基于形态学的可变权值匹配自适应图像增强算法  

Adaptive Image Enhancement Algorithm with Variable Weighted Matching Based on Morphology

作　　者：刘艳莉[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,太原030051 [2]仪器科学与动态测试教育部重点实验室,太原030051

出　　处：《电子与信息学报》2014年第6期1285-1291,共7页Journal of Electronics & Information Technology

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

摘　　要：为了精确提取图像的细节特征,改善图像增强的效果,该文提出了一种基于形态学的可变权值自适应增强算法。通过构造扩展全方位多尺度结构元素,并进行top-hat变换,分解了图像不同方向不同尺度的细节。该文打破了传统形态学增强算法中各方向细节取均值融合的思想,实现了根据图像的局部灰度特性,动态调整各方向不同尺度细节的权值。在图像增强过程中,根据提取到的细节的结构化特征,构造相应的自适应增益函数,从而实现图像的自适应增强。实验结果表明,该算法能较好地利用图像的自相关性,比传统的形态学增强方法更有效地突出图像的细节信息,且能抑制噪声的放大。In order to extract accurately the image details, and improve the effect of image enhancement, an adaptive image enhancement algorithm with variable weighted matching based on morphological is proposed. With this method, extension omni-directional multi-scale structure element is constructed, which is used to decompose image of different scale details in different direction through top-hat translation. The proposed algorithm brokes the idea of that the detail weighted in each direction is taken average in traditional morphology method, and adjusts the weight of the different detail direction based on the dynamic characteristic analysis of the local gray level. In the image enhancement process, according to the structured feature of extracted details, the corresponding adaptive gain function is constructed to realize the image adaptive enhancement. The experimental results show that, the algorithm can highlight more effective image details than the traditional morphological method of image enhancement by using the autocorrelation of image, and can suppress the noise in some extent.

关 键 词：图像增强 形态学 自适应 结构元素 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]