带色彩恢复的多尺度Retinex改进型算法研究  

Research on the improved multi-scale Retinex algorithm with color restoration

作　　者：蔡云[1] CAI Yun(School of Electronic and Electrical Engineering,Anhui Sanlian University,Hefei Anhui 230601)

机构地区：[1]安徽三联学院电子电气工程学院,安徽合肥230601

出　　处：《宁夏师范学院学报》2022年第10期67-73,100,共8页Journal of Ningxia Normal University

基　　金：安徽三联学院校级科研基金项目(KJZD2021007);安徽省教育厅自然科学基金项目(KJ2021A1174)。

摘　　要：提出了一种改进型低光照图像增强算法,即在对低光照图像处理前可以先对其进行取反操作.然后再进行带色彩恢复的多尺度视网膜增强算法处理,使原图像中颜色偏暗区域的亮度和细节得以提高,对处理后的结果再进行取反.最后利用线性加权的方法将二者处理后的结果进行叠加,得到最终结果输出.通过从主观评价、亮度和信息熵对处理图像进行评价,效果较为明显,有一定意义.An improved low-light image enhancement algorithm is proposed.The inversion operation is performed on the low-light image before processing.Then, through the MSRCR algorithm, the brightness and details of the dark areas in the original image are improved, and the processed result is inverted.Again, the processed result is inverted.Finally, the final result is obtained by superposing the processed results of the above two methods using a linear weighting method.The processed images are evaluated through three aspects: subjective evaluation, brightness, and information entropy.The evaluation results show that the image processing effect of this method is more obvious, which has certain reference significance for the image enhancement effect.

关 键 词：清晰度 线性加权 信息熵 

分 类 号：TN911.73[电子电信—通信与信息系统]