Intelligent CCTV for Urban Design: AI-Based Analysis of Soft Infrastructure at Intersections

Researchers at the University of Minnesota developed an AI-powered CCTV analytics framework to measure the effectiveness of soft infrastructure interventions (temporary pedestrian refuges, curb extensions) on traffic safety. The study found speed reductions of 16-20% at both signalized and unsignalized intersections in Minneapolis, demonstrating that computer vision-based traffic analysis enables rapid, cost-effective evaluation of urban design policies.

This research demonstrates a practical application of computer vision and deep learning to urban transportation challenges, moving beyond theoretical AI capabilities into measurable policy outcomes. The study leveraged existing CCTV infrastructure—a cost advantage that makes widespread deployment feasible for municipalities—to quantify how physical street design changes driver behavior. By employing perspective-based speed estimation and repeated monitoring across two weeks, the researchers established credible evidence that soft interventions reduce both mean speeds and high-percentile speeds, directly addressing safety concerns at intersections.

The findings arrive as cities globally grapple with pedestrian safety and traffic calming in dense urban areas. Traditional methods for evaluating such interventions rely on manual counts or expensive sensor deployments, creating friction in evidence-based policy iteration. This AI approach removes those barriers, enabling rapid experimentation and measurement cycles that could accelerate adoption of effective safety measures.

For municipal governments and urban planners, this work validates AI-powered monitoring as a legitimate tool for transport planning, potentially shifting capital allocation toward computer vision infrastructure rather than physical sensors. The 12-22% reductions in vehicle speed represent measurable safety gains that justify further investment in similar analytical frameworks. Technology vendors developing smart city solutions can build on this methodological template to expand into traffic management consulting.

• →AI-enabled CCTV analysis reduced vehicle speeds by 16-20% following soft infrastructure interventions at Minneapolis intersections.
• →Existing CCTV networks can be repurposed for traffic analysis, eliminating expensive sensor deployment costs.
• →Pass-through traffic decreased by up to 12.2%, indicating the interventions effectively discourage cut-through vehicle routes.
• →Repeated post-installation monitoring (Week 1 and Week 2) demonstrates sustained behavioral change rather than novelty effects.
• →Computer vision methods enable rapid, evidence-based evaluation of urban design policies at scale.