农业机械学报2025，Vol.56Issue(9)：547-556,565,11.DOI:10.6041/j.issn.1000-1298.2025.09.045
基于无人机多光谱遥感的覆膜冬小麦叶片含水率反演
Inversion of Leaf Water Content for Mulched Winter Wheat Based on Multi-spectral Remote Sensing of Unmanned Aerial Vehicle
摘要
Abstract
Efficient and accurate acquisition of water content of winter wheat leaves is of great significance in guiding farmland irrigation.Taking 2022-2023 mulched winter wheat as the research object,and a multi-spectral camera mounted on an unmanned aerial vehicle(UAV)was used to collect the spectral information of winter wheat during the greening,nodulation,tasseling,and filling periods,and the background elimination was carried out,and then the extreme learning machines(ELM),random forest(RF),and particle swarm optimization-support vector machine(PSO-SVM)were used to establish a model of leaf water content in winter wheat.Comparative analysis of the accuracy of leaf water content inversion was done by using"vegetation index+texture characteristics"and"vegetation index"as input variables,and re-modeled by filtering the characteristic variables for each fertility period through an iterative subset variable optimization method.The results showed that"texture features+vegetation index"as input variables improved the inversion accuracy of winter wheat leaf water content,and the R2 of the extreme learning machine,random forest,and particle swarm optimization vector machine was increased by 19.19％,16.85％,and 19.40％on average,respectively.Among them,the accuracy of the inversion of leaf water content in mulched winter wheat by using the random forest model was the highest in the regrowth-irrigation period(R2,MAE,and RMSE were 0.94,0.02 g/g,and 0.03 g/g for the training set,and 0.88,0.03 g/g,and 0.05 g/g for the test set,respectively);and R2 for the random forest model was the highest in the nodule pulling,tasseling,and irrigating periods after iterative subset of the variable optimization methods.and grouting stages improved the R2 by 19.09％,20.58％and 12.21％,respectively.The results can provide theoretical references for UAV multi-spectral remote sensing to monitor the water content of leaf blades of mulched winter wheat.关键词
覆膜冬小麦/叶片含水率/无人机多光谱/机器学习/纹理特征/反演Key words
mulched winter wheat/leaf water content/UAV multi-spectral/machine learning/texture feature/inversion分类
农业科技引用本文复制引用
谷晓博,徐洋,程智楷,周智辉,韦春宇,杜娅丹..基于无人机多光谱遥感的覆膜冬小麦叶片含水率反演[J].农业机械学报,2025,56(9):547-556,565,11.基金项目
国家重点研发计划项目(2021YFD1900700)和陕西省重点研发计划项目(2023-YBNY-040) （2021YFD1900700）
