Surface and subsurface damage segmentation using thermography and deep learning

Infrastructure facilities in Canada and many other places worldwide continuously deteriorate. Most structures either reached or exceeded their design service life. Bridges, buildings, roads, and other facilities deteriorate over time. To ensure the safety of these structures, visual inspections are routinely carried out by trained engineers. However, these visual inspections have some critical disadvantages, such as risks to the inspector, and visual inspections are time-consuming and erroneous. This study proposes an artificial intelligence-based method for surface and subsurface damage detection in structures using thermography to solve this issue. The multispectral image data will be used as input to the network to identify structural damage. The method will detect damage with high accuracy at the pixel level in the multispectral dynamic image.

Faculty Supervisor:

Young-jin Cha

Student:

Partner:

North Forge

Discipline:

Engineering

Sector:

Education; Management of companies and enterprises; Professional, scientific and technical services

University:

University of Manitoba

Program: