The advent of the charge-coupled device (CCD) sensor was hailed
as being the end
of maintenance of cameras. For a short while, this is actually
appeared to be true, but
only with the simple CCD chips of approximately 10 years ago.
CCD CHIPS
What is a CCD Chip?
The CCD chip is actually an array of thousands of photosensitive
diodes arranged in
a matrix. Associated with every photodiode is a metal oxide
semiconductor (MOS)
capacitor which is used to retain a charge, built up during the
field period of the video
signal. An analogue shift register also forms part of the CCD chip
and the shift register
forms the other part of what is often known as a bucket brigade
device. The buckets are
the MOS capacitors and the brigade refers to the analogue shift
register. The charge
is passed from bucket to bucket through the device.
The advertised advantages of CCDs are that they:
(a) are solid state
(b) have no electron beam
(c) are almost immune to magnetic or electric fields
(d) are not susceptible to vibration
(e) have no burn in
(f) need no adjustments
(g) need no changing of tubes
(h) have an unlimited lifetime.
However, the reality of the situation was less than perfect. The
CCD chips gave a lot
lower resolution (approximately 240 lines on the first devices) and
also had a lack of
ability to deal with the extremes of light level. Thus, all lenses
had to have auto iris
control, whereas the Vidicon tubed cameras did not require auto iris
controlled lenses,
but used manual iris lenses.
The early CCD chips were all frame transfer devices, which were
also less sensitive
than the tubed cameras chosen for outdoor use. Interline transfer
devices then
improved the sensitivity, but brought a crop of problems, such as
transfer smear (see
below). The unlimited lifetime of the CCD chips is true to some
extent, but the life is
limited by the very rapid advances in chip design — making
obsolescence the limit of
lifetime of the chips.
Types of CCD Chips
At present, there are both monochrome and
color CCD chips, in
both single and three
chip combinations. It is possible to achieve more than double the
early resolution, ie
up to 570 lines in monochrome and up to a supposed 480 lines in
color (only
achievable at the chip, which very few manufacturers mention). The
sensitivity of the
CCD chips has improved markedly since their introduction.
During the 1990s, there have been two basic types of CCD chip, ie
the frame transfer
and the interline types.
The frame transfer chip has two distinct sections, ie a
photosensitive area and a
storage section. The focused image is projected onto the
photosensitive area, thus
causing the charges to be defined in the chip. The storage area is
masked from the
light using an opaque aluminum optical mask. Every one-fiftieth of
a second, a field of
information is transferred from the photosensitive area to the
storage area, where it is
then processed during the next one-fiftieth of a second.
The interline transfer chips have shift registers between the
columns of pixels. This
enables better processing to be carried out, although light leaks
onto the shift registers.
Although screened, light does penetrate and generates electrons deep
within the chip
where there are large overloads. These deep electrons add to the
charges being
transferred in the registers, an effect which is typified by a
vertical red or white line
down the screen, known as transfer smear.
For many years the manufacture of CCD chips was dominated by one
manufacturer,
but there is now competition. The types that are now practical
devices are hyper hole
accumulated diode (hyper HAD), super dynamic, frame interline
transfer and the
exwave HAD.
Hyper hole accumulated diode
The hyper HAD was released in 1990 and introduced a number of
improvements in the
design of interline transfer devices. These improvements serve to
reduce the dark
noise which could be purely from:
dark current
reset noise
optical shot noise
fixed pattern noise
residual point noise.
Due to the improved sensitivity, which
is largely a result of the
on chip lens, it has
rendered the use of electronically-controlled shutter speeds, to be
a far more reliable
method of controlling the exposure. There is also an improvement in
dealing with
transfer smear, because of careful profiling in the design of the
chip.
Super dynamic
This is a recent refinement in the design and manufacture of
interline devices where
the number of vertical shift registers has been doubled, thus
enabling charge generated
by the photodiodes to be dealt with in two different ways.
