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2 实验一分类精度统计
|
Table2 Classification Accuracy Statistics of Experiment 1
|
|
类别 | 训练样本 | 测试样本 | Boosting | RF | ELM | TCRC | AdaBoost ELM | Boost TCRC |
---|
道路 | 15 | 1 272 | 82.60 | 82.20 | 89.67 | 88.91 | 91.13 | 92.44 | 草地 | 15 | 1 099 | 89.78 | 90.93 | 89.12 | 85.54 | 95.29 | 89.49 | 阴影 | 15 | 204 | 71.03 | 94.02 | 66.31 | 90.46 | 81.87 | 88.43 | 土壤 | 15 | 364 | 73.24 | 94.45 | 80.66 | 86.82 | 86.85 | 91.19 | 树木 | 15 | 1 336 | 98.68 | 93.49 | 99.15 | 98.18 | 99.33 | 98.89 | 建筑物 | 15 | 1 270 | 83.72 | 79.67 | 89.60 | 94.63 | 87.32 | 96.17 | 总体分类精度/% | 86.09 | 87.31 | 89.36 | 90.91 | 91.92 | 93.73 | 平均分类精度/% | 83.18 | 89.13 | 85.75 | 90.76 | 90.30 | 92.77 | Kappa系数 | 0.825 7 | 0.841 3 | 0.866 5 | 0.888 5 | 0.898 3 | 0.920 8 | 时间/s | 1.25 | 0.10 | 0.14 | 18.40 | 1.30 | 126.95 |
|
|
|