Speaker
Description
Synchrotron X-ray Micro-CT is an invaluable tool for non-destructive 3D morphological visualization of different kinds of small samples in high resolution. To extend this capability to large-scale comparative studies, or the systematic digitization of museum specimens, a radical shift in imaging workflow is required. We present a comprehensive, highly automated pipeline for “serial tomography” that supports routine high-throughput 3D imaging of large sample collections. The approach consolidates robotic sample preparation, self-directing scanning procedures, GPU-accelerated online 3D reconstruction, and AI-assisted quality assurance protocols within a coordinated workflow. Efficient processing is achieved through a decentralized DAQ pipeline, while metadata are centrally managed from sample preparation onward to maintain traceability and interoperability across all stages. The system consists of a UR20 robotic arm equipped with a suction cup gripper and placed on a motorized gantry. The sample tray can facilitate up to 1000 vials of volume 0.2 ml. The robotic arm contains a datamatrix code reader for sample identification. Current sample throughput in a high-quality setting is 15 samples/hour with pixel size 1.3 micron. The developed workflow offers the scientific community access to high-quality 3D data and supports comparative research addressing principal questions across life sciences and material research.
