Crack Detection in Concrete Bridges Using Machine Learning

Transport infrastructures are the backbone of our society and bridges are the bottleneck of the transport network. Climate change resulting in higher corrosion rates, and extreme climatic events are important threats to the reliability and safety of the transport network. This has led to growing concerns for the remaining service life of infrastructure under normal service conditions and the structural resilience under extreme climate events. We will develop a reliable crack detection tool that allows to detect and quantify the presence of crack in concrete infrastructure. The specific objectives are: (1) To annotate an existing 3D point cloud database of cracked and uncracked concrete surfaces, and (2) To train and validate a deep learning algorithm to detect cracks on concrete surfaces in real time. The project will lead to a safer, quicker and more accurate method for condition assessment of concrete infrastructure.

Alexander Mountain
Faculty Supervisor: 
Ali Mahdavi-Amiri
British Columbia
Partner University: