Superpixel-based Mean and Mean Square Deviation Dark Channel for Single Image Fog Removal
-
摘要: 雾、霾天气会引起图像严重降质.本文假设局部雾浓度恒定，认为暗通道优先的有效性随景深递增呈指数衰减，其比例大小可间接反映雾浓度高低.在此基础上提出一种基于超像素的均值-均方差暗通道单幅图像去雾算法.首先通过超像素分割得到景深恒定的小区域，接着在每个区域内计算均值-均方差暗通道：用均值替代最小值抑制景深突变处的光晕效应，同时采用均方差对其进行修正纠正景深无限远处的偏色问题，由此生成的透射率在超像素内保持不变且更加精细、准确.实验结果表明该算法在雾浓度较大时能够显著提高大景深图像的可见性.Abstract: Fog and haze can cause serious image degradation. In this paper the fog concentration is supposed to be locally constant, and the effectiveness of dark channel prior decays exponentially as the depth increases, with its proportion indirectly reflecting the level of fog density. Based on this, we propose an SMMD (superpixel-based mean-mean square deviation dark channel) algorithm for single image fog removald. Firstly, the small regions in which the depth is kept constant by superpixel segmentation are obtained. Then the mean-mean square deviation dark channel is computed in these regions. This dark channel can suppress the "halo effect" which occurs in depth mutation using mean instead of minimum. Moreover, the mean square deviation is employed to revise mean in order to rectify the cast problems which happen when depth is infinite. On the foundation of that the resulting transmittance keeps constant and is finer and more precise. Experimental results show that the algorithm can effectively improve the visibility images of large depth in field when fog concentration is big.
-
Key words:
- Image defogging /
- atmospheric scattering model /
- superpixel /
- mean /
- mean square deviation
-
-
表 1 光晕效应强度
Table 1 The intensity of halo effect
图像名称(分辨率) 原图 DCP Tarel SMMD 大山(551$\, \times\, $416) 0.578 0.448 0.567 0.321 黄山(184$\, \times\, $256) 0.657 0.425 0.575 0.262 
下载: 导出CSV
表 2 暗通道比例
Table 2 The ratio of dark channel
图像名称(分辨率) 原图(%) DCP (%) Tarel (%) SMMD (%) (参数取值) 大山(551$\, \times\, $832) 0.04 59.04 17.96 96.69 (${T}=3$, $n=64$, $k=1$) 黄山(739$\, \times\, $1 024) 18.57 38.29 41.72 59.56 (${T}=2$, $n=64$, $k=1$) 香港(431$\, \times\, $800) 2.18 60.97 34.12 82.83 (${T}=2$, $n=64$, $k=1$) 院子(693$\, \times\, $711) 37.27 51.29 75.11 78.33 (${T}=2$, $n=64$, $k=0$) 黄山(369$\, \times\, $1 024) 30.52 67.52 70.96 89.05 (${T}=2$, $n=64$, $k=1$)
下载: 导出CSV
-
[1] Choi L K, You J, Bovik A C. Referenceless prediction of perceptual fog density and perceptual image defogging. IEEE Transactions on Image Processing, 2015, 24(11):3888-3901 doi: 10.1109/TIP.2015.2456502 [2] Tan R T. Visibility in bad weather from a single image. In: Proceedings of the 2008 IEEE Conference on Computer Vision and Pattern Recognition. Anchorage, USA: IEEE, 2008. 1-8 https://www.computer.org/csdl/proceedings/cvpr/2008/2242/00/04587643-abs.html [3] Pan X Y, Xie F Y, Jiang Z G, Yin J H. Haze removal for a single remote sensing image based on deformed haze imaging model. IEEE Signal Processing Letters, 2015, 22(10):1806-1810 doi: 10.1109/LSP.2015.2432466 [4] Li Z G, Zheng J H. Edge-preserving decomposition-based single image haze removal. IEEE Transactions on Image Processing, 2015, 24(12):5432-5441 doi: 10.1109/TIP.2015.2482903 [5] Ancuti C O, Ancuti C. Single image dehazing by multi-scale fusion. IEEE Transactions on Image Processing, 2013, 22(8):3271-3282 doi: 10.1109/TIP.2013.2262284 [6] 南栋, 毕笃彦, 马时平, 何林远, 娄小龙.