Single Image Dehazing Based on Middle Channel Compensation

doi:10.12068/j.issn.1005-3026.2021.02.005

Abstract

Abstract: Focusing on foggy images with low contrast and blurry detail， this paper decomposes the images into texture layer and structural layer firstly. Then the structural layer containing most of the fog is dehazed， and the texture layer is enhanced. To avoid the effect of white objects in the scene on atmospheric light estimation， this paper proposes a kind of stereo decision map in RGB space， and designs an atmospheric light estimation method which can adaptively distinguish between sky and non-sky area based on a threshold constraining scheme. Due to the inefficiency of dark channel prior when dealing with large areas of sky and dense fog， a transmission estimation approach based on middle channel compensation is put forward， which can avoid color distortion of dehazed images. Furthermore， the transmission is refined by the guided filtering with side window， which can preserve image details. Extensive experimental results show that the proposed approach can realize haze removal thoroughly， and the dehazed image has natural colors and vivid details.

Key words: image dehazing; middle channel compensation; adaptive atmospheric light estimation; guided filtering with side window; image layering

CLC Number:

TP391

YANG Ai-ping， XING Jin-na， LIU Jin， LI Xiao-xiao. Single Image Dehazing Based on Middle Channel Compensation[J]. Journal of Northeastern University(Natural Science), 2021, 42(2): 180-188.

References

[1]Ai Y，Tsai P，Yao C，et al.Improved local histogram equalization with gradient based weighting process for edge preservation ［J］.Multimedia Tools and Applications，2017，76(1):1585-1613.
[2]Liu H B，Yang J，Wu Z P，et al.A fast single image dehazing method based on dark channel prior and Retinex theory［J］.Acta Automatica Sinica，2015，41(7):1264-1273.
[3]Li H，Xie W H，Wang X G.GPU implementation of multi-scale Retinex image enhancement algorithm［C］//IEEE/ACS International Conference of Computer Systems and Applications.Agadir，2016:1-5.
[4]Meng G，Wang Y，Duan J，et al.Efficient image dehazing with boundary constraint and contextual regularization［C］//Proceedings of the IEEE International Conference on Computer Vision.Sydney，2013:617-624.
[5]Fattal R.Single image dehazing［J］.ACM Transactions on Graphics，2008，27(3):1-9.
[6]Tarel J P，Hautière N.Fast visibility restoration from a single color or gray level image［C］//IEEE International Conference on Computer Vision.Kyoto，2009:2201-2208.
[7]He K M，Sun J，Tang X O.Single image haze removal using dark channel prior［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，2011，33(12):2341-2353.
[8]He K M，Sun J，Tang X O.Guided image filtering［J］.IEEE Transactions on Software Engineering，2013，35(6):1397-1409.
[9]Cai B L，Xu X M，Jia K，et al.DehazeNet:an end-to-end system for single image haze removal［J］.IEEE Transactions on Image Processing，2016，25(11):5187-5189.
[10]Li B Y，Peng X L，Wang Z Y，et al.AOD-Net:all-in-one dehazing network［C］//IEEE International Conference on Computer Vision(ICCV).Venice，2017:4780-4788.
[11]Li C Y，Guo J C，Porikli F，et al.A cascaded convolutional neural network for single image dehazing［J］.IEEE Access，2018，6:24877-24887.
[12]Ren W Q，Liu S，Zhang H，et al.Single image dehazing via multi-scale convolutional neural networks［C］ //Proceedings of 2016 European Conference on Computer Vision.Munich，2016:154-169.
[13]Narasimhan S，Nayar S.Vision and the atmosphere ［J］.International Journal of Computer Vision，2002，48(3):233-254.
[14]Li Y，Guo F F，Tan R T，et al.A contrast enhancement framework with JPEG artifacts suppression［M］.Lecture Notes in Computer Science.Heidelberg:Springer，2014，8690:174-188.
[15]曲晨，毕笃彦，严盛文，等.基于明暗像素先验的随机游走图像去雾［J］.系统工程与电子技术，2017(10):2368-2375.(Qu Chen，Bi Du-yan，Yan Sheng-wen，et al.Defogging of random walk images based on light and dark pixel priors［J］.Systems Engineering and Electronics，2017(10):2368-2375.)
[16]陈本豪，高涛，卢玮，等.基于雾天图像退化模型的自适应参数优化的去雾算法［J］.科学技术与工程，2019，19(21):219-227.(Chen Ben-hao，Gao Tao，Lu Wei，et al.Dehazing algorithm based on adaptive parameter optimization of foggy image degradation model［J］.Science Technology and Engineering，2019，19(21):219-227.)
[17]Yin H，Gong Y H，Qiu G P.Side window guided filtering［J］.Signal Processing，2019，165:315-330.
[18]Qin X，Wang Z L，Bai Y C，et al.FFA-Net:feature fusion attention network for single image dehazing［J/OL］.［2020-05-16］.https://www.researchgate.net/publication/337336627_FFA-Net_Feature_Fusion_Attention_Network_for_Single_Image_Dehazing.
[19]Yang H H，Yang C H H，Tsai Y C J.Y-net:multi-scale feature aggregation network with wavelet structure similarity loss function for single image dehazing［J/OL］.［2020-05-16］.https://arxiv.org/pdf/2003.13912.pdf.