Study On Precise Monitoring Of Wheat And Corn Growth Based On Remote Sensing Image Of Unmanned Aerial Vehicle

In order to monitor crop growth more quickly and accurately in actual agricultural production,estimate crop yields and provide scientific basis for management and regulation of agricultural production in small and medium-sized regions,in this study,we chose winter wheat and summer maize as the main crops.Firstly,winter wheat was set as the variety,and summer maize was treated with different nitrogen application rates.Secondly,based on the UAV platform,we obtained the high-resolution spectral images of the key growth stages of wheat and cornthrough which we calculatedspectral parameters with the highest correlation of growth information.Then using thespectral parameters weconstructedthe best fitting model of crop growth monitoring and yield prediction.Finally we tested the relevant models to verify the accuracy and stability of the model.The specific results were as follows: 1.Monitoring model of LAI of Wheat and maizeThe results of model fitting indicated that monitoring model constructed by the ratio vegetation index RVI(810,560)has the best monitoring effect on winter wheat LAI.The model is y =-1.1629 × RVI(810,560)+ 7.9207,and the determination coefficient is 0.71.The RMSE and MRE of prediction models were 0.68 and 13.31%,respectively.This model can be used as a fitting model for the accurate monitoring of winter wheat LAI.The results of model fitting indicated that monitoring model constructed by the difference vegetation index DVI(800,550)has the best monitoring effect on summer maize LAI.The model is y =19.128 × DVI(800,550)-0.8742,and thedetermination coefficient is 0.90.The RMSE and MRE of prediction models were 0.41 and 14.36%,respectively.This model can be used as a fitting model for the accurate monitoring of summer maize LAI.2.Monitoring model of dry matter accumulation in wheat and maizeabovegroundThe results of model fitting indicated that monitoring model constructed by the difference DVI(810,560)has the best monitoring effect on the dry matter accumulation of the upper part of winter wheat.The model is y =-8.4529 × DVI(810,560)+ 5.7085,and the determination coefficient is 0.74,the RMSE and MRE of the prediction model were 0.36 and 17.58%,respectively.The model can be used as a fitting model for accurate monitoring of dry matter accumulation in winter wheat.The results of model fitting indicated that monitoring model constructed by the normalized vegetation index NDVI(760,560)has the best monitoring effect on the dry matter accumulation of summer maize.The model is y = 323 × NDVI(760,560)-1.8303,and the determination coefficient is 0.95,the RMSE and MRE of the prediction model were 8.92 and 13.17%,respectively.The model can be used as a fitting model for accurate monitoring of dry matter accumulation in summer maize.3.Monitoring model of SPAD of Wheat and maizeThe results of model fitting indicated that monitoring model constructed by the normalized vegetation index NDVI(790,660)has the best monitoring effect on the winter wheat SPAD value.The model is y = 20.001 × NDVI(790,660)+ 42.366,and the determination coefficient is 0.72,the RMSE and MRE of the prediction model were 2.24 and 3.66 %,respectively.This model can be used as a fitting model for accurate monitoring of winter wheat SPAD values.The results of model fitting indicated that monitoring model constructed by the difference vegetation index DVI(735,550)has the best monitoring effect on the summer maize SPAD value.The model is y = 42.8046 × DVI(735,550)+ 48.219,and the determination coefficient is 0.69,the RMSE and MRE of the prediction model were 2.21 and 3.45%,respectively.This model can be used as a fitting model for accurate monitoring of summer maize SPAD values.4.Predict model of Yield of Wheat and maizeThe model fitting results showed that the fitting model of vegetation index at the flowering stage and winter wheat yield was the best.The model constructed by flowering NDVI(870,679)had the best effect on the winter wheat yield estimation model.The model was y = 2924× NDVI(870,679)+ 6534.6,and the determination coefficient is 0.66,the RMSE and MRE of the prediction model were 178.28 and 1.68 %,respectively.This model can be used as a fitting model for winter wheat yield estimation.The model fitting results showed that the fitting model of vegetation index at the milk stage and summer maize yield was the best.The model constructed by the milk maturity ratio vegetation index RVI(790,660)has the most estimation effect on summer maize yield.Well,the model is y =-2219.5 × RVI(790,660)+ 16836,and the determination coefficient is 0.75,the RMSE and MRE of the prediction model were 462.81 and 5.02%,respectively.This model can be used as a fitting model for summer corn yield estimation.