测试技术学报2025，Vol.39Issue(4)：415-423,9.DOI:10.62756/csjs.1671-7449.2025047
基于超像素分割的暗通道先验图像去雾算法
Dark Channel Prior Image Dehazing Algorithm Based on Superpixel Segmentation
摘要
Abstract
To solve the problem that the dark channel prior algorithm is susceptible to white objects or bright areas,which leads to inaccurate estimation of atmospheric light and transmission,a dark channel prior image fog removal algorithm based on superpixel segmentation is proposed.Firstly,the simple linear iterative clustering superpixel algorithm is used to improve the dark channel prior.Secondly,the foggy image is segmented by superpixel segmentation by using the improved dark channel before obtaining superpixel blocks,and then the local atmospheric light value is calculated for each superpixel block and the average value is taken.Then,the gamma correction is performed on the coarse transmission map,and the average gradient value is used as a weight to fuse the weights of the coarse transmission map and the corrected transmission map to obtain the final transmission map.Finally,the inverse process of the atmospheric scattering model is used to obtain the dehazing image.The experimental results show that superpixel segmentation solves the problem of the dark channel prior algorithm to estimate the atmospheric light's dependence on the brightest pixel.The proposed algorithm can improve the clarity and retain the texture details of the image well and is superior to other comparison algorithms.关键词
暗通道先验/超像素分割/简单线性迭代聚类算法/图像去雾/伽马校正Key words
dark channel prior/superpixel segmentation/simple linear iterative clustering algorithm/image dehazing/gamma correction分类
电子信息工程引用本文复制引用
李波,胡红萍,杨正民..基于超像素分割的暗通道先验图像去雾算法[J].测试技术学报,2025,39(4):415-423,9.基金项目
山西省基础研究计划资助项目(20210302123019,202103021224195,202103021224212,202103021223189) （20210302123019,202103021224195,202103021224212,202103021223189）
山西省回国留学人员科研项目(2021-108) （2021-108）
