THANKS for kind reply to my last email.... I was hoping that you could answer a question about a camera problem I have. The main problem I have, especially in Texas, is that my customer wants the camera to focus on the entrance of his building from the inside out; now the problem I come up with is the light situation, inside the store is dark and outside is very bright. Even with the auto-iris lens, the picture of the person that is at the entrance is not clear (dark). According to your expert opinion what is the solution.
Sign me Dark Sided in Texas
Dear Dark Sided
One of the hardest applications ever presented to a CCTV camera, black &
white or color, is looking into a bright background and coming up with a usable
image. The most common of these applications could be looking at a glass door to
the outside or looking toward a solid overhead door when it opens. Other
applications would include open sided garage type parking ramps and large rooms
with banks of windows on a single wall. Consequently, there have been many
different efforts to present solutions to these problems over the years. Today,
there are four different methods for meeting or matching the challenge (Auto
Iris lenses, peak average, masking, back-lighting and electronic iris). Some
more effective than others. The first step is to understand the beast. The beast
in this case as described is bright background lighting.
Even today, the most common method of controlling the brightness of an image
focused onto a chip or tube is the Auto Iris (A/I) lens. Unfortunately, there is
a general misconception by a large majority of CCTV designers as to exactly how
an A/I lens works. This ignorance of operation quite often leads, to camera
applications that are beyond the scope of the camera's ability to work properly.
The consequence of this misapplication of technology are images that are
silhouetted or extremely dark. A/I lenses are designed to respond to the average
amplitude of the raw video signal produced by a camera through a video sampling
circuit. As a chip or tube (imager) is exposed to more light, the raw video
signal increases. As the light decreases, the raw video signal decreases. The
A/I lens, when installed on a camera, is to be balanced, by the installer, in
such a way as to insure an average of .6 volts peak to peak (VPP) of raw video
image under optimal lighting conditions. As the raw video increases or
decreases, the A/I closes or opens in direct proportion, insuring a consistent
.6 VPP average raw video. For normal lighting conditions and fluctuations, this
method of control works just fine. However, since the video sampler works on
averages, a camera staring into bright areas with dark surroundings produces
silhouette images. A prime example would be looking toward a glass door, in the
center of a solid wall, facing the outside. The major problem is that the
portion of the video image representing the door overloads into the high peaks
and drives the remaining video image into sub-dark or black. Result? A washed
out center area and dark surrounding areas. If a person were to enter through
the door, their image would be silhouetted at best. This is why A/I lenses
cannot compete or keep up with this everyday, problem application.
Since tube type imagers were so susceptible to retained images and target
damage, there was little that could be done to rectify the above situation.
Three methods were used. The first way to insure an image in the door way was to
set the iris of the lens to a higher f-stop (closed iris) and watch the
surrounding walls become so dark that all detail of the image was lost.
Unfortunately the image of the person coming through the door was still a
silhouette in most cases. The second method was to increase the amount of light
that was aimed directly at the door from the inside. The objective was to place
enough light on the subject entering the door as to make their front as bright
or brighter than the surrounding outside light. The third method was to re-aim
the camera to look across or from the same direction as the light from the
outside. This would cost the designer the valuable information of a frontal
shot. Hence the introduction of a method of video signal manipulation called
"Peak Average" (P/A). Although developed for a few select tube
cameras, this form of electronic averaging of the video signal has been carried
over into some CCD cameras. It is a limited method of signal manipulation with
good results in a few cases. First, how does it work and when can it be used.
P/A works on the bases averaging the video signal, dark and bright, into one
overall signal. The results are obvious. If the dark are brought out, the bright
areas go into an ultimate overload and wash out. If the brights are brought out,
the darker areas go darker and consequently loose detail of image. The major
problem comes from the antiquated method of electronic averaging. P/A requires
an image with 2/3's dark, 1/3 bright or 2/3's bright with 1/3 dark area. P/A
will not work on a 50/50 image. For this reason, the applications are very
limited. Although still used, often in an automated mode, P/A is taking a back
seat to the more sophisticated methods of video signal manipulation.
