Title :
Motion estimation of a rigid planar patch: a robust computational technique
Author_Institution :
Dept. of Comput. Sci., Purdue Univ., West Lafayette, IN, USA
Abstract :
Robust computing procedures for recovering the motion parameters of a planar patch are presented. Two new concepts are introduced: a virtual plane that is normal to the translational vector and the ratio of the depths of a featured point at two different times. These two concepts drastically simplify the analysis and lead to simple and robust procedures for recovering the motion parameters. The robustness of the technique is discussed for each step involved. Twelve different sets of motion parameters are used to test the technique. Results using real images are provided
Keywords :
computer vision; position measurement; featured point; motion estimation; motion parameters; real images; rigid planar patch; robust computational technique; robust procedures; robustness; translational vector; virtual plane; Cameras; Computer science; Computer vision; Image motion analysis; Image sequence analysis; Layout; Motion analysis; Motion estimation; Noise robustness; Testing;
Conference_Titel :
AI Systems in Government Conference, 1990. Proceedings., Fifth Annual
Conference_Location :
Washington, DC
Print_ISBN :
0-8186-2044-7
DOI :
10.1109/AISIG.1990.63822
