基于无人机多光谱遥感的大豆叶面积指数反演

为给大豆科学管理提供基础数据,利用无人机多光谱遥感数据实现对大豆叶面积指数(LAI)的反演估值。从多种光谱植被指数中选出与LAI相关性较好的5种指数,分析探讨在田块尺度上,适用于东北地区的大豆叶面积指数的低空无人机遥感反演模型。结合田间实测LAI数据及模型精度及拟合效果,NDVI模型精度较好,但拟合效果较差,其余4种植被指数模型精度和拟合效果较好,拟合效果R²均达到了0.6以上;支持向量机模型决定系数R²达到0.688,均方根误差达0.016,具有更好的预测能力。2种模型均表明无人机多光谱遥感系统可以快速反演田间大豆叶面积指数,在指导精准农业生产方面具有实用意义。

Inversion of Soybean Leaf Area Index Based on UAV Multispectral Remote Sensing

In order to provide basic data for the scientific management of soybean, the inversion and estimation of LAI was realized by using the multispectral remote sensing data of UAV. Five indices with good correlation with LAI were selected from various spectral vegetation indices, and the remote sensing inversion model of soybean leaf area index in northeast China was analyzed and discussed. The results show that except NDVI, the other four vegetation index models have better precision, and the determination coefficient R² is more than 0.6; support vector machine model, the determination coefficient R² is 0.688, and the root mean square error is 0.016, which has better prediction ability. Both models show that the UAV multispectral remote sensing system can quickly retrieve the soybean leaf area index in the field, which has practical significance in guiding precision agricultural production.