Robot-guided multi-view point cloud fusion for accurate 3D reconstruction

In biomedical applications such as surgical navigation, conventional solutions like optical tracking and mechanical positioning often rely on bulky rack-fixed hardware, limiting adaptability in dynamic and space-constrained environments. To address these challenges, we present a compact multi-view 3D reconstruction system that integrates a stereo camera with a six-axis robot, eliminating dependence on external tracking devices. Within a hand–eye calibration framework, the system performs multi-view scanning, image segmentation of the target object, and point cloud fusion, producing dense, consistent, and accurate 3D reconstructions in real time. With its compact form factor and multi-view capability, the system offers a portable and practical solution for intraoperative imaging and broader biomedical guidance tasks, enhancing both accuracy and usability in constrained clinical settings.