Image Dehazing Network Based on Densely Connected Residual Block and Channel Pixel Attention

doi:10.16182/j.issn1004731x.joss.21-1160

Abstract

Abstract:

Abstruct: A lot of research achievements have been made in image dehazing based on neural network,but there aiming at the fog residue, even the color distortion and texture loss, in complex outdoor image dehazing, an image dehazing network based on densely connected residual block and channel pixel attention is proposed. Densely connected residual blocks are used to extract and fuse the features of foggy images,and the repair module with channel pixel attention mechanism is used to repair the color and texture of the feature maps. The experimental results show that, compared with the existing methods, the proposed method and significantly improves the objective evaluation index and subjective visual quality, effectively avoid the color distortion, texture loss and residual fog in the process of image dehazing.

Key words: image dehazing, densely connected residual block, attention mechanism, color distortion, detail texture

CLC Number:

TP391.9

Weidong Jin, Shuli Zhang, Peng Tang, Man Zhang. Image Dehazing Network Based on Densely Connected Residual Block and Channel Pixel Attention[J]. Journal of System Simulation, 2022, 34(8): 1663-1673.

Figures/Tables 13

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Table 1

Fig. 5

Fig. 6

Table 2

Fig. 7

Table 3

Fig. 8

Table 4

Fig. 9

References 23

1	蒲亨飞, 黄治勇. 图像去雾算法研究综述[J]. 软件工程, 2021, 24(5): 2-6.
	Pu Hengfei, Huang Zhiyong. An Overview of Reaserch on Image Dehazing Algorithms[J]. Software Engineering, 2021, 24(5): 2-6.
2	Narasimhan S G, Nayar S K. Vision and the Atmosphere[J]. International Journal of Computer Vision(S0920-5691), 2002, 48(3): 233-254.
3	He K, Sun J, Tang X. Single Image Haze Removal Using Dark Channel Prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S1939-3539), 2011, 33(12): 2341-2353.
4	Zhu Q, Mai J, Shao L. A Fast Single Image Haze Removal Algorithm Using Color Attenuation Prior[J]. IEEE Transactions on Image Processing (S1941-0042), 2015, 24(11): 3522-3533.
5	Zhao Dong, Xu Long, Yan Yihua, et al. Multi-Scale Optimal Fusion Model for Single Image Dehazing[J]. Signal Processing: Image Communication (S0923-5965), 2019, 74: 253-265.
6	金炜东, 胡燕花, 唐鹏, 等. 基于结构化森林的前向车载视频图像的细节强化[J]. 系统仿真学报, 2018, 30(12): 4602-4610.
	Jin Weidong, Hu Yanhua, Tang Peng, et al. Detailed Enhancement of Forward Vehicle Video Images Based on Structured Forest[J]. Journal of System Simulation, 2018, 30(12): 4602-4610.
7	胡燕花. 接触网巡检图像的细节强化与场景重构[D]. 成都: 西南交通大学, 2018.
	Hu Yanhua. Detailed Enhancement and Scene Reconstruction of Catenary Inspection Images [D]. Chengdu: Southwest Jiaotong University, 2018.
8	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 (S1941-0042), 2016, 25(11): 5187-5198.
9	Ren W Q, Liu S, Zhang H, et al. Single Image Dehazing Via Multi-Scale Convolutional Neural Networks[C]//European Conference on Computer Vision. Springer International Publishing AG, 2016: 154-169.
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. Venice, Italy: IEEE, 2017: 4780-4788.
11	Qin X, Wang Z L, Bai Y C, et al. FFA-Net: Feature Fusion Attention Network for Single Image Dehazing[C]//Association for the Advance of Artificial Intelligence. 2020: 11908-11915.
12	He K, Zhang X, Ren S, et al. Deep Residual Learning for Image Recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV: IEEE, 2016: 770-778.
13	Huang G, Liu Z, Van Der Maaten L, et al. Densely Connected Convolutional Networks[C]//IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI: IEEE, 2017: 2261-2269.
14	Hu J, Shen L, Albanie S, et al. Squeeze-and-Excitation Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S1939-3539), 2020, 42(8): 2011-2023.
15	Ioffe S, Szegedy C. Batch Normalization: Accelerating Deep Net-Work Training by Reducing Internal Covariate Shift[C]// IEEE International Conference Machine Learning. 2015: 1-11.
16	Roodschild M, Gotaysardias J, Will A. A New Approach for the Vanishing Gradient Problem on Sigmoid Activation[J]. Progress in Artificial Intelligence(S2192-6352), 2020, 9(4): 351-360.
17	Odena A, Dumoulin V. Deconvolution and Checkerboard Artifacts[J]. Distill(S2476-0757), 2016, 1(10): e3.
18	杨丽娟, 李利. 基于双线性插值的内容感知图像缩放算法仿真[J]. 计算机仿真, 2019, 36(12): 244-248.
	Yang Lijuan, Li li. Simulation of Content-Aware Image Scaling Algorithms Based on Bilinear Interpolation[J]. Computer Simulation, 2019, 36(12): 244-248.
19	Chen W, Ding J, Kuo S. PMS-Net: Robust Haze Removal Based on Patch Map for Single Images[C]//IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA: IEEE, 2019: 11673-11681.
20	Woo S, Park J, Lee J Y. CBAM: Convolutional Block Attention Module[C]//European Conference on Computer Vision. 2018: 3-19.
21	Zhou W, Bovik A C. Mean Squared Error: Love It or Leave It? A New Look at Signal Fidelity Measures[J]. IEEE Signal Processing Magazine(S1053-5888), 2009, 26(1): 98-117.
22	Li B Y, Ren W Q, Fu D P, et al. Benchmarking Single Image Dehazing and Beyond[J]. IEEE Transactions on Image Processing(S1941-0042), 2019, 28(1): 492-505.
23	Li L, Dong Y, Ren W, et al. Semi-Supervised Image Dehazing[J]. IEEE Transactions on Image Processing(S1941-0042), 2020, 29: 2766-2779.

