Automated Detection, Tracking, and Safety Analysis of Pedestrians and Cyclists Using YOLOv9

Banafsheh Rekabdar, Portland State University

Co-investigators:

Summary:

This applied research project aims to adapt and further train YOLOv9 to enhance its ability to accurately detect, track, and count cyclists and pedestrians in video data from the Portland Metro Area. The method will also extract surrogate safety metrics, such as Post Encroachment Time, which are crucial for assessing conflicts and designing safer, more efficient transportation facilities. By analyzing these interactions during daylight, the project seeks to deepen our understanding of active transportation behaviors. The expected outcomes include improved detection and analysis capabilities, leading to better infrastructure planning, proactive safety measures, and informed design elements that encourage biking and walking. Ultimately, this project will contribute to developing human-centered, multimodal transportation systems that promote safety and equity in the Pacific Northwest.

Project Details

Project Type:
Research
Project Status:
In Progress
End Date:
August 31,2026
UTC Grant Cycle:
PacTrans Year 2
UTC Funding:
$70,000