Abstract Manager

The City of Montréal aims to enhance the safety of vulnerable road users as part of its Vision Zero Action Plan. While collision data is a fundamental component of road safety assessments, it often provides an incomplete basis for proactive intervention. Traffic conflict analysis, capturing near‑miss events and interactions between road users, offers a valuable complementary approach for identifying risks before collisions occur.

Advances in automated video analytics and artificial intelligence have transformed the way conflict studies are conducted. Traditionally based on human observation over short time periods, conflict analysis can now be performed using longer-duration video recordings processed through automated detection algorithms. However, methodologies, detection criteria, and output formats differ significantly across commercial solutions, raising important questions about comparability, repeatability, and practical applicability.

As the first phase of a broader initiative to develop municipal guidelines for video-based conflict analysis, this project included: (1) a literature review of current practices and emerging techniques; (2) a survey of available technologies and data providers; and (3) a comparative analysis of conflict data collected and processed by three providers using two distinct technologies (video and LiDAR) at the same high-risk intersection in Montréal. This intersection, characterized by historically documented safety concerns, served as a relevant site for evaluating differences in analytical approaches.

The results highlight notable variations in how conflicts are detected, classified, and represented, as well as differences in interpretability. These findings provide essential insights for the City of Montréal as it moves toward developing practical guidelines to support consistent, transparent, and high-quality use of video conflict analysis. The recommended next steps aim to strengthen methodological clarity, improve decision-making, and position the City to proactively identify and mitigate safety risks in a sustainable and systematic way.