多尺度top-hat变换提取细节的对比度增强算法  

Contrast enhancement using extracted details based on multi-scale top-hat transformation

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,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：《计算机工程与设计》2014年第4期1332-1335,1340,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61071192;61271357;61171178);山西省自然科学基金项目(2009011020-2);山西省国际合作基金项目(2013081035);山西省研究生优秀创新基金项目(2009011020-2)

摘　　要：针对传统对比度增强算法对图像增强的不足,提出一种基于形态滤波重构原图像的对比度增强方法。该方法使用多尺度top-hat变换提取图像多尺度下的亮、暗细节特征,并根据多尺度下局部细节特征的重要性,利用非线性函数对这些特征进行反差增强,突出图像隐藏的信息。实验结果表明,与传统算法相比,该方法有效的增强了图像的对比度,且能抑制噪声放大,视觉效果更好,避免了传统对比度增强算法存在的过增强或细节增强不足的问题,适用范围较广。In view of the defects of traditional contrast enhancement algorithms, a new contrast enhancement method based on morphological filtering is proposed to reconstruct the original image. Multi-scale top-hat transform is used to extract the bright and dark features of the image. And according to the importance of these extracted local features, the contrast enhancement be- tween the extracted features is achieved using nonlinear function to highlight the hidden information of the image. The results of experiments show that compared with general methods, the new way can effectively enhance the image contrast, has certain noise suppression capability, and make the enhancement image better visual effects. Also, it can prevent over enhancement or in sufficient in details which is often observed in the traditional contrast enhancement algorithms, and this method has wide applica- tions.

关 键 词：数学形态学 多尺度形态学 TOP-HAT变换 对比度增强 细节特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]