In one of your classes, you mentioned in order to obtain positive I.D. with CCTV, the subject to be identified should be at least 1/10th of the scene. For some reason I have not been able to find a complete explanation of this in your manual "Professional's Guide To The Application & Design of CCTV".
I'm trying to deal with a major equipment supplier, from your neck of the woods, who has demonstrated a total lack of professionalism in the CCTV end of things. So, I am left with putting virtually everything in writing, even to the point of crisp focus.
Thanks for your help. Signed me Dogged by Detailed Ignorance
Dear Dogged
The general information that you are looking for can be found in the 1997
edition of the Professional's Guide to CCTV or the Application / Design books on
pages 1.6 - 1.8. Just in case you don't have either of those editions, I have
scanned the pages in and attached them to this email as an ASCII.txt file. I
have included the 1998 update portion of this section as well (the last two
paragraphs of page 1.8). If you have any problems downloading, let me know and
we will fax you the information.
I sincerely hope that this helps you out of the current situation. Ignorance
is always a difficult barrier to jump. Education is the only key. Good luck and
let me know if there is any thing else that I or my staff can be doing to help
you. We are in your Service.
The Only Purpose of CCTV in Security
From a Systems design and operational perspective there are only four things
to know, but they are the most important aspects of CCTV system design and
unfortunately the most commonly missed points of the design process.
Keeping everything in proper perspective is the first and foremost step to a
good CCTV system design. This means that you must remember that CCTV systems are
meant to be 'Visual Assessment" tools and nothing more! You should never
allow yourself or a hired consultant to design your CCTV system with anything
more or less in mind. Visual assessment refers to having information of a proper
identifiable and/or descriptive nature either during and/or after the fact. The
actual amount of visual information that you end up with is determined by the
lens and the angle or position of the view, not the camera.
The camera that you choose to use will be determined by either it's
sensitivity and/or the resolution of the final image. Sensitivity refers to the
amount of actual visible or Infrared light necessary to produce a quality image.
Resolution defines the image quality from a detail or reproduction perspective.
Therefore the camera type or model is always decided upon prior to the lens
selection.
Lenses on the other hand, determine the amount and type of image that you
will ultimately see on the screen of the monitor. For this reason, the lens
should always be treated as a peripheral to the system and picked based upon the
descriptive qualities necessary to produce the desired identification
information. To do this we first define the three theoretical identity views of
a CCTV system. They are "Personal Identification", "Action
Identification", and "Scene Identification".
Personal Identification refers to the ability of the viewer to personally
identify something within the scene beyond a shadow of a doubt. Notice I said
"to personally identify something verses someone. The reason
is really quite simple. Personal Identification does not reflect human
identification but rather the ability to personally identify specific
information or objects within an image. To demonstrate this form of
Identification, take a one hundred dollar bill from your pocket (a five or
twenty will work just as well). Tack the bill to a wall and stand about three
feet from it. Assuming that you have average to mediocre eyesight, you should be
able to identify what you are looking at, beyond a shadow of a doubt, as a one
hundred dollar bill. Remember the first most important point is that this is
strictly visual perspective and so must carry enough detail to guarantee that
the bill is what it appears to be. When dealing in the real world of three
dimension, touch, smell and physical assessment, your identification of objects
and subjects is easier. You have more assessment information. CCTV assessment is
limited to one form (sight) and so must be as detailed as possible according to
the subject of personal identification. Now step back thirty feet and look at
the bill on the wall. Can you. beyond a shadow of a doubt, give a visual
verification that what you are looking at is a one hundred dollar bill? With
normal eyes your answer should be no. This is a demonstration of what is
referred to as personal identification. The hundred dollar bill was the object
of personal identification, your eyes were the lenses with your mind as the
camera.
A second demonstration of personal identification will refer to and actually
demonstrate how the "Angle of View" of the camera affects the results
available from a CCTV system. This demonstration is also very easy and can be
done in your office. Stand (carefully) on top of your desk and look directly
down onto the top of various peer's heads as they come and go from your office.
Can you, beyond a shadow of a doubt, identify your associates? Yes? Well what if
you did not know them personally and you had to find them in a crowd at face
level? Could you pick them out based upon a video replay of the tops of their
heads? Probably not.
The point of both exercises was to demonstrate that personal identification
has two very important faces. The first being the relationship of size and
detail of an image and the second being the angle of view of from which a scene
is viewed. Without careful consideration to both details, the value of your
recording or CCTV system is lost in useless images.
The second form of identification that is used in CCTV is called Action
identification. This form of identification also works hand in hand with the
first two but is equally, if not more so, important to the overall design of the
CCTV system. To demonstrate, let's go back to the first example and once again
pin a one hundred dollar bill to a wall. Got it up? Good, now stand back far
enough so as to be able to identify the hundred dollar bill, but not so close as
to be able to touch it. Now invite an associate into the room. Have them walk up
to the bill, look at it, and admire it for all of it's value. Now close your
eyes for five seconds. If, when you open your eyes again, the bill is gone, can
you will you would you prosecute your associate for stealing it from you? I hope
not, because you will ultimately find yourself in a court room being sued for
"Defamation of character" or possibly slander at the very least.
