Aerial Vehicle Detection and DBSCAN-Based Interchange Region Identification for Urban Traffic Analysis

As urban traffic becomes increasingly complex, this study examines the dynamics at the Puente del Guambra Interchange in Quito, Ecuador, employing a novel methodology that integrates YOLO for vehicle detection, Voronoi analysis for spatial segmentation, and DBSCAN for clustering. Using drone-captured video footage, the study effectively identifies and distinguishes the intrinsic characteristics of the interchange’s upper and lower passes. A custom tracking solution is developed to accurately follow vehicles across multiple frames, addressing challenges such as missed detections and high-speed anomalies caused by off-center bounding boxes. The system consolidates multiple IDs for the same vehicle and calculates velocities, applying thresholds to correct false high-speed instances. A vehicle traffic inventory of the interchange is conducted. The findings reveal significant variations in vehicle counts and velocities across key urban arteries, particularly Av. 10 de Agosto and Av. Patria. This integrated approach not only enhances the accuracy of traffic assessments but also provides valuable insights for urban planners and traffic managers. The results underscore the potential of utilizing advanced analytical techniques in traffic monitoring and management to improve urban mobility.