ESTIMATION MODEL OF WINTER WHEAT YIELD BASED ON UAV HYPERSPECTRAL DATA

Winter wheat is one of the main food crops in China, accurately forecasting the yield of winter wheat is of great significance for agricultural management and decision. UAV remote sensing has the advantages of high spatial-time resolution, low cost, flexibility and repeatability. In this paper the growth condition remote sensing and yield estimation of winter wheat were carried out using UAV hyperspectral sensor in Xiaotangshan Town, Changping District, Beijing. Based on the DSD (directional second differential) method and AIVI (Angular Insensitivity Vegetation Index), LAI (leaf area index) and LNC (leaf nitrogen content) of winter wheat at heading and filling periods were retrieved, and according to the result of DSSAT simulation, the forecasting model between LAI, LNC at heading and filling periods and yield of winter wheat was established by random forest algorithm. The R-2 of yield estimation model is 0.787 and RMSE is 727.87 kg/ha, which shows the yield estimation model can accurately and effectively estimate winter wheat yield.