
® 16k CL
25
UM 16kCL – REVC – 11/14
e2v semiconductors SAS 2014
6.3.5
Flat Field Correction
How is performed the Flat Field Correction ?
What is the Flat Field correction (FFC) ?
The Flat Field Correction is a digital correction on each pixel which allows :
To correct the Pixel PRNU (Pixel Response Non Uniformity) and DSNU (Dark Signal Non Uniformity)
To Correct the shading due to the lens
To correct the Light source non uniformity
Before
After
How is calculated / Applied the FFC ?
The FFC is a digital correction on the pixel level for both Gain and Offset.
Each Pixel is corrected with :
o
An Offset on 10 bits (Signed Int S9.1). They cover a dynamic of
256LSB in 12bits with a resolution of 1/2 LSB
12bits. Offet : the MSB is the sign, the rest of 9bits is from 0 .. 256 with precision of 1/2
o
A Gain on 12 bits (Unsigned Int U2.12) with a max gain value of x5
(*)
The calculation of the new pixel value is : P’ = ( P + Off).(1 + Gain/1024
(*)
). Gain : 0 to 4095
The FFC processing can be completed with an automatic adjustment to a global target. This function is designed as
“FFC Adjust”. This adjustment to a User target is done by an internal hidden gain which is re-calculated each time the
FFC is processed while the FFC adjust function is enabled.
The FFC is always processed with the max pixel value of the line as reference. If enabled, the FFC adjust module
(located at the output of the FFC module) calculates the adjustment gain to reach the target defined by the User.
When the FFC result is saved in memory, the adjust gain and target are saved in the same time in order to associate
this gain value with the FFC result.