摘要
本文提出一种基于边缘似然性估计的边界检测算法，它包含三个主要步骤。首先，从图像中检测局部边缘；其次，通过统计一边缘元素成为局部极大的次数，计算其真正成为边缘的可能性；最后，对边缘元素进行跟踪处理，从而将离散的边缘元素连接成连续的边界。
A new algorithm to detect boundary based on the likelihood of being an edge.ispresented in this paper. It consists of three steps. Firstly, local edges are detected in images. Secondly, the possibility of the local edges being true edges is evaluated by countingthe number of them being local maximum. Finally, the local edges are followed and boundaries are formed.
出处
《小型微型计算机系统》
EI
CSCD
北大核心
1996年第12期47-52,共6页
Journal of Chinese Computer Systems
