Traditionally, motion detection has been done by pixel shift tracking. This causes a large number of false alarms produced by irrelevant elements, such as the movement of a leaf or a branch. Therefore, Safirehas developed TrueSense technology, capable of filtering detection by humans or vehicles. This improves the reliability and efficiency of recordings.
What is TrueSense?
The TrueSense algorithm is based on Deep Learning and its configuration is very simple:
When establishing a VCA rule you have to select if you want the alarms to be produced by vehicles, humans or both. At the time of drawing the rule, it will also be necessary to choose the maximum and minimum size of the elements to detect.
With this, it is possible to filter 90% of false alarms, allowing recordings by event in a reliable way. This ensures that the device memory will be used in relevant situations. In addition, it allows searches by element, either by vehicles or humans.
TrueSense is especially useful when establishing intrusion rules in area or line crossing, this combination achieves the protection and surveillance of perimeter zones.
Speed Dome Cameras
For the first time this technology will be included in Speed Dome cameras, this will improve the ability to detect and track objects.
These cameras can perform motorized trajectories with horizontal and vertical movement, and the lenses have an optical zoom of up to 32x.
All of this is essential in video surveillance of large and open spaces. However, covering such a large space can cause a multitude of false alarms. This is where TrueSense becomes the key to the situation.Thanks to this technology, rules can be created for perimeter protection with filters of humans and vehicles, and these rules will remain fixed for the chosen area, even if the camera performs different trajectories.In addition, most Speed Dome cameras have Autotracking, the camera is able to move following an object if it has been configured so. TrueSense makes the difference when it comes to detecting the object to follow, since it can do with precision thanks to its training Deep learning feature.
Login
