线性注意力与跨尺度融合的遥感语义分割网络

张景鸿,邵玉斌,唐菁敏,杜庆治,龙华,张赜涛

Remote sensing semantic segmentation network with linear attention and cross-scale fusion

Jinghong ZHANG,Yubin SHAO,Jingmin TANG,Qingzhi DU,Hua LONG,Zetao ZHANG

表 3 LoveDA测试集上不同方法的性能比较

Tab.3 Performance comparison of different methods on LoveDA test set %

方法 骨干网络 IoU mIoU 背景 建筑 道路 水体 裸地 森林 农田 FarSeg ResNet50 43.1 51.5 53.9 76.6 9.8 43.3 58.9 48.2 FactSeg ResNet50 42.6 53.6 52.8 76.9 16.2 42.9 57.5 48.9 BANet ResT-Lite 43.7 51.5 51.1 76.9 16.6 44.9 62.5 49.6 TransUNet ViT-R50 43.0 56.1 53.7 78.0 9.3 44.9 56.9 48.9 Segmenter ViT-Tiny 38.0 50.7 48.7 77.4 13.3 43.5 58.2 47.1 SwinUperNet Swin-Tiny 43.3 54.3 54.3 78.7 14.9 45.3 59.6 50.0 DC-Swin Swin-Tiny 41.3 54.5 56.2 78.1 14.5 47.2 62.4 50.6 UNetFormer ResNet18 44.7 58.8 54.9 79.6 20.1 46.0 62.5 52.4 MobileViT V3 MobileViTv3 43.0 60.1 56.9 81.3 17.6 48.1 56.1 51.9 SSFNet ResNet18 45.6 57.4 56.9 81.5 18.4 45.5 63.4 52.7 RS3Mamba R18-Mamba-T 39.7 58.8 57.9 61.0 37.2 39.7 34.0 50.9 LASC-Mamba ResNet18 46.1 58.4 55.5 79.9 19.2 47.6 63.2 52.9 ORLA-OCM UNetFormer(本文方法) ResNet18 45.6 60.0 55.9 81.1 20.8 47.1 63.1 53.4