Government agencies at all levels have deployed video surveillance cameras but find that monitoring live video footage is impractical. Manually locating individual incidents of interest after the fact is too time-consuming. Fortunately, recent advances in video analytics have enhanced this technology’s value to law enforcement agencies, making better and more safety and service applications available.
Video analytic applications automatically analyze video to detect and determine events based on multiple images. The best way to appreciate this is to compare analytics to traditional video motion detection. Unlike motion detection, which assumes every pixel change is significant, analytics filters out uninteresting activity or irrelevant scene changes, such as shadows or changes in weather, to ensure that only qualified alarms/events are delivered to users.
Individuals can immediately access video from any networked facility where an alarm was generated verifying the incident so they can respond appropriately. Alternatively, mobile units can receive automatic e-mail notifications and snapshots of alarms so they know what to expect when they arrive on the scene.
Beyond wasting valuable staff time, one of the biggest dangers of too many false alarms is that when real events happen, they may not be treated seriously. Compared with motion detection, video analytics drastically reduces false alarms, significantly increasing alarm integrity and responder readiness.
In one recent sensitive installation, false alarms were reduced from an unmanageable 300 per day to only four per day. Less challenging environments have reported one or fewer per day. By incorporating video analytics into video surveillance systems, government agencies have been able to address a wide range of issues. Reduce Crime
Video analytics can identify potentially troublesome events such as movement in alarm zones, perimeter breaches during off-hours, wrong-way movement, loitering and suspicious bags of objects left behind without putting staff at undue risk during the monitoring process. Video analytics not only provide alerts of suspicious events in real time, but also enable law enforcement staff to quickly pinpoint recorded activity and identify perpetrators of criminal activity.
For example, loitering in specific areas at certain times (such as city parks or loading docks) is a leading indicator of an impending crime. Using simple motion detection in such cases is likely to result in too many false alarms (someone just passing by or an animal tripping the alarm). But adding video analytics drastically reduces the false alarms, making remote video alarm verification practical.
It also enables police to be notified quickly and reliably if people in high-risk areas remain in a location for a specific period. Many localities have expanded their capabilities even further by combining analytics with mobile alert systems. In such cases, law enforcement staff may receive alerts on mobile devices while out on patrol so they can arrive at the scene of suspicious activity in minutes.
One locality has combined the use of cameras with video analytic functionality for its city parking lots. The analytics are configured to detect whether a person standing next to a car is its true owner, or a potential thief or vandal. The determination is made based on the average time it takes a person to open a car door. If a 45-second threshold is broken, the system sends an alert and the matter is investigated.
In the area of homeland security, similar functionality can be used to determine if suspicious objects are dropped or vehicles remain motionless in high-security areas for a certain period. Items as small as 15-by-15 pixels can be detected and tracked, although law enforcement agencies may experience higher rates of false positives tracking small objects in high traffic areas.
In each of these cases, real-time alerts can tip off law enforcement staff that suspicious activity is underway. And if mobile alerts are set up, first responders may be able to arrive at the scene before suspects can flee, or at least before witnesses disappear and evidence is compromised.
In the area of public safety, video analytics can alert security personnel to things that may compromise emergency procedures, such as objects blocking emergency exits. Many municipalities have also placed surveillance systems coupled with video analytics in transportation settings to protect passengers waiting on train platforms from theft, assault and other accidents.
Prevention as well as rapid response to these safety, security and liability threats can save transit organizations significant amounts of money in reduced operational down-time while increasing ridership. Among useful video analytic applications targeted for public transportation are those that sound alarms when people attempt to cross train tracks or walk into restricted areas.
While visibly installed cameras serve as crime deterrents, they may also encourage criminals to attempt to defeat the system by spray painting lenses, covering them with objects or blocking their views. In some communities, this has reduced their surveillance effectiveness. Fortunately, applications can now be deployed that alert authorities when a camera has been moved or its view obstructed. Optimize Operations Efficiency
Video analytics have also been successfully deployed by government agencies to gain valuable information regarding operational inefficiencies and decreasing costs. For example, video analytics enable people counting and identifying traffic flow patterns that can help plan government buildings, roads and intersections, and help to staff locations appropriately. Video analytics can also help enforce payment of parking tickets and other traffic violations.
In one municipality, video cameras using analytics are trained on traffic lights and alert authorities when vehicles go through a red light. Real-time mobile alerts are sent along with the video shot to identify transgressors. In cases where individuals can’t be pulled over, vehicles can often be identified by their license plates.
Analytics have also been deployed to analyze high-traffic roadways and intersections, cars stopped on the shoulder, and entry and exit traffic patterns at city lots. Further, by analyzing traffic anomalies caused by public holidays or events, municipalities can better plan for required traffic assistance. Conduct Rapid/Accurate Forensics
Real-time video viewing is generally impractical, and it is often impossible for law enforcement agents to check out every alert because of their frequency, the remote location of a site, etc. In such cases, it is crucial that law enforcement be able to quickly review pertinent video.
Without analytics, that process may take a day or longer to go through weeks of video. With analytics, by specifying the area of camera coverage and the event sought (e.g., cars entering and leaving a parking space or cars stopped for more than five minutes in a crowded city center), it may take only minutes.
Maintaining public safety remains a difficult and expensive problem. But with the use of video analytics, public safety organizations can cost-effectively reduce the number of false alarms while increasing security and operational efficiencies. In fact, when taking into account the total price of a video surveillance system, it may cost only 5 percent more to add sophisticated video analytic applications that exponentially increase public safety. Marc Holtenhoff is the chief executive officer at Aimetis, a leader in intelligent video surveillance products and services.