光学精密工程2025，Vol.33Issue(4)：610-623,14.DOI:10.37188/OPE.20253304.0610
面向遥感图像道路提取的多尺度上下文感知网络
Multi-scale context-aware network for road extraction in remote sensing images
摘要
Abstract
To address the issues of local feature loss and low extraction accuracy faced by deep neural net-works in remote sensing image road extraction,a multi-scale context-aware network was proposed based on the SwinUnet network for remote sensing image road extraction.Firstly,a branch with a context aggregation module was designed in the encoder to enhance the extraction of contextual information and alleviate the prob-lem of semantic ambiguity caused by occlusion.Secondly,to solve the problem of semantic information mis-match between the encoder and decoder and to improve the model's ability to extract spatial information,a spa-tial feature extraction module was introduced in the skip connections,replacing the direct copying of encoder features in SwinUnet.Finally,a feature compression module was designed in the down-sampling stage to re-duce information loss in the encoder and enhance the network's segmentation capability.The test results on the Massachusetts road dataset show that this method achieved F1,IoU,Pr,and Re scores of 80.91%,69.40%,78.03%,and 65.20%,respectively.In comparison with mainstream methods such as UNet and SwinUnet,the IoU improved by 4.45%and 2.72%,respectively,demonstrating that the proposed algorithm effectively improves the accuracy and performance of remote sensing image road extraction through global mod-eling,context enhancement,and information matching optimization.关键词
遥感图像/道路提取/语义分割/SwinUnetKey words
remote sensing/road extraction/semantic segmentation/SwinUnet分类
计算机与自动化引用本文复制引用
李智杰,惠爱婷,李昌华,董玮,张颉,介军..面向遥感图像道路提取的多尺度上下文感知网络[J].光学精密工程,2025,33(4):610-623,14.基金项目
国家自然科学基金(No.62276207) （No.62276207）
陕西省住房城乡建设科技计划项目(No.2020-K09) （No.2020-K09）
陕西省教育厅协同创新中心项目(No.23JY038) （No.23JY038）
