基于暗通道先验去雾算法的优化
|
摘要
在雾霾环境下获取的图像往往是不清晰的,研究人员提出了很多去雾算法以获得清晰图像。在暗通道先验去雾算法的基础上,针对现有算法在处理户外有雾图像后存在色彩偏移以及在景深突变区域出现光晕现象提出改进。首先分析之前一些经典算法易错选大气光值的原因,为此采取基于四叉树递归思想,最终选取一定大小的局部区域像素点平均值作为大气光值。接着分析了在景深突变区域产生光晕的原因提出了在估算透射率时,采用分区域最小值滤波,直接获得精细的透射率值,避免额外细化透射率的滤波算法,提升算法效率,最后结合大气耗散模型,利用改进优化后算法获得的大气光值和透射率图恢复去雾后图像。实验结果表明,改进算法不但能较好较快的去雾后效果图,并且在色彩还原和细节还原取得了很好效果,最后与其他几种经典的图像去雾算法分别进行客观与主观对比,证实了算法的可靠性。
Images acquired in a hazy environment are often unclear, and researchers have proposed many defogging algorithms to obtain clear images. On the basis of the dark channel prior defogging algorithm, the existing algorithm has improved the color shift after processing outdoor foggy images and the halo phenomenon in the sudden change of depth of field. To this end, firstly analyze the reasons why some classical algorithms are easy to choose the atmospheric light value. In this paper, based on the recursive idea of quadtree, the average value of the local area pixel is selected as the atmospheric light value. Then the reason for the generation of Halo in the sudden change of depth of field is analyzed. When estimating the transmittance, the sub-region minimum filter is used to directly obtain the fine transmittance value, which avoids the additional filtering algorithm for refining the transmittance and improves the efficiency of the algorithm. Finally, combined with the atmospheric dissipation model, the image after de-fogging is restored by using the optimized atmospheric light value and transmittance map. The experimental results show that the improved algorithm not only has a better post-fog effect map, but also achieves good results in color reproduction and detail reduction. It is objective and subjective comparison with several typical image dehazing algorithms, which proves the algorithm is feasibility.
Images acquired in a hazy environment are often unclear, and researchers have proposed many defogging algorithms to obtain clear images. On the basis of the dark channel prior defogging algorithm, the existing algorithm has improved the color shift after processing outdoor foggy images and the halo phenomenon in the sudden change of depth of field. To this end, firstly analyze the reasons why some classical algorithms are easy to choose the atmospheric light value. In this paper, based on the recursive idea of quadtree, the average value of the local area pixel is selected as the atmospheric light value. Then the reason for the generation of Halo in the sudden change of depth of field is analyzed. When estimating the transmittance, the sub-region minimum filter is used to directly obtain the fine transmittance value, which avoids the additional filtering algorithm for refining the transmittance and improves the efficiency of the algorithm. Finally, combined with the atmospheric dissipation model, the image after de-fogging is restored by using the optimized atmospheric light value and transmittance map. The experimental results show that the improved algorithm not only has a better post-fog effect map, but also achieves good results in color reproduction and detail reduction. It is objective and subjective comparison with several typical image dehazing algorithms, which proves the algorithm is feasibility.
引文
[1] 汪荣贵,张璇,张新龙,等.一种新型自适应Retinex图像增强方法研究[J].电子学报,2010,38(12):2933-2936.
[2] STARK J A.Adaptive image contrast enhancement using generalizations of histogram equalization [J].IEEE Transactions on Image Processing,2000,9(5):889-896.
[3] KIM J Y,KIM L S,HWANG S H.An advanced contrast enhancement using partially overlapped sub-block histogram equalization[J].IEEE Transactions on Circuits and Systems for Video Technology,2001,11(4):475-484.
[4] 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.
[5] 张驰宇,贾银亮,梁康武.基于暗通道的单幅图像快速去雾算法[J].电子测量技术,2017,40(10):143-147.
[6] TAN R T.Visibility in bad weather from a single image[C].Proceddings of IEEE Conference on Computer Vision and Pattern Recognition,Alaska,America:IEEE,2008:2201-2208.
[7] FATTAL R.Single image dehazing[J].ACM Transactions on Graphics,2008,27(3):1-9.
[8] TAREL J P,HAUTIERE N.Fast visibility restoration from a single color or gray level image[C].Proceedings of IEEE the 12th International Conference on Computer Vision.Tokyo,Japan:IEEE,2009:2201-2208.
[9] FATTAL R.Dehazing using color-lines[J].ACM Transactions on Graphics,2014,34(13):1-14.
[10] 张小刚,唐美玲,陈华,等.一种结合双区域滤波和图像融合的单幅图像去雾算法[J].自动化学报,2014,40(8):1733-1739.
[11] Xu H R,J M,LIU Q,et al.Fast image dehazingusing improved dark channel prior[C].Proceedings of the IEEE International Conference on Information Science and Technology.Hubei,China:IEEE,2012:663-667.
[12] GIBSON K B,VO D T,NGUYEN T Q.Aninvestigation of dehazing effects on image and video coding[J].IEEE Transactions on Image Processing,2012,21(2):662-673.
