About the project
This research focuses on the visual inspection of aircraft images in digital format. It involves identifying various types of defects and training these images using machine learning techniques. The goal is to develop a tool that assists maintenance engineers during checks by indicating what actions are needed next, such as assessing the criticality of a defect. In addition, the research will deliver a cost benefit analysis of the whole process, to evaluate the actual benefits from such digitalized and automated tool.
This research will be conducted within the Computational Engineering and Design Group, offering you the opportunity to work in a collaborative and stimulating environment, with access to state-of-the-art facilities, and necessary resources to conduct high-quality research. Furthermore, access to one of Europe鈥檚 first 鈥済reen鈥 hangars, for line and base maintenance services will be provided by .
The research, with the participation and interest from an industrial aircraft partner, aims to achieve the following key outcomes:
- record any detecting damage, as a function of damage type, damage size, and damage depth
- develop a software tool that allows for the accurate identification of defects during aircraft that has the ability to identify defects even with limited amount of data
- increase the capability to detect small type of defects, by investigating image enhancing technologies and implementing an automated process flow in order to achieve full Automated Defect Recognition (ADR)
- classify the detected defects, as well as categorise the types of the defects using different machine learning algorithms (built and tested)
- apply a Convolution Neural Network (CNN) in order to be trained with obtained data for the classification of the defects
- develop an advanced diagnostic defects detection algorithm of aircraft structures
- deliver a cost benefit analysis of the whole process, indicating the actual advantages from such developed approach
