User:Josef.uher/sandbox

Robotic Computed Tomography is an approach to Computed Tomography that uses for repositioning of the X-ray tube and detector robots.

Robotic X-ray imaging and CT scanner RadalyX. It uses a pair of collaborative robots

Design

Computed Tomography is a method of sample volumetric inspection using X-ray imaging. A CT scanner measures a set of X-ray images each from a different viewing angle. These images (projections) are then combined into a 3D data set that represents the internal structure of the object. The robotic CT uses for the repositioning of X-ray tube / R-ray detector a pair robotic arms. The robots are moved synchronously maintaining a precise mutual position of the X-ray tube and detector.

Benefits robots for Computed Tomography

Pair of robots X-ray inspecting a landing gear mounts of a small aircraft. The robot carrying an X-ray tube was installed inside the aircraft. The robot that positions the X-ray detector is outside of the fuselage.

The limitation of common medical CT and industrial CT scanners is that the sample has to "fit" inside the scanner. I.e. the scanners do not allow scanning only a region-of-interest (ROI) of a larger object. That limits applicability of the CT. Contrary to that, the robotic arms can be installed next to the object to be inspected or even inside it. Simply said, the robotic CT can be brought to the sample and not the sample to the CT machine. Robots then produce X-ray images of the ROI from a variety of angles which are then combined to the 3D volumetric data. The flexibility of robot movement allows also measurements such as tomosynthesis which is more suitable for large planar objects (e.g. aeroplane wing) that a rotational scan path around the object is not possible.

Arbitrary path CT

A typical CT scanner rotates the X-ray tube / detector pair only around a single axis. The flexibility of robots on the other allows nearly any direction of X-ray view (projection). That can be beneficial as some directions of view can proved more information that other; consider a cylindric object: many projections that are perpendicular to the object axis are required to determine that the object is cylindric. However, only one is needed when looking along the axis. The robots bring the imaging closer to approaches we humans do if we are looking at a new object. We are also looking from a variety of very different angles rather than rotating the object around single axis only.

Adaptive scanning

The X-ray tube / detector distance does not have to be maintained constant. It can vary depending on the shape of scanned object or it can zoom in using the geometrical X-ray magnification to image object details as required. The outer object shape information can be used to precisely navigate the robots along the object. The surface shape is obtained either from CAD model or from an optical 3D scan.

Multimodal imaging

X-ray imaging and CT is in many cases not capable of detecting all flaws or features of inspected objects. It cannot detect for instance crystalline structure of the object, defects like delaminations and disbonding in composite materials or even colour of the object. Therefore, it is beneficial to combine the X-ray image and CT data with other imaging modalities. It is common for robotic arms to change tools as need. The same approach is used for the multimodal imaging making possible the combined scans fully automatic.

Combined scan of a drone propeller. It was measured using reflectometry to get image of the surface, laser profiling for precise outer shape measurement, CT to get 3D volumetric information and with regular 2D X-ray imaging.

Limitations

The limitations of the robotic CT comes mainly from the reach of robots. A typical CT measures projections from 360 degrees around the object. This is not often possible with a single pair for robots. Two pairs of robots might be used to overcome this problem. However, that increases the complexity and cost of the system.