ITE 2040665 - PI Claudia Marin, Howard University
Computer vision (CV) enables the use of video cameras as non-contact type sensors (i.e., vision-based sensors) in structural testing to overcome the limitations of the conventional contact-type sensors (i.e., displacement sensors). The vision-based sensor system consists of a camera (or cameras) and a computer for applying algorithms to process the recorded images. The project focuses on validating the capabilities of a 2-D CV-based video-tracking program, developed by a research team led by Prof. Claudia Marin in Howard University, for displacement and deformation monitoring in structural testing. The program tracks the displacements at discrete points of interest through post-processing the recorded video frames using the Lucas-Kanade optical flow method.
Research Methodology and Outcomes:
The program was used to perform testing at the NHERI Lehigh Experimental Facility (EF) in conjunction with Professor Marin’s NSF Project OIA 2040665 – “NSF Convergence Accelerator Track D: Intelligent Surveillance Platform for Damage Detection and Localization of Civil Infrastructure.” The program was also incorporated into a project for a NHERI REU student hosted at the NHERI Lehigh EF. A series of tests were conducted to validate the capabilities of the program and study the effect of various parameters (i.e., the camera lens type, the camera resolution, etc) on the accuracy of the vision-based measurement feedback. Recommendations for structural testing procedures using vision-based measurement feedback and updates on the current program were established at the end of the testing program. Results from the testing program will transform the way structural displacement and deformation data is acquired and provide a readily available tool for laboratory and field testing.
The PIs have recruited Alia Amer, Ph.D., as a postdoctoral research associate to work on this project. She mentored a NHERI REU student during Summer 2023 to complete the structural testing and acquire the vision-based measurement feedback. The research team will host workshops and seminars to introduce the CV-based structural displacement and deformation monitoring to the research community. The educational seminars will be recorded and archived on the NHERI Lehigh EF website on DesignSafe