Bosch Video Analytics for Security and Beyond
By Sean Murphy
Despite the huge amount of video data collected, statistics show that only 10 percent of data is ever used and most loses its value within seconds of being generated. Why use such a limited amount of data?
Many security organizations are focused on delivering the right information in case of an emergency or providing the correct evidence after a criminal act. Yet the fact is, a new video security system can be a large investment. Together with the service, maintenance and management costs associated with it, most video systems are currently considered overhead.
So how can you maximize the value of your video system? Video analytics ensure surveillance images are continuously analyzed in real-time to alert users to things that need attention. This helps organizations make sense of video data and adds an extra layer of protection by providing alerts to potential security risks before or as they occur, such as detecting loitering in a parking lot or a perimeter breach after hours.
Following are a few of the conditions that video analytics can be programmed to alert on.
Line crossing: Alert operators if a person crosses a perimeter, whether it’s a fence or an invisible line at the edge of an un-fenced campus environment; analytics can detect a person or object crossing over a line in the scene.
Illegal parking: Alert operators if a car is parked or idling in an area where it should not be, such as near a loading dock door, in a fire lane, or in another restricted zone. Analytics can be configured to detect idle objects or objects left behind to alert in this type of situation.
Loitering: Notify if a person enters an area and does not leave after a specified time, while ignoring those that innocently pass through the scene. When programmed for loitering, analytics can alert security personnel that someone may be looking for an opportunity to enter a locked door when an employee exits or that someone may be vandalizing an area. With advanced warning, security personnel can send a patrol to the area before the person actually enters the property or building or to prevent further damage.
People Counting: Alert if the number of people within an area reaches a pre-defined limit to indicate a safety concern or identify when additional staff are needed in the area. Analytics can count people and/or detect crowds in a defined location.
Speeding: Notify on vehicles speeding in a parking lot to alert to unsafe conditions. Analytics can filter for speed and size. This enables the system to ignore all movement below a certain speed, yet alert the operator when the movement is faster than that speed and by an object that is at least the size of a small car.
Color Matching: Alert operators to the presence of an object that is a particular color, such as a red vehicle that is entering the property or a suspicious person wearing a yellow jacket that is spotted in a restricted area. This can be used to notify personnel to the presence of a known security concern.
These are just some of the ways video analytics can assist to enhance security and safety.
Using Bosch Video Analytics to Trigger System Events
Integrating analytic alerts with other security systems enables organizations to use the data to trigger responses from other components of the security solution. This can increase overall security, better mitigate risk, and reduce complexity for users to improve efficiency.
For example, intelligent cameras equipped with video analytics can initiate intrusion detection system events when alarms are triggered. The analytic alert can immediately fault a corresponding point on the panel. This can prompt the panel to communicate the alarm to the central station or to send video snapshots to security personnel.
At a school, this type of alert can be used at a door for exit and entry to and from the playground. The door leading out to the playground can automatically unlock when the request-to-exit detector senses a person approaching from the inside of the building, allowing for easy egress for recess, after-school activities, or in emergency situations. From the outside, teachers, aides and other authorized individuals can easily unlock the door with the proper access credentials. However, when an unauthorized person attempts to force open the door from the outside, a nearby camera with on-board video analytics that is programmed to alert on loitering, can fault a point on the control panel, triggering a text or email alert with a video snapshot to be sent to the principal or school resource officer (SRO). The notification can also include the IP camera’s DNS or IP address, allowing the principal or SRO to connect directly to the relevant camera simply by clicking the link. By delivering this type of meaningful information, users gain the ability to easily monitor security events in addition to the central station.
In a retail store or warehouse environment, video analytics can also alert to an emergency exit that has been blocked by a pallet of boxes or by another object. Using an idle object rule, the IP camera with analytics can alert when an object remains in the area for longer than a pre-defined amount of time. When this occurs, the alert can fault a point on the intrusion control panel, which can then send an email or text message with a video snapshot to the store or warehouse manager. This can help to prevent code violations and unsafe conditions.
Triggering Audio Communications
Video analytics can also be used to trigger communications via a public address or radio dispatch system. For example, at a critical infrastructure or other high-security site, cameras with on-board analytics can use a line crossing rule to detect a person crossing a perimeter or climbing a fence. When integrated with the intrusion control panel, the analytic event can create an alarm, which can then trigger the public address system to automatically make an announcement in the area advising the intruder that he or she has been detected and that the authorities will be dispatched. A notification to facility or security personnel through their two-way radios can also be triggered.
If the event occurs after hours and no personnel are onsite, the event alert can raise an alarm at the monitoring station where operators can use additional cameras to zoom in on the area for a closer look. Instead of the recorded message, a live announcement can override the recording, and the operator can address the intruder directly. This can be a powerful deterrent that may prevent possible damage and theft.
The end result of both scenarios is that security teams and authorities are sent to investigate and apprehend the intruder while the entire event is recorded and preserved for forensic evidence should it be needed.
Extending Beyond Security
Not only can video analytics detect threats, alert to security breaches, and help enforce health and safety regulations, it can also do much more. The examples described up to this point demonstrate how video analytics can be used by other security systems when those products are designed to work seamlessly together. But it is also important to understand that video analytics can also bring significant benefits beyond conventional security applications.
What else can be done with the data generated by video security solutions? Data gleaned by interpreting video with analytics can do far more for businesses than security alone. It can enable organizations to re-purpose the data for new uses for the business—offering valuable insights to other departments within an organization, such as providing the ability to analyze behavior in retail stores to help merchandizers create more effective display placements.
How is this possible? Analytics can provide organizations with the additional information about video data that they need to re-purpose it for a business advantage. The cameras can interpret data directly at the source and re-assign it to help organizations make smarter decisions. This includes monitoring presence to reduce utility bills, identifying patterns in customer activity to improve sales, and distinguishing road blocks to optimize retail shop layout and increase customer satisfaction.
People counting and crowd detection can also help organizations improve customer satisfaction by monitoring for long lines or people gathered in an area, indicating additional assistance may be needed. In some cases, an analytics alert could trigger an automatic announcement from the public address system requesting an additional cash register be opened because the number of people in a checkout line have exceeded the pre-defined threshold. People counting can also be used to determine the number of people entering a store to help retailers determine trends that aid in ensuring correct staffing levels on busier days.
By providing business information that goes beyond conventional security applications, new functionality can easily be added to a video security system. In this way, video analytics can help organizations reduce costs, increase efficiency, and improve sales. The organization gains greater value from the system and a return on investment that can be measured in tangible business results. Then, the video system is no longer considered overhead.
While you already understand how your organization currently works and what requirements you have now, understanding how your organization would like to work in the future is key to determining how video analytics can maximize the value of your system.
Start by asking functional leaders at your organization a few probing questions to get them thinking about their needs in a new way. Are there areas of the facility that require greater security? Are there currently manual processes that could be streamlined? Are there notifications or other technologies that could improve efficiency of personnel and the safety of employees? Are there regulations or government mandates that need to be adhered to? Are there uses for the video data beyond security?
Overall, using video analytics as part of an integrated security system and fully using the data gained from the analytics can help you better meet your organization’s needs for security and extend surveillance data to deliver additional business benefits. At Bosch we believe that every business should be able to take advantages of these advances in technology - without the need for additional investment or a license fee. Which is why the majority of our cameras now feature on-board video analytics as standard. And because every business is different, we can tailor our analytics solutions to meet your exact needs.
Thinking beyond security opens up video analytics to revolutionize how video data will be used in the future and can take data usage to a whole new level.
Deciding Where the Analytics Should Reside
Whether the analytics are embedded in the IP camera at the edge, run centrally on a server, or hosted in the cloud, there are significant benefits to be gained for the organization. While there is debate in the industry over which method is best, the application of the analytics should guide the selection.
Analytics at the edge, or embedded in the IP camera, enables the camera to understand what it sees. In effect, the camera becomes smart, so you have a constant eye on a scene that can instantly alert to conditions that require action. Edge analytics are ideal for perimeter protection, tracking moving objects, alerting to loitering or objects removed, as well as people counting and crowd detection. Metadata added to the video at the edge also enables fast searching through hours of stored video for instant retrieval of relevant data—even for events not originally set up as alerts.
Specialized analytics, such as facial or license plate recognition, require database matching, making them best suited for server-based solutions. Here, it is important to note that edge and centralized analytics are not mutually exclusive; an organization may want to use a combination. For example, edge analytics can be used for general perimeter protection while centrally-run license plate recognition can be implemented on video streams from the parking lot or facial recognition analytics can be run on video from cameras at entry points into a facility. These technologies have uses beyond security as well. For example, license plate recognition can help with parking lot capacity and usage analysis or for automating payments for regular patrons of the lot or garage.
Cloud-based analytics are a great fit in applications where vast quantities of metadata need to be analyzed for business support purposes. For example, in a retail store, security cameras equipped with edge analytics can send metadata from video images to the cloud. Once in the cloud, sophisticated algorithms can aggregate the metadata to create trajectories of the paths that shoppers take through a store and heat maps indicating where they stop and dwell. These capabilities enable merchandisers to evaluate the success of displays and store layouts, which directly impacts customer engagement. This can also be used by operations and store managers to understand shopper wait times at checkout to ensure an adequate number of open cashiers at peak times – thereby avoiding the possibility of abandoned purchases due to long lines.
Fast-forward to the future of video analytics – the possibilities and potential within and beyond security are unlimited. Armed with this knowledge, you can use the hidden potential of video for a business advantage – one that helps you achieve 100 percent utilization of your video security data. And that is how we maximize the value of video systems.
Sean Murphy is the director of regional marketing—video systems for Bosch Security Systems.