Automated Damage Assessment

When a disaster occurs, it is important to know as soon as possible how many people are affected, where they live, and the severity of the damage, to properly plan response operations.

The Automated Damage Assessment tool (ADA) is a deep learning model that identifies damaged buildings in satellite images after a disaster has happened. Automating this assessment makes it faster and far less dependent on human effort, therefore enabling 510 to quickly deliver the information that is needed. The assessment time decreases from weeks to hours, which makes a difference in the speed of operations and, ultimately, on the number of lives that can be saved.