吉林大学学报(工学版) ›› 2019, Vol. 49 ›› Issue (4): 1377-1384.doi: 10.13229/j.cnki.jdxbgxb20180208

• • 上一篇    

结合天空分割和超像素级暗通道的图像去雾算法

王柯俨1(),胡妍1,王怀2,李云松1   

  1. 1. 西安电子科技大学 综合业务网理论及关键技术国家重点实验室, 西安 710071
    2. 中国人民解放军31668部队, 西宁 810000
  • 收稿日期:2018-03-09 出版日期:2019-07-01 发布日期:2019-07-16
  • 作者简介:王柯俨(1980?),女,副教授,博士. 研究方向: 图像编码与处理. E?mail:kywang@mail.xidian.edu.cn
  • 基金资助:
    国家自然科学基金项目(61301291);“111”高等学校学科创新引智计划项目(B08038)

Image dehazing algorithm by sky segmentation and superpixel⁃level dark channel

Ke⁃yan WANG1(),Yan HU1,Huai WANG2,Yun⁃song LI1   

  1. 1. State Key Laboratory of Integrated Services Networks, Xidian University, Xi'an 710071, China
    2. 31668 Troops of the Chinese People's Liberation Army, Xining 810000, China
  • Received:2018-03-09 Online:2019-07-01 Published:2019-07-16

摘要:

为了提升去雾图像质量,提出了一种基于天空分割和超像素级暗通道的单幅图像去雾方法。首先提出一种简单有效的多阈值天空分割方法,可将图像划分为天空区域和非天空区域;其次,根据天空区域估计大气光值;然后分别估计两类区域的透射率,对天空区域利用该区域无雾和有雾时暗通道值间的线性关系直接估计其透射率,对非天空区域则通过计算超像素级暗通道值来估计透射率;最后根据大气散射模型恢复无雾图像。实验结果表明,与现有方法相比,本文方法能更准确地分割天空且阈值自适应性强,大气光和透射率估计也更准确高效。采用本文方法得到的去雾图像具有对比度高、颜色自然、细节清晰等优点。

关键词: 信息处理技术, 图像去雾, 透射率, 暗通道先验, 超像素

Abstract:

In order to improve the quality of restored images, a single image dehazing approach based on sky segmentation and superpixel-level dark channel model is proposed. First, a simple but effective multi-threshold sky segmentation method is presented, which divides the image into sky area and non-sky area. Second, the atmospheric light is estimated in sky area. Third, the transmission maps of the two types of areas are estimated respectively. For sky area, the transmission is directly estimated by the linear relationship between the dark channel values of clear images and the corresponding hazy images. For non-sky area, the transmission is estimated by superpixel-level dark channel values. Finally, the haze-free image can be restored by the atmospheric scattering model. Experimental results show that, compared with the existing methods, the proposed method can segment the sky more accurately with better adaption of the thresholds, and can estimate the atmospheric light and the transmission map with higher accuracy and efficiency. Moreover, the dehazed images obtained by the proposed method have many advantages such as high contrast, natural color and abundant details.

Key words: information processing technology, image dehazing, transmission, dark channel prior, super?pixel

中图分类号: 

  • TN919.8

图1

算法总体框图"

图2

天空分割流程"

图3

透射率估计的分步结果"

图4

天空分割结果对比"

图5

对天空区域采用不同透射率估计方法的去雾结果对比"

图6

去雾结果与大气光估计结果对比"

图7

去雾效果及细节对比"

表1

客观评价"

