Implementación de un prototipo con sistema de visión artificial para el procesamiento de imágenes que permita mejorar la productividad en el proceso de corte por plasma

Abstract

The objective of this study was the implementation of a prototype with an artificial vision system for image processing to improve the quality parameters in plasma cutting articles; in this way, it seeks to provide the metallurgical sector with a low-cost tool that provides an effective and fast solution to determine the quality of the product. For this purpose, a study of the inspection time and the analysis of the operator's activities in the current method was carried out utilizing the elaboration of a process diagram. With this study it was possible to analyze the productivity in the process of recognition of defects and the importance of eliminating manual activities, so a prototype is proposed to identify the physical defects of the parts quickly and efficiently, it consisted of a software, which was developed in the Python program, allowed the analysis and comparison between a standard image of a part without defects and another part to be analyzed, and additionally generates a report of the analysis of defects, thus eliminating one of the operations that took the operator more time at the time of inspection. In addition, the observation bench provides the appropriate environment for the identification of defects in the parts. Upon completion of the prototype, a time study was carried out to validate its effectiveness, preparing a time analysis of the proposed inspection process, resulting in a reduction in inspection time of 19.3 seconds. This time reduction is reflected in the increase of productivity, going from analyzing 2 pieces/minute to 5 pieces/minute.