基于不同空间分辨率无人机多光谱遥感影像的小麦倒伏区域识别方法

魏永康, 杨天聪, 丁信尧, 高越之, 袁鑫茹, 贺利, 王永华, 段剑钊, 冯伟

Wheat Lodging Area Recognition Method Based on Different Resolution UAV Multispectral Remote Sensing Images

WEI Yongkang, YANG Tiancong, DING Xinyao, GAO Yuezhi, YUAN Xinru, HE Li, WANG Yonghua, DUAN Jianzhao, FENG Wei

表7 不同倒伏分类模型的比较验证

Table 7 Comparison and validation of different lodging classification models

特征选择方法 分辨率/cm 特征数量 SVM RF KNN OA/% Kappa OA/% Kappa OA/% Kappa 全特征集 1.05 52 87.5 0.753 85.2 0.731 80.6 0.644 2.09 52 84.2 0.712 80.2 0.638 77.4 0.582 3.26 52 81.9 0.664 79.1 0.625 76.8 0.578 ReliefF算法 1.05 14 79.9 0.604 78.9 0.598 79.2 0.604 2.09 9 75.6 0.538 74.2 0.527 75.6 0.538 3.26 6 74.4 0.531 75.3 0.525 73.2 0.502 RF-RFE算法 1.05 11 85.7 0.724 83.9 0.710 77.9 0.592 2.09 13 84.1 0.702 82.6 0.682 81.4 0.651 3.26 6 82.7 0.686 80.3 0.642 77.0 0.577 Boruta-Shap算法 1.05 35 90.4 0.823 88.9 0.769 79.2 0.631 2.09 39 88.6 0.765 85.3 0.732 77.8 0.586 3.26 36 86.2 0.741 84.1 0.702 74.8 0.539