基于蒲公英优化随机森林模型的沙漠土壤Fe2O3 含量高光谱遥感反演

胡昕, 买买提·沙吾提, 张峰, 崔锦涛, 艾尼玩·艾买尔, 阿斯娅·曼力克

Hyperspectral remote sensing estimation of Fe2O3 content in desert soil based on dandelion-optimized random forest model

Xin Hu, Sawut Mamat, Feng Zhang, Jintao Cui, Aimaier Ainiwan, Manlike Asiya

表2 基于3种模型构建的反演结果评价指标

Table 2 Estimation model results based on three models

模型 变换类型 建模集 验证集 R2 RMSE RPD R2 RMSE RPD RF FOD1.2 0.864 0.432 2.660 0.873 0.395 2.926 CWT1 0.847 0.491 2.337 0.881 0.406 2.841 PLSR FOD1.2 0.734 0.577 1.989 0.690 0.610 1.893 CWT1 0.711 0.601 1.909 0.682 0.618 1.868 SVM FOD1.2 0.788 0.538 2.133 0.674 0.644 1.794 CWT1 0.829 0.475 2.417 0.868 0.427 2.704