[13] 崔倩男,田小平,吴成茂.基于引导滤波改进的暗原色去雾算法[J].计算机科学,2018,45(5):285-290.
[14] 沈逸云,刘春晓,张金栋,等.鲁棒图像去雾的大气光校正与透射率优化算法[J].计算机辅助设计与图形学学报,2017,29(9):1604-1612.
[15] 陈丹丹,陈莉,张永新,等.修正大气耗散函数的单幅图像去雾[J].中国图象图形学报,2017,22(6):787-796.
[16] CAI B,XU X,JIA K,et al.An end-to-end system for single image haze removal[J].IEEE Transactions on Image Processing,2016,25(11):5187-5198.
[17] REN W Q,LIU S,ZHANG H,et al.Single image dehazing vin multi-scale convolutionalneural networks[M].Computer Vision-ECCV 2016,Springer Internation Publishing,2016:154-169.
[18] 白荣鑫,杨峰.一种改进的暗通道先验去雾算法[J].山东师范大学学报(自然科学版),2017,32(4):32-36.
[19] MENG G F,WANG Y,DUAN J Y,et al.Efficient image dehazing with boundary constraint and contextual regularization[C].IEEE International Conference on Computer Vision.New York:IEEE Press,2014:617-624.
[20] He K M,SUN J,TANG X O.Guided image filtering[J].IEEE Transactions on Pattern Analysisand Machine Intelligence,2013,35(6):1397-1409.
[21] GUO F,CAI Z X.Objective assessment method for the clearness effect of Image defogging [J].Journal of Automatica sinica,2012,38(9):1410-1419.
[2] STARK J A.Adaptive image contrast enhancement using generalizations of histogram equalization [J].IEEE Transactions on Image Processing,2000,9(5):889-896.
[3] KIM J Y,KIM L S,HWANG S H.An advanced contrast enhancement using partially overlapped sub-block histogram equalization[J].IEEE Transactions on Circuits and Systems for Video Technology,2001,11(4):475-484.
[4] 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.
[5] 张驰宇,贾银亮,梁康武.基于暗通道的单幅图像快速去雾算法[J].电子测量技术,2017,40(10):143-147.
[6] TAN R T.Visibility in bad weather from a single image[C].Proceddings of IEEE Conference on Computer Vision and Pattern Recognition,Alaska,America:IEEE,2008:2201-2208.
[7] FATTAL R.Single image dehazing[J].ACM Transactions on Graphics,2008,27(3):1-9.
[8] TAREL J P,HAUTIERE N.Fast visibility restoration from a single color or gray level image[C].Proceedings of IEEE the 12th International Conference on Computer Vision.Tokyo,Japan:IEEE,2009:2201-2208.
[9] FATTAL R.Dehazing using color-lines[J].ACM Transactions on Graphics,2014,34(13):1-14.
[10] 张小刚,唐美玲,陈华,等.一种结合双区域滤波和图像融合的单幅图像去雾算法[J].自动化学报,2014,40(8):1733-1739.
[11] Xu H R,J M,LIU Q,et al.Fast image dehazingusing improved dark channel prior[C].Proceedings of the IEEE International Conference on Information Science and Technology.Hubei,China:IEEE,2012:663-667.
[12] GIBSON K B,VO D T,NGUYEN T Q.Aninvestigation of dehazing effects on image and video coding[J].IEEE Transactions on Image Processing,2012,21(2):662-673.
[13] 崔倩男,田小平,吴成茂.基于引导滤波改进的暗原色去雾算法[J].计算机科学,2018,45(5):285-290.
[14] 沈逸云,刘春晓,张金栋,等.鲁棒图像去雾的大气光校正与透射率优化算法[J].计算机辅助设计与图形学学报,2017,29(9):1604-1612.
[15] 陈丹丹,陈莉,张永新,等.修正大气耗散函数的单幅图像去雾[J].中国图象图形学报,2017,22(6):787-796.
[16] CAI B,XU X,JIA K,et al.An end-to-end system for single image haze removal[J].IEEE Transactions on Image Processing,2016,25(11):5187-5198.
[17] REN W Q,LIU S,ZHANG H,et al.Single image dehazing vin multi-scale convolutionalneural networks[M].Computer Vision-ECCV 2016,Springer Internation Publishing,2016:154-169.
[18] 白荣鑫,杨峰.一种改进的暗通道先验去雾算法[J].山东师范大学学报(自然科学版),2017,32(4):32-36.
[19] MENG G F,WANG Y,DUAN J Y,et al.Efficient image dehazing with boundary constraint and contextual regularization[C].IEEE International Conference on Computer Vision.New York:IEEE Press,2014:617-624.
[20] He K M,SUN J,TANG X O.Guided image filtering[J].IEEE Transactions on Pattern Analysisand Machine Intelligence,2013,35(6):1397-1409.
[21] GUO F,CAI Z X.Objective assessment method for the clearness effect of Image defogging [J].Journal of Automatica sinica,2012,38(9):1410-1419.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |