一种运动背景下目标快速定位方法

针对运动背景下目标检测算法计算量大,难以实现实时跟踪的问题,提出了一种目标快速定位方法.该方法采用图像配准补偿背景运动量,由多帧图像差分消除背景图像获得目标图像,通过自适应阈值对目标图像二值化,用形态学处理消除噪声斑点,最后通过连通区域像素个数统计判决,进一步去除伪目标,最终定位目标区域.试验结果表明该方法能在2-6 ms内准确定位平移运动背景下的目标.

A Fast Method of Locating Target under Moving Background

A fast method of locating the target under the moving background is proposed in light of the computation of the usual algorithm is too large to realize the real-time target tracking.Firstly,the shift of the moving background can be compensated by the image registration;then the mean and difference operating on the registered image sequences will be made to eliminate the background image;the corresponding target image can be got;furthermore,set an adapting threshold the binary target image can be achieved;finally,the binary target image will be morphologically processed to eliminate the noisy spots;and the number of the pixels including the linked regions will be judged,then the false targets will be eliminated and the true target will be located.The test results show that the method proposed can precisely locate the target under the parallel moving background in 2~6 ms.