Review of Research on 3D Vision-Based Robotic System for Unstructured Bin-Picking

In the manufacturing process, the Vision-Based Robotic System for Unstructured Bin-Picking can be used for parts sorting. Compared with two-dimensional (2D) vision technology, the 3D Vision-Based Robotic System for Unstructured Bin-Picking (3DV-BRUPS) has outstanding technical advantages in disorderly sorting scenes. Considering that existing reviews concerning 3DV-BRUPS are far from comprehensive, this paper adopts a narrative review research paradigm. It first analyzes the inherent characteristics of 3D vision technology. Second, it systematically sorts out the system configuration and core technologies of 3D-vision-based robot sorting platforms. Meanwhile, it thoroughly discusses the current challenges as well as the strengths and weaknesses of emerging artificial intelligence techniques for target recognition and pose estimation. Finally, it investigates state-of-the-art domestic and overseas applications of 3DV-BRUPS in logistics parcel sorting and scattered-part picking scenarios, and explores its future research directions, so as to provide references for further in-depth research and large-scale practical deployment of this technology.