UAV-based flexible measurement platform and deep learning approach for precise vehicular emission assessment

Urban transportation is a major contributor to air pollution and traffic congestion, yet traffic monitoring and vehicular emission assessments are often conducted separately using fixed systems with limited coverage. This study introduces a flexible, UAV-based mobile platform capable of simultaneously collecting traffic and emission data in urban environments. The UAV was equipped with cameras and emission sensors, enabling both dynamic and stationary measurements across key road segments. Aerial imagery was analyzed using the state-of-the-art deep learning algorithm YOLOv9 to generate high-resolution traffic density and emission maps. Data from UAV flights conducted at various times over a one-week period revealed dynamic pollution patterns and helped identify major emission hotspots. The results demonstrate the practicality, energy efficiency, and adaptability of the UAV-based platform as a sustainable alternative to conventional monitoring systems. The versatility and flexibility of the developed mobile measurement platform, compared to fixed systems, underscore its potential as a sustainable technology of the future. Furthermore, the employed method exhibits higher energy efficiency relative to current monitoring and measurement systems, surpassing them by multiple folds.