基于Boosting的高光谱遥感切空间协同表示集成学习方法

虞瑶,苏红军,姚文静

Boosting Ensemble Learning for Hyperspectral Image Classification Using Tangent Collaborative Representation

Yao Yu,Hongjun Su,Wenjing Yao

表3 实验二分类精度统计

Table 3 Classification Accuracy Statistics of Experiment 2

类别 训练样本 测试样本 Boosting RF ELM TCRC AdaBoost ELM Boost TCRC C1 72 1 356 55.41 57.69 62.81 77.25 64.70 84.85 C2 42 788 62.28 45.65 54.27 82.03 53.77 75.27 C3 25 458 73.51 73.44 87.20 97.59 90.84 95.83 C4 37 693 89.22 91.23 94.67 94.47 96.94 97.29 C5 24 454 94.05 97.05 99.04 99.52 99.80 100 C6 49 923 58.46 51.53 56.81 80.01 55.69 73.18 C7 123 2332 63.06 79.60 62.90 66.03 63.54 75.16 C8 30 563 52.91 35.99 62.93 81.43 71.67 84.19 C9 64 1 201 94.81 95.27 98.45 98.49 98.93 99.14 总体分类精度/% 69.48 71.05 71.51 80.14 72.09 84.11 平均分类精度/% 71.53 69.72 75.45 86.31 77.32 87.21 Kappa系数 0.637 1 0.655 3 0.662 1 0.762 8 0.668 4 0.812 0 时间/s 12.21 0.35 0.13 164.30 2.30 3 651.34