"Distribute," "Divide," and "Digitize"
When we utilize surveillance cameras, we usually want to distribute them for viewing, stream them for viewing and recording, and digitize the signal for some degree of posterity whether it’s 24 hours or 24 years.
Distribution prior to 2000 was mainly done through switchers, large cross-point devices which enabled you to decide which input goes to which monitor or recorder. The signal was then divided accordingly to those personnel who needed to view it and those machines that needed to record it.
While these switchers are still applicable in today’s digital realm, and still expensive, most switching is now done locally within a digital framework of servers which provide unique functionality and flexibility. This is strictly due to the signal processing capabilities of advanced electronics to distribute, amplify, switch, compress, divide and record in different formats many all at the same time.
What’s Its Use?
Is it theft, surveillance of gang activity, urban shot detection, or historical archiving to confirm that an officer did not attack an inmate? These are very important questions, so let’s make a schedule of important aspects that your particular case might need from digital surveillance and recording.
Here’s the list I would recommend:
- Image capture
- Passive surveillance
- Active surveillance for suspicious activity
- Identity matching
- Behavior analysis / Video signature
- Forensic analysis
Image capture is fairly straightforward and would be used to gather evidence such as shoplifting. While the video may show the person leaving a distribution center, it is the actual process of theft, such as putting it in a pocket or overcoat or physically carrying the item out, that’s important.
The image capture device is usually sufficient, while a full length video is rarely important. It’s the wide view and usually the short activity captured that is most important.
Passive surveillance is used in parking garages available, not often monitored, but only used at certain times and if someone happens to capture or be aware of an incident. This is similar to most surveillance systems, but becomes less passive when employed in a situation such as a subway.
For example, the subways in Philadelphia during the 1980s had a sophisticated camera system including hundreds of cameras throughout the subway stations. This was due to an assailant raping and severely beating a woman who ended up with brain damage. The surveillance system was used on an active basis and was able to capture many acts which would have otherwise gone unnoticed.
Active surveillance is used for active monitoring of suspicious activity and is a crime prevention method other than simply affixing cameras and hoping that the honest people will stay honest. While identity matching through active surveillance is the use of cameras in a facial recognition situation using advanced software in today’s marketplace. Casinos have used identity matching in their video surveillance since the 1970s.
When I was involved in the casino industry, the use of binoculars, advanced CCTV cameras and recording mechanisms would zoom-in from every angle to a player with suspicious activity or those who were playing in a conspiracy to outwit the casino. The identities would be matched through facial recognition, markings, dress, or accomplices, and compared to a crime database that was available through the Casino Crime Commission. Alert posters were hanging in front of the surveillance personnel who worked separately but in tandem with the Casino Control Commission and the enforcement officers.
My own personal experience was by watching a woman nicknamed “Mrs. Wong” (since no other alias or full recognition could be found) win approximately $225,000 from the Golden Nugget in the 1980s. In a period from 0700 hrs to 1130 hrs, Mrs. Wong won at the blackjack table one-on-one almost every time. The casino did everything in its power to attempt to identify this person through national crime information and other databases. They did not succeed.
Behavior analysis and video signatures are a new software mechanism by which certain algorithms are learned by the system software to remember activities which are acceptable and to analyze any differential to alarm those activities which are out of bounds.
It’s kind of like teaching a dog what is the normal thing to do over a period of time. But, if someone would act in a peculiar manner, say walking the wrong direction or walking across instead of along side, the dog would bark. Think of your digital software system as a barking dog. Unfortunately, analytics has not followed Moore’s Law. In fact, it is still quite pioneering.
Forensic analysis is the last use. Obviously, forensic analysis has taken great strides as we have seen with the subway bombings of London, the Boston Marathon bombing, and the lack thereof from Paris and Brussels. Digital data retrieval in forensic applications is powerful.
The importance of forensic analysis, historical archiving, and registering the digital data in a uniform and highly available manner is much more important today than it has ever been before due to the ubiquitous use of surveillance cameras.
The London subway system, installed in the late 1990s and of which many stakeholders took part, included over 9,000 networked surveillance cameras on dedicated secure networks that shared surveillance from multiple points throughout the perimeter of London, and transferred images back and forth through sophisticated ATM (Asynchronous Transmission Mode) networks. There were 30,000 cameras in London’s system by 2000.
Philadelphia now has over 1,200 cameras accessible through federated receipt of non-owned cameras by commercial and government entities; while in the City of New York; I expect that over 1,000 cameras are installed every day in homes, businesses, and government entities.