An Approach of Automated Assembly Evaluation Using AI - Based Computer Vision Methods for Human – Robot Collaboration

Artificial Intelligence (AI) and Machine Vision (MV) methods are increasingly applied in manufacturing systems as this enables automated processes that lead to both cost and time efficient solutions, minimizing errors and maintaining the quality of the products in high level. This paper discusses an AI – based vision system that is responsible for the evaluation of an assembly performed in the context of a hybrid manufacturing system. Different object detection models are evaluated for the detection of the parts, utilizing both real-life and synthetic datasets, while a fully integrated implementation using ROS is presented. Finally, the solution is applied and evaluated in a Human – Robot Collaboration (HRC) based assembly scenario.