基于遥感图像道路提取的全局指导多特征融合网络

宦海,盛宇,顾晨曦

Global guidance multi-feature fusion network based on remote sensing image road extraction

Hai HUAN,Yu SHENG,Chenxi GU

表 8 基于DeepGlobe数据集不同网络的分割性能对比

Tab.8 Segmentation performance comparison of different networks based on DeepGlobe dataset

% 网络 IoU MIoU 背景 道路 DeepLabV3 98.06 62.24 80.15 APCNet 98.03 59.78 78.91 CCNet 98.09 61.77 79.93 DANet 97.98 61.77 79.88 EMANet 98.06 61.45 79.76 DNLNet 98.10 62.19 80.15 CRANet 98.05 62.04 80.04 SANet 98.15 63.05 80.60 GGMNet 98.14 63.11 80.63