Machine Learning Based Agricultural Crop Yield Estimation in Yingcheng District, Hubei Province

Precision agricultural finance requires precise monitoring of agricultural conditions, including crop area and yield estimation. In this paper, Yingcheng City, Hubei Province, were selected to study the yield estimation based on the growth analysis combined with historical crop yield and meteorological data. There are many factors that affect crop yield, including meteorological conditions, geographical conditions, soil conditions, disaster situations, and the use of fertilizers. The artificial intelligence machine learning models based on multiple characteristic factors such as crop yield and temperature, precipitation, LAI, EVI, etc., as well as soil and meteorological conditions and other yield estimation features obtained from Google Earth Engine was constructed to conduct crops yield estimation so as to improve the precision of yield estimation in Yingcheng District, Hubei Province. The accuracy of the model was validated by the average yield estimation accuracy and goodness of fit compared to both the traditional regression model and the support vector machine model. Experiments show that machine learning model is better than regression model, and Random forest machine learning model is better than support vector machine model.