一种区域多直方图红外图像增强方法*

直方图均衡是一种简单有效的红外图像增强技术,但存在着细节信息损失较大的缺陷。为改进这一缺陷,使直方图均衡在增强图像对比度的同时不损失灰度级别,并能增强图像细节特征,提出一种基于区域的multi-HE红外图像增强方法。该方法通过聚类算法将图像分割成多目标区域,据此将直方图分割成多个子图,运用多直方图均衡化对图像进行处理,从而达到在不同目标范围内的图像增强。经过实验验证,该算法能有效地抑制背景区的过增强,扩大了目标区的灰度范围,增强细节部分。

Method of infrared image enhancement based on regional histogram

Histogram equalization is a simple and effective infrared image enhancement techniques,but it leads to missing a lot of the detail informations of the objects.To improve this defect,this paper proposed the infrared image enhancement technique based on regional multi-HE,which could enhance image contrast base on histogram equalization,could enhance images’ details and keep no loss of gray level.This method identified the target sub-regions of the image firstly by using a clustering algorithm.Then,it divided the original histogram into many sub-histograms based on the target sub-regions,and equa-lized the image by utilizing the sub-histograms ensuing.Finally,it enhanced the target sub-regions of the image respectively as the result of this method.Experimental results show that the proposed techniques can effectively restrain the excessive enhancement on background region,at the same time,the method can expand the objects’gray zone as well as enhance the objects’ details.