基于改进大气散射模型的图像去雾算法研究  

Research on Image Dehazing Algorithm Based on Improved Atmospheric Scattering Model

作　　者：龚晓婷 段锦[1] 黄冠婷 韩学辉[1] 王尧 GONG Xiaoting;DUAN Jin;HUANG Guanting;HAN Xuehui;WANG Yao(School of Electronics and Information Engineering,Changchun University of Science and Technology,Changchun 130022)

机构地区：[1]长春理工大学电子信息工程学院

出　　处：《长春理工大学学报（自然科学版）》2019年第3期85-90,共6页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：吉林省科技厅重点科技攻关项目(20160204066G X)

摘　　要：在成像的过程中,光的传播受到悬浮颗粒散射的影响,图像质量不佳。为了解决这种问题,提出了一种基于改进的大气散射模型的图像去雾算法。首先,运用开运算获取精确的大气光,避免白色物体误判大气光;其次,提出一种可调介质传输函数的估算方法,通过白平衡思想估计传输率的最大范围,并用可调系数来调整传输率;最后,由于暗通道先验方法得到的去雾图像整体偏暗,进行像素级融合得到最终复原图像。通过大量试验分析发现,该算法可以有效的去雾,同时提高有雾图像的对比度,细节信息等,使去雾后的图像纹理更丰富。In the process of imaging,the propagation of light is affected by the scattering of suspended particles in the air,the received images are bad. In order to solve this problem,an algorithm based on improved atmospheric scattering model is proposed. Firstly,because the estimation of atmospheric light is affected by the white object in the optical image,the open calculation is used to obtain accurate atmospheric light. Secondly,an estimation method of the transfer function of the tunable medium is proposed. The maximum range of the transmission rate is estimated by the white balance idea and the transmission rate is adjusted by the adjustable coefficient. Finally,the defogged image obtained by the dark channel prior is darker overall,the final restored image is obtained by pixel level fusion. Through a lot of experimental analysis,the algorithm can effectively defog,at the same time,the contrast and the detail information of the foggy image are improved,and the texture details of the image after defogging are more abundant.

关 键 词：图像去雾 可调介质传输率 暗通道先验 图像融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]