AI Near-Miss Camera Tracks Speed and Risk at Yokine Intersection
AI Near-Miss Camera Tracks Speed and Risk at Yokine Intersection

A busy intersection in Yokine has become the latest site for a new AI-powered camera designed to detect near-misses between vehicles, pedestrians, and cyclists.

Camera Installation and Purpose

The non-enforcement near-miss camera was installed at the intersection of Alexander Drive and Wordsworth Avenue on Tuesday, June 16, as part of a road safety initiative between the City of Stirling and the Road Safety Commission. It is one of six such cameras being deployed across the metropolitan area.

The portable camera uses AI software to analyze video in real time, providing insights into near-miss events, including severity levels and risk ratings. It can track drivers' speeds, measure distances between road users, and identify different transport types.

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

Community Concerns and Previous Proposals

The Yokine intersection has previously been the subject of potential upgrades and community consultation on a proposal to restrict right-hand turns from Wordsworth Avenue onto Alexander Drive. However, with 86 percent of the 357 respondents objecting to the proposal, the City of Stirling resolved not to proceed at its meeting in August 2025.

The information collected at the site is expected to help identify risk factors and guide the introduction of safety treatments if needed. The City of Stirling said the installation of the camera was supported by Mt Lawley MLA Frank Paolino after ongoing resident concerns about the safety of the intersection.

Data Privacy and Previous Trial Results

According to the Road Safety Commission, the data collected from the cameras is de-identified through facial, number plate, and vehicle blurring before being uploaded to a dashboard. It can then be studied and used by local councils to help plan road safety improvements and decide where upgrades are needed.

The first trial of the AI technology, delivered by Australian company LAB3, in North Perth last year revealed dozens of daily near-misses, prompting plans for safety upgrades at the busy inner-city intersection. The camera detected an average of 38 close calls every day at the intersection of Fitzgerald and Angove streets, with 99 percent being right-angle near misses or 'T-bone' crashes.

Duration and Next Steps

The camera at the intersection in Yokine is expected to remain in place until September. The data gathered will be used to assess risk factors and determine potential safety treatments for the intersection.

Pickt after-article banner — collaborative shopping lists app with family illustration