Computed Tomography Explained

Computed Tomography (CT) is increasingly being used for non-destructive inspection and dimensional measurement of manufactured prototype and production parts. Besides the determination of measurement uncertainty, the recent advancements in technology, both hardware and software, has enabled this technology to become common place in universities, research centers and has started to penetrate front line inspection of production parts. These advancements, coupled with a lowering of the average price of a CT work-station, has brought this interesting technology within the reach of industrial applications and can now start to be considered for some applications as an alternative to traditional Coordinate Measuring Machines (CMM) and non-destructive measurement devices.

Porosity analysis show position, distribution and size of voids. – Source Jesse Garrant Metrology Center

Dimensional measurement with CT has the advantage, in comparison with traditional measuring techniques, that it is not just limited to the measuring of features on the work-piece’s surface. With CT it’s possible to evaluate the complete volumetric model of the part with high point densities comprising millions of points containing detailed information of internal part features not accessible in any other method of measurement other than destructive testing. CT can be used for variety of manufactured components from very small electronic parts to large automotive castings. The application of CT to dimensional metrology is a recent advancement; CT had been traditionally used for checking for flaws (cracks, porosity etc) in parts and verification of assembly operations particularly in the electronics sector. Improvements in image quality and resolution now make dimensional metrology a reality.

CT for dimensional inspection usually uses the cone beam method of computed tomography whereby the part to be inspected is rotated automatically in the radiation field of the X-ray source. By rotating the part through its full circumference a large number of beam directions are created, which after passing through the part, are collected using a flat panel detector and stored digitally as 2D projections.   The volumetric data required to reconstruct a part 3D image can be acquired much faster using the cone beam method over the alternate fan beam method. After storing all of the projection data the 3D image construction can commence and the volumetric data evaluated. The part is static during each 2D imaging.

Typical cone beam x-ray CT setup consisting of a source and detector with a rotary stage for the object. Images are collected through a full 360 degrees rotation of the object that are then reconstructed to a 3D model. – Source Creative Commons

Computed Tomography (CT) for metrology can be simply defined as the process of imaging a part from numerous directions using X-rays and subsequent processing of the acquired projected 2D images to create a 3D model defining both external surfaces and internal structure. The extraction of the features and elements contained in the acquired 3D model using specialized software allows the part geometry to be fully defined and analyzed against nominal CAD data if available.

The cone-beam CT method creates two physical effects which impact both the quality and accuracy of the acquired data-set namely Beam Hardening and Scattering. The CT reconstruction process is based on the assumption that an attenuation at a certain point is independent from the path by which the X-rays have reached that point – however this is only for a beam consisting of mono-energetic radiation. With the poly-energetic X-ray beams used for metrology the sources spectrum becomes harder as the beam propagates through the object because lower energy photons are attenuated more strongly. This effect is called beam hardening and if not corrected (compensated) gives rise to artefacts (i.e. observable errors in the reconstructed volume).

Dark stripes and edge super elevations (so called cupping effects) lead to problems in determining the real work piece surface; consequences of which are higher errors of measurement. In the energy range used in CT scattering is the dominant interaction for attenuation. Forward scattered photons hit the detector, but do not contribute useful information and reduce the contrast. X-ray scatter is also energy dependent and leads to artefacts similar to beam hardening.

Steps in the CT Process

1) Image Acquisition

The part is placed on the rotary index device and rotated in typically 0.5o increments through a full 360o. A high resolution digital radiograph image is acquired at each position.  The accuracy of this data set determines the ultimate quality of the final 3D data.

2) Correction

Each projected image is processed to perform both geometrical and shading correction removing spatial and intensity non-linearity’s introduced by the data collection process.

3) Reconstruction

Combining all of the corrected images allows a geometrically correct 3D data cloud is to be created.

4) Results

The resulting 3D model displaying both surface and internal part geometry is displayed and allows manipulation of the data set.

5) Post Processing of Data

The data cloud can be processed as-is or output as a stereo lithography file (stl), a format accepted by most CAD packages. Once imported into a CAD system the radiographic information can be compared directly with the original design file to highlight differences. The data can be used for rapid prototyping and reverse engineering. In the metrology application the CAD file is imported into metrology software that has been engineered to extract geometry from the model and either generates dimensional data for the part or compares the scanned parts geometry against nominal CAD model data.

CT Scan with CAD overlay – Source Jesse Garrant Metrology Center

Images courtesy of Jesse Garrant Metrology Center:



One thought on “Computed Tomography Explained

Comments are closed.