基于直方图均衡和幂次变换的灰度图像增强算法

摘要

Traditional histogram equalisation ( HE ) algorithm loses image details of bright regions and produces noise in dark regions while enhancing images with uneven illumination, which lead to the descent of information entropy. To solve this problem, in the paper we develop an HE-based image enhancement algorithm. It makes use of the characteristics of HE algorithm in high execution efficiency and good enhancement effect on processing the images with low contrast, and incorporates the feature of power transform that it is able to balance the brightness of image with uneven illumination, and adopts these two algorithms respectively to process the primitive image, after the nonlinear superposition of both results is done, the new algorithm obtains the final outcome by piecewise linear processing. Experimental results show that it has preferable enhancement effect on the gray image with uneven illumination and ample dark regions information.%针对传统直方图均衡化算法增强光照不均匀图像时出现暗部噪点过多、亮部细节丢失,导致图像信息量下降的问题,提出一种基于直方图均衡化的增强算法,该算法利用直方图均衡化算法执行效率高、对低对比度图像增强效果好的特点,结合幂次变换可以平衡光照不均匀图像的亮度的特性,采用两种算法分别对原图进行处理,将结果进行非线性叠加后,经分段线性处理得到最终结果.实验结果表明,对于光照不均匀、暗部信息丰富的灰度图像具有较好的增强效果.