Assessing the Yield of Wheat Using Satellite Remote Sensing-Based Machine Learning Algorithms and Simulation Modeling

Globally, estimating crop acreage and yield is one of the most critical issues that policy decision makers need for assessing annual productivity food supply. Nowadays, satellite remote sensing geographic information system (GIS) can enable estimation these production parameters over large areas. The present work aims to estimate wheat (Triticum aestivum) Maharajganj, Uttar Pradesh, India, using satellite-based data products Carnegie-Ames-Stanford Approach (CASA) model. Pradesh largest wheat-producing state in this district well known its quality organic wheat. India leader grain export, and, hence, monitoring growth top economic priorities country. For calculation acreage, we performed supervised classification Random Forest (RF) Support Vector Machine classifiers compared their accuracy based on ground-truthing. We found RF a significantly accurate assessment (kappa coefficient 0.84) SVM (0.68). CASA model was then used calculate winter (Rabi, winter-sown, summer harvested) net primary (NPP) study area 2020–2021 season RF-based product. NPP-yield conversion showed 3100.27 5000.44 kg/ha 148,866 ha total area. results growing season, all districts had similar trends. A 30 observational points were verify model-based estimates yield. Field-based verification shows estimated correlates with observed (R2 = 0.554, RMSE 3.36 Q/ha, MAE ?0.56 t ha?1, MRE ?4.61%). Such an regional prove be promising methods calculating whole region’s agricultural concludes classifier-based has shown more meet requirements regional-scale thus, highly beneficial making.