自适应图像边缘检测算法

A Self-adaptable Method to Detect Edge in Images

对LOG算子边缘检测方法的性能进行了分析和评价。针对LOG算子的缺陷,提出了依据图像灰度的一阶导数极大值和二阶导数零穿相结合的边缘检测方法,成功抑制了图像中的大部分噪声;并通过用图像灰度共生矩阵的惯性矩特征值自适应调整高斯空间系数和边缘检测阈值,实现了图像边缘的自动提取。

The capability of edge detection using LOG operators is analyzed and evaluated i n this paper. In order to avoid the defects of LOG operator. This thesis propose d an new edge detection method here. Firstly, by combining the first derivative s maximum and the second derivatives zerocrossing of image grayscale. It has made a good success in eliminating most of noise in image. Subsequently, by using variance of grayscale to selfadaptably definitize Gassian space coeff icient and edge detection threshold, the image edge can be automatically detecte d.