基于景深约束的单幅雾天图像去雾算法.电子学报, 2015, 43(3):500-504 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=dianzixb201503013Nan Dong, Bi Du-Yan, Ma Shi-Ping, He Lin-Yuan, Lou Xiao-Long. Single image dehazing method based on scene depth constraint. Acta Electronica Sinica, 2015, 43(3):500-504 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=dianzixb201503013 [7] 胡伟, 袁国栋, 董朝, 疏学明.基于暗通道优先的单幅图像去雾新方法.计算机研究与发展, 2010, 47(12):2132-2140 http://www.cqvip.com/QK/94913X/201012/36262101.htmlHu Wei, Yuan Guo-Dong, Dong Zhao, Shu Xue-Ming. Improved single image dehazing using dark channel prior. Journal of Computer Research and Development, 2010, 47(12):2132-2140 http://www.cqvip.com/QK/94913X/201012/36262101.html [8] Fattal R. Single image dehazing. ACM Transactions on Graphics, 2008, 27(3):Article No.72 http://dl.acm.org/citation.cfm?id=1360671 [9] Kratz L, Nishino K. Factorizing scene albedo and depth from a single foggy image. In: Proceedings of the 12th IEEE International Conference on Computer Vision. Kyoto, Japan: IEEE, 2009. 1701-1708 http://ieeexplore.ieee.org/xpls/icp.jsp?arnumber=5459382 [10] Tarel J P, Hautiére N. Fast visibility restoration from a single color or gray level image. In: Proceedings of the 12th IEEE International Conference on Computer Vision. Kyoto, Japan: IEEE, 2009. 2201-2208 http://ieeexplore.ieee.org/xpls/icp.jsp?arnumber=5459251 [11] He K M, Sun J, Tang X O. Single image haze removal using dark channel prior. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12):2341-2353 doi: 10.1109/TPAMI.2010.168 [12] He K M, Sun J, Tang X O. Guided image filtering. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(6):1397-1409 doi: 10.1109/TPAMI.2012.213 [13] Zhu S, Cao D H, Jiang S X, Wu Y B, Hu P. Fast superpixel segmentation by iterative edge refinement. Electronics Letters, 2015, 51(3):230-232 doi: 10.1049/el.2014.3379 [14] Shi J B, Du Y F, Wang W G, Li X L. Lazy random walks for superpixel segmentation. IEEE Transactions on Image Processing, 2014, 23(4):1451-1462 doi: 10.1109/TIP.2014.2302892 [15] Achanta R, Shaji A, Smith K, Lucchi A, Fua P, Süsstrunk S. SLIC superpixels compared to state-of-the-art superpixel methods. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(11):2274-2282 doi: 10.1109/TPAMI.2012.120 [16] Arbelaez P, Maire M, Fowlkes C, Malik J. Contour detection and hierarchical image segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(5):898-916 doi: 10.1109/TPAMI.2010.161 [17] Neubert P, Protzel P. Superpixel benchmark and comparison[Online], available: http://www.tu-chemnitz.de/etit/proaut/forschung/superpixel.html, May 3, 2012 [18] 汪云飞, 毕笃彦, 刘华伟, 刘凌, 赵晓林.一种局部受限的规则聚类超像素算法.