可保留图像细节的直方图修正法

Histogram Modification Methods Without Image Details Losing

针对图像经直方图修正 (Histogram Modification,HM)后因部分灰度级被合并而导致细节丢失的问题 ,提出了两种可保留图像细节的直方图修正方法 :一种是扩展灰度范围的 HM方法 ,另一种为局部灰度修正 HM方法。对多幅不同类型图像进行处理后的实验结果表明 ,这两种方法不仅可有效保留图像细节信息 ,而且算法简洁 ,计算量小。

The problem of gray level incorporation during image histogram modification is analyzed in the paper. And two histogram modification methods which can prevent image details losing are presented. One is addressed by expanding the gray-level range of image, so less gray levels are integrated and more details of the image can be hold. The other method is realized by modifying local gray level. According to the method, the image's histogram is modified by way of general histogram modification method firstly. Then the gradient of modified image and original image is calculated and compared. The pixels whose gradient decreased markedly are found. And their gray levels are modified according to the gradient of the original image finally. So the details lost during histogram modification can be resumed. Both the methods can enhance the image contrast and hold the details simultaneously. Further more, the proposed approach is simple and easy to be performed. Experimental results are given at the end of the paper. Four images are adopted to demonstrate the performance of the methods.