基于无人机多光谱遥感的玉米草地贪夜蛾虫害等级监测OA

Fall armyworm(Spodoptera frugiperda)is a major migratory pest that poses a serious threat to maize production.Rapid and accurate monitoring of this pest is essential for scientific pre-vention and precision pesticide application.To enable the identification and quantitative assessment of fall armyworm damage severity at maize whorl stage,this study focused on infested maize fields and developed a damage-level monitoring model using UAV-based multi-spectral remote sensing.From the UAV multi-spectral imagery,an optimal subset of spectral features sensitive to pest stress was selected through a cross-validation recursive feature elimination algorithm.Subsequently,three classification models:K-Nearest Neighbors(KNN),Support Vector Machine(SVM),and Random Forest(RF)were constructed,and grid search was employed for hyperparameter optimization.The re-sults demonstrated that the random forest model achieved the best overall performance,with a valida-tion accuracy of 0.871 8,an F1 score of 0.872 4,and a Kappa coefficient of 0.828 8,confirming its ca-pability to accurately monitor fall armyworm infestation levels.