改进的Mean Shift跟踪算法在车辆跟踪上的应用

Application of Vehicle Tracking Based on Improved Mean Shift Algorithm

使用传统Mean Shift目标跟踪算法实现运动目标跟踪时,跟踪框大小不变,可能会导致跟踪过程中运动目标跟丢的情况发生。因此,提出了一种结合背景差分法的Mean Shift跟踪算法,从而实时地提取出大小合适的运动目标跟踪框。实际应用中,通过DM642数字信号处理器采集D1格式的视频,然后对视频帧图像进行改进的Mean Shift跟踪算法实现后可以发现,改进的Mean Shift跟踪算法可以实时地实现目标跟踪框大小的变化。在跟踪效果上,改进的跟踪算法具有较好的效果。

Mean-shift object tracking algorithm is used to tracking the moving object. In the traditional mean-shift tracking algorithm, the fixed size of tracking box brings the effect, which will lose the moving object during the tacking process. Therefore, an improved mean-shift algorithm is proposed, which combines with background subtraction. By the way, the right tracking box can be extracted. The video is captured that is D1 format by the DM642 digital signal processor, and the frames of images are handled though the improved mean-shift tracking algorithm. It can be conclude that the improved algorithm can get varying tracking box in real time. The improved algorithm has a better effect in tracking process.