Multi-Modal Deep Learning for Crop Yield Forecasting Using IoT, Satellite, Weather, and Soil Data
Contributors
Dr Vaishali Khobragade
PAWAN KUMAR CHAURASIA
Keywords
Proceeding
Track
General Track
License
Copyright (c) 2026 Sustainable Global Societies Initiative

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Abstract
Precise crop yield forecasting is vital for sustainable agricultural planning and food security. Traditional approaches often depend on individual data sources and therefore cannot completely capture the spatial and temporal variability of crop growth. This paper suggests a multi-modal deep learning framework integrating IoT sensor observations, weather information, Sentinel-2 satellite imagery and soil data for crop yield forecasting. Satellite-derived indices such as NDVI, GNDVI, and EVI provide spatial crop information, though IoT and weather observations capture localized and temporal conditions. A feature-fusion approach is employed to combine mixed data, followed by CNN–LSTM and Transformer-based models for yield prediction. The proposed framework is compared with conventional machine learning and deep learning models using RMSE, MAE, R², and MAPE. The projected approach offers a foundation for more precise, scalable, and climate-resilient precision agriculture.