Single Image Dehazing based on Two-stage Optimal Fusion Method

10.6919/ICJE.202205_8(5).0096

Weina Zhou；Chengxiang Ouyang

Image Processing ； Adaptive Filter Window ； Bright Channel Prior ； Atmospheric Scattering Model

International Core Journal of Engineering

8卷5期（2022 / 05 / 01）

750 - 759

英文

Fog is the result of the accumulation of dust and smoke particles in the air, which can lead to a decrease in the visibility and contrast of images, and has a huge impact on outdoor video surveillance, daily camera photography, and more. Based on the improved DCP algorithm, this paper proposes a new two-stage single image removal algorithm. The two stages are the optimization of the transmission map and atmospheric light, respectively. For the transmission map, a rough transmission map is obtained by weight fusion, which consists of sky and non-sky regions. The non-sky area is obtained by the improved dark channel algorithm based on the adaptive filter window, and the sky area is obtained by the classical color attenuation prior algorithm. Finally, a weighted fusion method is proposed to combine the transmission maps of the two regions, so that the intersection of the combined coarse transmission maps is more natural. For atmospheric light, we combine the advantages of the bright channel and propose to replace the global atmospheric light with the local atmospheric light value, so that the estimated atmospheric light can better reflect the real ocean environment. The experimental results show that the proposed dehazing algorithm can greatly improve the visibility of the image and effectively avoid the color distortion of the hazy image.

1. Land, Edwin H. The Retinex Theory of Color Vision[J]. Scientific American, 1977, 237(6):108-128.
   連結：
2. Doo Hyun Choi; Ick Hoon Jang; Mi Hye Kim. Color Image Enhancement Based on single-scale Retinex With a JND-Based Nonlinear Filter[C]. IEEE International Symposium on Circuits and Systems, 2007.
   連結：
3. Jobson D J, Rahman Z, Woodell G A. Multiscale retinex for bridging the gap between color images and the human observation of scenes[J]. IEEE Transactions on Image Processing, 1997, 6(7):965-976.
   連結：
4. Z.Rahman, D.J.Jobson, G.A.Woodell, Multiscale retinex for color rendition and dynamic range compression, in Proceedings of SPIE2847, Society of Photo-Optical Instrumentation Engineers Press, Bellingham, 1996, pp. 183–191.
   連結：
5. S. G. Narasimhan and S. K. Nayar, Contrast restoration of weather degraded images, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 6, pp. 713-724, Jun. 2003.
   連結：
6. J.B. Wang, N.He, L.L.Zhang, et al., Single image dehazing with a physical model and dark channel prior, Neurocomputing 149(2015)718–728.
   連結：
7. Nayar S K, Narasimhan S G. Vision in Bad Weather[C]// Seventh IEEE International Conference on Computer Vision. 2002.
   連結：
8. He K, Jian S, Tang X. Single image haze removal using dark channel prior[C] IEEE Conference on Computer Vision & Pattern Recognition. 2009.
   連結：
9. Berman, D., Treibitz, T., & Avidan, S. (2016). Non-local Image Dehazing. In: Computer Vision and Pattern Recognition. Las Vegas. pp:1674-1682.
   連結：
10. Zhu Q S, Mai J M, Shao L. A Fast Single Image Haze Removal Algorithm Using Color Attenuation Prior[J]. IEEE Trans. on Image Processing. 2015-24(11): 3522-3533.
    連結：
11. K. He, J.Sun, X.Tang, Guided image filtering, Pattern Anal. Mach. Intell. 35(6) (2013)1397–1409.
    連結：
12. Q. Shu, C. Wu, R. W. Liu, et al., Two-phase transmission map estimation for robust image dehazing, in Proc. ICONIP, Siem Reap, Cambodia, Dec. 2018, pp. 529-541.
    連結：
13. Cai, B., Xu, X., Jia, K., Qing, C., & Tao, D. (2016). DehazeNet: An End-to-End System for Single Image Haze Removal. IEEE Transactions on Image Processing, 25(11), 5187-5198.
    連結：
14. W. Ren, S. Liu, H. Zhang, J. Pan, X. Cao, and M. H. Yang, Single image dehazing via multi-scale convolutional neural networks, in Proc. ECCV, Amsterdam, The Netherlands, Oct. 2016, pp. 154-169.
    連結：
15. B. Li, X. Peng, Z. Wang, J. Xu, and D. Feng, Aodnet: All-in-one dehazing network, in Proc. IEEE ICCV, Venice, Italy, Oct. 2017, pp. 4780-4788.
    連結：
16. S. Zhang, W. Ren, and J. Yao, FEED-Net: Fully end-to-end dehazing, in Proc. IEEE ICME, San Diego, CA, USA, Oct. 2018, pp. 1-6.
    連結：
17. Jiang Y T Sun C M Zhao Y, et al., Image dehazing using adaptive bi-channel priors on superpixels[J[, Computer Vision and Image Understanding 2017,165, 17-32.
    連結：
18. Achanta R, Shaji A, Smith K, et al., SLIC superpixels compared to state-of-the-art superpixel methods[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(11): 2274-2282.
    連結：
19. Duo Qiu; Hu Zhu; Xiongjun Zhang, An Implementable Accelerated Alternating Direction Method of Multipliers for Low-Rank Tensor Completion, 2018 IEEE International Conference on Internet of Things, 2018.
    連結：
20. Xu Y, Guo X, Wang H, et al. Single image haze removal using light and dark channel prior [C], IEEE International Conference on Communications in China, 2016: 1- 6.
    連結：
21. Hai-Miao Hu, Qiang Guo, Jin Zheng, et al., Single Image Defogging Based on Illumination Decomposition for Visual Maritime Surveillance. 2019.
    連結：
22. N. Hautiere, J.P. Tarel, D. Aubert, et al., Blind contrast enhancement assessment by gradient rationing at visible edges, Image Anal. Stereol[J]. 27 (2) (2008)87–95.
    連結：
23. L. Chen, B.L.Guo, J.Bi, et al., Algorithm of single image fog removal based on joint bilateral filter, J.Beijing Univ. Posts Telecommun. 35(4)(2012)19–23.
24. Zhou W, Zhou Y. An attention nested U-Structure suitable for salient ship detection in complex maritime environment[J]. IEICE Transactions on Electronics, Vol.E105-D,No.6,pp.-,Jun. 2022.
25. E. J. McCartney, Optics of the atmosphere, Science Press, 1988.