Search

Air-to-ground target identification was used to select simulated target characteristics that resulted in performance similar to that found in real-world tasks. for a wide variety of ground target-vehicles, it was found that simulated-target identification levels could be matched to any desired real-world level by making visually realistic modifications to target contrast.

Perceptual motion blur was studied using imagery presented on an LCoS projector equipped with a mechanical shutter to reduce pixel hold-time. Perceptual measures of image blur were obtained with a simple test stimulus, as well as imagery similar to that used in Air Force flight simulation and training. Measured pixel hold-time was found to accurately predict perceived blur.

The ratio (τ) of a target's angular size to the rate of change in its angular size can be used by observers to judge the time remaining before they will collide with the target. We consider here whether texture flow rates (i.e., optical flow rates), consistent with simulated observer motion, affect judgments of time-to-collision (TTC). TTC was estimated for various relative velocities of the observer and a circular target. Relative velocity was varied such that the proportion of the total closing velocity attributable to the observer was either 0.0, 0.33, 0.67, or 1.0. Since the closing velocity between the observer and target was held constant, visual information for τ was the same under all test conditions. However, visual information (i.e., texture flow rate) for simulated self-motion was varied. The Weber fraction for estimated TTC progressively decreased as the proportion of observer motion increased, indicating that estimated TTC decreased as texture flow rate increased. This result has practical implications for judging TTC in high-fidelity flight simulators, since the rendering of terrain texture detail directly affects both texture flow rates and the perception of self-motion, either of which may effect estimated TTC.

Perceptual motion blur was studied in imagery presented on an LCD projector equipped with a mechanical shutter to reduce pixel hold-time. Perceptual measures of image blur were obtained with both a simple test stimulus, as well as real-world imagery. Both were found to correlate well with the measured pixel hold-time.

Air-to-ground target identification was found to improve with decreases in background resolution as defined by a satellite-imaging metric. This metric was then compared to a more conventional measure of target and image spatial resolution based on gaussian blur. 1. Introduction Photorealistic flight-simulator databases are produced by mapping satellite imagery onto a terrain height-map. The resolution of these databases is typically defined by the inverse of the spacing of the pixels representing the satellite imagery. This definition assumes that the imaging optics are sufficient to resolve the individual image pixels. Based as it is on an optical criterion, database resolution, as defined above, is not directly comparable to resolution measures typically used to evaluate imagery used in, for instance, flight-simulation (Geri, et al. , 2004) and night-vision (Driggers, et al. , 1999) applications. One way to relate the two measures is to compare performance using flight-simulator imagery derived from each of them. It has previously been shown (Geri

Simulated imagery was used to determine the effect of target and background blur on target identification performance. An interaction between these two variables was found, indicating that greater image detail may not improve performance, and hence may not be required in applications for which high-resolution databases are not readily available.

— In this study, the effects of manipulating the properties of 3-D objects and terrain texture on the control of altitude in simulated flight has been investigated. 3-D objects were found to be as effective as terrain texture for controlling altitude, and this result was interpreted as suggesting that both terrain texture and 3-D objects can serve as effective carriers of information about motion parallax and optical expansion and contraction. The present results, which were obtained using a vertically defined flight task, are inconsistent with the findings of Patterson et al., who reported that 3-D objects were more effective than terrain texture for controlling a horizontally defined heading task. The present results indicate that, when terrain texture is present, the motion parallax or optical expansion associated with the presence of 3-D objects does not improve altitude control.

The effects of terrain texture type and the height and density of 3D objects on heading control were studied in two experiments conducted in a high-performance flight simulator. The results suggest an orthogonal-extension principle which relates performance to the number and length of contours that are perpendicular to the ground plane, and also begin to define the stimulus conditions for which the principle is valid.

Increasing Display line rate did not improve aspect angle recognition performance beyond a level predicted by measured display resolution. Image antialiasing improved performance even though it did not increase the measured spatial resolution. Finally, the threshold for aspect angle recognition was found to be consistent with that obtained for other visual tasks dependent on target spatial detail.

— Stereoscopic perception of tilt about a vertical or horizontal axis is influenced by size and shear disparities, respectively. Other researchers have reported that, under certain conditions, stereoscopic perception deficits occur when the dots in a random-dot stereogram move at a velocity that produces optokinetic nystagmus. Here we examine how size disparity and shear disparity affect stereoscopic tilt perception under various motion conditions. We hypothesized that visual stimulus motion may interact with these disparities to affect tilt perception. Our results indicate that shear disparity and size disparity effects under static conditions are maintained under motion conditions. A possible explanation for the conflict between the current and previous results is discussed, as are implications for binocular head-mounted display applications.

Slant perception was studied under viewing conditions that combined horizontal-size, verticalsize, and overall-size binocular disparities with motion and static image conditions. For both motion and static conditions, results indicated similar effects of size disparity on perceived slants of disparity and zero-disparity stimuli. Implications for binocular head mounted display (HMD) systems are considered.

The apparent size of a simple stimulus was measured in order to quantify the effect of viewing distance on the perception of real (i.e., uncollimated) CRT imagery. The viewing distances tested were those used in several operational, real-image flight simulators. All identifiable distance cues to object distance except for ocular vergence and accommodation were eliminated. For six observers, it was found that perceived size changed significantly with changes in stimulus distance between 800 cm and either 61 cm or 94 cm, or vice versa. The data are consistent with the changes in perceived size reported for other stimuli and viewing conditions, and suggest that simulator imagery presented within about 1 m of the observer must be magnified by 15-25% in order to provide size cues consistent with those provided by objects located at optical infinity. For two of the six observers, measured vergence levels were consistent with the viewing distances tested, whereas measured accommodation levels were less than expected. Thus, at least for these two observers, it appears that differences in the level of ocular vergence were primarily responsible for the observed changes in perceived size.