UAV-based Traffic Intensity Analysis Framework: A Case Study on Pedestrian Crossings

• Published in: Baltic Journal of Modern Computing Vol. 12; no. 1; pp. 110 - 125
• Main Authors: Gudauskas, Julius, Petkutė, Gintarė, Trakšelis, Kristupas, Kriščiūnas, Andrius
• Format: Journal Article
• Language: English
• Published: Riga University of Latvia 01-01-2024
• Subjects: Accuracy; Algorithms; Architecture; Case studies; Data collection; Datasets; Deep learning; Drone vehicles; Drones; Neural networks; Object recognition; Pedestrian crossings; Pedestrians; Road humps; Roads & highways; Traffic flow; Unmanned aerial vehicles; Vehicles; Lithuania
• ISSN: 2255-8942; 2255-8950
• DOI: 10.22364/bjmc.2024.12.l.07

In this research, a novel framework is presented for traffic monitoring and analysis that leverages drone footage and georeferencing technology. The study focuses on the analysis of various pedestrian crossing configurations, including standard zebra crossings, crossings with islands, and crossings with speed bumps, with a particular emphasis on their impact on vehicle and pedestrian behaviour. High-resolution drone footage was captured along selected roads in Kaunas, Lithuania, enabling a comprehensive examination of the interactions between vehicles and pedestrians at these diverse crossings. The proposed framework incorporates advanced object detection models, such as Faster R-CNN, to effectively identify and analyse vehicular movements. The experimental validation of the framework is demonstrated through a specific case study involving pedestrian crossing analysis.