Research On Remote Sensing Estimation Of Maize Yield In Plots In West Jilin Based On Crop Growth Model

The crop growth model is a comprehensive crop planting model that combines regional climate conditions,soil characteristics,crop variety characteristics,field management data and other modules.Among them,the simulation of crop yield is an important function of the crop growth model,and also an important indicator for decision-making and risk assessment of actual crop production labor.Although the crop growth model can simulate the growth status of crops day by day,it lacks a certain amount of adjustment and control for a wide range of crop growth and crop yield reduction after disasters.Using the reflectance value of the remote sensing data,the crop reflectance in the crop growth model can be assimilated,and the simulated annealing algorithm is used to make the simulated value closer to the real value to form a coupled model.This article is based on the crop growth model,combined with data assimilation related theories,using multi-source remote sensing data and multiple remote sensing data processing methods and remote sensing image supervision classification methods,to gather multi-year weather data,precipitation data,soil survey data,planting data in the study area 1.Field management data,and remote sensing estimation of 5 corn plots in 5 counties in the west of Jilin including Baicheng and Songyuan.Select the meteorological data of 6 meteorological stations from 2017,2018,January to October 2019 within the study area,early July,late July,early August,late August,early September,September In the later period,the multi-temporal remote sensing image data covering the research area is extracted,and the reflectance value of the image is extracted as the input of the crop growth model to drive the crop growth model.The sensitivity analysis of the input data of the CERES_MAIZE corn model in the DSSAT system was carried out,which carried out the model localization experiment on the variety parameter data of the five sites.Using experimental data and remote sensing data to build a remote sensing estimation model of corn yield,the average absolute percentage error is 6.71%,the maximum relative error is-14.29%,and the minimum relative error is 0.28%.The model has certain adaptability and reliability.On the basis that single-point corn yield simulation can be achieved,according to the principle of statistical multiple stepwise regression,multi-temporal remote sensing vegetation index values are extracted to establish a regression yield estimation model,so that the crop growth model is not limited to simulating the corn yield of pixels,but to expand To the extent of the plot.The single point corn yield simulation value and plot corn yield distribution are evaluated for accuracy,combined with field corn yield assessment model accuracy.The simulation result is close to the measured output value,indicating that the model has certain feasibility.The distribution of corn yield in plots is close to the growth status of high-resolution remote sensing data during the growth period of corn,and the yield estimation method has certain reliability.