摘要
鉴于暗原色先验算法能复原不同雾浓度和场景深度的图像,而基于非局部算子概念的NL-CTV(Non-Local Color Total Variation)模型能较好地保持图像边缘和纹理等特征,融合暗原色先验与NL-CTV模型,提出了一种新型单幅彩色图像去雾模型。通过暗原色先验得到精确的大气光强度和大气传输函数,然后推导包含大气光强度和大气传输函数的非局部能量泛函,再通过引入辅助变量和Bregman迭代参数,为其设计相应的快速split Bregman算法来求解该模型。将该算法与He算法、暗原色先验和Retinex算法的实验结果进行分析比较,从而验证了该模型不论从视觉上,还是客观数据上都要优于其他两种算法。
Since the dark channel prior algorithm can recover images with different degrees of fog and scene depth,and NL-CTV(Non-Local Color Total Variation)model based on the non-local operators can maintain the image features better,a single color image dehazing algorithm on the basis of dark channel prior and NL-CTV model is proposed.Firstly,obtaining the atmospheric light and the transmission from the dark channel prior accurately.Then,deriving the non-local energy function which contains the atmospheric light and transmission.Finally,using auxiliary variables and Bregman iterative parameters to calculate the energy function,and designing its split algorithm.Comparing the proposed model with He algorithm,dark channel prior and Retinex algorithm,experimental results show that the proposed algorithm is superior to the traditional methods visually and objectively.
引文
[1]Hautière N,Tarel J P,Halmaoui H,et al.Enhanced fog detection and free-space segmentation for car navigation[J].Machine Vision and Applications,2014,25(3):667-679.
[2]Tang Zixing.Image dehazing based on haziness analysis[J].International Journal of Automation and Computing,2014,11(1):78-86.
[3]He K,Sun J,Tang X.Single image haze removal using dark channel prior[C]//IEEE Conference on Computer Vision and Pattern Recognition,2011:1956-1963.
[4]Levin A,Lischinski D,Weiss Y.A closed form solution to natural image matting[C]//IEEE Conference on Computer Vision and Pattern Recognition,2006:61-68.
[5]He K,Sun J,Tang X.Guided image filtering[C]//European Conference on Computer Vision.Berlin,Heidelberg:Springer,2010:1397-1409.
[6]Parthasarathy S,Sankaran P.A RETINEX based haze removal method[C]//IEEE International Conference on Industrial and Information Systems,2012:1-6.
[7]刘海波,杨杰,吴正平,等.基于暗通道先验和Retinex理论的快速单幅图像去雾方法[J].Acta Automatica Sinica,2015,41(7):1264-1273
[8]聂慧,彭娇.基于KL变换和TV模型的彩色图像修复算法[J].硅谷,2015(1):66-66.
[9]Zhou Li,Bi Duyan,He Linyuan.Single color image dehazing using variational partial differential equation[J].Optics and Precision Engineering,2015,23(5):1465-1472.
[10]Narasimhan S G,Nayar S K.Contrast restoration of weather degraded images[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2003,25(6):713-724.
[11]Blomgren P,Chan T F.Color TV:total variation methods for restoration of vector-valued images[J].IEEE Transactions on Image Processing,1998,7(3):304-309.
[12]Yu Y,Pan Z,Wei W,et al.Comparison of the edge preservation capabilities of different variational models for vectorial image denoising[J].Journal of Image&Graphics,2011,16(12):2223-2230.
[13]Schechner Y Y,Narasimhan S G,Nayar S K.Instant dehazing of images using polarization[C]//Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition,2001:325-332.
[14]Schechner Y Y,Narasimhan S G,Nayar S K.Polarizationbased vision through haze[J].Applied Optics,2003,42(3):511-525.
[15]Zuo Z,Lan X,Zhou G,et al.A time dependent model via non-local operator for image restoration[C]//The Seventh International Conference on Intelligent System and Knowledge Engineering,The 1st International Conference on Cognitive System and Information Processing,2012:217-224.
[16]冯文强.非局部算法在图像去噪中的应用[D].合肥:中国科学技术大学,2011.
[17]Duan J,Pan Z.Non-local TV models for restoration of color texture images[J].Journal of Image&Graphics,2013.
[18]Tan R T.Visibility in bad weather from a single image[C]//IEEE Conference on Computer Vision and Pattern Recognition,2008:1-8.
[19]杜宏博,王丽会.基于改进暗原色先验模型的快速图像去雾方法[J].计算机工程与应用,2016,52(1):178-184.
[20]Brown E S,Chan T F,Bresson X.Completely convex formulation of the Chan-Vese image segmentation model[J].International Journal of Computer Vision,2012,98(1):103-121.
[21]Long W,Liang X,Cai S,et al.A modified augmented Lagrangian with improved grey wolf optimization to constrained optimization problems[J].Neural Computing&Applications,2016:1-18.