human retina is developed based on the retina's cellular structure. Two types of retinal cells, horizontals and amacrines, perform spatial averaging across the retina to form a low-pass image channel. This low spatial frequency information is fed back to the retina's receptor cells to form a difference channel of high-passed spatial frequencies. Such a model is suggested by electro-physiological as wll as psychophysical evidence. Analysis of the model characterizes the low-pass channel as a constrast channel and the difference channel as an edge detection channel. Application of the model to image quality assessment suggests a two factor approach involving metrics in the model's eye domain.
Each picture element (pixel) of a digitally-controlled display can be commanded to a finite number (typically 256) of luminance levels. The relation between the digital command levels and the resulting luminances is called a "gamma" function, using a loose adaptation of photographic terminology. The purpose of the technique described here is to obtain a gamma function which results in the contrast between each of the steps being equally visually discriminable. The procedure starts with a set of 51 photometric measures of a specific set of digitally-controlled targets designed to determine the interaction between the display's target and surround luminances, and the display's useful luminance range. These data are used in the Rogers-Carel Formula for predicting visual threshold of within-target modulation contrast, m as a function of average target luminance, t, average surround luminance, s, visual size of the target area, a, and the visual spatial frequency of detail within the target, f, (i.e., m = f(t, s, a, f)). Through curve fitting, optimization of viewing distances and other operations, the equation is reduced to m = f(ℓ) where ℓ is the luminance when all the display pixels are commanded to the same level. This equation is used to determine the number of just noticeable differences (JND's) in target contrast within the useful display luminance range and then to determine the luminance value for each command level so that each succeeding step represents the same number of JND's.
CRT (Cathode Ray Tube) softcopy image display systems have found widespread acceptance in the image exploitation area in applications ranging from remote sensing to medical diagnosis. However, to extract the maximum performance from these display systems, careful setup, control and monitoring of important display parameters is required. This paper describes a key set of quality assessment parameters needed to quantify the performance of CRT-based image display systems, and a hardware analyzer to measure them. The analyzer is a portable device based on a linear photodiode array and small minicomputer which quickly allows both calibration and assessment of the overall "health" of the CRT display to be made.
The advent of lead lanthanum zirconate titanate (PLZT) ceramics as electro-optic shutters in light weight, head-worn viewing devices has led to many recent applications for stereoscopic displays. PLZT stereoscopic viewers offer the advantage of essentially unrestricted freedom of head movement while achieving strong binocular depth-of-field sensations using a single CRT display. A survey of recent applications for field-sequential stereoscopic video and computer generated imagery is presented. Video applications include real-time color stereoscopic video for the Remote Underwater Work System (RUWS), a prototype stereoscopic x-ray imagery system, and moving platform stereoscopic imagery. Computer generated stereoscopic imagery applications include computer graphics simulations of an aircraft carrier landing sequence, computer modeling of double helix DNA molecular structures, and stereoscopic displays derived from multispectral satellite imagery.
The remote sensing, transmission, recording and display of visual information using holograms continues to be a desired goal, The fundamental problem is that conventional off-axis holograms have high space-spatial frequency bandwidth products which require extremely high resolution recording media. However, much of the essential visual information of a scene is intrinsically low spatial frequency information. In general, high spatial frequency has been required to separate the real and virtual images from the self-interference terms in the hologram. A viable alternative is to selectively eliminate the self-interference terms of the hologram during the recording process. One such method is electronic heterodyne recording where electro-optic modulation is used to introduce a sinusoidal phase shift between the object and reference wave. The resulting temporally modulated holographic interference pattern is scanned by a commercial image dissector camera and rejection of the self-interference terms is accomplished by heterodyne detection at the camera output. The electrical signal representing this processed hologram can then be used to modify the properties of a liquid-crystal light valve or similar device. Such display devices, when illuminated by coherent light, transform the displayed interference pattern into a phase modulated wave front rendering a three-dimensional image. System requirements of holographic display systems will be considered as well as specific requirements for television rate and format systems.
A display terminal architecture has been developed by the Boeing Aerospace Company that affords significant design flexibility to meet the requirements of a number of diverse interactive applications. The architecture is dominated by an image refresh memory and a local database memory. These memories are interconnected by a variable number of highspeed general purpose microprocessors and hardware processors that translate local database and highspeed interface information into appropriate image list formats. Images are formed using a high performance directed beam graphics generator and indicator. The console may be connected to its host computer using either standard serial interfaces or DMA. Serial interfaces link directly to one of the microprocessors while DMA is provided to the database and the refresh memories. By using a combination of stored program algorithms, special microcoded instruction sequences, special hardware processors and tailored graphics command formats; realtime, application dependent, processes are implemented within the context of a performance hierarchy. This paper describes hardware and software issues as related to the console architecture and describes two contrasting display systems: an airborne electronic counter counter-measure (ECCM) display and a mobile ground based air surveillance display. In particular, hardware and software techniques that implement concurrent and pipelined processes, display database sharing, and critical regions for these applications are described.
A digital image analysis system has been developed at the Remote Sensing Research Program, Space Sciences Laboratory, University of California, Berkeley. The system is based upon high speed random access semiconductor memory used in an interactive video display configuration. The primary use of the system is for remote sensing applications and, as such, has been designed for multi-sensor and multi-temporal data display. Control of all variables allows the system configuration to be tailored to meet almost any user's needs. The video memory is modular, allowing two different intensity resolutions. Individual mapping tables allow functions to be applied to data without changing original data values. A selected portion of a memory module may be scrolled and viewed alongside another module. Simultaneously, a pseudo-color representation of a third or fourth module could be displayed and used in the image analysis process. Two types of cursors aid in object delineation and can be used to direct a module to module feedback operation for "quick look" manipulation of data. These system capabilities will be described and an example using Landsat data presented.
Stereoscopic oscilloscope displays have been slow in gaining acceptance due in part to the past scarcity of applications, but perhaps mainly due to the apparent unavailability of a general-purpose instrument. But now the increasing use of computer-generated displays has begun to focus attention on them. The stereovectorscope, developed in our laboratories, is a new stereo oscilloscope with an initial 4-6 month availability. In addition to stereo, it provides linear perspective and two axes of manually-controlled spaceform rotation. Use of a proprietary spaceform rotation mechanism results in a simple, compact, lightweight instrument. One embodiment of the stereovectorscope is as a compact desktop laboratory instrument, designed to be viewed in the same manner as a stereo microscope. It has provisions for connecting a large-screen monoscopic monitor for group viewing. One immediate application is as a device for displaying vectorcardiograms. Another immediate application is as an OEM device for displaying computer-generated spaceforms. In this application, the stereovectorscope frees the computer and programmer from the requirement to provide rotation, perspective, and stereo transformations. In all applications it provides a highly natural man-machine interface for spaceform rotation via its "integrated" rotator.
An operational description of an intelligent image processing display terminal is given. The terminal acts as an alphanumeric display, a raster graphics terminal, a color image display, and a real time image processor. The image processing function is accomplished by a unique form of spatial convolution.
In this paper a two-dimensional measure of image quality is proposed and discussed. The measure can be used as an objective specification for imaging systems, can be readily related to several forms of subjective data, and is based on easily-understood definitions. The paper includes a brief summary of the concepts, with examples; and a discussion of some critical details.