关键词: 智能交通, 车辆跟踪, 特征点提取, 光流法, SURF特征, KTL算法

Abstract:

To handle the dynamic illumination changes, scale variation and partial occlusion in complex traffic scene, a stable and reliable vehicle tracking method based on feature point is proposed. With the characteristics of large interframe motion of moving vehicle, the pyramid model of KLT algorithm is constructed and the feature points with poor stability will be deleted according to forward and backward tracking offset. Meanwhile, SURF algorithm is used as a compensation mechanism for updating and adjusting the feature point sets. Then, according to the relative location and relative angle of feature points in the first frame, the scale and rotation variation of objects in the current frame is determined. Finally, the object region is determined in the current frame. The experimental results show that the proposed method effectively solves the problems of object deformation and partial occlusion, and it is robust to scale and rotation variation.

Key words: intelligent transportation, vehicle tracking, feature point extraction, optical flow, SURF features, KLT algorithm