• contact:

    Matthias Balzer

  • funding:

    BMBF / “Joint Declaration on a Strategic Partnership in Education, Research and Innovation” between the Federal Republic of Germany and the Russian Federation

  • Partner:

    KIT-LAS, KIT-IPS, KIT-IEKP, KIT-IPE, Saint Petersburg State University of Civil Aviation / RUS, Shubnikov Crystallography Institute / RUS, Tomsk Polytechnic University / RUS

  • startdate:


  • enddate:


Progress in recent years concerning X-ray optics, detector technology, and the tremendous increase of processing speed of commodity computational architectures gives rise to a paradigm shift in synchrotron X-ray imaging. UFO (Ultra-fast X-ray Imaging of Scientific Processes with On-line Assessment and Data-driven Process Control) aims to enable a new type of smart experiment using the vast computational power of massively parallel computation units: On-line assessment of sample dynamics will enable active image-based control, allow an unprecedented image quality, and will provide insight into so far inaccessible scientific processes.

UFO – Ultra-fast X-ray Imaging of Scientific Processes with On-line Assessment and Data-driven Process Control

The UFO team at IPE has contributed to ultrafast and high-throughput X-ray tomography with:

  • Accelerated reconstruction with GPUs enables online reconstruction and image quality control; the data processing pipeline can be configured by high-level description in the UFO Parallel Processing Framework; continuous processing of data streams in the order of Gigabytes per second.
  • The novel fast control system “Concert” integrates beamline instrumentation and the GPU reconstruction pipeline and achieves sample throughput as high as 1000 tomograms per week of beamtime.
  • A versatile high-throughput FPGA-based modular smart camera platform has been developed. Key features are: continuous data streaming by PCIe link, the option to realize online data processing and a number of pluggable image sensors from 2 to 20 Megapixel and various frame rates.