Deep learning-based steel corrosion and fatigue crack detection using a depth camera

In order to overcome the drawbacks of the traditional approaches to steel structure damage detection and to improve reliability and safety of existing steel structures, this research proposes to develop an innovative autonomous SHM system that offers transformative enhancements over existing steel structure monitoring systems.
This will involve:
1. Collecting
Collecting image data from various steel corrosion and cracks to establish databank to train a deep faster region-based convolutional neural network.
2. Developing Algorithm
Developing a deep faster R-CNN-based steel damage detection system using a 3-dimensional depth camera that is able to provide first warnings of structural damage by localizing and quantifying damage.
3. Application
Developing an automated system to apply this research o real-world problems.
We will install a web camera on a steel bridge to providing monitoring in real-time using a wireless system that will transfer the image data to a base workstation.

Faculty Supervisor:

Young-jin Cha

Student:

Partner:

Korea University

Discipline:

Engineering

Sector:

Construction; Technology; Sustainability & the Environment

University:

University of Manitoba

Program: