Sistema de clasificación autónoma de objetos mediante visión artificial y un brazo robot industrial

Abstract

Abstract: This thesis develops an automated system for classifying and manipulating objects using computer vision, artificial intelligence, and robotic control. The proposed solution employs a detection model based on YOLOv8 to identify and classify objects in real time, using a previously annotated and optimized dataset. From the detection point, coordinates are converted from the image plane to the reference frame of the work environment, allowing for the generation of precise trajectories for pick-and-place tasks. The robotic manipulator's control is implemented in an industrial simulation environment, ensuring safe, repeatable, and efficient movements. The results obtained demonstrate that the system is capable of improving the automation of the classification process, reducing errors, and increasing operational efficiency. This research highlights the potential of artificial intelligence applied to industrial robotics as a viable alternative for automated manufacturing and recycling processes.