基于多结构元的数学形态学的边缘检测  

Morphological Edge Detection Based on Multi-Structuring Elements

作　　者：罗湘[1] 韦玉科[1] 

机构地区：[1]广东工业大学计算机学院,广东广州510006

出　　处：《计算机技术与发展》2012年第4期85-87,92,共4页Computer Technology and Development

基　　金：国家"十一五"科技支撑计划项目(2006BAI08B01-3)

摘　　要：针对传统边缘检测算子具有对噪声敏感、提取出的边缘细节特征没有得到很好的保持等不足,文中在Canny算法启发下提出了一种基于多结构元的数学形态学边缘检测算法,并经过数学形态学细化算法提取图像边缘,并与Sobel边缘检测算法和Canny边缘检测算法得到的图像边缘仿真结果进行对比,分析比较其优点不足,得出结论:文中算法得到的图像边缘其定位准确且较平滑,轮廓清晰,很好地保留了边缘细节特征,比较容易并行实现,而且由于此种边缘检测算法抗噪能力较强,能较好解决性噪比和单边缘响应之间的矛盾。According to the traditional edge detection operators to noise sensitive,the extracted features edge details are not well maintained,it presents a multi-structuring elements based on morphological edge detection algorithm inspired by the Canny algorithm,and the edge got through morphological thinning algorithm,and compare the simulation results of the Sobel edge detection algorithm and Canny edge detection algorithm,analyse the advantages and disadvantages to a drawn conclusion： the edge of this paper＇s algorithm extracted has a more accurate orientation and relatively smoother contour,more clear outline,easy to parallel implementation,well maintained features edge detail,and because this kind of edge detection algorithm has a strong antinoise ability,it can well solved the conflict between performance/noise ratio and single edge response.

关 键 词：边缘检测 数学形态学 多结构元 细化算法 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]