全局信息提取与重建的遥感图像语义分割网络
|
|
梁龙学,贺成龙,吴小所,闫浩文
|
Remote sensing image semantic segmentation network based on global information extraction and reconstruction
|
|
Longxue LIANG,Chenglong HE,Xiaosuo WU,Haowen YAN
|
|
| 表 3 在Potsdam测试集上与先进的高精度网络的定量比较结果 |
| Tab.3 Quantitative comparison result with advanced high-precision network on Potsdam test set |
|
| 方法 | 骨干 | F1/% | F1mean/% | OA/% | mIoU/% | | 不可渗透 | 建筑 | 低矮植被 | 树 | 车 | | DANet(2019) | Resnet18 | 92.3 | 96.0 | 86.6 | 88.4 | 90.3 | 90.7 | 90.2 | 83.1 | | MAResUNet(2021) | Resnet18 | 93.3 | 96.8 | 87.9 | 89.0 | 96.6 | 92.7 | 91.3 | 86.7 | | ABCNet(2021) | Resnet18 | 93.0 | 96.5 | 87.5 | 88.2 | 96.2 | 92.2 | 90.8 | 85.8 | | BANet(2021) | ResT-Lite | 92.5 | 96.1 | 87.1 | 88.8 | 96.0 | 92.1 | 90.6 | 85.6 | | MANet(2022) | Resnet18 | 92.9 | 96.1 | 87.5 | 88.8 | 96.6 | 92.4 | 90.9 | 86.1 | | UnetFormer(2022) | Resnet18 | 93.1 | 96.5 | 87.8 | 89.2 | 96.7 | 92.7 | 91.1 | 86.6 | | DCswin(2022) | Swin-tiny | 93.3 | 96.7 | 88.1 | 89.7 | 96.6 | 92.9 | 91.3 | 86.9 | | MAGIFormer | MSCAN-tiny | 93.3 | 97.1 | 88.1 | 89.5 | 96.9 | 93.0 | 91.4 | 87.1 |
|
|
|