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 5 在Massachusetts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.5 Comparison between proposed model and several other state-of-the-art road extraction methods on Massachusetts dataset

方法 P/% R/% F1/% IoU/% APLS/% Unet++(2018） 78.79 60.69 68.61 52.08 59.25 DeepLabV3+(2018） 75.88 76.34 75.36 61.29 55.11 SGCNet(2022) 74.05 63.76 68.52 52.12 49.53 TransRoadNet(2022) 78.75 81.24 79.98 66.67 56.67 CARNet(2023) 79.02 81.73 80.35 67.21 56.80 RoadFormer(2024） 79.95 81.79 80.87 68.46 57.16 StripUnet(2024) 80.81 81.52 81.16 68.18 59.81 MSRSF-Net(本文模型) 80.80 82.14 81.46 68.57 60.27