UFC 4-021-02NF
27 September 2006
change 1, 23 October 2006
4-17.2
Video Motion Detection. (VMD) makes use of software usually provided in
the camera (or in the DVR) to analyze differences in contrast between the image pixels
over time. If the contrast changes from one moment to the next, the associated software
is capable of detecting the difference. Video Motion Detection was the industry's first
attempt to provide automated alarm notification on detecting motion. It was also initiated
to reduce the tremendous storage requirements of CCTV information when closed
circuit television systems converted to digital images. Reduced memory requirements
were achieved because if there were no pixel changes, the system was programmed
not to store video. This technique saved some hard disk drive space. The downside of
Video Motion Detection is that it alerts on everything in a scene that moves (or causes
pixel changes). This could be an object blowing across the field-of-view, snow falling,
rain, waves on water, tree leaves moving or simple lighting changes as a cloud moves
overhead. Initial Video Motion Detection systems experienced many false alarms or
alarms of pixel changes that were not actual intrusion events.
4-17.3
evolutionary step for CCTV system improvement for intrusion detection. Video Content
Analysis (synonymous with Artificial Intelligence is sometimes referred to as "Smart
Video: or "Intelligent Video". Video Content Analysis classifies images within a camera's
field-of-view. Programming algorithms differentiate an animal from a person. The
algorithms start by noticing pixel changes. Additional programming takes into account
color changes, speed changes, types of motion, patterns within the motions, and
degrees of freedom within each motion type. Video Content Analysis can filter out
potential causes of nuisance alarms such as animal movement, rain, snow, birds, and
so forth.
4-17.3.1 A variant of Video Content Analysis pertains to asset protection as opposed
to intrusion detection. In this method, the camera focuses on a specific asset such as
a safe. As long as the camera continues to see the safe than no alarm signal is sent.
Obscuration or removal of the safe can be programmed to generate an alarm signal.
4-17.4
Motion Path Analysis systems use complex computer algorithms to not only
detect motion within a scene, but also to determine the rate and direction of motion.
With these systems it is possible to identify things such as vehicles traveling the wrong
direction on a one-way street, vehicles moving faster than allowable, and vehicles that
stop moving. Motion path analysis has the ability of installing virtual trip wires with
complex rules and can be programmed to differentiate between nuisances and targets.
An example of this is vehicle within a site moving in a normal path of operational traffic
(assume left to right) can be programmed as a normal event. Movement of a vehicle in
the opposite direction (right to left in this case) can be programmed to generate an
alarm for subsequent assessment and action. This technology may be suitable for PTZ
or long-range thermal imaging systems. When motion is detected, the system zooms to
the target and tracks it.
68
Previous Page
