This research project focuses on the visual inspection of aeronautical mechanical parts by relying on a robotic system equipped with cameras which provide 2D images and 3D point clouds. The objective is to verify that a complex aeronautical assembly (typically an airplane engine) is well assembled and that all the individual parts (brackets, screws, cables, etc.) are present in their respective pre-specified locations. Certain anomalies can be detected by analyzing 2D images (ex. absence of a bracket or a line of screws), while the others require 3D point cloud analysis (ex. verification that 2 electrical cables or tubes are at a safe distance to each other according to the security standards). Both mentioned inspection modes (2D or 3D camera) can be facilitated if a 3D CAD model of the inspected element is at disposal.
The objective of the PhD is to develop an inspection method based on the combined utilization of the 3D CAD model, 2D images analysis, 3D point cloud analysis and Machine Learning techniques (on 2D images and 3D point clouds).
A detailed description of the subject can be downloaded here:
(c) GdR 720 ISIS - CNRS - 2011-2019.