Industrial X-ray CT scanning or Computed Tomography

An illustration of the complete process

X-ray tomography or CT scanning is a form of non-destructive 3D scanning. It allows you to capture the complete outside surface, but also the internal structure of an object without the need to destroy it.

We use this technique a lot in our metrology services for the quality inspection of plastics. You can also have a look at this case of a 2-pin connector. It’s another good example of a typical CT scan and the results that you can get with it.

Since most of the other samples that we scan for our clients are under an NDA agreement, we picked a toy car for children built out of different pieces of plastic to show the complete process. With the scan data of this object we can illustrate some of the principles and dive a bit deeper in the different results of industrial CT scanning.

Have a look at our case study CT scan of a plastic toy car to discover some more results of this scan.

The toy car consists out of different plastic blocks that click together. This means that it is a fairly complex assembly and there is a lot that can be discovered with CT scanning.

Typical inspections include:

Deformations of the free state components vs. the assembled versions
Dimensions of both the individual elements as well as the complete assembly
Porosity or inclusions in the material
Fittings and gaps (or sealing) between different components

A CT scan also provides a very complete and highly detailed watertight data set.
With a more traditional optical scanning system it would be impossible to capture the complete shape due to line-of-sight restrictions. This means that you will end up with an incomplete mesh with holes and it will take same time to make it watertight for further processing with additive manufacturing for example.

The X-ray CT scanning process

The complete X-ray CT scanning process consists out of 3 different steps.

Collection of the X-ray images
3D CT volume reconstruction
Polygonisation or extraction of the surface data

Collection of the X-ray images

X-ray imaging is a well known technique from the medical world. When you have a broken arm, you go to the hospital and there they take an X-ray image to see where and how your bone is broken. The result is a typical black and white image where the more dense materials (your bone structure) appear in a white color and the surrounding air for example appear as a black color.
In fact it is not the density which plays a role here, but the attenuation coefficient of a material. This tells you how much the X-ray beam intensity will be reduced when it travels through a certain material. It’s closely related to the density of that material however.
There are also more X-rays absorbed when you have more material versus less material of the same attenuation. This also means that this spot will be whiter on the image.

In a CT scanner you have a X-ray source, a rotation table and a detector.

The X-ray source produces a cone beam of X-rays which fall onto the detector.
In between these two components you have a rotation table.
The object that needs to be scanned is placed on this rotation table.

When you start a scan, the system will take a first X-ray image and then rotate the object a little bit using the rotation table before taking the next image.
Usually there are somewhere between 1.000 and 3.000 X-ray images taken over a full rotation of 360 degrees.

3D CT volume reconstruction

The drawback of an X-ray image is that the geometry of the complete object is mapped on this 2D image. Features that are on the back of the object will overlap with features that are on the front. This makes it sometimes difficult to interpret a single X-ray image and therefor it’s also not very useful for metrology purposes.

When reconstructing a 3D CT volume the computer will go from a bundle of flat 2D images to one complete 3D volume.

This 3D volume is like a big box filled with tiny cubes. Each cube is called a voxel and has a certain size (the voxel size or resolution of the CT scan). Each voxel has a certain grayscale color that corresponds to the attenuation coefficient of that specific piece of material.

Whatever is in the 3D volume can be rendered in 3D, but you can also slice the volume virtually and have a look at a 2D representation of this slice.

Polygonisation or extraction of the surface data

In the last step of the process we go from a 3D volume of voxels to a surface of triangles or a mesh.

To do this the software looks for the different materials in a CT scan and calculates the border between them. In the image below you can see a black (air) and a white (plastic) region. There is also a certain gradient when going from black to white. Wherever this gradient is the steepest, that’s the spot where the surface lies. This is visualized with a red line in the screenshot. This red line is calculated at a sub-voxel level, so accuracy wise you are not bound to the resolution of a scan.

Polygonisation or extraction of the surface data from a CT scan

The result is a watertight, highly detailed, 3D surface mesh.

Industrial X-ray CT scanning or Computed Tomography

The X-ray CT scanning process

Collection of the X-ray images

3D CT volume reconstruction

Polygonisation or extraction of the surface data

TetraVision bvba

Social

Legal