There are a couple of different methods out there to measure "the" SFR (Spatial Frequency Response) of a digital imaging system. The problem is, that you might get different result using different methods.
And different does not mean that one is right and others are wrong, it just means that you measure the different behavior of the camera based on different structures.
For modern signal processors as used in digital cameras, it is relatively easy to maintain an edge while reducing noise on a flat field. So the system will behave differently if measuring on an edge or on other structures.
To get a more detailed insight into the differences based on the different methods, we made a test.
If we want to test the method, we have to make sure that all other factors that could influence a SFR measurement are constant. So with the same camera, same illumination, same object distance, same setting and same operator, we tested different cameras. Additionally, we made some tests on artificial created and modified data. So we wrote a simple image processing pipeline, so we can make sure that we know if there is image enhancement applied or not.
The detailed description is shown here »
ua