Development of an information system for organizing the work of a production workshop

The article discusses the development of an information system for organizing the work of a production workshop, capable of automatically assessing the quality of products, recognizing defects on products using computer vision and machine learning methods. To select the optimal computer vision algorithm, a comparative analysis of computer vision methods was carried out, to select the optimal machine learning model, a comparative analysis of neural network models of pattern recognition was carried out. As a result of the comparative analysis, the GrabCut computer vision algorithm and the ResUNet neural network model were selected. The parameters of the GrabCut algorithm were chosen in such a way that it segmented the products on the conveyor belt. The neural network model was trained to recognize defects on these products.