评价指标 文献[3]结果 文献[7]结果 文献[8]结果

本文

结果

新增可见边之比 e

1

2

3

-0.1817

0.1250

0.4887

-0.4355

0.1432

0.3020

0.4645

0.1129

0.3120

0.4842

0.1517

0.4866

可见边梯度均值 r

1

2

3

1.0671

0.8624

1.2806

0.9397

0.6440

1.0272

1.1471

0.9443

1.0975

1.8351

1.3181

1.5168

1 Fattal R . Single image dehazing[J]. ACM Transactions on Graphics, 2008, 27(3): 1⁃9.
2 Tarel J P , Hautiere N . Fast visibility restoration from a single color or gray level image[C]∥IEEE International Conference on Computer Vision, Xi’an, China, 2009: 2201⁃2208.
3 He Kai⁃ming , Sun Jian , Tang Xiao⁃ou . Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341⁃2353.
4 Zhu Qing⁃song , Jia⁃ming Mai , Shao Ling . A fast single image haze removal algorithm using color attenuation prior[J]. IEEE Transactions on Image Processing, 2015, 24(11): 3522⁃3533.
5 Cui Tong , Qu Liang⁃qiong , Tian Jian⁃dong , et al . Single image haze removal based on luminance weight prior[C]∥IEEE International Conference on Cyber Technology in Automation, Control, and Intelligent Systems, Chengdu, China,2016: 332⁃336.
6 毕笃彦, 眭萍, 何林远, 等 . 基于Color Lines先验的高阶马尔科夫随机场去雾[J]. 电子与信息学报, 2016, 38(9): 2405⁃2409.
Bi Du⁃yan , Sui Ping, He Lin⁃yuan , et al . Higher⁃order Markov random fields defogging based on Color Lines[J]. Journal of Electronics & Information Technology, 2016, 38(9): 2405⁃2409.
7 Bui T M , Tran H N , Kim W , et al . Segmenting dark channel prior in single image dehazing[J]. Electronics Letters, 2014, 50(7): 516⁃518.
8 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.
Wang Yun⁃fei , Liu Hua⁃wei , Zhao Bo⁃xin . Superpixels⁃based golden dark channel for single image fog removal[J]. Journal of Xidian University, 2017, 44(5): 147⁃152.
9 Zhu Qing⁃song , Wu Di , Xie Yao⁃qin , et al . Quick shift segmentation guided single image haze removal algorithm[C]∥IEEE International Conference on Robotics and Biomimetics, Bali, 2014:113⁃117.
10 郑良缘, 王平, 高颖慧 . 基于超像素分割的图像去雾算法[J]. 重庆理工大学学报, 2015, 29(6): 100⁃106.
Zheng Liang⁃yuan , Wang Ping , Gao Ying⁃hui . Single image dehazing based on superpixels segmentation[J]. Journal of Chongqing University of Technology, 2015, 29(6): 100⁃106.
11 王睿, 李蕊, 廉小亲 . 基于大气多散射模型和超像素分割的图像去雾[J]. 光子学报, 2016, 45(4): 124⁃130.
Wang Rui , Li Rui , Lian Xiao⁃qin . Multiple scattering model based image dehazing with superpixel[J]. Acta Photonica Sinica, 2016, 45(4): 124⁃130.
12 蒋建国, 侯天峰, 齐美彬 . 改进的基于暗原色先验的图像去雾算法[J]. 电路与系统学报, 2011, 16(2): 7⁃12.
Jiang Jian⁃guo , Hou Tian⁃feng , Qi Mei⁃bin . Improved algorithm on image haze removal using dark channel prior[J]. Journal of Circuits and Systems, 2011, 16(2): 7⁃12.
