Video analytics work by detecting movement, but when they are used outdoors without image stabilization, the movement of the camera can be a big source of nuisance alarms. That’s because it’s difficult for smart cameras to detect intruders in a scene when its whole field of view is also moving from wind or vibrations. For indoor applications, such camera shake is rarely a problem. In the outdoors, where cameras are mounted high on poles, even a slight movement can cause a huge number of nuisance alarms.
Of course, wind and vibrations are rarely slight in the outdoors. Many video-based intrusion detection systems are used along open areas that are difficult to patrol — railways or railyards, airport perimeters, national borders, seaports facing open water, and others — areas that are naturally impacted by high winds or vibrations from planes, trains, weather and machinery. Without effective image stabilization, these applications can be overwhelmed by nuisance alarms or worse, outright misdetects.
The best way to overcome the impact from wind or vibrations is to first stabilize the image electronically, before the video analytics take place. Cameras that use sufficient image processing can first stabilize the image for translation/rotation and zoom effects, sometimes refered to 3D stabilization, before the video analysis takes place.
You can see this in action with the following video, taken with a smart thermal detection camera. On the left, you’ll see the raw video as it enters the security camera and the effect of wind and motion. On the right, you see the image after stabilization with the video analytics detection. Notice how the camera is still able to detect the second intruder in the distance without triggering nuisance alerts.
httpv://youtu.be/54DopsLA3J0