This study collected multispectral imagery of the Simao pine forest area in Mojiang Hani Autonomous County,Yunnan Province,using a UAV-mounted multispectral imaging system.With the aid of remote sensing image processing techniques and ground measurement data,two machine learning models,Random Forest and Support Vector Machine,were constructed to accurately analyze and predict the occurrence of pine forest insect damage.The results showed that both models effectively identified and predicted insect damage areas with AUC values above 0.9.The Random Forest model demonstrated superior accuracy,with a training set accuracy of 97.14％and a test set accuracy of 89.16％.The predicted areas of insect damage in the study area by the Random Forest and Support Vector Machine models were 224 264.135 m2 and 212 078.258 m2,respectively.The inversion accuracy of the Random Forest model was 92.86％,and that of the Support Vector Machine model was 87.82％,indicating overall good inversion effects.
关键词
无人机/多光谱影像/思茅松林/虫害监测/反演
Key words
UAV/multispectral imagery/Simao pine forest/insect damage monitoring/inversion
维普
万方数据