A study of optimal angle of color vector for optical flow extraction

摘要

The image processing technique of optical flow is very useful for recognition of objects moving in captured images. This paper describes a way for generating optical flow from colored images. The colored image provides more rich information for the object recognition than the gray scaled one. In our approach, RGB data in the captured colored images are converted to the color vectors. The simultaneous equation consisting of derivation of the vector based on gradient-based method resolves the optical flows. Since the color vector has arbitrary orientation, the analysis of the optimal angle of the vector is reported in this paper.