Exploring the potential and limitations of Artificial Intelligence based Structural Health Inspections
Project Description
Advances in AI have unlocked new capabilities towards structural health monitoring that are expected to drive the next revolution in information modeling and decision making in the built environment. This project will provide the opportunity to learn and explore AI for structural damage (such as cracking and corrosion) detection. In this UROP project, we will have hands-on experience over the whole AI model development process, starting from infrastructure damage data collection, processing, to AI model training and evaluation. We will also discuss about the limitations of AI-based approaches (i.e., can AI beat human for structural damage detection?)
Supervisor
ZHANG, Jize
Quota
4
Course type
UROP1100
UROP2100
UROP3100
UROP3200
UROP4100
Applicant's Roles
1. Hand-on experience of collecting, labelling, and processing real-world infrastructure crack data using cameras;
2. Thorough exploration of using advanced AI methods to build autonomous, vision-based infrastructure damage detection systems.
2. Thorough exploration of using advanced AI methods to build autonomous, vision-based infrastructure damage detection systems.
Applicant's Learning Objectives
1. Learn the fundamentals concepts and experience hands-on applications of machine learning to smart infrastructure systems;
2. Understand the domain knowledge of AI-based structural damage detection, including data labeling, AI segmentation algorithms, and AI performance evaluations.
2. Understand the domain knowledge of AI-based structural damage detection, including data labeling, AI segmentation algorithms, and AI performance evaluations.
Complexity of the project
Moderate