Abstract
A person moving through the world must be able to detect moving objects to avoid collisions. In previous studies, we examined human ability to find an object moving at a different angle from the radial optic flow field generated by observer motion. In this study, we examined whether differences in image speed could be used to detect moving objects within a radial flow field. Observers viewed a radial flow field simulating observer motion of 200 cm/sec in a straight line toward a set of 25 circular objects initially at a distance of 800 cm from the observer. In half the trials a randomly chosen target circle moved at an increased or decreased image speed. The target speed was computed by multiplying the image speed calculated for an object at the target's image position by a given factor. Each trial lasted 1 sec. In every trial, observers pressed a key to indicate whether or not the target circle was present. We measured the percentage of correct responses for speed factors of 0.3 to 0.9 for slower moving objects and 1.1 to 1.5 for faster moving objects. Thresholds were computed by fitting the data with a sigmoidal curve and determining the speed factor at which the observers performed with 75% accuracy. The results for 10 observers show that for slower moving objects (speed factor 1), accuracy increased with increasing speed factor, with an average threshold factor of 1.4. Thus, observers can detect a moving object within a radial flow field based on its image speed if its speed is sufficiently slower or faster than the image speeds of objects that are part of the stationary scene.
Supported by NSF grant #IBN-0343825.