Title :
In search of the elusive flow field
Author :
Perrone, John A.
Author_Institution :
NASA Ames Res. Center, Moffett Field, CA, USA
Abstract :
A specific set of problems must be overcome before an optical-flow field can be derived from an image sequence. These problems are discussed and an outline of a scheme for overcoming these obstacles is presented. A 2-D motion sensor is presented which is capable of measuring edge velocity in twelve separate directions at any point in the image. However, because of the aperture problem, the sensors can only determine the velocity component in the direction orthogonal to the edge. A second stage which uses a cosine-weighted voting scheme is used to derive the true flow field from the edge velocities. The whole system is based on a parallel network of connections and weights and is potentially a very fast method of obtaining the 2-D flow field
Keywords :
computer vision; optical information processing; 2-D motion sensor; aperture problem; cosine-weighted voting scheme; edge velocity; elusive flow field; image sequence; optical-flow field; parallel network; Fluid flow measurement; Humans; Image generation; Image motion analysis; Motion measurement; Navigation; Optical sensors; Psychology; Retina; Velocity measurement;
Conference_Titel :
Visual Motion, 1989.,Proceedings. Workshop on
Conference_Location :
Irvine, CA
Print_ISBN :
0-8186-1903-1
DOI :
10.1109/WVM.1989.47108
