CAJ | 학술논문

传统的直方图均衡化算法在增强图像的同时可能会引入一些视觉退化效应，如一些图像的部分区域出现过度增强。为了克服这个缺点，已有一些灰度均值保持算法，但是这些算法并不能很好地保持图像处理前后灰度均值的稳定性。提出了一种自适应动态峰值剪切直方图均衡化算法：使用滤波器对原图像的直方图进行滤波操作，并且根据图像的信息来确定分割区间及区间数目；对分割的区间进行重新映射；对区间的直方图进行剪切操作，然后分别地进行均衡化处理，并对处理后的图像进行灰度归一化操作。实验结果表明，该算法可以很好地在保持原图像均值的前提下实现图像增强。
전통적직방도균형화산법재증강도상적동시가능회인입일사시각퇴화효응，여일사도상적부분구역출현과도증강。위료극복저개결점，이유일사회도균치보지산법，단시저사산법병불능흔호지보지도상처리전후회도균치적은정성。제출료일충자괄응동태봉치전절직방도균형화산법：사용려파기대원도상적직방도진행려파조작，병차근거도상적신식래학정분할구간급구간수목；대분할적구간진행중신영사；대구간적직방도진행전절조작，연후분별지진행균형화처리，병대처리후적도상진행회도귀일화조작。실험결과표명，해산법가이흔호지재보지원도상균치적전제하실현도상증강。
Conventional HE may introduce some visual degradation effect while enhancing image. In order to overcome such drawback, many methods for brightness preserving and contrast enhancement have been proposed, but the methods can’t ensure the aim. This paper proposes a new method, known as Adaptive Dynamic Clipped Histogram Equalization (ADCHE). Firstly, the method smoothes the histogram and partitions the histogram. Next each partition will be assigned to a new dynamic range. And then clipping process is implemented to each sub-histogram. Finally, the conventional HE is implemented to each sub-histogram and normalize the out image to the input mean brightness. The results of the experiments show that the proposed method enhances the contrast while preserving mean brightness.