28
•
The Piranha XL XDR Camera
Note that when adjusting the row selection, a momentary loss of LVAL will occur. This will also
happen when changing user sets that include row selection
.
Image Response Uniformity
See the section Flat Field Category in Appendix A for GenICam features associated with this section
and how to use them
Relevant Features: Calibrate FPN, Calibrate PRNU, Calibration Algorithm, Calibration Target
It is common to find an image with lower response at the edges of the camera’s field of view
compared to its center. This is typically the result of a combination of lens vignetting (cos
4th
) roll-
off and the beam structure of the illumination source. A more diffused light may reduce this roll-off
effect. However, if decreasing the lens aperture improves the edge roll-off, then barrel vignetting (a
shadow cast on the sensor by the focus helical or extension tubes) may also be present.
The camera can compensate for edge roll-off and other optical non-uniformities through flat field
calibration.
•
When performing Flat Field (PRNU) calibration, the camera should be imaging a front
illuminated white target or rear bright field illumination source. The optical setup should be
as per the inspection system, including lens magnification, aperture, and illumination
intensity, plus illuminator beam structure.
•
Flat field calibration should be performed for each row selection that will be used by the
system and saved as a distinct user set (see the Saving & Restoring Camera Setup
Configurations section). Whenever a different number of rows are selected the camera setup
will default to the factory settings. The user must select and load the associated user set for
the number of rows in use that was previously saved in the user set
•
Flat field calibration should be performed when the camera temperature has stabilized.
•
When the camera is commanded to execute a flat field calibration it will adjust all pixels to
have the same value as that of the peak pixel value or target level, as per the calibration
mode selected.
•
If flat field calibration is being set to a target level that is lower than the peak value and the
system gain is set to a low value, then it is possible that the sensor will maximize its output
before the camera’s output reaches 255 DN. This can be seen when a portion of the output
stops increasing with increasing illumination and the PRNU deteriorates. This effect can be
resolved by reducing the light level or exposure control time.
On completion of a flat field calibration, all pixels should be at their un-calibrated peak value or
target value. Subsequent changes in gain allow the user to make refinements to the operating
responsivity level.
Note that the best flat field calibration can be achieved by performing it at the mid DN level of the
working range used in the operation. Any flat field error associated with residual non linearity in the
pixel will be halved as compared to performing a calibration at the peak value of the operating
range. A simple way of performing this is to reduce exposure time to half what is used in the
operation in order to get the mid DN level for flat field calibration. Once complete, return the
exposure time to its original setting.
Those areas of the image where high roll-off is present will show higher noise levels after flat field
calibration due to the higher gain values of the correction coefficients. Flat field calibration can only
compensate for up to an 8:1 variation. If the variation exceeds 8:1, then the line profile after
calibration will include pixels that are below the un-calibrated peak level.
