Prediction Of Maize Yield Based On High-resolution Remote Sensing Images

Maize has a high growth capacity,a wide range of cultivation,and a wealth of uses,which can bring enormous benefits.Globally,the planting area and yield are leading positions.Gongzhuling City,Changchun City,Jilin Province,is one of the golden corn belts in China.It is an important grain production base and commodity grain export base in Changchun City,occupying an important position in the city’s agricultural system.During the maize planting process,the growth of crops is crucial for agricultural management.By observing crop yield data,it is possible to understand the productivity of farmland and predict economic benefits.Using advanced technical means to collect corn growth status and accurate yield data can not only provide guidance,supervision,and disaster assessment for agricultural production,but also provide the government with more scientific and reasonable policies to ensure national food security.In this study,we collected a large number of corn samples from Huaide Town,Gongzhuling City,Changchun City.After analyzing the remote sensing image and geological survey data of Gaofen-1,NDVI time series information was obtained,and combined with various factors affecting crop growth,the growth status of corn was evaluated using statistical methods.In addition,combining NDVI values,meteorological conditions,topographic characteristics,and field investigations,key factors affecting corn yield have also been identified.Using regression analysis techniques,an accurate remote sensing estimation model for corn yield has been constructed to better grasp the growth of corn.Through accurate analysis of corn production in Huaide Town,Gongzhuling City,accurately predict its development trend.Through experimental research,the main conclusions of this article are as follows:(1)By using high-precision remote sensing images,we can conduct a complete study of the four months from early June to late September while maintaining the original geographical location and soil type.Using standardized processing,NDVI grading,geological survey,and other technologies,we can more accurately predict the growth of corn in 2022.A total of 657258.26 mu of arable land and 4907227.55 mu of corn planting area are proposed,accounting for 74.66%.Referring to the 2022 corn planting data of the "3S+Internet of Things" precision underwriting and claims settlement service new model,the 2021 financial support agricultural service innovation project of the Ministry of Agriculture and Rural Affairs,it is shown that the accuracy of the extraction results is 96.93%,and the accuracy of the growth monitoring results is 95.82%.This indicates that the method is suitable for extracting agricultural planting area and monitoring growth in large scale and long time series.The remote sensing index of the planting area,yield situation,and yield gradient area of each village were calculated,and the overall yield situation of each village was analyzed in depth.(2)Based on the results of growth monitoring and geological survey sampling,it is concluded that the main disasters suffered by the region in 2022 are waterlogging and wind disasters.In combination with the geographical location,meteorological conditions,topography,and other factors of the study area,scientific disaster prevention and loss reduction measures are proposed.(3)Analyze the correlation between sensitive factors and yield prediction,combining elevation data factors,meteorological factors,longitude and latitude factors,and NDVI accumulation value factors.Three types are selected from various different factor combinations for modeling,and 10 sets of actual production measurement data are brought into the model for accuracy verification.The experimental results show that the model with the NDVI accumulation value on August 20 th as the sensitive factor of the yield estimation model has the best accuracy,which is closest to the measured yield in geological exploration.The error range is-9.61% to 8.75%,the accuracy rate is 90.39% to 99.66%,the standard error is34.9261 kg,and the tolerance is 69.8522 kg,all within the tolerance range.Therefore,this model can be used to estimate the output of Huaide Town.The results showed that using multi temporal and high resolution remote sensing images,through standardized processing,NDVI grading,disaster information,and other technical methods,is suitable for large-scale,long-term series corn yield estimation.