融合残差及通道注意力机制的单幅图像去雨方法

doi:10.3969/j.issn.1000-1158.2021.01.04

摘要
图/表
参考文献(18)
相关文章 (15)

全文: PDF (4567 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要为了去除雨天图像上附着的雨滴并恢复图像的清晰度,提出一种基于深度学习思想结合图像增强技术融合残差及通道注意力机制来实现的单幅图像去雨方法。首先,利用导向滤波将有雨图像分解为平滑基本层和高频细节层;其次,提出自适应Gamma校正算法增强平滑基本层以提高对比度;然后,构建融合残差块和通道注意力机制的深度神经网络实现高频细节层去雨;最后,将去雨后的高频细节层与增强后的平滑基本层融合实现单幅图像去雨功能。实验结果表明:与具有代表性的单幅图像去雨方法相比,所提方法效果较好并可保留更多的图像细节信息。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	张世辉
	闫晓蕊
	桑榆

关键词 ：计量学, 单幅图像去雨, 图像处理, 压缩和激励残差网络, 注意力机制, 深度学习, Gamma校正

Abstract：In order to remove the raindrops and restore the image sharpness,a single image rain removal method based on depth learning and image enhancement technology combined with residual and channel attention mechanism is proposed.Firstly, the rainy image is decomposed into the smooth base layer and the high-frequency detail layer by using the guided filter. Secondly,an adaptive gamma correction algorithm is proposed to enhance the smooth base layer to improve contrast. Thirdly, the deep neural network with residual block and the channel attention mechanism is constructed to remove rain in the high-frequency detail layer. Finally, the high-frequency detail layer after rain removal is combined with the enhanced smooth base layer to realize the single image rain removal. The experimental results show that compared with the representative single image rain removal method, the proposed method works well and can retain more image detail information.

Key words： metrology single image rain removal image processing squeeze and excitation residual network attention mechanism deep learning Gamma correction

收稿日期: 2019-10-11 发布日期: 2021-01-19

PACS:

TB96

基金资助:国家自然科学基金(61379065);河北省自然科学基金(F2019203285)

通讯作者: 张世辉 E-mail: sshhzz@ysu.edu.cn

作者简介: 张世辉(1973-),男,河北赞皇人,燕山大学教授,博士生导师,主要从事视觉信息处理、模式识别方面的研究。Email:sshhzz@ysu.edu.cn

引用本文:

张世辉,闫晓蕊,桑榆. 融合残差及通道注意力机制的单幅图像去雨方法[J]. 计量学报, 2021, 42(1): 20-28.
ZHANG Shi-hui,YAN Xiao-rui,SANG Yu. Single Image Rain Removal Method by Fusing Residual and Channel Attention Mechanism. Acta Metrologica Sinica, 2021, 42(1): 20-28.

链接本文:

http://jlxb.china-csm.org:81/Jwk_jlxb/CN/10.3969/j.issn.1000-1158.2021.01.04 或 http://jlxb.china-csm.org:81/Jwk_jlxb/CN/Y2021/V42/I1/20

［1］Zhi T C, Pires B R, Hebert M, et al. Multispectral Imaging for Fine-Grained Recognition of Powders on Complex Backgrounds ［C］// Proceedings of 2019 32nd IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, USA, 2019: 8691-8700.
［2］Chen Y, Chen X D, Zhu J J, et al. Development of an autonomous unmanned surface vehicle with object detection using deep learning ［C］// Proceedings of 2018 44th Annual Conference of the IEEE Industrial Electronics Society, Washington, USA, 2018: 5636-5641.
［3］郭茂森, 王宇, 张建国, 等. 最小控制递归平均算法对光纤声音传感系统的降噪作用［J］. 计量学报, 2019, 40 (5): 880-886.
Guo M, Wang Y, Zhang J, et al.Noise Reduction Effect Of Minimum Control Recursive Average Algorithm On Optical Fiber Acoustic Sensor System ［J］. Acta Metrologica Sinica, 2019, 40 (5): 880-886.
［4］温丽梅, 周苗苗, 李明, 等. 改进的Tikhonov正则化图像重建算法［J］. 计量学报, 2018, 39 (5): 679-683.
Wen L, Zhou M, Li M, et al. Improved Tikhonov Regularization Method for Image Reconstruction ［J］. Acta Metrologica Sinica, 2018, 39 (5): 679-683.
［5］Garg K, Nayar S K. Detection and removal of rain from videos ［C］// Proceedings of 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Washington, USA, 2004, 1: 1528-1535.
［6］Zhang X P, Li H, Qi Y Y, et al. Rain Removal in Video by Combining Temporal and Chromatic Properties ［C］// Proceedings of 2006 IEEE International Conference on Multimedia and Expo, Toronto, Canada, 2006: 461-464.
［7］Kang L L, Lin C W, Fu Y H. Automatic Single-Image-Based Rain Streaks Removal via Image Decomposition ［J］. IEEE Transactions on Image Processing, 2011, 21 (4): 1742-1755.
［8］Huang D A, Kang L W, Wang Y F. Self-Learning Based Image Decomposition With Applications to Single Image Denoising ［J］. IEEE Transactions on Multimedia, 2014, 16 (1): 83-93.
［9］Luo Y, Xu Y, Ji H. Removing Rain From a Single Image via Discriminative Sparse Coding ［C］// Proceedings of 2015 IEEE International Conference on Computer Vision, Santiago, Chile, 2015: 3397-3405.
［10］Fu X Y, Huang J B, Ding X H, et al. Clearing the Skies: A Deep Network Architecture for Single-Image Rain Removal ［J］. IEEE Transactions on Image Processing, 2017, 26 (6): 2944-2956.
［11］Fu X Y, Huang J B, Zeng D L, et al. Removing Rain from Single Images via a Deep Detail Network ［C］// Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, USA, 2017: 1715-1723.
［12］Yang W, Tan R T, Feng J, et al. Deep Joint Rain Detection and Removal from a Single Image ［C］// Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, USA, 2017: 1685-1694.
［13］Zhang H, Patel V M. Density-aware Single Image De-raining using a Multi-stream Dense Network ［C］// Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, USA, 2018: 695-704.
［14］Fan Z W, Wu H F, Fu X Y, et al. Residual-Guide Network for Single Image Deraining ［C］// Proceedings of 2018 ACM Multimedia Conference, Seoul, Korea, 2018: 1751-1759.
［15］He K M, Zhang X Y, Ren S Q, et al. Deep Residual Learning for Image Recognition ［C］// Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 2016: 770-778.
［16］Ledig C, Theis L, Huszar F, et al. Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network ［C］// Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, USA, 2017: 105-114.
［17］Hu J, Shen L, Sun G. Squeeze-and-Excitation Networks ［C］// Proceedings of 2018 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Salt Lake City, USA, 2018: 7132-7141.
［18］Wang Z, Bovik A C, Sheikh H R, et al. Image Quality Assessment: From Error Visibility to Structural Similarity［J］. IEEE Transactions on Image Processing, 2004, 13 (4): 600-612.