基于轨道特征的电子稳像算法研究

he monocular vision is the key technology for the locomotive anti-collision warning system. The range precision influence the system’s performance. In this paper according to the question of video jitter result in the accuracy reducing, proposed a new EIS algorithm based on the orbital characteristic, through extracting and matching partial feature template obtained the global movement vector. Treated the partial feature template, instead of treating the global image, the speed of system was improved obviously. The result of simulation indicated that this algorithm can effectively eliminate image migration which produces because of the video jitter, has solved the deviation of ranging precision, and satisfies the real-time request of system.