" But", you argue, "I saw my associate steal the bill!" No,
actually you did not. What you saw was your associate enter the room, take
notice of the bill, admire the bill and lastly have access to the bill. You did
not however, see your associate take the bill or stuff it into their pocket.
What you have is circumstantial evidence at best. You did not record or witness
the action of the person taking the bill. In fact, if you had been standing in
the back of the room, looking down you probably would have not had enough
personal identification to prove that your associate even took a one hundred
dollar bill. Enough information to say that they took something perhaps, but a
one hundred dollar bill? Now you see the intricacies of personal and action
identification and how they ultimately and intimately interact together.
"OK", you say "that was easy." As a matter of fact, it
makes sense even if you never actually verbalized it before. So what is
"Scene Identification"? My favorite story about scene identification
comes from a disgruntled employee, working for a firm on the East coast. Seems
that this particular man was passed over for a promotion for what ever reason.
He was not happy about the situation and decided to spout his opinion to the
President of the company at his first opportunity. Since the employee had been
working with the firm for many years, he had what you could call a general lay
of the land. One day the security cameras located in the main hallway recorded
the President of the company walking out of his office, turning left into the
main hall way and then walking away from the camera. One minute later, our
disgruntled employee was recorded walking towards the Presidents office. Coming
from the doorway directly under the camera, the view of the employee was
primarily the top of his head as well as the back of his body as he walked down
the hallway. When he arrived at the President's office door, he turned left into
the office giving a full, personal identification side view of himself to the
camera). Once inside the office, he climbed onto the President's desk, squatted
and proceeded to leave a rather large deposit of his personal opinion. He was
then recorded leaving the President's office, turning right and walking toward
the camera, again presenting a good personal identification of his person. One
minute later, the President was recorded returning to his office. Once inside,
it's easy to understand why he was upset and the proverbial crap hit the fan.
Security went into immediate action determining who the perpetrator was, and the
employee was called in for interrogation and summary sentencing. When approached
with the accusations of the incident however, the employee immediately went into
his act that he was innocent and that because everyone knew that he was unhappy,
he was being framed. The Personnel Director pointed out that they had a video
recording of him doing the deed. The employee was quick to point out that all
they had was potential circumstantial evidence since they had no proof that he
actually entered the president's office, let alone did anything wrong inside it.
Again they referred to the video tape recording. The employee stated that he had
seen the President that day leaving the cafeteria just prior to his entering,
but otherwise had not been near the President's office. The interviewers then
pointed out the character generated lettering on the tape identifying the
hallway as the president's area, not the cafeteria. The employee pointed out
that both hallways appeared the same (white floors, white walls, white ceilings,
brown doors with equal spacing). He then accused the personnel director of
having tampered with the video tape by changing the character identification on
the cafeteria camera, making a recording of him coming and going, and then
changing everything back. This obviously was a plot to entrap him with false
evidence in order to frame him and get him out of the company. Basically, the
employee won this round because two cameras, within the same compound had scene
views that could not be distinguished as individual based upon their own merit.
He was later dismissed for other reasons and did not return to work. He was not
however, prosecuted for the actual deed.
The whole point of scene identification is that each scene must stand upon
it's own merit. In the above case, one hallway could have had a fire
extinguisher, a picture, a potted plant, or something else that would make it
stand alone and identifiable from the rest. Having such independent
identification would not have given enough information to convict the employee,
since there was no camera actually recording the action of the crime, but it
would have provided enough evidence of available intent that the employee could
have been released without much ado. Scene identification is one of those all
important, but most often missed forms of identification that is so vital to
video systems today. You simply cannot rely on the discretion of character
generated words to provide your sole method of scene discrimination. This is
especially true in those cases where multiple cameras within a single compound
have the same basic angle of view of similar scenes.
Now that we have discussed the three forms of video identification and the
importance of angle of views, it should be easy to understand why I started by
saying there are more important issues to consider than what camera or lens to
use. And these prior to choosing a lens and/or location of a camera. Be certain
that you understand exactly what it is that you are trying to accomplish with a
camera prior to installing it. Determine first the personal identification of
what you need to see, secondly that action of what it is that you are trying to
prevent, thirdly what the best angle of view is for both ID forms and lastly,
what the final view is in relationship to other such camera scenes within a
system.
In the end, and according to the Police Chiefs Association of America, they
would prefer to see personal identification of an object or person take up at
least ten percent (10%) of the over scene width. Dependent upon the size of the
object, you can usually stretch that to fifteen or twenty percent (15% - 20%).
Therefore, if the average human being is two feet (2') wide, you would place
them in a scene of a maximum screen width of twenty feet (20'). A hundred dollar
bill (US) is six inches long. Therefore, the maximum scene width as displayed on
the monitor would be sixty inches (60")
Action identification should take up no less than twenty percent of the
overall scene. That is unless it is tied to a personal identification object of
some sort. In that case, the personal identification takes precedence and you
may be forced to used two cameras interfaced together through a quad,
multiplexer, or picture in picture unit.
