基于局部复杂度和粗糙度的形态学空间锐化法  

Image Sharpening Algorithm Based on Local Coarseness and Complexity in Morphological Space

作　　者：薛英娟[1] 张权[1] 刘艳莉[1] 郝慧艳[1] 桂志国[1,2] 

机构地区：[1]中北大学电子测试技术国家重点实验室,山西太原030051 [2]中北大学仪器科学与动态测试教育部重点实验室,山西太原030051

出　　处：《中北大学学报（自然科学版）》2014年第3期342-347,共6页Journal of North University of China(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61071192;61271357;61171178);山西省自然科学基金资助项目(2009011020-2);山西省高等学校优秀青年学术带头人支持计划项目;山西省国际合作项目(2013081035);山西省研究生优秀创新项目(20123098)

摘　　要：传统锐化算法对灰度突变的强边缘响应强烈,对噪声敏感,易产生过冲效应,且对灰度变化较小的微弱细节锐度不足,为此提出一种新的图像锐化算法.利用顺序形态变换的相关性质和概念,构造了一种局部加权均值滤波器,克服了传统线性滤波在平滑图像的细节和噪声时,一些重要边缘也被平滑而易产生过冲的问题;同时针对图像灰度剧变区和级别丰富区,应用图像局部粗糙度和复杂度自适应调节增益函数,有效地提升了图像中弱边缘和纹理细节的表现力.实验结果表明:该算法有效地避免了强边缘过冲现象,且在增强图像边缘和微弱细节的同时,保持了整体背景噪声与原图像一致,抑制了噪声的放大.Traditional sharpening algorithm is sensitive to noise,it is intense to strong edges of gray mutation respond,and easy to suffer from overshoot effects on strong edges,moreover the acutance for weak details is insufficient,for this problem,a novel image sharpening approach was proposed.According to correlative properties and conception of order morphology transformation to construct a local weighted mean filter to smooth the original image details and noise,and the problem of overshoot effects which was often caused in traditional linear filter by smoothing essential edges was avoided.Meanwhile,for the regions in image where gray value is rich or change intensely,local complexity and local coarseness were adopted to dynamically control gain function,so as to effectively improve the weak edges or texture details of image.Experimental results show that the algorithm not only avoids overshoot effects on strong edges,but also improves the weak edges and texture details when keeping the overall background noise and the original image consistent,and effectively restraining the noise amplification.

关 键 词：锐化算法 顺序形态变换 结构元素 局部复杂度 局部粗糙度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]