Automated Internal Inspection of Wind Turbine Blades Using Computer Vision and Machine Learning Techniques
External reference persons Prof Alessandro Sabato, University of Massachusetts Lowell
Thesis type RESEARCH / EXPERIMENTAL
Description Computer Vision technologies capable of recreating point clouds are becoming more important in the field of structural health monitoring. The possibility of using mobile systems such as drones or robots capable of reaching points otherwise unreachable has increased the scientific community's interest in these systems.
The aim of the thesis is the definition of a methodology for the creation of a system to inspect the inside of wind turbine blades, recreate reconstructions of the internal surface of the blades, and identify changes in the structure over time.
Specifically, starting from data acquisitions carried out with drones with RGB and infrared cameras and LiDAR scanners, the candidate will have to define a procedure to reconstruct a point cloud of the geometry of the inside of wind turbine blades and use Machine Learning techniques to automate the recognition of defects such as splits, delaminations and cleavages.
The work involves an evaluation of the system in the laboratory (with predefined and known boundary conditions) and the possibility of testing the developed system in situ on wind turbines with a length of up to 75 metres.
Required skills MATLAB, possibly prior knowledge of Machine Learning algorithms
Notes Average grade required: >= 27/30 and knowledge of the English language at least B2/C1 level
Deadline 23/11/2024 PROPONI LA TUA CANDIDATURA