The purpose of video surveillance is to monitor and control scenes to ensure the safety of people and property. The visual component is ideal for many purposes but has the problem of generating many false alarms. The new Safire devices equipped with Truesense technology make it possible to achieve a high level of reliability in the detection of incidents even compared to some intrusion systems.
Motion detection and smart detection rules
Historically in CCTV the only way to detect motion was by checking the changes between one frame and the next, which is called a video sensor. This detection method is very sensitive to false alarms generated by environmental changes or changes in lighting, as well as movement of leaves, branches, animals or other irrelevant elements.
In more recent times, the ability to analyze certain events, known as smart detection or video analysis, has also been added. In Safire we find the VCA rules in IP cameras and DVRs. These rules work like motion detection, but under an imposed logic, such as a line crossing, an intrusion in an area, a sudden change of scene or even a face detection or an audio exception.
Deep Learning and Artificial Intelligence applied to CCTV
The era of Big Data and quantum computing has brought artificial intelligence closer to video surveillance systems. Safire has successfully created an intelligent detection algorithm using Deep Learning to analyze hundreds of cases of real detections. The result is the ability to apply a filter to VCA detections that allows humans and vehicles to be distinguished from other objects that generate false alarms.
IP Cameras, NVRs and DVRs with TrueSense technology
In order to understand the detection of data analysis, neural networks are used in powerful computers. This technology in a recorder or in an IP camera would be extremely expensive, however, it is possible to extract a detection algorithm from this analysis that has already been "trained". Which, along with a powerful graphics processor built into all TrueSense equipment, manages to reduce the costs close to what could be a standard device.
The simplicity of TrueSense
Unlike video analytics software that requires complex and repeated configurations, with Safire TrueSense everything's much easier. A VCA rule is created and selected if human and/or vehicle filtering is desired. Additionally, since the cameras do not perceive depth, it is necessary to establish the minimum size of the objective in the background of the image and the maximum size of the objective in front of the image, by drawing simple rectangles.
In other words, instead of receiving multiple notifications of false alarms, you will only receive notifications generated by objects that are important, such as humans and vehicles, which increases the reliability of the security system.
Login
