October 6, 2020
November 2, 2020
October 6, 2020 - October 19, 2020
COVID-19 has forced many people to work remotely, including the city staff responsible for inspecting a community's infrastructure as well as private buildings. The City of Austin is experimenting with using AI to more efficiently inspect publicly visible infrastructure so that its staff members can perform more timely inspections of private property and buildings.
The AI uses data collected via lidar and other cameras on city vehicles while on duty. The AI machines are trained when real inspectors identify code violations in the images.
In normal times there is a large backlog of tasks for code inspectors. Property owners or developers can wait a long time to get an inspection. The City of Austin is interested in testing if we use AI to accelerate the process by carrying out virtual inspections or have virtual training.
Current training for code inspectors is with PDFs, PPTs, and on-site. It is expensive, not scalable, and has problems in terms of its repeatability.
The research team will have access to the following resources:
The city of Austin would like to achieve the following outcomes:
In the short term:
In the long term: