基于改进的数学形态学边缘检测算法研究  

Edge Detection Algorithm Based on Improved Mathematical Morphology

作　　者：王东霞[1] 许伟昶[1] 

机构地区：[1]济源职业技术学院,河南济源459000

出　　处：《实验室研究与探索》2014年第2期89-92,178,共5页Research and Exploration In Laboratory

基　　金：国家自然科学基金项目(11001075);河南省科技攻关项目(132102210229)

摘　　要：研究了彩色图像边缘检测问题。针对传统边缘检测算法存在的边缘分辨率较低、低强度边缘保护能力较差等情况,提出了基于数学形态学彩色图像边缘检测改进算法。该方法将利用数学形态四运算,即膨胀、腐蚀、开、闭等变换以及它们的组合,并根据不同的结构元素的尺度大小和结构元类型,给出了一种改进的形态学抗噪型边缘检测算子,有效地检测出完整的图像边缘信息,并保持图像边缘的平滑性。实验结果证明了,与传统的边缘检测算法相比,本文提出的算法能有效提取准确的边缘信息,而且又具有很强的抗噪性,是一种有效的边缘检测方法。This paper studies the problem of color image edge detection. The existence of traditional edge detection algorithms were lots disadvantages such as the lower-resolution, low-intensity edge of the protection, hence a color image based on mathematical morphology was proposed to improve edge detection. This method used the mathematical form of four operations, dilation, erosion, opening, closing and other transformations and combinations. Depending on different sizes and structures of the element type, the algorithm gave an improved noise canceling edge morphology detection operator. It could effective detect a complete edge information, and maintain the smoothness of edges. Experimental results showed, comparing with the traditional edge detection algorithms, the proposed algorithm could extract the exact edge of the information, and held a strong noise immunity, was an effective edge detection method.

关 键 词：边缘检测 数学形态学 彩色图像 多结构元 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]