一种改进的Canny算子边缘检测算法  

An improved Canny edge detecuion algorithm

作　　者：周晓明[1] 马秋禾[1] 肖蓉[1] 杨娜[1] 

机构地区：[1]信息工程大学测绘学院,河南郑州450052

出　　处：《测绘工程》2008年第1期28-31,共4页Engineering of Surveying and Mapping

摘　　要：介绍Canny算子边缘检测的基本原理,并对其性能进行分析和评价。针对传统Canny算子在滤波过程中存在的缺陷,提出一种基于自适应平滑滤波的改进Canny边缘检测算子,该算法根据图像中像元灰度值的突变特性,自适应的改变滤波器的权值,在平滑图像的过程中使图像的边缘锐化。在计算梯度幅值的时候采用了邻域的梯度幅值计算方法,考虑了像素对角线方向的梯度计算,进一步抑制了噪声的影响。通过对实验图像的分析表明,改进的检测算法对图像边缘提取具有较好的检测精度和准确性。This paper introduces fundamental theory of Canny operator, and carrys on its analysis and evaluation. On this foundation, an improved Canny edge detection algorithm based on adaptive smooth filtering is proposed. According to characteristic of revulsion every point gray scale, this algorithm adaptively changescoefficients of the filter, according to gradient on each point. The results of the experiment pictures indicate that the improved algorithm has the better accuracy and precision in the edge detection.

关 键 词：图像处理 边缘检测 CANNY算子 自适应滤波 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]