Deep learning-based vegetation monitoring using the NDVI index

This research aligns with the Saudi Vision 2030, as it presents an innovative approach to environmental conservation in the Kingdom, focusing on improving vegetation cover by contributing to increasing and maintaining the percentage of cultivated areas and utilizing AI-based insights to support the national program for sustainable environmental management. This project seeks to use advanced geospatial analysis and integrate the Natural Variety Vegetation Index (NDVI) with artificial intelligence techniques to improve environmental health and biodiversity, By developing an AI model capable of detecting NDVI values, an indicator of vegetation health, using standard RGB images, without relying on massive sensors. The expected results include increased plant quality, improved biodiversity, and the development of predictive environmental models. We implemented the project using data from the city of Riyadh to train the U-net architecture to extract vegetation gradients and track them through NDVI values, and we obtained an accuracy of 97%.