航空科学技术2025，Vol.36Issue(6)：98-104,7.DOI:10.19452/j.issn1007-5453.2025.06.011
面向航拍图像去雾的高效模型研究
Research on an Efficient Model for Dehazing in Aerial Images
摘要
Abstract
Image dehazing technology is dedicated to recovering clear haze-free images from those affected by haze.Due to the influence of atmospheric conditions,aerial images are often subjected to haze interference,which degrades image quality.Existing deep learning approaches often fall short in addressing the vast scenes and intricate details found in aerial images.To tackle this challenge,an efficient dehazing model suitable for aerial images has been developed.This model comprises two key components:a detail enhancement module and a haze adaptation module.The detail enhancement module employs a multi-branch structure to enhance the network's ability to capture detail features and simplifies the inference process through the reparameterization technique.The haze adaptation module,on the other hand,utilizes local and global attention mechanisms to flexibly handle unevenly distributed haze in the image,achieving better dehazing effects.Experimental validation has demonstrated that the proposed model excels in dehazing performance and possesses high operational efficiency.The research outcomes not only provide a fast and efficient solution for the field of aerial image dehazing but also hold significant implications for promoting the application of subsequent related algorithms on aerial images.关键词
图像去雾/航拍图像/深度学习/图像恢复/计算机视觉Key words
image dehazing/aerial images/deep learning/image restoration/computer vision分类
计算机与自动化引用本文复制引用
杨易乾,孙维超..面向航拍图像去雾的高效模型研究[J].航空科学技术,2025,36(6):98-104,7.基金项目
航空科学基金(2023Z071077008) Aeronautical Science Foundation of China(2023Z071077008) （2023Z071077008）
