Identifying Urban Decay with AI

Yikes! When is the city gonna do something about these potholes?

This is Sandra Tsing Loh with the Loh Down on Science.

Cities are growing – FAST – and urban planners need better data to keep track of where repairs are needed. How about keeping an EYE on them with A-I?

Enter Andrea Vallebueno and Yong Suk Lee of Stanford and Notre Dame University.

Researchers trained an A-I model to detect decaying objects in a city. These included graffiti, discolored building facades, garbage, and potholes. A-I used street view imagery from cities like San Francisco, Mexico City and South Bend, Indiana.

Surprisingly, the A-I model recognized patterns of decay and inequality across the different neighborhoods. It even provided information on homelessness patterns in San Francisco! The hope is A-I detection can be used to help policymakers respond quickly to community needs.

Soon we won’t need to call potholes in. Maybe phone a friend instead. You know… your friend… big brother…


Reference: Vallebueno, A., & Lee, Y. S. (2023). Measuring urban quality and change through the detection of physical attributes of decay. Scientific Reports, 13(1), 17316.