基于相对熵和边缘检测的图象分割算法  

A Relative Entropy and Edge Detection-based Algorithm to Image Segmentation

作　　者：王建军[1] 苑玮琦[2] 张石[3] 张宏勋[3] 

机构地区：[1]郑州大学计算机系,450052 [2]沈阳建工学院 [3]东北大学

出　　处：《控制与决策》1997年第5期581-584,601,共5页Control and Decision

摘　　要：利用相对熵选择阈值和边缘检测提出一种图象分割算法。其主要思想是通过相对熵来选择最佳阈值,然后用任何一种边缘检测对图象进行分割。将所提出的算法和基于局部熵的算法分别用于现场颗粒物料图象的分割,实验结果表明,该算法优干基于局部熵的图象分割算法。A algorithm of image segmentation is proposed based on relative entropy selection thresholding and edge detection. The main idea of this algorithm is to find a threshold by using relative entropy, and then any edge detection is used to segment image. In order to see the performance of the proposed algorithm. the local entropy algorithm and it are used to segment the plant scene particle material image , respectively. Some ex-perimental results are provided to show the proposed algorithm to be superior to the local entropy-based al-gorithm of image segmentation.

关 键 词：图象分割 相对熵 边缘检测 算法 图象识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]