Automated Aerial Inspection of Wind Turbines

Wind turbines need to be regularly visually inspected, notably their blades, in order to detect as early as possible potential cracks and thus minimize downtime and repair costs. The goal of this project is to develop a vision-based navigation system for a multicopter, enabling automated inspection of wind turbines. The navigation system will be integrated with the flight control system and on-board sensors, and will provide relative positioning of the multicopter with respect to the structure to the path planner and trajectory tracker. A component will also be added to perform automated detection of cracks on the captured images. Finally, a simple reactive path planner will be developed to follow the structure based on the vision system outputs, and attempt to provide full coverage of the turbine. A parallel Engage grant application focuses more deeply on the trajectory optimization aspects of the problem. TO BE CONT’D

Faculty Supervisor:

Jerome Le Ny

Student:

Partner:

microdrones Canada Inc

Discipline:

Engineering

Sector:

Manufacturing

University:

École Polytechnique de Montréal

Program: