Geospatially-integrated surveillance systems

Everything is going geospatial…Amazon, Twitter, Google, Face Book, Pinterest. Although it may manifest itself under other names - geo-intelligence, GEOINT, map-based, geo-referencing, geo-analysis - all are terms that refer to the concept of geospatial … one that is defined as relating to, or denoting data that is associated with a particular location. When used for physical security, geospatial refers to WHAT and WHERE, thereby providing security professionals with an enhanced situational awareness that allows them to react to situations quickly and more effectively.

Map-based presentation

So what is a geospatial surveillance system and what value does it provide to a security provider? In a nutshell, such a system is one that relies heavily on location data for primary display (aka a map-based user interface) and for analysis and integration of sensor signals.

One of the biggest advantages of a geospatial system is that it provides a high level of situational awareness, allowing for a map-based representation of security sensors. A user can view the user interface and quickly gain an understanding of the current state of events. This includes the ability to dynamically show a security event including the intruder’s location and the associated track.

A geospatial system also provides for user interaction through this same interface (Fig.1 is an example of such an interface in use at a seaport). This means sensors can be queried and controlled by interaction with a map-based display. Sensors that are used to interrogate or react to events, such as fixed cameras, PTZ (Pan–Tilt–Zoom) cameras, acoustic devices and spotlights, can be controlled from the user interface, using the real-time intrusion track as means to direct sensors to the exact geographic location.

Geospatial sensors

Key components of a geospatial surveillance system are its sensors. When we think about those that understand inputs and outputs in terms in real space and time, perhaps the most obvious that come to mind are radars and GPS. These sensors yield large volumes of spatio-temporal data including speed and even acceleration as well as location.

A less obvious group of geospatial sensors are smart phones, most of which have a GPS capability but are often overlooked from a security standpoint. Smart phones in the hands of first responders can be a basis for coordinating assets during an event and their cameras can be employed to gather location-based images that can be documented and shared within the context of the geospatial surveillance system. Fig.2 gives examples of these geospatial and "pseudo-geospatial sensors

There are yet other sensors that are partially geospatial or can be geospatially-enabled. For example, fence intrusion systems can provide an exact linear position along the fence or, at a minimum, identify a specific “zone” where an intrusion occurred. Cameras also fall into this category. Out of the box, they are not a geospatial sensor, but they can be enabled with this capability through a process called “geo-referencing.” This process involves superimposing the camera’s field of view onto a terrain map, thereby allowing the camera to understand where each pixel exists in latitude, longitude and elevation.

A geospatially-enabled camera essentially acts as a short range visual radar. Once a camera understands where a specific pixel resides in real space, it has the ability to identify an object’s real size and real speed (Fig.3). Combining that with visual information, such as aspect ratio, compactness, colour and texture, cameras can now classify objects in categories, such as, animal, person, car, truck, boat, aircraft, etc.

Another group of sensors can be considered “pseudo-geospatial.” These include those that have a physical location that has meaning to the security or operations manager. This would include access control systems and where the location of a specific door or gate has value. Other sensors that fall into this category are proximity sensors, RFID and, of course, sensors that detect heat, water or gas emissions.

Map-based sensor control

In addition to being able to display sensor location, sensor state and target information on a map-based GUI, a geospatial system also enables user interaction through the same interface. This means sensors can be queried and controlled by interaction with the map-based display. Sensors typically used for event interrogation: PTZ cameras, laser illuminator, spotlights, range finders and acoustic devices, can also be geospatially-enabled for monitoring and control purposes using latitude and longitude information.

So in addition to dynamically displaying the type of intrusion, the location and dynamic tracks, a geospatial system can use these same tracks as cues to point interrogation sensors directly to the location of the security event. Some sensors can then automatically follow the target as it moves within the scene. This ability to both monitor and react to situations in a map-based content makes these interfaces very intuitive and efficient for security personnel to interact with.

Sensor collaboration

A second advantage of a geospatial surveillance system is the ease with which sensors can share information and collaborate. Those that are geospatially-enabled share a common operating picture based mainly on a physical coordinate system. This means they can effectively communicate with each other using a common geospatial command set. As a result, geospatial cameras do not need to understand radars’ range & bearing and a LRAD (Long Range Acoustic Device) does not need to compute the reflection distance of a fence sensor. They exchange position information in terms of elevation, longitude and latitude. This helps automate many tasks that previously required operator intervention such as steering cameras to manually track targets and assets. As an example, in Fig.4, the location of an intrusion, detected by a fence sensor, is communicated to the geospatial surveillance system. This automatically swings a PTZ camera to the location; triggers automatic camera following, and classifies and confirms the target as human. An LRAD is then automatically steered to the intruder to provide a deterrent message – its urgency changing based on the location, direction and speed of the intruder. Those monitoring the event then use the map-based GUI to locate the nearest guard

Track fusion

In the context of geospatial command and control, both radar and video have the ability to detect and track targets in real time. But what happens when a single target passes through several sensors, potentially creating multiple tracks on the map for the same target? Geospatial systems allow for what is referred to as “track fusion.” This is a means by which several sensors with overlapping area coverages take account of accuracy and update rates of each sensor to determine whether it is a single target, or multiple targets. The data from the various sensors is then fused to provide the user with a best case, single track and ID for that event.

If one of these targets is a GPS-enabled asset, such as a company vehicle or patrolling officer, the system can automatically identify the object as a friend and avoid tripping any nuisance alarms as part of, say, regular patrols. The system then denotes the object as a “friend” through the use of an appropriate icon on the map and it passes this information to other sensors as the “friend” moves through the property. This capability is typically referred to as “IFF” or ”Identify Friend or Foe.”

Forensic capability

Finally, a geospatial surveillance system can provide enhanced forensic value. The ability to assign locations to events carries over into alarm searches and forensic efforts. Each alarm contains location information, so events may be investigated and reviewed based on a region or location, versus trying to identify the sensors involved. This can be extremely helpful when investigating specific incidents or exploring trends.

Geospatial surveillance systems also have the ability to replay a full map-based scenario; essentially replaying everything that an operator sees on-screen, including alarms, sensor positions and map locations. At any point the reviewer can drill down to get the detailed data associated with the event or sensor and perform map-based activity searches, as in Fig.5.

Conclusion

“Everything happens somewhere” and today’s security sensors and applications understand the value of location data, being able to track and share it with other sensors. As security surveillance systems become more complex, and as security professionals are required to process and assess more and more data, user-friendly surveillance systems with geospatial capabilities provide an enhanced level of situational awareness.

Eric Olson is Vice President of Marketing at geospatial video analytics developer PureTech Systems (www.puretechsystems.com).