AI-Based Structural Feature Recognition for Robotic Grasping of Additive Manufacturing Parts

Additive manufacturing factories have emerged in recent years to provide printing services to a wide range of individuals and businesses. Although 3D printing itself is primarily automatic, many other operations in such production factories are still manual, hindering the overall operational efficiency. Therefore, this article aims to improve the automation level of additive manufacturing factories by developing a mobile robot system that integrates a vision system and a two-finger flexible manipulator to automatically complete the model removal work after 3D printing. The system uses a camera to scan and identify the best grasping point of the print through an improved multi-layer convolutional neural network (CNN), while using image segmentation technology to accurately locate the target grasping position. The optimal grasping angle is determined through multi-view analysis to achieve efficient and precise grasping operations. The performance of the system was verified through CNN training on a self-built data set and successful crawling on unknown models, demonstrating its potential in intelligent operation and maintenance of 3D printing equipment.