Satellite-based Crop Monitoring and Yield Prediction: An Integrated Approach using Sentinel-2 and Soil Health Data for Precision Agriculture

This article provides a hybrid model to perform Remote Sensing and Machine Learning tools to handle farming better through Hybrid and decision tree models that inspect Sentinel-2 data for crop identification alongside weed and disease detection. This research shows that the Hybrid Model delivers higher performance than single models in 5km agricultural regions by achieving 97.07% accuracy along with 0.0293 as its best error measures. Based on these results the proposed model shows better sensitivity for detecting real crop conditions when compared to CNN Model which produced lower accuracy with higher errors. The hybrid approach proves efficient when applied widely in precision agriculture with small resource needs to keep food production running.