多结构元素的数学形态学图像边缘检测算法

A Multi-structure Elements Algorithm for Image Edge-detection Based on Mathematical Morphology

数学形态学作为一门新兴的、以形态为基础对图像进行分析的学科,已得到人们的广泛关注,并应用于图像处理的许多方面。本文提出一种基于多结构元素的数学形态学图像边缘检测算法,该算法通过构造4个不同方向的结构元素,应用形态学梯度算子得到图像4个方向的边缘检测结果,并将这些结果进行加权平均,得到最终的图像边缘。为验证算法的效果,给出了本文算法和几种传统算子对Lena图像进行边缘提取的实验结果。结果表明:本文算法成功地完成了图像的边缘检测,且检测效果明显优于经典的Sobel算子,Laplace算子和Canny算子。

As a newly emerging subject,mathematical morphology is used for analyzing image based on morphology.In this paper,based on mathematical morphology,a new multi-structre element algorithm for image edge-detection is proposed,it builds four structure elements,uses morphologic operator to detect multi-directional edges of image,and averages these edges with different coefficients to gain the final detecting result.In order to test the effect of algorithm,the results of this algorithm and common algorithms to detect edge of image Lena are given.Experiments show that the algorithm detect image edge successfully,and effect of edge-detection is evidently better than classical operators,such as Sobel operator,Laplace operator and Canny operator.