Abstract: One of the most fundamental problems in vision is segmentation; the way in which parts of an image are perceived as a meaningful whole. Recent work has shown how to calculate images of physical parameters from raw intensity data. Such images are known as intrinsic images, and examples are images of velocity (optical flow), surface orientation, occluding contour, and disparity. While intrinsic images are not segmented, they are distinctly easier to segment than the original intensity image. Segments can be detected by a general Hough transform technique. Networks of feature parameters are appended to the intrinsic image organization. Then the intrinsic image points are mapped into these networks. This mapping will be many-to-one onto parameter values that represent segments. This basic method is extended into a general representation and control technique with the addition of three main ideas: abstraction levels; sequential search; and tight counting These ideas are a nucleus of a connectionist theory of low 'eve and m'ermediate-level vision. This theory explains segmentation in terms of massively parallel cooperative computation among intrinsic images and a set of parameter spaces at different levels of abstraction.

Abstract: One of the most fundamental problems in vision is segmentation; the way in which parts of an image are perceived as a meaningful whole. Recent work has shown how to calculate images of physical parameters from raw intensity data. Such images are known as intrinsic images, and examples are images of velocity (optical flow), surface orientation, occluding contour, and disparity. While intrinsic images are not segmented, they are distinctly easier to segment than the original intensity image. Segments can be detected by a general Hough transform technique. Networks of feature parameters are appended to the intrinsic image organization. Then the intrinsic image points are mapped into these networks. This mapping will be many-to-one onto parameter values that represent segments. This basic method is extended into a general representation and control technique with the addition of three main ideas: abstraction levels; sequential search; and tight counting These ideas are a nucleus of a connectionist theory of low 'eve and m'ermediate-level vision. This theory explains segmentation in terms of massively parallel cooperative computation among intrinsic images and a set of parameter spaces at different levels of abstraction.

Abstract: An apparatus for use in the cockpit of an aircraft which provides sufficient illumination to read the cockpit display indicators, controls and charts but does not destroy the pilot's night vision. A casing, which is normally worn on the wrist of the pilot, contains a microencapsulated electroluminescent phosphorus lamp which provides illumination that will not affect the pilot's night vision. The lamp is powered by an inverter which is connected through a switch to a power supply, all of which are also contained within the case.

Abstract: Both anecdotal reports from pilots and theories of visual cues would predict lower approaches to narrow or long runways than to wide and short runways. Similar, practice approaches made by pilots to a particular width of runway should lead to an increase in subsequent approach angles flown to wider runways and a decrease in approach angles flown to narrower runways. Two experiments with instrument-rated pilots made quantitative tests and these predictions. In Experiment I, three pilots flew simulated approaches and landings in a fixed-base simulator with a computer-generated image visual display. Practice approaches were flown with an 8,000-ft long runway that was either 75, 150, or 300 ft wide; test approaches were to runways with widths or 75, 100, 150, 200, and 300 ft. In Experiment II, 40 pilots controlled the slant of a moving model runway during simulated night visual approaches. Five different models simulated runways from 100 to 300 ft wide and 3,000 to 9,000 ft long. As predicted, training on a wide runway in Experiment I lowered approach angle in approaches to narrower runways; a narrow practice runway also raised approach angles to wider runways. The magnitude of these practice effects increased as distance from runway threshold decreased. There was also a general tendency for approach angles to decrease as runway width decreased. The latter effect was corroborated in Experiment II; in addition, generated approach angles decreased with increasing runway length. Giving half the pilots information about runway size prior to each approach had no effect on responses. These findings add to the quantitative evidence of danger in night visual approaches due to visual illusions and large variability in the visual perception of approach angle. Language: en

Abstract: Several types of gray-weighted ``medial axes'' have been defined. This paper shows that one of them, the min-max medial axis, can be used to reconstruct good approximations to the original image based on a relatively small amount of information.

Abstract: A night-vision instrument for presenting a 1:1 image comprises an objective, an image intensifier or converter tube, an ocular, a flat double-sided input mirror, and an optical system adapted to transfer the image appearing on the phosphor screen of the intensifier to the ocular. The ocular axis is in line with the viewed object, while the objective and the intensifier have their axis positioned at a right angle to the viewing axis. The input mirror is positioned in front of the ocular inclined at 45° to its axis and to the axis of the objective, so as to direct the light rays received from the viewed object through the objective onto the cathode of the intensifier. The intensified image appearing on the phosphor screen is directed to the rear side of the input mirror by one single prism, one inclined mirror, one double prism and an optical lens system. The rear side of the input mirror projects the image through the ocular into the eye of the observer. The geometry of the instrument is such that the image of the object appears in the eye under the same visual angle as if viewed from the same distance by the unaided eye, which is obtained by making the distance between the optical center of the input mirror and the eye lens equal to the distance between the said mirror center and the optical center of the objective, and by designing the optical system accordingly.

Abstract: This article discusses the problems involved in trying to see adequately while driving at night and the limitations of all headlighting systems. Despite substantial improvement in headlamp performance low beams, which are the ones used by most drivers, do not provide adequate visibility distance under many driving conditions. No improvement in design is foreseen in the near future since the major limitations are the human visual system and some practical problems associated with headlamps mounting and conditions of use.

Abstract: The present invention provides a training aid and method that produces restic colored images of targets, as seen through a thermal viewer, using black and white video or photographic techniques with filters and light modulators.

Abstract: An easy retrofitting of a night vision thermal imager to mount a televisionTV) camera thereon which allows the thermal imager outputs to be viewed remotely by a CRT display. The retrofitting provides a compact and utilitarian remote viewing capability without altering the imager itself. An unpluggable mirror unit, positioned at the output of the imager, provides a folded light path allowing the TV camera to be mounted in an inverted position over the viewer and operate as a remote observation station with the mirror unit attached, or be viewed directly with the mirror unit removed.

Abstract: By using a binocular telescope system (6 to 11), which is constructed of an objective (6), a highly refracting prism system (8, 9) having a refractive index exceeding 1.65 and any physical beam splitter (7), as well as two eyepieces (10, 11), a light, slim and low-cost instrument becomes possible. In consequence of approximately equal free lens diameters in the telescope system in front of and behind the prism system (8, 9), optimal use is made of the prisms (8, 9).

Abstract: To respond effectively to significant roadway conditions, drivers must make timely and accurate detections and identifications. Detection and identification are different processes, and neither is necessarily simple or automatic. This paper discusses both processes; it points out problems with each, and suggests means for improvement.

Abstract: Recovery of night vision at altitude has been examined in relationship to changes of 2,3,diphosphoglycerate (DPG). It appears that, if DPG influences night adaptation, its effect is small. Inter-individual variability is large, both in terms of DPG and performance on the MESOP, the device used to measure dark adaptation.

Abstract: A case is made in this paper for the importance of peripheral visual detection in the nighttime bicycle detection problem. Based upon the findings of this field study, a reflector arrangement consisting of two horizontally separated high performance rear reflectors in combination with an improved taillight and improved pedal reflectors would seem highly promising and beneficial with regard to the rear conspicuity (detection) problem of bicycles at night. More research is needed to get a better understanding of the human visual detection mechanisms involved in foveal vs. peripheral detection at night. Also more research is needed to examine the detection distance statistics and threshold multiples for more than two reflectors, for various reflector arrangements and separation distances, for various reflector or visual stimulus intensity levels and to determine appropriate correction factors to be used when performing engineering illumination or visibility calculations involving peripheral visual detection. (Author)

Abstract: The visual acuity of 30 male and female drivers aged 20-25, 40-45, and 60-65 years was examined under conditions of “high” background luminance, “low” background luminance, and “high” background lu...

Abstract: Target recognition performance for an RCA Shottky barrier staring FLIR with and without target motion is presented together with various minimum resolvable temperature (MRT) measurements obtained using both a 4-bar and a modulation perception requirement, i.e. using the strict requirement to see the 4-bars in the MRT bar pattern and the less restrictive requirement to see some modulation in the displayed image. In addition, both recognition and MRT results for a hand-held thermal viewer having approximately the same limiting spatial frequency as the staring system are presented. The various results are compared and examined to determine if the two dimensional aliasing in staring systems significantly degrades recognition performance relative to conventional scanning systems with similar size detectors and to establish which MRT, vertical 4-bar or modulation, or horizontal 4-bar or modulation is the best predictor of recognition performance when used in conjunction with the Night Vision & Electro-Optics Laboratory (NV&EOL) static recognition model.© (1981) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Abstract: Historical background is presented of the U. S. Army's requirement for a high performance, lightweight, night vision goggle for use by helicopter pilots. System requirements are outlined and a current program for development of a third generation image intensification device is described. Primary emphasis is on the use of lightweight, precision molded, aspheric plastic optical elements and molded plastic mechanical components. System concept, design, and manufacturing considerations are presented.

