Leveraging Satellite Imagery and Machine Learning for Urban Green Space Assessment: A Case Study from Riyadh City

The “Green Riyadh” project in Saudi Arabia represents a major initiative to enhance urban sustainability by expanding green spaces throughout Riyadh City. The initiative aims to improve air and water quality, increase tree and plant coverage, and promote environmental well-being for city residents. However, accurately assessing the extent and quality of green spaces remains a significant challenge. Current methods for evaluating green areas and measuring tree density are limited in precision and reliability, preventing effective monitoring and planning. This paper proposes an innovative solution that leverages live satellite imagery and advanced deep learning techniques to address these challenges. We collect extensive satellite data from two sources and then build two separate analytical pipelines. These pipelines process high-resolution satellite imagery to identify trees and measure green density in vegetated areas. The experimental results show significant improvements in accuracy and efficiency, with the YOLOv11 model achieving a mAP@50 of 95.4%, precision of 94.6%, and recall of 90.2%. These findings offer a scalable and reliable alternative to traditional methods, enabling comprehensive progress evaluation and facilitating informed decision-making for urban planning. The proposed methodology not only supports the objectives of the “Green Riyadh” project but also sets a benchmark for green space evaluation that can be adopted by cities worldwide.