非对称结构的高光谱与激光雷达图像分类模型

李明婉,房胜,李哲

Asymmetric structure based hyperspectral and LiDAR image classification model

Mingwan LI,Sheng FANG,Zhe LI

表 2 3个数据集上不同方法的定量对比分析

Tab.2 Quantitative comparison of different methods on three datasets

模型结构 模型名称 Houston2013 MUUFL Augsburg OA/% AA/% Kappa/% OA/% AA/% Kappa/% OA/% AA/% Kappa/% CNN ENDNet 88.05 87.86 87.07 80.75 80.33 75.24 65.83 54.14 55.14 HybridSN 86.22 87.40 85.17 62.46 58.36 54.25 58.73 53.85 46.63 S2ENet 94.59 95.40 94.16 79.23 79.72 73.57 74.75 66.20 66.14 ViT SpectralFormer 69.33 70.66 66.89 76.34 76.12 69.94 39.76 53.09 28.94 MFT 92.31 93.42 91.70 73.04 73.46 66.53 71.54 65.81 62.61 CNN-ViT S2EFT 86.94 86.30 85.82 79.19 75.13 73.07 62.57 57.19 49.63 HCTNet 94.72 95.68 94.30 74.52 73.42 67.93 73.94 66.64 65.29 MHST 94.22 95.18 93.75 76.85 77.25 70.71 66.54 66.55 56.95 ViT-CNN 本研究方法 97.60 97.97 97.41 83.95 84.83 79.47 75.47 67.97 67.18