一种改进的LoG图像边缘检测方法  

An Improved LoG Image Edge Detection Method

作　　者：赵景秀[1] 韩君君 王菁[1] 赵昭[1] 

机构地区：[1]曲阜师范大学计算机科学学院,山东日照276826

出　　处：《计算技术与自动化》2009年第3期74-77,共4页Computing Technology and Automation

基　　金：山东教委科技创新项目(J06b05)

摘　　要：对LoG算子边缘检测算法的性能进行分析,指出LoG算子在实际应用中的不足。针对LoG算子的缺点,通过实验数据得出图像灰度共生矩阵的熵与高斯空间系数的关系。实现使用LoG算子对图像进行检测边缘时能根据具体的图像的熵值自动获取适合该图像的高斯空间系数的值。因此,提出的改进算法既能有效的抑制噪声又能较精确的定位边缘。The capability of edge detection algorithm using LoG operator is analyzed and the deficiency of the LoG operator in the practical application is proposed in this paper. The relation between the entropy of the gray level co- occurrence matrix and the Gaussian space coefficient based on the date of experimentation is deduced, which is aimed at voiding of the defects of the LoG operator. Optimum Gaussian space coefficient of LoG operator can be acquired adaptively based on the entropy of the concrete image. So the improved method in the paper not only effectively restrains most noise of the images, but also locates the edge accurately.

关 键 词：边缘检测 灰度共生矩阵 熵 LOG算子 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]