COMPUTER VISION ALGORITHMS FOR THE IDENTIFICATION OF DAMAGES ON FULL-SCALE AIRCRAFT COMPONENTS

The objective of this study is to present a novel approach for airplane inspection to identify skin deterioration on the fuselage. Algorithms for computer vision were used as an instrument for automating the process of inspection and detection, decreasing human error, and increasing productivity and security. An overview of the problems, methods, and recent developments in the field of computer vision algorithms used for general damage detection on aircraft components is provided in this research work. Data were collected using a highquality acquisition system. The data set was created by gathering photographs to highlight different sorts of defects and have the greatest possible variety of instances, images collected on two separate aeronautical demonstrations: a commercial partial full-scale aircraft fuselage section in primer paint and a general aviation aircraft fuselage white painted. In particular, 964 images and more than 6000 regions of interest were manually annotated. Datasets that accurately represent various types of damages can be limited, making it difficult to train accurate and reliable models. The Convolutional Neural Networks and machine learning models were trained on large datasets of annotated images, enabling them to learn complex patterns and features associated with different types of damage. Data augmentation techniques were adopted to add diversity to the training data. Transfer learning techniques, which leverage pre-trained models on large-scale image datasets, have also proved to be effective in achieving accurate and robust detection results