融合多尺度分辨率和带状特征的遥感道路提取

李国燕,李鹏辉,刘榕,梅玉鹏,张明辉

Remote sensing road extraction by fusing multi-scale resolution and strip feature

Guoyan LI,Penghui LI,Rong LIU,Yupeng MEI,Minghui ZHANG

表 4 在DeepGlobe数据集上所提模型与其他几种先进的道路提取方法对比

Tab.4 Comparison between proposed model and several other state-of-the-art road extraction methods on DeepGlobe dataset

方法 P/% R/% F1/% IoU/% APLS/% Unet++(2018） 76.42 74.44 75.41 60.53 64.50 DeepLabV3+(2018） 79.60 85.92 81.43 70.21 67.96 SGCNet(2022) 72.78 66.98 69.76 53.57 57.39 TransRoadNet(2022) 81.01 84.13 82.53 70.20 67.00 CARNet(2023) 81.08 86.25 83.58 71.76 68.69 RoadFormer(2024） 83.86 85.36 84.61 73.38 69.62 StripUnet(2024) 82.82 85.32 84.05 72.04 68.46 MSRSF-Net(本文模型) 83.92 85.85 84.87 73.76 69.78