Moving object tracking method based on improved lucas-kanade sparse optical flow algorithm
Dan Li, Daihong Jiang, Bao Rong, Sun Jin-Ping, Zhao Wen-jing, WANG Chao · 2017
Improved the traditional L-K algorithm by lifting the wavelet multi-resolution algorithm, and the tracking speed of system was greatly enhanced while combined with SURF matching algorithm. On the basis of detecting feature points, reduced the probability of the exterior points. Tracking local feature points by multi resolution wavelet Pyramid optical flow algorithm solved the problems of object deformation, high speed, fog and haze, uneven illumination, partial occlusion in complex environment. The new method can improve the anti noise ability and improve the efficiency and accuracy of the algorithm. In addition, an adaptive template updating strategy is proposed to avoid tracking failures due to long time tracking errors.