Automated Change Detection in Geospatial Imagery

Partners: U.S. Census Bureau - Geography Division
Project Status: Ongoing
An isometric grid of cubes is arranged to look like a city block. One of the cubes is taller and a different color than the rest.

Are there emerging technologies and methods that might serve us as we work to ensure we have timely and complete collection of residential addresses from which we create critical data? We’re employing machine learning models which correlate changes in housing unit data with satellite imagery to improve automation around change detection and bias reduction. If successful, this project will reduce algorithmic bias and bias caused by human error in current workflows, to improve operational efficiencies and representation in surveys.

We’re Hiring

Join us in discovering new and exciting ways to make a difference.

Apply Now