The resolution capabilities of computer image generators (CIG) used for simulation and training have advanced to the degree that they exceed the capabilities of existing shadow mask, direct view color TV displays and color TV projectors. One solution to this problem is the modern day implementation of a trinoscope color display that uses the optical merging of three high resolution monochrome cathode ray tubes--red, green, blue color phosphors, respectively--to produce a full color image. Such systems are particularly suitable for telescopes and periscopes needed in tank or submarine simulations where the color-combining optics can be integrated into the simulated sight optics. This type of system can provide color resolution equal to the best of CIG systems. Systems operating at over 2000 TV scan lines have already been delivered. This paper describes the many technical advances required to assure maximum resolution and, more importantly, superior color convergence (i.e., the capability to make the three CRT images fall exactly on top of each other so that the resulting full color image is produced without undesirable color fringing).

A full color, Computer Image Generation (CIG) raster visual system has been developed which provides a high level of training sophistication by utilizing advanced semiconductor technology and innovative hardware and firmware techniques. Double buffered refresh memory and efficient algorithms eliminate the problem of conventional raster line ordering by allowing the generated image to be stored in a random fashion. Modular design techniques and simplified architecture provide significant advantages in reduced system cost, standardization of parts, and high reliability. The major system components are a general purpose computer to perform interfacing and data base functions; a geometric processor to define the instantaneous scene image; a display generator to convert the image to a video signal; an illumination control unit which provides final image processing; and a CRT monitor for display of the completed image. Additional optional enhancements include texture generators, increased edge and occultation capability, curved surface shading, and data base extensions.

This paper describes the development of a 25-inch, high resolution RGB shadow mask color display which has the capability to operate in either a raster or calligraphic (stroke) format. It is also possible to intermix the two formats to achieve a level of performance not possible with either individually. The display utilizes a new high resolution shadow mask color CRT capable of resolution in excess of 1100 TV lines. Some of the performance goals set for the display are: operation in a TV raster format, 1023 lines with a standard 7 microsecond blanking time; precision vector scanning at a rate of one inch per microsecond; a jump and settle mode where the beam can be positioned from any point to any other point on the screen in less than 12 microseconds; and TV raster rotation of ±45 degrees with externally generated resolved sweep signals. The display has many features, including DC coupled focus which provides the user with a dynamic spot size capability; an internal raster generator selectable by a TTL level control; and an optional internal sync/pattern generator which allows off-line maintenance and calibration. In addition, busy signals are provided from each deflection amplifier as well as the dynamic focus amplifier to provide a handshake capability with computer based digital image generation systems. The display is being developed specifically for use in simulation and takes into account the mechanical and optical constraints imposed by a simulator. For example, the convergence controls are provided in a remote control box which allows the display to be adjusted from the subject's viewpoint.

This paper describes the impact on the training of air refueling boom operators by a very unique and a very realistic visual system. This system was conceived and developed by the Farrand Optical Co., Inc. for the Boom Operator Part Task Trainer which was fabricated by the Aeronautical Systems Division of the Air Force Systems Command. The visual system is truly three-dimensional, reproducing all real world angles accurately. Two years of training experience with this simulator has proven the value of realism in simulation when the proper cues are provided.

Large dome simulators for air-to-air and air-to-ground aircraft weapons tactics trainers are coming into prominence. In such devices, the training objective is to improve pilot proficiency and coordination by allowing the pilot to train in the use of weapons in realistic operational and threat environments. Typically, the pilot will fight against a TV projected image that is slewed across the field of view. From geometrical considerations, a standard rectangular or square TV image looks distorted to the pilot. The degree and shape of the distortion changes with the shift in location of the "target" on the screen relative to the pilot as it is slewed either by servo pointing the entire projector or by optical means. This paper describes the design and development of a TV projector that is capable of predistorting the TV raster such that from the pilot's viewpoint the image will look rectilinear at all times. This dynamic raster shaping can be updated at the TV field rate (typically 60 times per second) so that there is no perceptible jumpiness in the image as the shape is varied. In addition, the raster can be zoomed to create the appearance of distance change to the target and rotated to compensate the effects caused by the mirror steering of the target image across the dome.

With today's technology, the art of generating a visual image is still a very complex and sophisticated process. The intent of this paper is to describe for the reader two similar visual systems which relate to complexities, similarities, and requirements for an end product. The systems selected for this comparison are visual simulation systems and avionic display systems. This broad base comparison is intended to relate the major components of each system relative to the output image and allow the reader to appreciate their complexities and similarities with minimum involvement and effort.

The paper examines the relationship between human evaluation of the sharpness of an image and the objective degree of edge enhancement, defined in terms of a modulation transfer function derived from a physical light source and photographic materials. Human perception of the sharpness of the image was measured by the modified method of constant stimuli. The sharpness of the images of four pictures was correlated with the modulation transfer function expressed by information volume and photographic materials. The sample correlation coefficient was r = 0.83. Theoretical implications of less than perfect correlation between the human evaluation and physical units are thoroughly discussed.

Computer Image Generation (CIG) visual systems provide real time scenes for state-of-the-art flight training simulators. The visual system reauires a greater understanding of training tasks, human factors, and the concept of image realism to produce an effective and efficient training scene than is required by other types of visual systems. Image realism must be defined in terms of pilot visual information reauirements. Human factors analysis of training and perception is necessary to determine the pilot's information requirements. System analysis then determines how the CIG and display device can best provide essential information to the pilot. This analysis procedure ensures optimum training effectiveness and system performance.

This paper presents a method for converting Landsat imagery of natural rural scenes to horizontal viewing perspectives in a digital image processing system. The technique uses digital terrain images for a three-dimensional representation of the scene. Full color pixel-by-pixel (as opposed to skeletal or graphical) images are synthesized, and hidden pixels are eliminated. A sequence of synthesized images of the Colorado River basin is shown. Examples of panoramic and orthographic projections are also shown. An appendix presents a method for converting a contour map into a digital terrain map in raster format.

Real time image subtraction techniques are employed to obtain positive and negative versions of the same object. In white light, the method codifies, through contrast reversal, gray level information into colors.

The concept of spatial frequency has been a useful tool in evaluating two-dimensional imaging systems. In three-dimensional imaging the spatial frequency concept must be applied to the disparity function as well as to the luminance function. Sensitivity functions of human observers to various combinations of luminance and disparity spatial frequencies have been obtained. Taken together, these constitute a three-dimensional transfer function which should be useful in optimizing stereo display systems.

The Spanish Surrealist artist Salvador Dali has recently perfected the art of producing two paintings which are stereo pairs. Each painting is separately quite remarkable, presenting a subject with the vivid realism and clarity for which Dali is famous. Due to the surrealistic themes of Dali's art, however, the subjects preser.ted with such naturalism only exist in his imagination. Despite this considerable obstacle to producing stereo art, Dali has managed to paint stereo pairs that display subtle differences of coloring and lighting, in addition to the essential perspective differences. These stereo paintings require a display method that will allow the viewer to experience stereo fusion, but which will not degrade the high quality of the art work. This paper gives a review of several display methods that seem promising in terms of economy, size, adjustability, and image quality.

This paper presents the principle, the realisation and the results of a system simulating in real-time extremely realistic views of 3D scenes. The simulation considers two factors, namely the seasonal and diurnal illumination variation and the change in the atmospheric visibility. Such effects are responsible for the main aspect changes in real scenes like for instance mountainous terrains. The scene aspect simulation is based on information from a digital terrain model and albedo maps previously derived from aerial photographs. A scene model is assembled in a stage of preprocessing and is then used as the data base for the simulation. The generation of the image itself is performed in real-time. Any modification of the scene illumination and the atmospheric visibility within the scene can be entered interactively and the new aspect of the scene can be seen instantly on the display where the new images are updated at video rate. This simulation system is implemented on a display processor making an intensive use of its transform capabilities via video processor and look-up-tables. The results show typical illumination and visibility effects in various scenes simulating real natural terrain.