Falcon 4M Camera Manual
03-032-20044-03
Teledyne DALSA
48
H ow can one m atch gain and offset valu es on m u ltip le cam eras?
One w ay is of cou rse to u se flat field correction. All cam eras w ou ld be set u p u nd er the
sam e cond itions inclu d ing lighting and then calibrated w ith
ccf
and
cpa
. This can be
tim e-consu m ing and com p licated (esp ecially the w hite target). Another w ay is to u se
analog offset and system gain (d igital gain):
1)
Starting from factory settings (
sao 0 0, ssg 0 4096, epc 1 1
), take note w hat
the highest d ark offset is am ong the set of cam eras. If the highest d ark offset is h igher
than abou t 16 DN (10 bit) you m ight w ant to consid er recalibrating the FPN
correction (
ccf
). Large d ifferences in d ark offset betw een the factory and u ser are
typ ically cau sed by d ifferences in tem p eratu re from factory to u ser. Large d ark offsets
w ill resu lt in PRN U-correction -ind u ced FPN and shou ld therefore be avoid ed .
2)
Increase the offset (cam era in d ark) on all cam eras (
sao
com m and ) u ntil they are the
sam e and reach at least 4 DN (10 bit).
3)
Illu m inate to abou t 80 % satu ration (820 DN , 10 bit) and n ote the highest signal level
am ong the set of cam eras.
4)
Increase the d igital gain (
ssg
) on the cam eras u ntil they all reach the sam e ou tp u t
level (highest of all cam eras).
5)
Place cam era in the d ark and rep eat step 2 to 4 u ntil both d ark offset and 80 % sat
signal levels are equ al on all cam eras.
An important note on w indow blemishes:
When flat field correction is p erform ed , w ind ow cleanliness is p aram ou nt. The figu re
below show s an exam p le of w hat can hap p en if a blem ish is p resent on the sensor
w ind ow w hen flat field correction is p erform ed . The blem ish w ill cast a shad ow on the
w afer. FFC w ill com p ensate for this shad ow by increasing the gain. Essentially FFC w ill
create a w hite sp ot to com p ensate for the d ark sp ot (shad ow ). As long as the angle of the
incid ent light rem ains u nchanged then FFC w orks w ell. H ow ever w hen the angle of
incid ence changes significantly (i.e. w hen a lens is ad d ed ) then the shad ow w ill shift and
FFC w ill m akes things w orse by not correcting the new shad ow (d ark sp ot) and
overcorrectin g w here the shad ow u sed to be (w hite sp ot). While the d ark sp ot can be
p otentially cleaned , the w hite sp ot is an FFC artifact that can only be corrected by another
FFC calibration.