GUO Hongguang， CHEN Yong
（School of Electronics and Information Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China）
Abstract： To solve the problem of color distortion after dehazing in the sky region by using the classical dark channel prior method to process the hazy images with large regions of sky, an improved dark channel image dehazing method based on Gaussian mixture model is proposed. Firstly, we use the Gaussian mixture model to model the hazy image, and then use the expectation maximization (EM) algorithm to optimize the parameters, so that the hazy image can be divided into the sky region and the non-sky region. Secondly, the sky region is divided into a light haze region, a medium haze region and a heavy haze region according to the different dark channel values to estimate the transmission respectively. Thirdly, the restored image is obtained by combining the atmospheric scattering model. Finally, adaptive local tone mapping for high dynamic range images is used to adjust the brightness of the restored image. The experimental results show that the proposed method can effectively eliminate the color distortion in the sky region, and the restored image is clearer and has better visual effect.
Key words： image processing; image dehazing; Gaussian mixture model; expectation maximization (EM) algorithm; dark channel theory
References
［1］Wu D, Zhu Q S. The latest research progress of image dehazing. Acta Automatica Sinica, 2015, 41(2)： 221-239.
［2］Guo F, Cai Z X, Xie B, et al. Review and prospect of image dehazing techniques. Journal of Computer Applications, 2010, 30(9)： 2417-2421.
［3］Yu P, Hao C C. Foggy image enhancement by combined fractional differential and multi-scale retinex. Laser & Optoelectronics Progress, 2018, 55(1)： 011012.
［4］Jiang B J, Zhong M X. Improved histogram equalization algorithm in the image enhancement. Laser & Infrared, 2014, 44(6)： 702-706.
［5］Fattal R. Single image dehazing. ACM Transactions on Graphics, 2008, 27(3)： 721-729.
［6］Tarel J P, Hautiere N. Fast visibility restoration from a single color or gray level image. In： Proceedings of IEEE International Conference on Computer Vision, Kyoto, Japan, 2009： 2201-2208.
［7］He K M, Sun J, Tang X O. Single image haze removal using dark channel prior. In： Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, San Francisco, CA, USA, 2010： 1956-1963.
［8］He K M, Sun J, Tang X O. Guided image filtering. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(6)： 1397-1409.
［9］Xing X M, Liu W. Haze removal for single traffic image. Journal of Image and Graphics, 2016, 21(11)：1440-1447.
［10］Yin F, Chen T T, Fu Z R, et al. Improved method of image dehazing using dark channel prior. Journal of Frontiers of Computer Science and Technology, 2017, 11(7)： 1131-1139.
［11］Meng G F, Wang Y, Duan J Y, et al. Effcient image dehazing with boundary constraint and contextual regularization. In： Proceedings of IEEE International Conference on Computer Vision, Sydney, NSW, Australia, 2013： 617-624.
［12］Zhang X G, Tang M L, Chen H, et al. A dehazing method in single image based on double-area filter and image fusion. Acta Automatica Sinica, 2014, 40(8)： 1733-1739.
［13］Chen L Q, Wang W X. Single image dehazing based on dark colors image reference. Application Research of Computers, 2017, 34(6)： 1871-1875.
［14］Guo H, Xu X T, Li B. Study on image dehazing methods based on dark channel prior. Acta Optica Sinica, 2018, 38(4)： 0410002.
［15］Ou Y J. Study on image segmentation based on Gaussian Mixture Model. Beijing： Beijing Jiaotong University, 2015： 11-33.
［16］Lu Jie, Yang X Z, Lang W H, et al. SAR image segmentation with region-based GMM. Journal of Image and Graphics, 2011, 16(11)： 2088-2094.
［17］Du Y Q. Improved algorithm for image haze removing based on dark channel prior and retinex. Lanzhou： Northwest Normal University, 2015： 14-22.
［18］Ahn H, Keum B, Kim D, et al. Adaptive local tone mapping based on retinex for high dynamic range images. In： Proceedings of International Conference on Consumer Electronics, 2013： 6486837.
结合高斯混合模型的改进暗通道图像去雾方法
郭红光， 陈永
（兰州交通大学 电子与信息工程学院， 甘肃 兰州 730070）
摘要：针对经典的暗通道先验方法在处理含有大面积天空的有雾图像时， 去雾图像的天空区域出现不同程度的颜色失真等问题， 提出了一种结合高斯混合模型的改进暗通道图像去雾方法。首先， 采用高斯混合模型对有雾图像进行建模， 然后用期望最大化（Expectation maximization, EM）算法优化模型参数， 从而将有雾图像分割成天空区域和非天空区域。 其次， 根据天空区域暗通道值的不同将其分为淡雾区、 中雾区和浓雾区， 分别估计透射率。 并结合大气散射模型得到复原图像。 最后， 采用高动态范围图像自适应局部色调映射方法提升复原图像的亮度。实验结果表明， 该方法有效地解决了经典暗通道先验方法去雾时产生的天空失真问题， 且复原后的图像更清晰、 视觉效果更好。
关键词：图像处理； 图像去雾； 高斯混合模型； 期望最大化算法； 暗通道理论
引用格式：GUO Hongguang， CHEN Yong. Improved dark channel image dehazing method based on Gaussian mixture model. Journal of Measurement Science and Instrumentation, 2021, 12（1）： 53-60. DOI： 10．3969／j．issn．1674-8042．2021．01．007
[full text view]
