Machine Vision-Based Damage Assessment

Dr. Ioannis Brilakis, assistant professor in the School of Civil and Environmental Engineering, and his team of researchers recently returned from Port-Au-Prince, Haiti. The trip was the result of an NSF-funded RAPID proposal that included principal investegators (PIs) Laura Lowes (UW), Ioannis Brilakis (GT), Reginald DesRoches (GT), Marc Eberhard (UW), and Gustavo Parra-Montesinos (UM). Dr. Brilakis spent six days with researchers and structural engineers in Port-Au-Prince collecting data in various damaged reinforced concrete buildings. His team included Stephanie German, Georgia Tech student; Gustavo Parra-Montesinos, associate professor, University of Michigan; Kurt Swensson, consulting engineer with KSi; Dave Swanson, consulting engineer with Reid Middleton; Jonathan Weigand, UW PhD student; and Joshua Pugh, UW PhD student. The data they collected was used to validate the ongoing research of the PIs which will be used to address current limitations of post-earthquake assessment procedures by coupling the procedures with a quick, machine vision-based quantitative assessment through video data.

Dr. Brilakis' team surveyed several buildings, obtaining manual and video data of the structural damage at each location. Specifically, the team assessed the CDTI Hospital, Union Elementary School and Learning Center, and the Digicel Building. The video data consisted of a detailed walk-through of the damaged portions of the buildings, and the team slo collected manual data similar to that obtained during an ATC 20-type assessment procedure. Now that the team has returned to the U.S., they will use the video data with the manual data to perform and validate the objectives of the research. The team has previously achieved the automatic detection of reinforcement concrete columns (shown in figure below), and is currently working on automating the detection of exposed reinforcement in concrete columns.