基于二进制掩码的单幅图像天空区域识别去雾算法  

An Image Defogging Algorithm Based on Binary Mask for Sky Area Recognition

作　　者：史佳霖 刘鹏[1] 张天翼 毕誉轩 王彩霞[1] SHI Jialin;LIU Peng;ZHANG Tianyi;BI Yuxuan;WANG Caixia(School of Electronic 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,长春130022

出　　处：《长春理工大学学报（自然科学版）》2024年第1期105-111,共7页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：吉林省科技发展计划资助项目(20210201021GX)。

摘　　要：针对暗通道先验去雾算法处理得到的去雾图片存在颜色失真、偏移等问题,提出了一种基于天空区域识别的改进算法。首先,利用区域窗口自适应最小化滤波器获得暗通道图像,并使用二进制掩模图进行天空区域识别。其次,对天空区域进行透射率和大气光值的设定,同时引入引导滤波器对透射率图进行优化,避免了图像恢复时天空部分色彩失真的问题。最后,利用大气散射模型对原始图像进行复原。实验结果表明,本算法能够有效对图像中天空等明亮区域进行去雾,使处理后的图像具有良好的视觉效果。This paper presents an improved algorithm that addresses the problems of color distortion and offset in the dark channel prior defogging algorithm through sky region recognition.The proposed approach involves several key steps.Firstly,the dark channel image is obtained using the region window adaptive minimization filter.Subsequently,the binary mask map is utilized to identify the sky region accurately.To overcome color distortion in the sky,a guiding filter is introduced to optimize the transmittance map.This step ensures that the algorithm preserves the true color of the sky during the defogging process.The experimental results confirm the algorithm's effectiveness in effectively defogging bright areas such as the sky,leading to visually pleasing and natural-looking results.The proposed algorithm offers a promising solution to address color distortion and offset,enhancing the overall quality of defogged images.

关 键 词：去雾算法 自适应 天空识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]