Teaching Cameras to Zoom … “Intelligently”

Intelligent CameraMap Based video management systems have the ability to control security cameras directly from the map-based user interface.  These map-based VMS, also referred to as geospatial VMS, have many advantages, but one often-overlooked advantage is the idea of “intelligent camera zoom.”

Scenario Consideration

A typical security scenario is the need to redirect a camera from a distant zoomed-in view at one location to a closer zoomed-out view at another location.  For example, a security operator is monitoring the main vehicle entrance, approximately 200 meters from the camera.  He is then notified of an invalid badge attempt at the main employee entrance, only 30 meters from the camera.

Positioning a camera from one scene to another in response to an event may first appear a simple task, but it actually entails several non-trivial steps.

Positioning a camera from one scene to another in response to an event may first appear a simple task, but it actually entails several non-trivial steps.

Your initial reaction might be that this is a trivial action, just redirect the camera to the new position.  However, if you merely pan the camera to the employee door, you will likely get a very zoomed in view of a rock or a tree.  You will have no feel for the view in relationship to the employee door, so your next action will most likely be to zoom out to gain your bearings.  Now with a better understanding of the current camera view relative to where you want to be looking, you will make a slower pan move to get to the main employee door.  Once you have the scene in view, a final tilt and zoom adjustment are probably in order.  Now that you have negotiated the camera to its intended location, you hope that the target has not left the scene.

Camera Control – What Seems Trivial, May Not be So Trivial

As this scenario illustrates, this deceptively simple sounding maneuver actually entails three non-trivial actions: 1) Panning the camera to the direction of the new location, 2) Tilting the camera to the correct angle to see the person at that location and 3) Zooming the camera to a zoom level that affords an acceptable view of the scene.  Talented security operators are trained to perform these types of maneuvers, but as the number of sensors and the complexity of the security systems increase, it is more difficult to insure you have the luxury of this type of talented user control.  Even then, the infinite number of scenarios still makes it a challenge for even the best operators.

Intelligent Pan, Tilt and Zoom

Geospatial VMS systems understand how every pixel in the camera’s view relates to a specific latitude, longitude and elevation.

Geospatial VMS systems understand how every pixel in the camera’s view relates to a specific latitude, longitude and elevation.

One way to address this operational need is equipping your VMS with geospatial capabilities.  Simply put, geospatial in the context of video management systems means that the system understands how every pixel in the camera’s view relates to a specific latitude, longitude and elevation on the map.  This allows the system to point a camera directly to point of the map.  Without getting into the underlying geometry and video analytics, this same understanding of camera pixel to geographic location also provides the basis for understanding not only the desired tilt, but also an appropriate zoom to use for a target at the desired location.  So in essence, adding intelligence to pan, tilt and zoom.

Simplifying Camera Maneuvering with Map-Based Control

So let us go back and revisit our earlier scenario.  In a map-based (geospatial) VMS, when a security operator wants to redirect a camera from a distant zoomed-in position at 200m (the vehicle entrance) to assess a potential target at the main employee door, 30 meters away, the operator merely selects the camera and clicks on the location on the map that corresponds to the main entrance.  The system calculates the appropriate, pan, tilt and zoom and redirects the camera in one move.  In a fraction of the time required for the operator to manually maneuver the camera, it is now in the desired position providing video and waiting for further action from the operator.

Enabling Camera Automation

It is easy to see the usefulness of map-based controlled, with its automated adjustment of pan, tilt and zoom.  By combining this capability with other technologies, such as slew to cue and camera auto follow, one can see how far surveillance systems have advanced technically.   Now a VMS can received alarm information for a variety of perimeter sensors (Fence, Radar, Proximity, Access Control, RFID), provide the camera the appropriate pan, tilt and zoom commands to center that event in its field of view, and then engage camera auto follow to maintain surveillance of the target.  The value proposition being that none of the actions requires operator intervention.

Adding the Capability

If you are in search of a new or replacement video management system and have the desire to automate some of your security efforts, certainly take the time to inquire about map-based systems.  Perhaps you are not in the market for a new system.  Not a problem.   Geospatial capability can be added to existing video management systems using edge devices and add-on servers.  In most cases, existing cameras augmented with geospatial capability can achieve this level of automation as well.

This entry was posted in Geospatial, Perimeter Surveillance, Security Sensor Integration, Video Analytics, VMS and tagged , , , , , , , , , , , , , , , , . Bookmark the permalink.