实现对运动目标的跟踪主要依据的是运动估计技术。运动估计，大体来说，就是将图像序列的每一帧分成许多互不重叠的宏块，在下一帧图像中根据一定的匹配准则找出与特定块最相似的块。所以，运动估计的核心在于匹配。常见的运动估计匹配准则有：MAD、MSE和NCCF，由于MAD没有乘除操作，实现简单方便，所以使用较多。但通常使用求和绝对误差（SAD）代替MAD。 本文对YUV色彩空间做了具体的研究。首先利用RGB与YUV的转换关系实现了将AVI格式的视频转换为YUV格式的视频。然后具体研究了YUV视频如何播放。最后，基于运动估计的基本原理，研究如何利用图像序列的每一帧中像素块的亮度信息并对其进行处理——应用绝对差和最小准则进行像素块的匹配，实现了对选定的特定点的跟踪，并勾画出这一点运动的轨迹。同时，实现了在团队比赛视频中对某位运动员的跟踪。

The achievement of moving object tracking mainly depends on motion estimation technology. Motion estimation, generally speaking, is to divide each frame in the image sequence into many macro blocks which don’t overlap each other. Then identify the blocks which match the specific blocks in the next frame according to a certain matching criterion. Therefore, the essential of motion estimation is to find a match algorithm. The common match algorithms used in motion estimation are MAD, MSE and NCCF. As MAD does not need multiple and division operation, it is simpler and more commonly used. However, we usually use SAD instead of MAD. This paper makes a detailed study on YUV color space. First, it demonstrates how to transform the AVI to YUV based AVI. Second, it discusses how to play the YUV based AVI. Lastly, based on the basic theory of motion estimation, it studies how to use the brightness of every pixel in the frame to match the blocks with SAD, and finally how to track the specific points and plot the route. At the final part, we use it to track the athlete in a video as an example.