天空区域分割修正的图像去雾新算法

doi:10.19665/j.issn1001-2400.2019.02.027

摘要/Abstract

摘要：

为了改进基于暗通道先验假设图像去雾算法的细节信息丢失、大气光强值估计偏低、天空区域去雾效果不佳等不足,笔者提出一种天空区域分割修正的彩色图像去雾新算法。新算法在暗通道先验算法的基础上,对暗通道与亮通道先验模型进行带参线性加权运算,提出加权平均融合COPLIP模型和MSR模型的天空区域修正新模型及实现算法。与现有去雾算法比较,实验结果表明新算法能够克服现有算法对于天空区域去雾效果不佳的问题,同时通过客观评价指标验证了新算法的有效性。

关键词: 图像去雾, 参数驱动, 天空区域分割修正, 图像融合

Abstract:

In order to solve the problems of the detailed information loss, the low atmospheric light intensity estimation and the poor defogging effect in the sky region based on the dark channel prior image defogging algorithm, this paper proposes a new color image defogging algorithm for sky region segmentation correction. Based on the dark channel prior algorithm, the dark channel and bright channel prior model is linearly weighted with a parameter. This paper also proposes a new sky region correction model based on weighted average fusion of the COPLIP model and the MSR model. Compared with the existing dehazing algorithm, experimental results show that the new algorithm can overcome the problem of poor effect in the sky region. In addition, the paper also validates the effectiveness of the new algorithm by objective evaluation indicators.

Key words: image defogging, parameter-driven, sky region segmentation correction, image fusion

中图分类号:

TP391.41

暴婉婷,王俊平,魏书蕾,李艳波,周勇. 天空区域分割修正的图像去雾新算法[J]. 西安电子科技大学学报, 2019, 46(2): 164-169.

BAO Wanting,WANG Junping,WEI Shulei,LI Yanbo,ZHOU Yong. Novel image defogging algorithm for sky region segmentation correction[J]. Journal of Xidian University, 2019, 46(2): 164-169.

图/表 5

图1

表1

图2

表2

表3

参考文献 18

[1]	WU D, ZHU Q S, WANG J J , et al. Image Haze Removal: Status, Challenges and Prospects[C]// Proceedings of the 2014 4th IEEE International Conference on Information Science and Technology. Piscataway: IEEE, 2014: 492-497.
[2]	李锦, 王俊平, 万国挺 , 等. 一种结合直方图均衡化和MSRCR的图像增强新算法[J]. 西安电子科技大学学报, 2014,41(3):103-109. doi: 10.3969/j.issn.1001-2400.2014.03.015
	LI Jin, WANG Junping, WAN Guoting , et al. Novel Algorithm for Image Enhancement with Histogram Equalization and MSRCR [J ]. Journal of Xidian University, 2014,41(3):103-109. doi: 10.3969/j.issn.1001-2400.2014.03.015
[3]	LENG X G, JI K F, XING X W , et al. Hybrid Bilateral Filtering Algorithm Based on Edge Detection[J]. IET Image Processing, 2016,10(11):809-816. doi: 10.1049/iet-ipr.2015.0574
[4]	QIAO T, REN J C, WANG Z , et al. Effective Denoising and Classification of Hyperspectral Images Using Curvelet Transform and Singular Spectrum Analysis[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017,55(1):119-133. doi: 10.1109/TGRS.2016.2598065
[5]	LI Y, WANG J P, GAO K , et al. Fast Morphological Filtering Haze Removal Method from A Single Image[J]. Journal of Computational Information Systems, 2015,11(16):5799-5806.
[6]	YU T H, MENG X, ZHU M , et al. An Improved Multi-scale Retinex Fog and Haze Image Enhancement Method[C]// Proceedings of the 2016 International Conference on Information System and Artificial Intelligence. Piscataway: IEEE, 2017: 557-560.
[7]	ANCUTY C O, ANCUTI C, BEKAERT P . Effective Single Image Dehazing by Fusion[C]// Proceedings of the 2010 International Conference on Image Processing. Washington: IEEE Computer Society, 2010: 3541-3544.
[8]	ALHARBI E M, GE P, WANG H . A Research on Single Image Dehazing Algorithms Based on Dark Channel Prior[J]. Journal of Computer and Communications, 2016,4(2):47-55. doi: 10.4236/jcc.2016.42006
[9]	WANG W C, CHANG F L, JI T , et al. A Fast Single-Image Dehazing Method Based on A Physical Model and Gray Projection[J]. IEEE Access, 2018,6:5641-5653. doi: 10.1109/ACCESS.2018.2794340
[10]	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. doi: 10.1109/TPAMI.2010.168 pmid: 20820075
[11]	BI G L, REN J Y, FU T J , et al. Image Dehazing Based on Accurate Estimation of Transmission in the Atmospheric Scattering Model[J]. IEEE Photonics Journal, 2017,9(4):797629. doi: 10.1109/JPHOT.2017.2726107
[12]	LIU Y, Li H, WANG M . Single Image Dehazing via Large Sky Region Segmentation and Multi-scale Opening Dark Channel Model[J]. IEEE Access, 2017,5:8890-8903. doi: 10.1109/ACCESS.2017.2710305
[13]	汪云飞, 刘华伟, 赵搏欣 . 融合超像素的黄金暗通道单幅图像去雾方法[J]. 西安电子科技大学学报, 2017,44(5):147-152. doi: 10.3969/j.issn.1001-2400.2017.05.025
	WANG Yunfei, LIU Huawei, ZHAO Boxin . Superpixels-based Golden Dark Channel for Single Image Fog Removal[J]. Journal of Xidian University, 2017,44(5):147-152. doi: 10.3969/j.issn.1001-2400.2017.05.025
[14]	JACKDON J, ARIYO O, ACHEAMPONG K , et al. Hybrid Single Image Dehazing with Bright Channel and Dark Channel Priors[C]// Proceedings of the 2017 2nd International Conference on Image, Vision and Computing. Piscataway: IEEE, 2017: 381-385.
[15]	XIANG D, BAN Y, WANG W , et al. Adaptive Superpixel Generation for Polarimetric SAR Images With Local Iterative Clustering and SIRV Model[J ]. IEEE Transactions on Geoscience and Remote Sensing, 2017,55(6):3115-3131. doi: 10.1109/TGRS.2017.2662010
[16]	冯玉颖, 王俊平, 马塾亮 , 等. COPLIP新模型及其图像增强新算法[J]. 西安电子科技大学学报, 2018,45(1):66-71. doi: 10.3969/j.issn.1001-2400.2018.01.012
	FENG Yuying, WANG Junping, MA Shuliang , et al. New COPLIP model and its image enhancement algorithm[J]. Journal of Xidian University, 2018,45(1):66-71. doi: 10.3969/j.issn.1001-2400.2018.01.012
[17]	WANG W C, YUAN X H . Recent Advances in Image Dehazing[J]. IEEE/CAA Journal of Automatica Sinica, 2017,4(3):410-436. doi: 10.1109/JAS.2017.7510532
[18]	HE K M, SUN J, TANG X O . Guided Image Filtering[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013,35(6):1397-1409. doi: 10.1109/TPAMI.2012.213

M	STD	AG	IE	MSE	PSNR	SSIM	EPI	PAIDM
1.0	1.0000	1.0000	0.0000	1.0000	0.0000	0.0000	1.0000	0.0000
0.9	0.8962	0.8606	0.0532	0.7405	0.0100	0.1907	0.8660	0.4042
0.8	0.7923	0.7394	0.1039	0.5268	0.2130	0.3677	0.7457	0.9438
0.7	0.7444	0.6182	0.2496	0.3548	0.3392	0.5326	0.6349	1.4943
0.6	0.5801	0.5091	0.4123	0.7322	0.4906	0.6821	0.5324	1.4096
0.5	0.4692	0.4121	0.5256	0.1234	0.5763	0.8093	0.4357	2.2334
0.4	0.3517	0.3152	0.6248	0.0574	0.6746	0.9107	0.3376	2.4820
0.3	0.2231	0.2061	0.7781	0.0161	0.8098	0.9777	0.2280	2.7507
0.2	0.1145	0.1152	0.8967	0.0000	0.9449	1.0000	0.1262	2.9451
0.1	0.0462	0.0485	0.9527	0.0010	1.0000	0.9983	0.0541	2.9906
0.0	0.0000	0.0000	1.0000	0.0073	0.9875	0.9863	0.0000	2.9665

算法	IE	PSNR	SSIM	EPI
FATTAL	8.3852	16.6203	0.6451	1.9277
MSR	14.7031	17.6591	0.7774	3.4981
HE	11.4709	53.7415	0.9386	2.9339
新算法	14.3740	68.8075	0.9994	1.0940