Spectral Instruments Series 800, Руководство пользователя, Страница: 59 / 70

Spectral Instruments
Pt # 2500- 59
A3 These are extra rows that are read out after the imaging rows. These rows
traverse the entire imaging area and all of these extra rows spend equal time on the CCD
so they should all see exactly the same dark signal and/or light leak. They should all
look alike! They won’t! The first “extra” row (and possibly one or two more rows)
capture any charge that didn’t get shifted out with the row that preceded it. This is where
trapped charge shows up. A plot of the first overscan row should look like a dark signal
row-plot. It shows some columns that are “high” because of the deferred charge that
leaks out after the readout. At this column address the next few rows may also show
some of this deferred charge. There is a limit on the amount of deferred charge that is
allowed before the trap in that column is statutorily a defective column.
B
This sets the total number of readout pixels expected in the serial direction.
There is a lot of interesting information in the serial overscan. This is because some
CCDs are designed to incorporate masked pixels for dark signal determination and
overscan readout shows them up. The low-cost TV CCD cameras typically read this
signal as a voltage to be subtracted from the rest of the image so as to correct for the DC
offset due to the thermal image.
B1
The serial register is really a separate structure from - although it is intimately
connected with - the parallel register. It is usually longer than the number of columns in
the parallel register. The extra pixels in the serial register are typically called pre-scan
and post-scan. However, this terminology often includes the dark masked pixels which
are actually on the parallel register. Spectral Instruments calls the extra pixels in the
serial register pre-extension and post-extension. Note that it is quite possible to extend
the post-extension into imaginary non-existent pixels and the difference between post-
extension and imaginary pixels is usually negligible. So, B1 is the number of serial
register pixels read before any parallel pixels (masked or not) are encountered.
B2:
The serial pre-mask pixels are actually physical pixels in the parallel array which
are covered with some sort of opaque mask so as to exclude light. The location of and
the degree to which these masked pixels are truly dark varies by CCD manufacturer.
They are usually included so that TV CCD readout can adjust the dark reference offset
before reading the row.
B3:
This is the dimension of the illuminated pixels in each row read from the sensor.
B4:
These pixels are also masked so they don’t see incoming light.
B5:
The serial register post extension combines both physical serial register pixels for
which there are no corresponding columns on the CCD sensor with “imaginary” pixels
that result from the readout circuits clocking more times than there are net serial pixels
to clock. The analog system does not care where pixels come from. It just reads out an
array of N x M pixels and you can make the size suit your own purposes. It is not
possible to add extra “imaginary” pixels at the beginning of the array but you can have
as many as you like at the end of the array.