Greenville City pilots 'City Detect' AI cameras on sanitation trucks to flag code violations

Greenville City is testing an artificial‑intelligence program called City Detect that uses cameras mounted on city waste‑collection vehicles to capture images of properties and flag potential code violations, Buddy, Director of Building and Codes and Chief Building Official, told the City Council at a work session.

"It's called City Detect. And what it does is it uses artificial intelligence to identify code violations in pictures," Buddy said. The department mounts cameras on trucks that run all city routes; images are processed by the AI and then presented on an interactive GIS map that highlights street segments and individual parcels with possible violations.

Buddy said staff set the AI parameters to identify items such as boarded windows, tarps on roofs, overgrown grass and other signs of structural damage. The system produces both a map view and a list, and staff can click each flagged parcel to view the original image. Buddy described the program as "another set of eyes" to help the city prioritize inspections and public‑works pickups after events such as Hurricane Helene.

Betty, a staff member involved in inspections and review, told council the system will not replace human review. In response to a council question about false positives, Betty confirmed the city will have people verify the images: "Absolutely," she said when asked whether human eyes will screen the results.

Buddy said the program builds a historical picture of parcels so staff can track whether a condition improves over time. He showed an example where a July image showed grass cut but the roof still had a tarp; the system lets staff review prior months’ images to see progress or recurrence. The pilot was initially used to help locate storm‑related debris and damaged properties and has also been used to identify illegal dumping so public works can clear sites proactively.

City Detect is being provided by a private vendor the presentation named as City of Tech. Buddy said the city is still in a testing phase and is waiting for network keys that will allow the team to push flagged items into the city’s code‑enforcement software to create cases. Staff emphasized that many AI hits will be filtered out by staff — for example, a dumpster with trash visible will not become a code case — and that staff will decide which items merit inspection or a work order.

Council members asked about the pilot’s privacy safeguards and about criteria for what the AI flags. Buddy and Betty said they are configuring the system to minimize irrelevant hits and to ensure human review before any enforcement action. Buddy credited former staff member John McDonough with helping start the program.

Staff said they will continue testing the system, tune the parameters, obtain the network keys needed to integrate results into case management and report back as the pilot progresses.