Vehicle Detection Using YOLO and Mobility Tracking During COVID-19 Pandemic Lockdowns

• Main Authors: Tan, Marion Ivan L, Calgo, Clyde J, Cabantac, Sheanne Eric P, Honrado, Jaime Luis E, Libatique, Nathaniel Joseph C, Tangonan, Gregory L
• Format: text
• Published: Archīum Ateneo 2021
• Subjects: COVID-19; Pandemics; Vehicle detection; Conferences; Pipelines; Internet; Classification algorithms; traffic videos; object detection; COVID-19 pandemic; SARS-COV-2; lockdown; community mobility; Electrical and Computer Engineering; Transportation
• Online Access: https://archium.ateneo.edu/ecce-faculty-pubs/120; https://ieeexplore.ieee.org/document/9612481
• Institution: Ateneo De Manila University

Data-driven approaches to traffic monitoring have important applications in tracking vehicle mobility in the COVID-19 pandemic lockdowns. We report preliminary results of a pipeline that uses the You Only Look Once (YOLOv3) and the Simple Online and Realtime Tracking (SORT) algorithms to count and classify vehicles in traffic videos. We correlate vehicle counts in Katipunan Avenue, Metro Manila and Google COVID-19 Community Mobility Reports from May to August 2020 and we show that vehicle detection data may be considered for monitoring community response to changes in COVID-19 lockdown stringency levels.