AI-based anomaly detection for defect detection during surface inspection of aerospace components
During aircraft maintenance, components are visually inspected for defects. There is no photographic documentation of the defects. In future, components are to be inspected using sensor technology in order to increase the quality and digitisation of the diagnostic data. A promising approach for defect classification in image recordings are neural networks, which, however, require a large amount of training data of the defect types for this purpose. An alternative approach is anomaly detection, which is able to detect any defects in the image as anomalies on the basis of error-free images. The aim of this work is to develop and test different anomaly detection approaches. For this purpose, a database with defect-free images will first be created, with the help of which different anomaly detection approaches will be tested on the basis of defect images. The different approaches will be compared with each other and evaluated with regard to their suitability, reliability and potential for improvement.
Your subtasks
- Research on texture data sets, anomaly detection
- Development of a training database, design and testing of anomaly detection approaches for surface defects
- Comparison of the approaches and investigation with regard to reliability, robustness, optimisation potentials
Your profile
- You are studying mechanical engineering, mechatronics or a comparable subject
- You have knowledge, skills or an interest in digitalisation, AI
- You have programming skills (Python, Matlab)
If you are interested, please contact us with your CV and overview of grades: Falko Kähler, M.Sc. | f.kaehler@tuhh.de
