基于多分辨率分级分块处理的边缘检测方法

为了有效提取运动目标的特征,在目标检测的基础上,要对其进行边缘检测.由于传统逐点扫描方法运算量大,提出一种基于多分辨率分级分块思想的边缘检测算法.该方法先把图像分为多个子块后根据其四个顶点判断该子块是否包含目标,再对目标子块进行扫描确定目标的边缘.实验结果表明,此方法准确地确定目标的边缘,同时检测速度比传统逐点扫描法有一定程度的提高.

A Edge detecting algorithm based on multi-resolution blocked processing

In order to get moving object feature,edge detection should be done based on motion detection.Because that a direct use of the method of traditional scanning points is computationally very expensive,a new algorithm is presented based on multi-resolution blocked processing.The method first divides the denary image into many blocks,which are judged whether to include object according to the four angular vertexes,and then scan the objective blocks to find out the edge of moving object.Experimental results demonstrate the performance of the proposed algorithm,and the detective rapidity is faster than the traditional method.