动态直方图双向均衡化的图像增强方法  

Dynamic histogram double equalization for image enhancement

作　　者：曾炜赫 杨俊炜[1] 邢宇翔[1] 

机构地区：[1]清华大学工程物理系,粒子技术与辐射成像教育部重点实验室,北京100084

出　　处：《中国体视学与图像分析》2014年第2期129-139,共11页Chinese Journal of Stereology and Image Analysis

基　　金：国家自然科学基金(11275104);国防科工局技术基础十二五科研项目(2011869)

摘　　要：由于图像增强的对象丰富,现有的方法很难在不同的领域达到很好的效果。本文提出了一种新的基于直方图均衡方法的图像增强技术:动态直方图双向均衡。该方法通过对图像直方图动态分区,独立地进行各个分区的直方图均衡,并均衡了灰度间距,具有很强的适应性和稳定性。经过测试,在不同类型的图像中,该方法在视觉上和量化指标评估上都得到了较好的增强效果。Since images to be enhanced can differ dramatically, existing image enhancement methods could not reach a ubiquitously satisfying effect in different fields. In this paper, a new image enhancement technique based on histogram equalization, dynamic histogram double equalization （DHDE）, is proposed. This method partitions the image histogram dynamically, equalizes them independently and also equalizes the spacing of gray - levels. Therefore it has advantages in adaptability and stability. Experimen- tal studies on different images show that this method achieves outstanding enhancement results on both visual and index assessment.

关 键 词：图像增强 直方图均衡 动态分区 灰度均衡 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]