Image Processing Reference
In-Depth Information
upper and central right parts of Figure 5.27. In this example, low voltage
V
L
is applied to
two neighboring gates and higher voltage
V
H
is applied to the other two adjacent gate
electrodes to form a potential well to integrate signal charges during the exposure period.
While it works as a sensor part, incident light is irradiated on the gate electrodes, as shown
in the right bottom part of Figure 5.27; that is, it is a photogate-type sensor. Although poly-
crystal silicon is mainly used as transfer electrodes in CCDs, it is also a kind of silicon.
Therefore, the spectral response is similar to that of crystal silicon. As can be understood
from penetration depth in Figure 2.20, light transmittance is high enough if the gate elec-
trode is thinner than 0.1 μm. But in real situations, it is difficult to realize thin enough
electrodes because they should have sufficiently low electrical resistance to be driven in
high frequency. So the sensitivity is apt to be lower under the 500 nm wavelength region.
The operation is explained by Figure 5.28. Figure 5.28a shows the situation where the
exposure period has just finished and signal charges have been integrated in all pixels labeled
11-33. All the signal charges are transferred to the storage area at high speed simultaneously,
as shown in Figure 5.28b. Because whole charges are transferred as one frame, this transfer is
called frameshift, and is the origin of the name of frame transfer CCD. By this transfer, charges
are separated from the sensor part (imaging area), and signal integration for the next exposure
starts, although this is not shown in the figure. This situation corresponds to Figure 5.9b in
IT-CCD. Hereafter, the sequence where signal charges in the storage area are transferred to the
HCCD line by line is shown in Figure 5.28c. Each charge packet in the HCCD is transferred to
the output part, which is the FDA, one by one to be converted to voltage signal output, such as
S11 shown in Figure 5.28d. This is the same situation as IT-CCD in Figure 5.9d. Thus, each line
is transferred to the HCCD and output serially. One frame output is completed at Figure 5.28f,
and signal integration for the next frame is also completed. In the case of interlace drive mode,
two pairs of adjacent electrodes alternate, forming a potential well.
Some part of light or charges come in the VCCD when there is smear in IT-CCDs, but
what about in FT-CCDs? During exposure, as signal charges being transferred for output
are in the storage area, which is covered with a light shield film, smear does not occur.
But in the frameshift periods, the image area is receiving incident light as well as in the
exposure periods, that is, all generated charges become smear charges, so the structure
factor is 100%. The only thing that can be done is to reduce the time factor. Therefore,
frameshift in high frequency is necessary to shorten the time factor. There have been pro-
posals to block out incident light using a mechanical or electro-optical shutters to avoid
smear phenomenon completely during the frameshift period.
Although pixels have a simple structure composed of only CCD and isolation, FT-CCDs
need an additional storage area, such as frame memory, separate from the image area,
because pixels work as sensors and transfer CCDs at each different operating time, and
the two operations are not concomitant in pixels. This causes an increase in sensor die
size. On the other hand, in the case of IT-CCDs, since the VCCD in the pixel plays the role
of frame memory, no additional storage area is necessary, although the pixel structure is
more complicated than that of FT-CCDs. From the viewpoint of production, this means
that IT-CCDs need finer processing technologies, although FT-CCDs need larger die size
by avoiding fine processing.
5.1.3.2 Frame-Interline Transfer CCD
In this section, we discuss the frame-interline transfer CCD (FIT-CCD),
17
which was the
standard sensor for applications requiring very high image quality such as studio cameras
for broadcasting. An FIT-CCD is constituted by installing a storage area like an FT-CCD
