The influence of gamma-correction on the total accuracy
Gamma correction originates from the imperfections of television
tubes. The intensity is not a linear function of the input signal.
Therefor the camera signal is corrected for the non linearity of
the sensor and for the receiving end (the tube). This is done by
raising the signal to the power 0.5. This is the official CCIR
value. It can vary from 0.45 to 0.7 however.
The (cheap) observation cameras do not have the option to switch
this correction off. Usually the frame-grabber does not perform the inverse
correction. For the picture of the CCD_Rasnik mask this means that
the white areas become bigger and the black areas smaller.
On average the effect is reduced to a large extend. But on the
edges of the picture (in X and Y direction) there is a problem.
When the picture moves, edges appear and disapear. Sometimes the
effects cancels out, sometimes not. Especially when code-lines
appear or disappear the effect will notable. For X and Y the effect is small.
However, for the Z calculations the outer contours are very important.
Therefor the Z coordinate can change when scanning over X and/ or Y.
To observe the effect an actual grabber-file is used in a Spice simulation.
The video signal is bandwidth limited and differentiated
(the same is done in the reconstruction software).
The same signal is gamma corrected and then fed through the same circuit.
One can clearly see the shifts of the output signals.
You can have a look at:
A test set-up is needed to find out how big the problem actually is...
go to
CCD_Rasnik home page.
You can mail
suggestions for (or questions about) this page to:
Henk Groenstege.
october 1996
