HSV空间的明度对数斜率变换的低照图像增强  

Low Light Image Enhancement Based on Logarithmic Slope Transformation of Lightness in HSV Space

作　　者：万里勇 汪浩[1,3] 罗文兵 WAN Liyong;WANG Hao;LUO Wenbing(Management Science and Engineering Research Center,Jiangxi Normal University,Nanchang Jiangxi 330022,China;School of Information and Artificial Intelligence,Nanchang Institute of Science and Technology,Nanchang Jiangxi 330108,China;School of Software,Jiangxi Normal University,Nanchang Jiangxi 330027,China;School of Computer and Information Engineering,Jiangxi Normal University,Nanchang Jiangxi 330027,China)

机构地区：[1]江西师范大学管理科学与工程研究中心,江西南昌330022 [2]南昌工学院信息与人工智能学院,江西南昌330108 [3]江西师范大学软件学院,江西南昌330027 [4]江西师范大学计算机信息工程学院,江西南昌330027

出　　处：《传感技术学报》2024年第7期1202-1209,共8页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金项目(61562063);江西省教育厅科学技术研究项目(GJJ212517,GJJ191105);南昌工学院机器人与智能系统研究中心项目(NGYZY-20-005);南昌工学院科技计划博士专项项目(NGKJ-22-01)江西省科技厅重点研发计划项目(20192BBEL50031);江西省高校人文社会科学研究项目青年项目(JY21242);江西省教育科学“十四五”规划课题项目(23YB327)。

摘　　要：为了在保持图像颜色自然的同时,适宜地提升图像的亮度和对比度,丰富图像的细节信息,提出了HSV空间的明度对数斜率变换的低照图像增强方法。将低照图像转到HSV颜色空间,将明度分量V进行多尺度Retinex分解,得到多尺度的细节层图像和光照图像;对多尺度的细节层图像进行自适应分数拉普拉斯增强,而用自适应的对数斜率变换对光照图像进行增强;将增强后的明度分量V与色调H和饱和度S分量组合,得到亮度适宜、高对比度和高清晰度的图像。与部分现有方法进行比较,实验数据显示所提方法的增强图像的视觉效果更佳、信息熵和平均梯度更高,因此具有更优的图像增强性能。In order to maintain the color naturalness,properly improve the brightness and contrast,and enrich the details for the image,a low light image enhancement method based on logarithmic slope transformation of lightness in HSV space is proposed,which transforms the low light image into HSV color space,and performs multi-scale Retinex decomposition on the lightness component V,so as to achieve multi-scale detail layer images and illumination image.The multi-scale detail layer images are subjected to adaptive fractional Laplace enhancement,and the illumination image is enhanced by adaptive logarithmic slope transformation.The combination of the enhanced lightness component V,the hue H and saturation S obtains an image with appropriate brightness,high contrast and high definition.Compared with some existing methods,the data show that the enhanced image of the proposed method has better visual effect,higher information entropy and higher average gradient.Thus,the proposed method has superior image enhancement performance.

关 键 词：图像增强 HSV空间 RETINEX 分数拉普拉斯增强 对数斜率变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]