视频流中的自适应阈值模板匹配车辆检测算法

Algorithm of Adaptive Thresholds of Template Matching for Vehicle Detection in Video Flow

在传统帧差法的基础上,通过对帧差法得到的感兴趣区域进行阈值处理,得到用白色代表前景、黑色代表背景的黑白二值图像,并设计了自适应阈值模板匹配车辆检测算法从黑白二值图像中定位出运动车辆.实验结果表明,该算法简单,容易实现,能准确定位车辆.

On the basis of traditional successive images difference, threshold processing was made to the regionof interest which was obtained by the method of successive images difference, getting foreground which is represented by white pix and background which is represented by black pix, then a new algorithm was designed to detect the vehicle -- algorithm of adaptive thresholds of template matching for vehicle detection in video flow. Experiments demonstrate that the algorithm is simple, implemented easily, and can detect the vehicle exactly.