The overall scene is processed and the exposure of the photodiode
is determined by
the light falling on the photodiode. For dark parts of the picture,
the photodiodes are
exposed for almost one-fiftieth of a second. The parts of the
picture with highlights
(where light levels are high) will be exposed for a far shorter
period of time, perhaps as
short as 1/2000th of a second. The short charge and the long charge
are processed
together which results in the output picture. This technique gives
dynamic light
handling, better than 20-40 times that of the conventional hyper
HAD.
Frame interline transfer
This type of device combines the benefits of both the frame
transfer and the interline
transfer devices.
The chip design features a masked storage area exactly along the
lines of the frame
transfer chip. Every one-fiftieth of a second, a field of the
generated charges is
transferred straight into this storage area. From here, the
information is then the
processed line by line, in the manner of the interline transfer
device.
In dealing with the charges in this manner, almost all the
problems of generation of
deep electrons within the semiconductor are avoided and this results
in an
improvement of the transfer smear by reducing the levels by a factor
of approximately
60. This technique was designed a few years ago and was used
exclusively in
broadcast cameras. It was considered too expensive to use in CCTV,
but it is now
possible to obtain the appropriate chips.
Exwave HAD
Exwave, in this case, means extended wavelength, because the chip
has an improved
spectral response. It gives approximately double the sensitivity at
800nm and
approximately four times the sensitivity at 900nm (both being in the
infra-red
wavelengths). This diode has greater sensitivity in part because of
the on chip lens
design and the improved masking of the vertical shift registers,
giving less transfer
smear.
CAMERA FACILITIES
What comes out of the back of a camera? Answer — a mains
cable! This is a genuine
answer given by a delegate on a training course, but it does raise
the question of what
is the right answer? The real answer could have been any one of the
following:
(a) composite video
(b) horizontal synchronization pulses
(c) vertical synchronization pulses
(d) luminance
(e) chrominance
(f) RS 485 data
(g) data, as defined by the Data Protection Registrar.
Is it all necessary? With tube cameras, the controls that were
available were all
necessary to enable the maintenance of picture quality over the two
year period or so
between tube replacements. Such controls were electrical
focus, target voltage, the
current and pedestal. For a while CCD cameras were very
simple, ie they were MOS
cameras that had no controls that were user configurable. However,
many cameras
now contain such features as line phase, genlock, chrominance, peak
white blanking,
auto white balance, auto white tracing, text and clocks, auto iris
settings, direct drive
iris outputs, peak/average adjustment, electronic iris, manual
shutter selection, high
light suppression, Y & C video outputs, picture memory, movement
detection, backlight
compensation, kangaroo lens drive, output level adjustment, user
memory settings,
motorised back focus, field integration, DSP, security code, remote
controllable
functions, on-screen menus, and restore factory default. A brief
overview of these
functions is given below.
1. Line phase enables a.c. powered cameras to have the line phase
adjusted to enable
multiple cameras to be switched through a simple analogue video
switch without the
resultant picture bounce of unsynchronized cameras.
2. Genlock is an input to enable all cameras to be synchronized
to a
single
synchronization source.
3. Chrominance usually enables the increase and decrease of both red
and blue,
manually (there is not usually adjustment of green, because
adjustment of red and blue
will give the impression of increasing and the decreasing the level
of green, ie red and
blue and green = white).
4. Peak white blanking describes the situation where highlights of
approximately 150%
or more will be changed from a peak white to either black or a shade
of grey to enable
darker parts of the picture to be seen.
5. Auto white balance means that the camera decides on what is white
within a picture
and compensates by adjusting the colors to achieve this white
level.
6. Manual white balance is the facility whereby at the commissioning
stage, the camera
is provided with a “white target” filling the screen this
is set as white, normally by
pressing a set button.
7. Auto white tracing is useful where the color temperature of the
lighting may change
with time or as external lighting changes.
8. Text and clocks can be programmed either at the camera only or
from the control
room; it should be noted that the remote programming of text can be
considered a
potential problem when evidential material is being collected.
9. Auto (video) iris settings involve an output known as a digital
auto iris; this output
used to be a video signal, but is now an integrated signal which can
cause problems
when used in combination with some lenses — making the setting
of iris level a very
twitchy thing to do.
10. Direct drive iris removes the control electronics from within
the lens and relocates
them to within the camera, thus allowing the purchase of cheaper,
lighter lenses.
11. Peak/average is an adjustment which gives the commissioning
engineer the
opportunity to change detection between peak within any part of the
picture or an
average over the whole picture.
12. Electronic iris enables the use of a manual iris lens, so the
camera now controls the
exposure of the CCD chip by changing the shutter speed continuously
to suit the
ambient lighting conditions.
13. Shutter is the equivalent of the shutter adjustment on a
photographic camera, which
helps to stop blur due to motion of the object. The higher the
shutter speed the greater
the amount of light that is required. It can also be used to
overcome the problems with
having a 60Hz mains frequency, but using systems at 50Hz, such as in
Saudi Arabia,
produces the effect of a visible pulsing of the lighting.
14. Y-C video outputs enable the highest resolution color
signal to
be transmitted and
recorded (recording as SVHS). However, it requires two transmission
paths and two
matrix inputs per camera.
15. Picture memory can enable one field of video to be stored in the
camera in response
to a stimulus.
16. Movement detection uses relatively simple movement detection
criteria, such as
video level detection, to determine if there is movement within the
field of view; it will
then signal (via telemetry) that there is movement within the
picture, but should never
be considered as an alarm device.
17. Backlight compensation is used to deal with the problem of
backlighting that causes
people and objects to be silhouetted.
18. A kangaroo lens drive is an output to drive the latest two
position iris lenses being
used with electronic iris and field integration.
19. Output level adjustment is used to boost the video signal prior
to transmission on
longer lengths of coaxial cable.
20. Memory settings — in some cases, up to four different
memories can be available
to memories the multitude of settings required for the operation of
the camera.
21. Motorized back focus can simplify the pre-assembly or the
installation of cameras
and can aid where infra-red lighting is used, causing a change in
focus.
22. Field integration is very useful when lighting levels are very
low, but should be used
with caution and (realistically) only with fixed cameras, because of
the amount of blur of
even slow motions that can be seen when using long integration
periods.
23. Digital signal processing (DSP) can improve the resultant image
by defining outlines
by introducing a one pixel wide black line around any objects within
the scene to improve
visibility where the levels of contrast are low.
24.Security mode ensures that if a camera has been stolen, it cannot
be used unless a
code number is entered to prove ownership.
25. On-screen menus have become more essential with the increase in
the number of
features of modern cameras.
26. Restore factory defaults can be an essential item where an
engineer has finger
trouble (they all do) and has forgotten where he was when he
started; this feature should
only be used as last resort and it is essential for all the settings
to have been
documented carefully for each camera within the system.
FUTURE DEVELOPMENTS
Suitability
The above list includes 26 different features, but users should
be careful that their
manufacturers give independent views. Who have the marketing
strategists targeted?
Have they persuaded the consultants that there are particular sets
of features which are
essential for every job they may be specifying? Alternatively, has
the control room
manager been given demonstrations of the camera in the controlled
environment of a
dark room or perhaps even a demonstration vehicle in his or her own
locality, but with
the highly trained manufacturer’s operator driving the
demonstration camera? At this
stage of the proceedings, it is very unlikely that the control room
operators have been
interviewed, let alone had some training and commented which
features may or may not
be necessary.
It is inevitable that as camera features become more varied, so
the operation of the
system will become more complex. There will be more adjustments to
make to the
system on a time-profile basis. This is one particular aspect that
can be simplified by the
availability of user configurable memories within the camera to
enable different settings
to be chosen, depending upon the ambient conditions. It must be
remembered that this
will also involve the operator having a good memory or clear concise
documentation
within a busy control room environment.
