基于人眼亮度感知的S型函数图像对比度增强算法  

Image Contrast Enhancement Algorithm with S-Type Function Based on Human Visual System

作　　者：王一竹[1] 李渊 杨宇[3] WANG Yizhu;LI Yuan;YANG Yu(School of Control Technology,Wuxi Institute of Technology,Wuxi Jiangsu 214121;State Grid Qinghai Electric Power Company Maintenance Company,Xining 810001;School of Instrument Science and Engineering,Southeast University,Nanjing 210096)

机构地区：[1]无锡职业技术学院控制技术学院,江苏无锡214121 [2]国网青海省电力公司检修公司,西宁810001 [3]东南大学仪器科学与工程学院,南京210096

出　　处：《电子科技大学学报》2022年第4期600-607,共8页Journal of University of Electronic Science and Technology of China

基　　金：国家自然科学基金(51477028)。

摘　　要：为了克服传统变换函数在低照度情况下的局限性,在间接对比度增强领域,提出了一种基于人眼亮度感知对比度灵敏性的S型函数。对于不同的图像亮度,存在不同的视网膜响应值,因此将人眼视网膜的对比度灵敏性建模为对数参数的指数函数。该方法以灵敏度模型作为Steven幂律的指数,推导出一个感知亮度的转换函数。同时还提出了一种参数优化方法,在保持输入图像的平均亮度和直方图的同时,保持信息损失最小。实验结果表明,该方法在保持输入图像平均亮度的情况下,具有更少的信息损失和更低的计算复杂度。在对比度增强、平均亮度保持和细节保持方面具有优势。In the field of indirect contrast enhancement, an S-type function based on the contrast sensitivity of human visual system is proposed, in order to overcome the limitations of traditional transformation functions under low illumination. This method models the contrast sensitivity of the human retina as an exponential function of logarithmic intensity, since there are different retinal response values for different stimulus intensities. This method uses the sensitivity model as the exponent of Steven’s power law to derive a transfer function for perceived brightness. At the same time, a parameter optimization method is proposed, which maintains the average brightness of the input image, and stretches the image histogram while ensuring minimal information loss. Experimental results show that this method has less information loss and better computational complexity performance while maintaining the average brightness of the input image. It has certain advantages in existing methods in terms of contrast enhancement performance, average brightness and preservation of details.

关 键 词：对比度增强 灵敏度模型 S型函数 Steven幂律 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]