Rubbish trucks detect neighbourhood blight in Tuscaloosa

The City of Tuscaloosa and the University of Alabama (UA) have developed a system that can automatically detect blight-related issues such as dilapidated property, abandoned vehicles, litter, illegal parking and dumped furniture.

The patent-pending system, which was created by Dr Erik Johnson, Assistant Professor of Economics in the Culverhouse College of Business at UA, and Brendan Moore, Executive Director of Urban Development for the City of Tuscaloosa, uses images of property captured by cameras mounted on city rubbish collection trucks.

These images feed into a computer model which uses a scoring system to provide an assessment and information on potential remedies. The artificial intelligence (AI) model was trained using data and images of past blighted properties and code violations.

The pilot phase is ongoing and the goal is to eventually expand the detection to all parts of Tuscaloosa to help the city reduce staff time spent on inspections and take a more proactive approach to addressing property issues early before they escalate and potentially become contentious and expensive to fix.

Moore said: “Efforts to address blight are not new or distinct to Tuscaloosa but it is a constant problem that is difficult to appropriately staff and address. This technology allows us to create early, equitable interventions that can enhance communities, prevent neighbourhood decline, and connect underserved populations to social services to generate long-lasting change.”

Scale

As well as providing a blight score, the system can also determine the issue that is driving the score – such as a broken fence, damaged windows, etc.

Johnson said: “Then we are able to offer bundles of low-cost remediation which will help drive that blight score down. That would be a lot harder for a code enforcement officer.”

The insights could also help to identify people who may be  at risk from living in substandard conditions and connect them with the right support agencies.

In 2018, Tuscaloosa conducted a blight inventory in the West Tuscaloosa area of the city as part of a broader effort to revive the district. Blight issues have been significantly reduced since then, Moore said, but notes that traditional processes can be “extremely expensive” and only capture a “snapshot in time”. The automated detection technology on waste trucks offers a much cheaper way to track issues on an ongoing basis over the whole city, he said.

The system takes still images rather than video. “This isn’t about monitoring people; it’s just about getting an idea what the property looks like. The goal is not surveillance,” Johnson said.

In future, the solution could be made available to other cities facing similar challenges. “It has the ability to be commercially scaled,” Moore said.

This is the latest example of cities using data and building on existing systems to become more proactive.

Earlier this year, Cities Today reported on research from Ohio State University which found that data about service requests to city 311 lines could help identify potential hotspots for opioid use and overdoses.

Findings showed that calls about issues such as municipal code violations, public health matters like pest management and broken streetlights were the best indicators of opioid use in Columbus communities.