一种自适应红外图像增强新算法

针对红外图像经直方图均衡后因部分灰度级被合并而导致细节丢失的问题,提出了一种新的红外图像增强算法——双阈值区分增强算法.该算法首先在原始图像直方图中剔除截断阈值以下的冗余灰度级,接着根据重建直方图的包络确定分界阈值用以区分背景和目标,最后对背景进行直方图均衡处理,对目标进行局部对比度增强及线性灰度变换处理.算法定义了一个自适应参数来适当地抑制背景和噪声,并放大目标细节.相对于传统的直方图均衡算法,所提算法能在增强整体对比度的同时保留图像细节,并且计算量小、实时性好.

New Self-adaptive Image Enhancement Algorithm for Infrared Image

As a new image enhancement algorithm for infrared image,bi-threshold differentiated enhancement algorithm was presented to prevent image details losing because of gray level incorporation during histogram equalization.Firstly,redundant gray levels below the truncation-threshold were eliminated in the original histogram.And then,dividing-threshold was defined based on the enve-lope of rebuilt-histogram to differentiate the background and objects.Finally,histogram equalization was applied to background,while local contrast enhancement and linear gray level transformation were applied to objects.An adaptive parameter Was defined to restrain the background and noise appropriately and to magnify the objects and details.Compared with traditional histogram equalization,the proposed algorithm can enhanee image contrast and hold details simultaneously.Furthermore,it has the features of small calculation and real-time performance.