Intelligent Endoscopic Examination of Internal Openings for Drilling Quality Control

The paper presents an intelligent computer vision complex for drilling quality control using an endoscopic video camera to examine the internal openings and artificial neural networks to identify holes and their defects. The concept is based on a quality guarantor paradigm, which provides statistical analysis of the trajectories of repeatable actions that characterize manufacturing process. The hardware part of the complex is built on a 5-axis machine, the working area of which allows access to the surfaces of the controlled part. As part of this solution, two new algorithms were developed: an algorithm for precise positioning of the endoscope to collect the necessary initial data, and a computer vision algorithm for detecting defects. For the second algorithm, an array of initial data was specially collected, which made it possible to train and subsequently use the neural network to monitor the inner surface of the holes. The resulting complex is designated to be used in manufacturing as a part of a quality control decision-making support system. #COMESYSO1120.