视频监控中运动物体的检测与跟踪

针对固定摄像头下的交通监控场景，首先给出一种基于分块原理的背景重建算法，克服了平均法重建的背景图像模糊的缺点。然后用减背景方法检测运动物体，并利用数学形态学方法对得到原始前景点作处理，填补了运动物体内部的空洞，减少了噪声点，改善了检测性能。为适应背景的变化，对背景进行自适应更新，并且通过对Meanshift算法的改进提高了跟踪的准确性。实验结果表明，算法在有效检测到运动物体的同时能够快速准确地跟踪运动物体。

Moving object detection and tracking in video surveillance

For the stationary scene, a block based background reconstruction algorithm is presented which can overcome fuzzy shortcoming by mean method.Then the background-subtraction method is used to detect the moving objects.After that the morphological process is adopted to fill the holes in moving objects,wipe out the noise spots and improve the detection. In order to fit to the background change, the adaptive background updating is used.To improve the accuracy of tracking, the modified Meanshift algorithm is carried out.Simulations show that this algorithm can detect objects effectively and track the objects rapidly and accurately.