方法	β=0.06		β=0.10		β=0.16		β=0.20		Average
方法	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM
DCP	16.25	0.813 3	16.01	0.818 4	15.89	0.799 8	15.91	0.790 7	16.02	0.805 6
CAP	21.61	0.881 9	22.15	0.893 2	21.98	0.874 4	20.77	0.843 2	21.63	0.873 2
MOF	18.57	0.861 2	18.65	0.872 2	18.96	0.857 3	18.63	0.833 3	18.70	0.855 1
MSCNN	20.38	0.851 0	23.61	0 894 1	22.79	0.876 9	20.17	0.838 8	21.74	0.865 2
AOD-Net	21.82	0.880 5	24.80	0.915 7	22.28	0.867 5	19.52	0.836 8	22.11	0.875 1
本文	28.46	0.921 8	27.71	0.920 1	25.34	0.888 5	24.07	0.870 2	26.40	0.900 2

方法	L_dc
DCP	0.136 4
CAP	0.172 5
MOF	0.162 3
MSCNN	0.189 3
AOD-Net	0.187 0
本文	0.149 4

方法	PSNR/dB	SSIM
DRCA-CTR	26.40	0.900 2
RCA-CTR	25.19	0.890 0
DR-CTR	25.54	0.893 2
DRCA	24.90	0.882 4

图像	BGR比重
原始图像	(0.18, 0.36, 0.26)
有雾图像	(0.47, 0.58, 0.51)
DRCA	(0.19, 0.30, 0.21)
DRCA-CTR	(0.19, 0.38, 0.29)

[1]	Junjie Qiu, Hong Zheng, Yunhui Cheng. Research on Prediction of Model Based on Multi-scale LSTM [J]. Journal of System Simulation, 2022, 34(7): 1593-1604.
[2]	Yin Wang, Feixiang Wang, Qianlai Sun. Vehicle Detection Method Based on Multi Scale Feature Fusion [J]. Journal of System Simulation, 2022, 34(6): 1219-1229.
[3]	Yin Shi, Hou Guolian, Chi Yan, Gong Linjuan, Hu Xiaodong. Prediction Method for Health Degree of Front Bearing of Wind Turbine Generator and Implementation [J]. Journal of System Simulation, 2021, 33(6): 1323-1333.
[4]	Yang Weilong, Xu Kai, Xie Xu, Sun Lin. Research on CGF-oriented Virtual Human Perceptual Attention Model [J]. Journal of System Simulation, 2021, 33(2): 262-270.
[5]	Kong Yisi, Hu Xiaofeng, Zhu Feng, Tao Jiuyang. Attention Mechanism in Battlefield Situation Awareness [J]. Journal of System Simulation, 2017, 29(10): 2233-2241.