Validação de peças com furos: implementação de um sistema baseado em visão computacional

This study presents the development of an automated computer vision system for inspecting 3D-printed parts with circular holes, addressing challenges related to manual inspection and the complexity of the parts. Using advanced signal and image processing techniques such as segmentation, filtering, edge detection, and the Hough Transform, the system aims to identify manufacturing defects through an objective comparison between images of inspected parts and reference standard images. This innovative approach enhances the accuracy and efficiency of inspection processes, reducing production costs and time, and contributing to industrial automation and the optimization of additive manufacturing. With functions developed to verify the quantity, position, and dimensions of holes, the system proved effective in positional and dimensional analysis but showed errors in identifying the number of holes, attributed to the Hough Transform, which generated false positives on square geometric shapes, highlighting the need to improve algorithms to better differentiate non-circular shapes.