5 –Camera Controller
A6600/A6650 User’s Manual
26
Detector
14-bit A/D
x
+
NUC Table
Bad Pixel
Replacement
Algorithm
Corrected Data
Uncorrected Data
x
+
Analog
Gain/Offset
Figure 4-12: Digital Process Showing NUC Table Application
To create the NUC table, the camera images either one or two uniform temperature sources. The
source can be an external source provided by the user or the camera’s internal NUC flag which is
basically a shutter the camera places in front of the detector. If the source is external it should be
uniform and large enough to overfill the cameras field-of-view (FOV). By analyzing the pixel data from
these constant sources, the non-uniformity of the pixels can be determined and corrected. There are
three types of processes which are used to create the NUC table; One-Point, Two-Point, and Offset
Update.
5.1.5.1.1
One-Point Correction Process
A One-Point Correction Process requires one uniform source, which is typically in the middle of the
usable range. The One-Point Correction replaces all gain coefficients in the NUC table with a value of
one (“1”) as seen in Figure 4-13. The offset coefficients are computed uniquely for each pixel.
x
+
NUC Table
Bad Pixel
Replacement
Algorithm
Corrected Data
Uncorrected Data
“1”
Offset
Coefficients
Figure 4-13: One-Point Correction
5.1.5.1.2
Two-Point Correction Process
The Two-Point Correction Process builds a NUC table that contains an individually computed gain
and offset coefficient for each pixel as seen in Figure 4-14. Two uniform sources are required for this
correction. One source at the low end and a second source at the upper end of the usable detector
input range. Because of the use of two images at either end of the input range, the Two-Point
Correction yields better correction results verses the One-Point process. A 2-point correct will also
work better over a wider range of scene temperatures than a 1-point correction.
