对无人机拍摄视频中的地面运动目标提出一种实时检测跟踪算法。该算法利用特征点的对应关系将图像对配准，再对配准图像进行变化检测，根据变化和运动信息检测目标并消除虚警，将检测与跟踪相结合，对目标跟踪失效的情况，能重新正确定位目标，从而获取目标的完整运动轨迹。采用无人机拍摄的地面车辆图像的测试结果表明，算法能有效检测跟踪运动目标，并能达到25 f/s以上的实时处理速度。

A real-time algorithm for detecting and tracking ground moving target by using unmanned aerial vehicle(UAV) video was proposed. Two images were matched based on the correspondence of feature points, and the change detection was proceeded on the basis of the matched images.Then according to the changing and moving information, target was detected and part of false alarm was eliminated. If tracking failed,the target could be correctly relocated to obtain full trace of moving target finally by combining detecting with tracking. The test result of groud vehicle image by UAV implies that the moving target can be detected and tracked at a real-time processing speed of more than 25 f/s by using our algorithm.