Dalsa Piranha ES Series, Руководство пользователя

Страница: 65 / 106

Piranha ES-xx User Manual
Teledyne DALSA 03-032-20070-02
65
algorithm ad ju sts the analog gain, it also affects FPN . If FPN is
calibrated p rior to ru nning this algorithm , FPN w ill be
observable in d ark cond itions and an incorrect FPN valu e w ill be
u sed d u ring PRN U calibration resu lting in incorrect PRN U
coefficients.
2
= Calcu lates the PRN U coefficients u sing the entered target
valu e as show n below :
PRNU Coefficient =
Target
(AVG Pixel Value ) - (FPN + value)
sdo
i
i
i
The calcu lation is p erform ed for all sensor p ixels bu t w arnings
are only ap p lied to p ixels in the reg ion of interest. This
algorithm is u sefu l for achieving u niform ou tp u t across m u ltip le
cam eras. It is im p ortant that the target valu e (set w ith the next
p aram eter) is set to be at least equ al to the highest p ixel across all
cam eras so that all p ixels can r each the highest p ixel valu e
d u ring calibration.
3
= This algorithm inclu d es an analog gain ad ju stm ent p rior to
PRN U calibration. Analog gain is first ad ju sted s o that the p eak
pixel value in tap’s ROI is w ithin 97 to 99% of the specified
target valu e. It then calcu lates the PRN U coefficients u sing the
target valu e as show n below :
PRNU Coefficient =
Target
(AVG Pixel Value ) - (FPN + value)
sdo
i i
i
The calcu lation is p erform ed for all sensor p ixels bu t w arnings
are only ap p lied to p ixels in the region of interest. This
algorithm is u sefu l for achieving u niform ou tp u t acro ss m u ltip le
cam eras.
This algorithm is u sefu l for achieving u niform ou tp u t across
m u ltip le cam eras by first ad ju sting analog gain and then
p erform ing PRN U calibration. This algorithm is recom m end ed
for u se only w hen FPN is negligible and FPN coefficients are set
to zero. Since this algorithm ad ju sts the analog gain, it also
affects FPN . If FPN is calibrated p rior to ru nning this algorithm ,
FPN w ill be observable in d ark cond itions and an incorrect FPN
valu e w ill be u sed d u ring PRN U calibration resu lting in
incorrect PRN U coefficients.
This algorithm is m ore robu st and rep eatable than algorithm 1
becau se it u ses an average p ixel valu e rather than a nu m ber
above target. H ow ever, this algorithm is slow er.
4 = Calcu lates the PRN U coefficient in the sam e w ay as cpa 2
w ith the excep tion that this com m and only calcu lates PRN U for
p ixels w ith in the cu rrent Region of Interest (ROI).
i
Peak target valu e in a range from 1024 to 4055DN . The target
valu e m u st be greater than the cu rrent p eak ou tp u t valu e.
N otes:

Perform all analog ad ju stm ents before calibrating PRN U.

This com m and p erform s the sam e fu nction as the
cpp
com m and bu t forces you to enter a target valu e.

Calibrate FPN before calibrating PRN U. If you are not
p erform ing FPN calibration then issu e the
rpc
(reset p ixel
coefficients) com m and and set the
sdo
(set d igital offset)