Many lenses have been designed with a form of P/A control built in. This is
accomplished by simple A/I control. Through micro adjustments in the Iris
position, darks can be brought into view by opening the iris slightly, while
bright areas can be toned down by closing the iris slightly. Since A/I lenses
already work with a "level" control (video sampler adjustment), a
second adjustment (P/A or ALC) was easy. A second pot was added that gave fine
tune or minute iris adjustment control. I have always referred to this second
control on lenses as the PFM (Pure Magic, you figure out the "F")
adjustment, since sometimes it works and sometimes it doesn't. It's really a
matter of turning the pot and looking for visible differences in the video image
at the camera.
In direct response to the problem, "masking" was developed. This is
a method of digital interfacing with the video signal and is built into specific
cameras or controllers. The first step is to digitally divide the video image
into grid sections. This is done by putting the camera or controller into a
programming mode and inserting a generated grid over the video image. Once done,
various sections of the image can be chosen to be "ignored" by
highlighting the associated grid points. Once programmed, the grid overlay is
turned off and the image is viewed unobstructed. Although a glorified, more
accurate method of P/A, masking is much more effective and controllable. No
longer is a 2/3, 1/3 bright/dark image required to implement image control. The
final effects are somewhat the same as P/A however. There will still be dark and
bright areas according to the masked areas. Detail of important image
information is retrieved however, so the cost is affordable and the enhancement
is done at the camera or controller versus the lens.
The third method of foreground enhancement is called back-lighting.
Back-lighting is a second, automated or manual form of A/I control. The
strongest point of difference is that back-lighting is automatic and does not
fit the camera into a fixed image of either the dark or bright areas when not
needed. The camera that has back-lighting turned on is a camera that opens the
A/I above the normal setting to compensate for extreme back-light situations.
This is done automatically when the back-lighting circuitry senses a portion of
the video image being driven into sub-black or super-white. The advantage is
that the image is highlighted only when necessary. This is a feature that has
been available on many different camcorders for the past few years and will
ultimately be available for a few years to come.
The fourth method of back-lighting control to come along is referred to as
"Electronic Iris" (E/I). This is the first true method of video signal
enhancement that broaches the need for A/I lenses. Unlike it's predecessor, E/I
works on true video signal averaging. This form of electronic enhancement
literally pulls the super-brights down and the sub-blacks up, forming a final,
equal video image. The advantage? Simple, a person standing in our earlier
described door is fully visible with tight detail while the surrounding features
of the image are also in true perspective. No overloaded bright background and
no sub-black walls, just a nice even picture. E/I is probably the most
electronically advanced method of back-light control and is, unfortunately,
available only on a few select models of cameras.
The fifth and newest method of electronic compensation for extreme back light
problems comes from Panasonic with the development of their "Super
Dynamic" CCD technology. What they have done is taken a super dynamic chip
and utilized two separate shutter speeds simultaneously. That is, half of the
chip works at 1/60th of a second exposure and the other half works at 1/10,000th
of a second exposure. The end result is that you end up with two images from one
camera. The 1/60th exposure image sees into the shadows or dark areas of the
image while the 1/10,000th exposure sees into the bight areas. They then overlap
the two images electronically and the net result is a very nicely balanced color
image of a brightly back lit area.
The final word is that there are multiple methods of handling the most
difficult of video situations facing the security world today. There are few
however, that are more effective than aiming the camera away from direct light
sources. All of the techniques listed are labeled with various names according
to the manufacturer using them, so beware of what you are truly trying to do and
with which technique. Will there someday be cameras that will handle bright/dark
situations as well as the human eye? Yes, there will be and are already cameras
that compensate for simultaneous bright/dark scenes as well as the human eye.
For the moment however, the few that can, are limited. Advance planning of
camera angles and views will continue for the next year or two as the best
method of prevention of poor images. Eventually, good video in multiple lighting
situations will be as simple as hanging a camera on the wall and aiming it at
what you want to see.
