Slide 15 of 17
Notes:
It should be said that there is no best way to measure spatial resolution, it depends on the type and frequency of your quality control checks.
Each of the methods has its pros and cons. The edge suffers from poorer repeatability than the wire, but the wire method can only be used where the pixel size is less than the Nyquist frequency. As spatial resolutions increase this may become a problem (although manufacturers are moving to 1024 x 1024 reconstruction matrices).
The standard deviation methods are the least computationally complex, and the ‘easiest’ of the assessments to perform since they do not require direct access to pixel data. A region of interest can very easily be drawn over a bar pattern using manufacturers software.
Of the visual methods, the CATPHAN (or any method with a large number of discrete frequencies) is preferable since it allows a more accurate estimate of where the cut-off frequency lies. But any method is only as good as it’s application and thus whichever method is to be used the consequences of changing the alignment, positioning, field of view size etc…should be understood.