Abstract: The Autonomous Acquisition Simulator System (AAS) is a modular real time image processing system that is currently being developed by the Ford Aerospace & Communications Corporation (FACC) for the U.S. Army Night Vision & Electro-Optics Laboratory (NV&EOL). As a cooperative effort between FACC and NV&EOL, the AAS Program has as its primary objectives, the development of a flexible real time image processing algo-rithm simulation device that may be used as a tool for the evaluation of artificial intelligence and image processing algorithms for application to tactical systems over the next few years. This paper summarizes the operational and system requirements for a testbed that would function as such a simulation device. The architecture of both the hardware and software designs for the AAS are then described in the context of resulting data structures and computational requirements for implementing these algorithms in real time. Also included is an approach for providing a compiler level language as a highly desirable feature to facilitate the development of software for implementing new algorithms.© (1981) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Abstract: The user of any imaging system is primarily concerned with securing some desired level of perception performance against real world objects of interest under anticipated typical operating conditions. Develop­ ing validated predictive models for this T field1 performance involves a number of complexities and sources of uncertainty which are not encountered when modelling hardware performance from objective laboratory measurements.These problems become apparent in the course of describing the basis of a program for generating the perception probability versus range curve for intensifier sights and a statistical trial which was conducted to validate its predictions.IntroductionThe requirement for a predictive field performance model for image intensifier systems has only arisen relatively recently. This was partly a response to a greater sophistication in user specification of field performance, and partly because of the greater range of variants and alternatives available for those components affecting intensifier performance. Previously it had been possible to confidently predict the field performance of proposed new designs by extrapolating from direct field measurements using readily available systems based on the same types of component.This contrasts with the situation in thermal imaging, where the greater complexity and range of design variables of these systems and the difficulties in procuring them for evaluation made it necessary to develop performance measures and modelling techniques in parallel with the systems.Field performance modellingThe higher perceptual processes of recognition and identification are not so well defined as that of detection. Moreover, the underlying mechanisms and image quality requirements relating to the achievement of a specified level of performance are not yet understood in detail. Consequently field performance modelling in the present state of the art relies on empirically derived relationships.Analysis of a large and growing bank of trials data at the Night Vision Laboratories, Port Belvoir, U.S.A. l indicates that the perception probability, P, of military vehicle targets can be related to the number of spatial cycles, f, which could be discriminated.by an observer using the imager or* a bar target of equivalent size and mean modulus of stimulus signal with respect to the immediate background. Clearly, accommodating human observer variability and using this very simple characterisation for a variety of complex stimulus profiles necessarily implies a limited precision. This appears to be of the order of + 15% in f, which is acceptable for many situations.The empirical fit shown in Figure 1 which is valid over the operationally significant range of P is modelled by:(f)p -^ .dy (1)r-with:This P is a standard normal cumulative distribution in the transformed variable x. a has a value of 0.21 and the value of f (P= 0.5) takes values of 1.0 for detection in medium to low clutter, 3.5 for recognition and 7«0 for identification.

Abstract: The relative efficiency of sticky traps and cylindrical electrical grid traps for capturing male beet armyworm moths,Spodoptera exigua (Hubner), was studied using night vision equipment. The sticky and grid traps captured 34 and 40%, respectively, of the insects which oriented to them. The grid trap's capacity was limited only by the size of the receptacle (3.8-liter bucket) at the base of the funnel. The maximum capacity of the sticky trap was ca. 110 insects, and the efficiency of this trap declined rapidly when captures exceeded 50% capacity. This decline was caused by the physical blockage of the sticky surface by the captured insects. There was no evidence to indicate that captured males had any repelling effect on other males subsequently attracted to the sticky trap.

Abstract: By using a binocular telescope system (6 to 11), which is constructed of an objective (6), a highly refracting prism system (8, 9) having a refractive index exceeding 1.65 and any physical beam splitter (7), as well as two eyepieces (10, 11), a light, slim and low-cost instrument becomes possible. In consequence of approximately equal free lens diameters in the telescope system in front of and behind the prism system (8, 9), optimal use is made of the prisms (8, 9).