Controlling the Features
Another word of caution is that it must be clearly remembered
that to have all these
advanced features available for your system, that you must have the
camera
manufacturer’s telemetry control system. This choice of the
control system may lead to
restrictions of hardware that may not (at first) be evident. In
other words if we fitted one
camera with remote control of back focus into another
manufacturer’s control system,
you cannot, without modification to the system, control the back
focus. This design and
modification can be carried out, but puts a premium on to the price
of the control system.
Digital Cameras
There is much talk of digital cameras, and some of the commonly
available cameras
actually have the word mentioned on the body of the camera.
However, there are only
a very limited numbers of true digital cameras available and they
are attached to a
computer, used in videoconferencing applications. The current
restriction is the data
transfer rate; for high-quality video in a serial data stream,
transfer rates of greater than
250 Mega bits per second are needed. Clearly, the current transfer
rates over ISDN are
totally inadequate, but if there is a parallel data bus system of,
say, 32 bits’ width, then
the data rate reduces to approximately eight Mega bits per second.
There is currently much work being done with asynchronous
transfer mode (ATM)
networks, Ethernet and the like, for both serial and parallel data
transmission. However,
for the immediate future, it is unlikely that digital cameras will
be
appearing in town
centers until the fiber optic equipment manufacturers can provide an
economic bearer
solution.
Similarly, there is no established standard yet for the
compression of video for both
transmission and storage. Will it be MPEG, JPEG, wavelet or even
fractal compression?
Any digital cameras that purport to use IEE1394 bus are cameras for
machine vision and
will feed into a PC and not a video matrix. The transmission
distance on copper cable
is rather limited as well — only 4.5m!
Some people have been slightly misled to believe that digital
signal processing (DSP)
cameras are digital cameras, but all DSP cameras still have an
analogue video output.
The DSP aspect relates to signal processing within the camera, that
is used to enhance
various aspects of the video output. DSP can be used to give
sophisticated backlight
compensation and the aspects of motion detection that some
manufacturers include.
When an area being observed is low contrast between the objects and
the background,
DSP cameras can enhance the outlines by drawing a one pixel wide
black line around
all of the objects, both foreground and background, thus giving
well-defined edges. It
should be noted that some cameras can introduce so much processing
that post-
production video enhancement, such as used by courts for evidential
purposes, cannot
actually improve the pictures further.
Camera Size
There has been a great tendency to reduce the size of the CCD
chip and the cameras,
driven largely by the camcorder market, which has led to the
development of the covert
spy-type cameras. With large-scale integration, the CCD chip and
lens can be separated
from the rest of the electronics to produce cameras that are fitted
into pens and
spectacles, which then combine with miniature video transmitters to
enable the remote
recording of those pictures. It should also be remembered that under
the Data Protection
Act 1998, all personal data must be fairly obtained.
Specialist Cameras
Another area of development relates to thermal imaging cameras
which, only a few
years ago, required the detector devices to be cooled by the liquid
nitrogen. It is possible
now to purchase a hand-held thermal imaging camera, with a fixed
lens, and no external
cooler, for under £14,000. Similarly, there are now image
intensifiers that can be fitted
between the lens and the CCD chip to give performance in lighting
conditions down to
starlight (0.0001 lux) and they cost less than £10,000.
There are also cameras that can give up to 700 lines of
resolution, but these are limited
to broadcast applications with cameras being priced from £20,000
upwards (prices and
sizes are dropping fast). Caution is needed when looking for these
high resolutions,
because they are obtained using three separate CCD chips and lenses
with prisms fitted
to split the light. These cameras give four output signals, as red,
green and blue and
synchronization (sometimes three, ie RGB with synch on green) and,
therefore, require
four separate cables and four channels per camera in the matrix
switch. A standard
industrial timelapse VCR will not be able to record this form of
signal.
Users should consider carefully the ramifications of installing
increasingly complex
cameras and control systems. The man machine interface or graphical
user interface
or keyboard needs to be as simple as possible, to enable operation
and management
of the system to be as effective as possible.
