Title :
Estimating the motion direction from brightness gradient on lines
Author :
Guerrero, J.J. ; Sagüés, C.
Author_Institution :
Departamento de Informatica e Ingenieria de Sistemas, Zaragoza Univ., Spain
fDate :
8/1/2001 12:00:00 AM
Abstract :
In previous work we combined feature-based techniques and optical flow methods to obtain depth or motion. The expressions relating the brightness constraint to the three-dimensional (3D) localization and motion of a line and its projection were established. In this paper, those expressions have been used to obtain the motion direction of a camera when the rotation velocity is bounded without assumptions about the depth of lines. Our approach exploits the visibility constraint and it allows us to make use of a priori information about the scene or the motion. With the proposed technique, the topology (easily extracted in the image) that relates adjacent edge elements into line segments is exploited to better compute camera motion. Besides the motion direction, our method allows also to compute the rotation velocity when lines in prominent 3D directions are available
Keywords :
brightness; edge detection; image sequences; motion estimation; stereo image processing; 3D line localization; adjacent edge elements; brightness gradient; camera; line motion; motion direction estimation; rotation velocity; tine segments; topology; visibility constraint; Adaptive control; Automatic control; Brightness; Control systems; Fuzzy control; Fuzzy logic; Fuzzy systems; Motion estimation; Neural networks; Nonlinear systems;
Journal_Title :
Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on
DOI :
10.1109/5326.971671