Spectral Imaging: Methods, Design, and Applications - Biomedical Optical Imaging Technologies: Design and Applications - page 113

required, while in other application, the brightness of the object is very low,
requiring long exposure times. It is therefore preferable to have a wide range
of exposure times that can be used according to the application.
6. Readout rate
Readout rate is a performance factor related also to the frame rate parameter; both
imply on the ability of the system to record data at high temporal frequencies. The
readout rate, also referred to as the pixel rate, is dominated by the time it takes to
convert an analogue signal (voltage) into a digital signal (count) and is defined
as the inverse of it. Thus, the frame rate, is governed by the number of pixels
acquired (or the field of view), the readout rate and also the exposure time. The
frame rate (typically measured in frames per second) has immense importance in
cases where recording of fast transient dynamics combined with low signals is
required.
The trade-off in high readout rates is that it also contributes to the intrinsic
noise in the system. The analogue-to-digital converter contributes noise also
coined as readout noise, typically growing respectively to the readout rate.
7. Binning
The physical pixels of the array detector can be “binned” so that effectively each
pixel that collects photoelectrons is bigger in size. This leads to a significantly
larger dynamic range because the number of electrons that can be stored in the
bigger pixel is much higher. Also, once this amount of photons is converted from
an analogue value to a digital number, the readout noise is added only once to
all the binned pixels, and therefore, the signal-to-noise ratio is improved. On the
other hand, the binning certainly comes with a loss of spatial resolution, as the
number of pixels is reduced. Some of the CCDs enable to select many different
binning values.
8. Gain factor
Photons hitting the detector are converted to electron charge which is converted
to a voltage which is finally translated to a digital value. This number is stored
in the computer and describes the light intensity that illuminated that pixel. The
gain factor is the ratio parameter between the amount of electrons per pixel to the
counts per pixel generated. In some cameras it is a fixed parameter while in others
it is a variable parameter that can be determined based on the actual amount of
electrons that filled the pixels of the camera. Each pixel has a full well depth
that determines the maximal capacity of electrons, say 10,000. The standard gain
factor converts this value to the maximal possible count number (depending on
the analogue-to-digital converter - e.g., 12 bits camera consists of 2 12 D 4,096

counts). On the other hand, in a given measurement, the light intensity may be
low and the wells may not be full at all.

Most of the spectral imaging systems are based on an array detector, typically
a CCD combined with fore-optics, and few others on point detectors such as
photomultipliers (PMTs). While excellent quality CCDs are available, the light itself
has intrinsic noise due to the Poissonian nature of photon emission (Shot noise)
that cannot be avoided. If n photons are counted, then shot noise, with its Poisson