Nevliudov, I.Botsman, I.Chala, O.Khrustalev, K.2021-11-222021-11-222021Automated System Development for the Printed Circuit Boards Optical Inspection Using Machine Learning Methods / I. Nevliudov, I. Botsman, O. Chala, K. Khrustalev // INFORMATION SYSTEMS AND TECHNOLOGIES (IST'2021) : proceedings of the 10-th International Scientific and Technical Conference, September 13-19, Odessa, 2021. – Р. 234-238.978-617-7519-59-0https://openarchive.nure.ua/handle/document/18350The problem of printed circuit boards (PCB) quality optical inspection at their production stage is considered. The automated method of PCB optical inspection on the machine learning methods basis is proposed. The necessary neural network parameters to develop an automated PCB inspection method are calculated. The main capabilities of the created artificial neural network for identifying PCB under test defects are analyzed. The results of the conducted neural network testing that confirm its operability and possibility of use for PCB inspection at the stage of production are presented. The software program was developed that is used for transformations over images, such as converting an image to a grayscale color space and image binarization, which speeds up the neural network by reducing the size of the input matrix to a binary value per pixel of the image. The accuracy of finding each of the PCB defects types was also investigated.enautomated systemoptical inspectionprinted circuit boardsmachine learningintelligent manufacturingneural networksConference proceedings