西安电子科技大学学报(自然科学版), 2016, 43(3):95-100 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=xadzkjdx201603017Wang Yun-Fei, Bi Du-Yan, Liu Hua-Wei, Liu Ling, Zhao Xiao-Lin. Locally-restricted regular clustering superpixel algorithm. Journal of Xidian University (Natural Science), 2016, 43(3):95-100 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=xadzkjdx201603017 [19] Noh S W, Ahn B, Kweon I S. Haze removal on superpixel domain. In: Proceedings of the 10th IEEE International Conference on Ubiquitous Robots and Ambient Intelligence (URAI). Jeju, Korea: IEEE, 2013. 597-598 http://ieeexplore.ieee.org/document/6677400/ [20] Singh G. Evaluation of various digital image fog removal algorithms. International Journal of Advanced Research in Computer and Communication Engineering, 2014, 7(3):7536-7540 [21] 李大鹏, 禹晶, 肖创柏.图像去雾的无参考客观质量评测方法.中国图象图形学报, 2011, 16(9):1753-1757 http://d.wanfangdata.com.cn/Periodical_zgtxtxxb-a201109027.aspxLi Da-Peng, Yu Jing, Xiao Chuang-Bai. No-reference quality assessment method for defogged images. Journal of Image and Graphics, 2011, 16(9):1753-1757 http://d.wanfangdata.com.cn/Periodical_zgtxtxxb-a201109027.aspx [22] Drago F, Myszkowski K, Annen T, Chiba N. Adaptive logarithmic mapping for displaying high contrast scenes. Computer Graphics Forum, 2003, 22(3):419-426 doi: 10.1111/cgf.2003.22.issue-3 [23] 吴迪, 朱青松.图像去雾的最新研究进展.自动化学报, 2015, 41(2):221-239 http://www.aas.net.cn/CN/abstract/abstract18603.shtmlWu Di, Zhu Qing-Song. The latest research progress of image dehazing. Acta Automatica Sinica, 2015, 41(2):221-239 http://www.aas.net.cn/CN/abstract/abstract18603.shtml [24] 陈书贞, 任占广, 练秋生.基于改进暗通道和导向滤波的单幅图像去雾算法.自动化学报, 2016, 42(3):455-465 http://www.aas.net.cn/CN/abstract/abstract18833.shtmlChen Shu-Zhen, Ren Zhan-Guang, Lian Qiu-Sheng. Single image dehazing algorithm based on improved dark channel prior and guided filter. Acta Automatica Sinica, 2016, 42(3):455-465 http://www.aas.net.cn/CN/abstract/abstract18833.shtml 期刊类型引用(12)
1. 宁君，彭周华，李铁山，陈俊龙. 带有输入量化的分布式多无人船舶自适应模糊编队控制. 控制与决策. 2024(08): 2588-2596 . 
百度学术2. 宁君，陈汉民，李伟，王启福，李铁山，陈俊龙. 基于扩张状态观测器的有限时间船舶编队控制. 中国舰船研究. 2023(01): 60-66 . 百度学术3. 马俊达，谭冲，刘可. 欠驱动无人船非奇异固定时间鲁棒包容控制. 控制与决策. 2023(05): 1286-1294 . 百度学术4. 王宁，高颖，王仁慧. 状态测量不确定和动力学未知的无人艇固定时间容错控制. 自动化学报. 2023(05): 1050-1061 . 本站查看5. 黄兵，肖云飞，冯元，于秀萍. 无人艇全分布式动态事件触发编队控制. 控制理论与应用. 2023(08): 1479-1487 . 百度学术6. 马俊达，谭冲，范佳佳. 基于齐次系统理论的多船有限时间饱和包容控制. 哈尔滨工程大学学报. 2023(10): 1779-1787 . 百度学术7. 高阳，吴文海，王子健. 具有输入约束和输出噪声的不确定系统级联线性自抗扰控制. 自动化学报. 2022(03): 843-852 . 本站查看8. 柳晨光，贺治卜，初秀民，吴文祥，李松龙，谢朔. 船舶编队控制综述. 交通运输工程学报. 2022(04): 10-27 . 百度学术9. 李莉莉，董凯，郭戈. 纵向速度和艏向角受限的水面艇有限时间协同路径跟踪. 控制与决策. 2021(02): 363-370 . 百度学术10. 张慧凤，董乐伟，魏新江. 一类带有输入饱和的随机系统抗干扰控制. 自动化学报. 2021(06): 1453-1459 . 本站查看11. 檀盼龙，秦华阳，孙明玮，刘俊杰，孙青林，陈增强. 欠驱动RTAC的滑模自抗扰镇定控制. 控制理论与应用. 2021(12): 2085-2093 . 百度学术12. 武宪青，徐可心，张益波. 基于输出反馈的欠驱动TORA系统的有界输入控制. 自动化学报. 2020(01): 200-204 . 本站查看其他类型引用(26)
-
图(5) / 表(2)
计量
- 文章访问数: 2739
- HTML全文浏览量: 398
- PDF下载量: 842
- 被引次数: 38
