Monitoring Tobacco Crop Health using Geospatial Data

This study examines a method that uses the Normalized Difference Vegetation Index (NDVI) to track the health of tobacco crops and provide farmers practical advice. The NDVI was calculated using Sentinel-2 and Landsat satellite image data. Google Earth Engine (GEE) was used to process the images. The method demonstrated a 91.7% accuracy in categorising crops, and a very small mean absolute error of 0.02 in NDVI computation. Nevertheless, drawbacks were observed, including the limits of a rule-based recommendation system, possible spectrum misinterpretation, and data gaps under foggy situations. Comparative field tests conducted over a month showed a 91.25% agreement between the system's suggestions and actual results. Future research is advised to incorporate other data sources, such as Synthetic Aperture Radar (SAR), improve data imputation methods, and switch to a dynamic recommendation model. The results show the system's potential as a scalable tool for precision agriculture. This means the system can empower farmers with timely insights to optimise crop management, reduce costs, and improve productivity.