The simple motion detection built into every modern video surveillance device aims to assist operators, but it is restricted in its effectiveness. By definition, it only detects motion. For many surveillance system customers, the limitations of this technology are clear.
Today, video content analysis (VCA) offers new hope. These systems extend well beyond motion detection to identifying behavior. VCA learns the typical appearance of a normal scene, so it can raise an alarm if something is removed or left behind. It also can adapt as it is used, learning to ignore noisesâ€”such as rain or snowâ€”or adjusting if something in the scene changes. It will re-learn the new “normal” scene without any human intervention.
- Illegal or dangerous parking – An end-user may want to know if a car is parked in an area where it should not be, such as in front of an emergency exit, fire lane or other restricted zone. VCA can be configured to detect idle objects or objects left behind.
- Speeding – VCA can filter for speed and size if a customer would like to be alerted to cars speeding in his parking lot. The VCA system can be programmed to ignore movement below a certain speed, but alert the operator when the movement is faster and made by an object the size of a car. The technology also uses perspective to automatically compensate for remote objects that appear to move more slowly.
- Theft or unauthorized removal of objects – Your customer may want to know if a container or box is removed from a storage area, indicating a possible theft. VCA can be configured to detect when part of a scene is removed. After it is removed and the alarm is raised, the VCA system automatically will re-learn the new scene in case another object is removed. <p>
- Trip-wire – An end-user, particularly at a critical infrastructure site, may want to know if a person crosses a perimeter set as an invisible line in the scene. In this scenario, the end-user could gain visual verification of sensor alarms on an exterior fence where the sensor detects motion. The camera can send an alarm if the behavior indicates a person attempting to scale or cut a hole in the fence. VCA also can serve as an early warning system if parameters are set to indicate, for example, if a car parks near the fence. <p>
- Loitering – If your customer is concerned about graffiti, vandalism or other criminal activity on their property, VCA can be configured to detect if a person enters an area and does not leave after a specified time. In this scenario, a camera pointing at a wall where graffiti is common can be used to detect people loitering in the vicinity.
Retrofitting for Behavioral Detection
In many cases, VCA devices can be “bolted” on and require no replacement of analog cameras, DVRs, intrusion control panels or central station receivers and software. With embedded VCA, intelligence is built into an IP encoderâ€”the “bolt-on” deviceâ€”giving the encoder behavior detection capabilities. This allows the encoder to add VCA to an existing analog camera by looping video from a DVR to the encoder (see illustration above).
DVRs are not the only devices that may need to be notified when an unusual event occurs. If a central monitoring station needs to be alerted by a VCA alarm, the relay output from the encoder would go to the intrusion control panel instead of to the DVR.
Another twist is the use of an add-on device that records metadata about everything it sees. Recording metadata, which is very small in size, gives your customers a way to search by day, week or month of video recordings within minutes. Access to metadata allows users to hunt through recordings to pinpoint the moment that something happened. The point of this type of forensic search is that your customer may want to add a particular rule after initial installation, such as “warn me if someone stops here.” Using metadata allows them to search the video for that occurrence.
By only requiring a PC for initial configuration of the system, adding intelligence to an existing CCTV system becomes a much more cost-effective solution when compared to PC-based VCA.
PC-based VCA systems require operating systems and anti-virus software, both of which require continual updates. The time and expense to administer this technology quickly increases the total cost of ownership (TCO) of the system over its lifetime. Some estimate that for every dollar spent on computer hardware, nearly double the amount of time is spent maintaining the systems. Required real estate, energy consumption and power used for cooling the PC server space also must be considered in the TCO equation. In total, the cost of ownership of server hardware and software can be up to five times the initial purchase cost per year. This equation means that over three years, for example, the TCO could be as much as 15 times the purchase price.
Choosing embedded VCA eliminates the added expenses of server technology, resulting in a cost per channel that can drop to as little as $400, making it a more affordable option.
Overall, recent improvements in the accuracy of VCA have made it a more sought after feature of video surveillance systems by security directors for government agencies, casinos, airports, retail stores and other private sector businesses. While VCA is not the silver bullet that can solve all of the security-related challenges they encounter, it may be an appropriate technology to assist these monitoring personnel. And, now with embedded video content analysis, it’s easier and more cost-effective to add it to your customers’ existing CCTV systems.