基于改进的形态学算子的灰度图像边缘检测  

A Method to Improve Edge Detecting Algorithm Based on Mathematical Morphology

作　　者：徐建东[1] 蒋野[1] 孙迎春[1] 李帅[1] 孙凯[2] 姜国千 

机构地区：[1]佳木斯大学信息电子技术学院,黑龙江佳木斯154007 [2]哈尔滨理工大学测控技术与通信工程学院,黑龙江哈尔滨150040

出　　处：《佳木斯大学学报（自然科学版）》2009年第6期857-859,共3页Journal of Jiamusi University：Natural Science Edition

摘　　要：提出一种改进的形态学图像边缘检测算法.针对单结构元素在检测时出现检测结果边缘线条粗、不连续性的问题,该算法提出了利用6种具有代表性的结构元素进行膨胀组合,形成多尺度结构元素,从而弥补了传统边缘检测算法提取灰度图像细节边缘少及抑制噪声能力差的缺点.An improved edge detecting algorithm based on mathematical morphology was proposed to overcome the limitation which embarrasses the performance of the traditional mathematical morphological methods.The new algorithm inherited the merit of the traditional methods,and introduced the new idea considering about the orientation of the edge.To get thinner edge in the morphological gradient graph,multi-structure element was put forward.After using the new edge connecting method based on gradient direction,the ultimate edge was obtained.

关 键 词：图像分割 数学形态学 多尺度结构元素 边缘检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]