基于Retinex-Net的低照度图像增强算法研究
Research on Low-Light Image Enhancement Algorithm Based on Retinex-Net
摘要: 视频监控图像在光照条件不足的情况下，会出现图像黑暗、噪声大、颜色失真等问题，严重干扰了人们对于图像内容的判断，使用图像增强算法可以有效改善图像的质量。在对现有流行的图像增强技术研究基础上，提出改进的Retinex-Net算法。该方法将残差网络用于特征的融合，从而得到更加详细的特征图，并通过加入色彩损失解决增强图像出现色彩偏差的问题。实验表明，改进后的Retinex-Net算法提高了图像的亮度和改善图像色彩失真，整体效果有所提升。
Abstract: In the case of insufficient lighting conditions, video surveillance images will have problems such as dark images, large noise, and color distortion, which seriously interfere with people’s judgment of image content. The use of image enhancement algorithms can effectively improve the quality of images. Based on the research on the existing popular image enhancement techniques, an improved Retinex-Net algorithm is proposed. This method uses the residual network for feature fusion to obtain more detailed feature maps, and solves the problem of color deviation in enhanced images by adding color loss. Experiments show that the improved Retinex-Net algorithm improves the brightness of the image and improves the color distortion of the image, and the overall effect is improved.
文章引用：豆世豪, 陈国瑞, 谭林立. 基于Retinex-Net的低照度图像增强算法研究[J]. 计算机科学与应用, 2022, 12(6): 1658-1664. https://doi.org/10.12677/CSA.2022.126166

参考文献

[1] Cheng, H.D. and Shi, X.J. (2004) A Simple and Effective Histogram Equalization Approach to Image Enhancement. Digital Signal Processing, 14, 158-170. [Google Scholar] [CrossRef] 
[2] Abdullah-Al-Wadud, M., Kabir, M.H., Dewan, M.A.A. and Chae, O. (2007) A Dynamic Histogram Equalization for Image Contrast Enhancement. IEEE Transactions on Consumer Electronics, 53, 593-600. [Google Scholar] [CrossRef] 
[3] 侯雷, 饶云波. 一种亮度融合的视频增强方法[J]. 光电工程, 2011, 38(8): 132-138.
[4] Land, E.H. (1977) The Retinex Theory of Color Vision. Scientific America, 237, 108-129. [Google Scholar] [CrossRef] [PubMed]
[5] He, K., Sun, J. and Tang, X. (2010) Single Image Haze Removal Using Dark Channel Prior. IEEE Transactions on Pattern Analysis and Machine Intelligence, 33, 2341-2353. [Google Scholar] [CrossRef] 
[6] Lore, K.G., Akintayo, A. and Sarkar, S. (2017) LLNet: A Deep Autoencoder Approach to Natural Low-Light Image Enhancement. Pattern Recognition, 61, 650-662. [Google Scholar] [CrossRef] 
[7] Wei, C., Wang, W., Yang, W. and Liu, J. (2018) Deep Retinex Decomposition for Low-Light Enhancement. arXiv preprint arXiv:1808.04560.
[8] Zhang, Y., Zhang, J. and Guo, X. (2019) Kindling the Darkness: A Practical Low-Light Image Enhancer. Proceedings of the 27th ACM International Con-ference on Multimedia, Nice, 21-25 October 2019, 1632-1640. [Google Scholar] [CrossRef] 
[9] Jiang, Y., Gong, X., Liu, D., Cheng, Y., Fang, C., Shen, X., et al. (2021) Enlightengan: Deep Light Enhancement without Paired Supervision. IEEE Transactions on Image Processing, 30, 2340-2349. [Google Scholar] [CrossRef] 
[10] 欧嘉敏, 胡晓, 杨佳信. 改进Retinex-Net的低光照图像增强算法[J]. 模式识别与人工智能, 2021, 34(1): 77-86. [Google Scholar] [CrossRef] 
[11] 李江华, 王坤. 一种基于卷积神经网络的低光照图像增强方法[J]. 江西理工大学学报, 2020, 41(5): 73-79. [Google Scholar] [CrossRef] 
[12] 黄辉先, 陈凡浩. 基于注意力机制和Retinex的低照度图像增强方法[J]. 激光与光电子学进展, 2020, 57(20): 45-52.
[13] 刘佳敏, 何宁, 尹晓杰. 基于Retinex-UNet算法的低照度图像增强[J]. 计算机工程与应用, 2020, 56(22): 211-216.
[14] 仇瑞阳, 吴章铭, 王宇辰. 基于深度学习的低照度图像增强[J]. 电脑知识与技术, 2022, 18(7): 76-77+88. [Google Scholar] [CrossRef] 
[15] 李博文, 唐贵进, 崔子冠. 基于Retinex模型的弱光照图像增强算法[J]. 计算机技术与发展, 2021, 31(5): 79-84.
[16] Dony, R.D. and Wesolkowski, S. (1999) Edge Detection on Color Images Using RGB Vector Angles. Engineering Solutions for the Next Millennium. 1999 IEEE Canadian Conference on Electrical and Computer Engineering (Cat. No. 99TH8411), Vol. 2, Edmonton, 9-12 May 1999, 687-692. [Google Scholar] [CrossRef]