昏暗光照下人脸图像校正算法研究  

Research on Face Image Correction Algorithm in Dim Light

作　　者：王卓琳 郑宇[1] 冯洋亿 WANG Zhuolin;ZHENG Yu;FENG Yangyi(University of South China,Hengyang Hunan 421000,China)

机构地区：[1]南华大学,湖南衡阳421000

出　　处：《信息与电脑》2022年第8期101-104,共4页Information & Computer

基　　金：2020年湖南省大学生实践创新项目(项目编号:S202010555108)。

摘　　要：笔者选取暗通道先验去雾算法对图像进行预处理,在此基础上研究Retinex算法和带色彩恢复的多尺度视网膜增强(Multi-Scale Retinex with Color Restoration,MSRCR)算法对人像增强效果的区别。首先采用暗通道先验、图像去雾算法补偿图像照度,其次通过Retinex和MSRCR算法进行人像增强与色彩修复,最后通过尺度不变特征变换(Scale Invariant Feature Transform,SIFT)算法识别质量增强的图像。The dark channel prior defogging algorithm is selected to preprocess the image. The difference between Retinex algorithm and MSRCR algorithm in portrait enhancement effect is studied. Firstly, dark channel prior and image defogging algorithm are used to compensate the image illumination. Secondly, the image enhancement and color restoration are carried out by Retinex and MSRCR algorithms respectively. Finally, the processed image with enhanced quality is identified by SIFT algorithm.

关 键 词：暗通道去雾 MSRCR算法 特征提取 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]