一种改进的基于Canny算子的图像边缘提取算法  

Improved Image Edge Extraction Algorithm Based on Canny Operator

作　　者：张震[1] 马驷良[1] 张忠波[1] 刘辉[1] 宫跃欣[1] 孙秋成[1] 

机构地区：[1]吉林大学数学研究所

出　　处：《吉林大学学报（理学版）》2007年第2期244-248,共5页Journal of Jilin University:Science Edition

基　　金：国家863项目基金(批准号:2003AA1Z2141)

摘　　要：对Canny边缘提取算法进行改进,采用二维高斯函数的一阶偏导数构造滤波器计算梯度幅值,利用一种四阈值边缘检测定位方法定位边缘,最后引入数学形态学方法对边缘细化处理.改进算法对噪声抑制效果明显,能够删除伪边缘,检出边缘更加精细,实验结果表明了改进算法的可行性和有效性.This paper introduces a new operator constructed with the first order approach of improved-Canny operator for edge detection. The filter partial derivative of two-dimensional Gaussian function is employed to compute gradient. Then the edge was located via the edge detection and localization method with four threshold values. And finally, edge thin operation was conducted by introducing a morphological operator. The improved algorithm can enhance the capability of suppressing noise, delete false edges and obtain exact edges. Experimental results indicate the feasibility and validity of the improved algorithm.

关 键 词：CANNY算子 边缘提取 图像处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]