13 Zhu Qing⁃song , Jia⁃ming Mai , Song Zhan , et al . Mean shift⁃based single image dehazing with re⁃refined transmission map[C]∥IEEE International Conference on Systems, Man and Cybernetics, San Diego, USA, 2014: 4058⁃4064.
14 Liu Hai⁃tao , Guo Jin , Sun Chang⁃yin . An improved dehazing method based on the transmission compensation[C]∥World Congress on Intelligent Control and Automation, Guilin, China, 2016: 2732⁃2737.
15 Han Xiao⁃xu , Feng Hong⁃wei , Bu Qi⁃rong , et al . Image dehazing based on two⁃peak channel prior[C]∥IEEE International Conference on Image Processing, Phoenix, 2016: 2236⁃2240.
16 Yuan Hui , Liu Chang⁃chun , Guo Zhi⁃xin , et al . A region⁃wised medium transmission based image dehazing method[J]. IEEE Access, 2017, 5: 1735⁃1742.
17 He Kai⁃ming , Sun Jian , Tang Xiao⁃ou . Guided image filtering[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(6): 1397⁃1409.
18 Wang Ke⁃yan , Zhang Shan⁃shan , Li Yun⁃song . Haze removal via edge weighted pixel⁃to⁃patch fusion[J]. Mobile Networks and Applications, 2017, 22(3): 464⁃477.
19 Achanta R , Shaji A , Smith K . SLIC superpixels compared to state⁃of⁃the⁃art superpixel methods[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2012, 34(11): 2274⁃2281.
20 Hautière N , Tarel J P , Aubert D , et al . Blind contrast enhancement assessment by gradient ratioing at visible edges[J]. Image Analysis and Stereology, 2008, 27(2): 87⁃95.
[1] 董超,刘晶红,徐芳,王仁浩. 光学遥感图像舰船目标快速检测方法[J]. 吉林大学学报(工学版), 2019, 49(4): 1369-1376.
[2] 托乎提努尔,张海龙,王杰,王娜,冶鑫晨,王万琼. 基于图形处理器的高速中值滤波算法[J]. 吉林大学学报(工学版), 2019, 49(3): 979-985.
[3] 付银娟,李勇,徐丽琴,张昆辉. NLFM⁃Costas射频隐身雷达信号设计及分析[J]. 吉林大学学报(工学版), 2019, 49(3): 994-999.
[4] 苏寒松,代志涛,刘高华,张倩芳. 结合吸收Markov链和流行排序的显著性区域检测[J]. 吉林大学学报(工学版), 2018, 48(6): 1887-1894.
[5] 徐岩,孙美双. 基于卷积神经网络的水下图像增强方法[J]. 吉林大学学报(工学版), 2018, 48(6): 1895-1903.
[6] 黄勇,杨德运,乔赛,慕振国. 高分辨合成孔径雷达图像的耦合传统恒虚警目标检测[J]. 吉林大学学报(工学版), 2018, 48(6): 1904-1909.
[7] 李居朋,张祖成,李墨羽,缪德芳. 基于Kalman滤波的电容屏触控轨迹平滑算法[J]. 吉林大学学报(工学版), 2018, 48(6): 1910-1916.
[8] 应欢,刘松华,唐博文,韩丽芳,周亮. 基于自适应释放策略的低开销确定性重放方法[J]. 吉林大学学报(工学版), 2018, 48(6): 1917-1924.
[9] 陆智俊,钟超,吴敬玉. 星载合成孔径雷达图像小特征的准确分割方法[J]. 吉林大学学报(工学版), 2018, 48(6): 1925-1930.
[10] 刘仲民,王阳,李战明,胡文瑾. 基于简单线性迭代聚类和快速最近邻区域合并的图像分割算法[J]. 吉林大学学报(工学版), 2018, 48(6): 1931-1937.
[11] 单泽彪,刘小松,史红伟,王春阳,石要武. 动态压缩感知波达方向跟踪算法[J]. 吉林大学学报(工学版), 2018, 48(6): 1938-1944.
[12] 姚海洋, 王海燕, 张之琛, 申晓红. 双Duffing振子逆向联合信号检测模型[J]. 吉林大学学报(工学版), 2018, 48(4): 1282-1290.
[13] 全薇, 郝晓明, 孙雅东, 柏葆华, 王禹亭. 基于实际眼结构的个性化投影式头盔物镜研制[J]. 吉林大学学报(工学版), 2018, 48(4): 1291-1297.
[14] 陈绵书, 苏越, 桑爱军, 李培鹏. 基于空间矢量模型的图像分类方法[J]. 吉林大学学报(工学版), 2018, 48(3): 943-951.
[15] 陈涛, 崔岳寒, 郭立民. 适用于单快拍的多重信号分类改进算法[J]. 吉林大学学报(工学版), 2018, 48(3): 952-956.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!