ALLIED Vision Technologies Pike F-1600 - Manual (Page 198)

Description of the data path
PIKE Technical Manual V5.1.2
198
Color correction
Why color correction?
The spectral response of a CCD is different of those of an output device or the
human eye. This is the reason for the fact that perfect color reproduction is not
possible. In each PIKE camera there is a factory setting for the color correction
coefficients, see Chapter
GretagMacbeth ColorChecker on page 198.
Color correction is needed to eliminate the overlap in the color channels. This
overlap is caused by the fact that:
• Blue light: is seen by the red and green pixels on the CCD
• Red light: is seen by the blue and green pixels on the CCD
• Green light: is seen by the red and blue pixels on the CCD
The color correction matrix subtracts out this overlap.
Color correction in AVT cameras
In AVT cameras the color correction is realized as an additional step in the pro-
cess from the sensor data to color output.
Color correction is used to harmonize colors for the human eye. With other AVT
(color) cameras so far, you had the opportunity to use it or to switch it off.
Pike cameras introduce for the first time the so-called color correction matrix.
This means: you are now able to manipulate the color-correction coefficients
yourself.
Color correction: formula
Before converting to the YUV format, color correction on all color models is car-
ried out after BAYER demosaicing via a matrix as follows:
GretagMacbeth ColorChecker
Sensor-specific coefficients C
xy
are scientifically generated to ensure that Gre-
tagMacbeth™ ColorChecker®-colors are displayed with highest color fidelity
and color balance.
These coefficients are stored in user set 0 and can not be overwritten (factory
setting).
Formula 4: Color correction
red* Crr red

Cgr green

Cbr blue
 + + =
green* Crg red

Cgg green

Cbg blue
 + + =
blue* Crb red

Cgb green

Cbb blue
 + + =