Estimating Surface Soil Moisture based on Remote Sensing Data and Crop Yield Classification using Machine Learning

Soil moisture plays a critical role in environmental monitoring and sustainable agricultural practices due to its significant impact on crop growth, water availability, and overall ecosystem health. Properly understanding and managing soil moisture is vital for efficient water usage, especially in agriculture. Traditional methods in measuring soil moisture have been confined to ground-based small-scale methods, which are high price due to high labor intensity and limited spatial coverage. This project addresses its limitation in the spatial view by integrating remote sensing data from the Sentinel-1 satellite. Sentinel-1 is particularly effective in monitoring the soil moisture as it captures data in all climate conditions. By combining the satellite data with Google Earth Engine (GEE) we create a soil moisture dataset. For analyzing the joint satellite and ground truth data, the machine learning technique using Random Forest (RF) is applied and classification of crop yield is done.