printer friendly version

Visual intruder and perimeter security

Physical perimeter security can be defined as a system (or systems) together with technologies that protect people and assets within a facility and its grounds, by blocking unauthorised physical intrusions across the set perimeter.

Numerous defence ‘layers’ should be equipped to protect the boundary and should comprise of the following: the holistic site and property perimeter, e.g. the fence line; the inner territory perimeter, specific buildings or key infrastructure; the building fascia perimeter, the external building shell and the internal perimeter, and the internal space where restricted access is necessary. Each layer should help delay, deter and detect intrusion.

Axxonsoft uses neural network analytics to develop products that are proficient in object classification. One of the neural network filters is the Object Tracker that analyses moving or abandoned objects to filter out false alarms from situation analysis detection tools. The filter recognises specified types of objects, e.g., humans or vehicles. And now with AxxonSoft’s deep learning technology, using artificial neural networks, it has been adapted to tell the difference between humans and animals too.

The neural network-based object classification forms part of the analytical system. This automates the decision making for standard situations but sends high-risk events/information to an operator in a CCTV control room to take action.

Using Intellect PSIM

A wildlife residential estate, in the east of Gauteng, is in close proximity to the national highway which posed a security risk of intruders. Axxonsoft Intellect PSIM software was installed to address the risk of unwanted access around the perimeter and also allow ease of access for residents. Intellect software platform allowed for the integration of existing CCTV hardware as well as new hardware. It was able to secure the perimeter, which meant less physical guarding, and consistent monitoring of the site.

Share this article:

Further reading: