Advanced Diagnostics of Aircraft Structures Using Automated Non-Invasive Imaging Techniques

There is an innovative need for more efficient and reliable damage inspection, reducing the time and cost of maintenance without compromising the safety of passengers and goods transported. The ultimate aim of the research is to develop an automated, vision-based damage evaluation system capable of detecting and characterising defects in metallic and composite aircraft specimens, using RGB and infrared thermography (IRT) data. The system will generate 3D models through photogrammetry from RGB images and integrate them with IRT data for comprehensive analysis of surface and subsurface defects. The respective scientific results of the research are:

1. To develop a dataset comprising at least 1000 real and 500 synthetic images derived from the 3D models.
2. To train and develop a machine learning model capable of detecting and classifying common defects for each material, achieving a minimum precision of 80%, a recall of 70%, and an F1 score of at least 0.75.
3. To address the challenges an unbalanced dataset poses and enhance the model's performance by at least 5%.
4. To develop a defect measurement tool capable of estimating the size and depth of detected defects with an error rate of less than 10%.