Integration of remote sensing and machine learning for real-time monitoring of Soil Health Parameters in response to Climate Variability in Karnataka

A remote-sensing and artificial neural network model is proposed for real-time data collection of soil health indices in Karnataka to counter climate change. The main objective is to monitor and understand the seasonal variations in moisture, nutrients, pH, organic carbon, texture and the impact of climate variability on these soil parameters. Data collection included the satellite images from Sentinel-2 and other optical bands such as Lands at 8 and additional high-resolution imagery from UAVs or drones, and the soil moisture was continuously streamed using IoT sensors. Random Forest, Gradient Boosting, and Neural Networks were used to forecast the changes of soil health parameters. These models were built on preprocessed data and were verified based on the results obtained. The climate models incorporated with the streaming technology allowed the system to estimate current and future climate conditions. The study shows the efficiency, significance and usefulness of the integrated approach in the assessment of soil health, and can be used as a tool in educating farmers for boosting performance and productivity of their crops. The system enhances the climate resilience for farmers in reaction to variability within a production process as well as minimizing adverse impacts, thus encouraging sustainable farming in Karnataka.