Machine Learning Techniques in Wheat Crop Yield Prediction Using NDVI Indices and Meteorological Parameters

In this study, an amalgamation of NDVI with different meteorological entities were taken in use to predict the wheat crop yield for the study area of Saharanpur region of Uttar Pradesh. Cloud free crop season's images of Sentinel satellite for a period of year 2021–22 containing Saharanpur along with its nearby areas were considered for the discrimination of the wheat crop. SNAP tool was used to derive NDVI images of the study area. Six different models including SVM, Random Forest, Neural Network, Linear Regression, Gradient Boosting and AdaBoost were used to calculate the values of MSE, RMSE, MAE and R2 score. The dataset including the NDVI values along with meteorological data is used in this analysis. The meteorological data includes minimum and maximum temperature, pH, humidity and sun shine hours. The observed yield prediction of the study area is calculated and compared with the predicted values. The minimum value of RMSE of 0.10 was observed for Random Forest classifier.