Scope of application
The use of neural networks in access control systems and at checkpoints
Main tasks to be solved
- License plates / brands / models of vehicles recognition
- Initiating sensor data recognition
- Opening upon license plate recognition (black / white lists)
- Event log storage and on-demand access
- Formation of reports on the number of registered events for the period
- Integration with third-party systems via Wiegand
- Sending a passage signal to external systems
- Fixing manual opening
- Watching videos in LIVE mode
In addition to the vehicle detection system and license plate recognition, a hardware-software complex for automation of vehicle entry / exit zones has been implemented. Let's consider its functionality in more detail.
Now the most common solution is a booth with a security guard and a barrier or gate at the entrance area. Sometimes, the systems are equipped with video analytics and license plate recognition, but the white list is managed within the ACS and is not available to various user groups, in fact, participating in the process.
Often the process is structured as follows:
- The tenant is waiting for some vehicle
- He asks for the plate number from the vehicle owner
- Calls the security and dictates the number
- The license plate number is added to the white list or, upon the appearance of the vehicle, it is verified by a security
But what if the process is automated? For example, to provide all residents of the HOA with the opportunity to independently form temporary and permanent passes, as well as manage them. In fact, this makes it possible to implement fully autonomous access systems, both within the framework of the house / cottage infrastructure, and within the framework of the automation of industrial enterprises.
We have implemented such a system and would like to share its functionality. The demo version is located at the link - http://edgetsz.ttcsoft.ru:3891/, login 79283672556, password 12345.
The system consists of the following blocks:
- Object video analytics (vehicle detection, license plate detection, license plate recognition, trajectory analysis of the driving direction)
- REST API for connecting various clients (web, mobile applications, etc.)
- Web interfaces for the following user groups: clients (tenants), HOA (manages users, providing tenants with access on demand) and MC (manages HOA)
- Mobile client for iOS (we plan to implement Android in the process of system development, but as a temporary solution, the adaptive web interface allows using the system on any platforms)
It is important to note the modularity of the system, that is, we are currently using our own video analytics, but if necessary, we can connect solutions from other vendors. Usually, the question arises here about the quality of detection and recognition of license plates. We will answer it with a short demonstrative video:
Due to all the same modularity, the system can be easily and quickly expanded and allows integration with any consumer. Now it is an iOS mobile application and a web interface, but if necessary, it can be any previously installed and integrated systems.
What does the application look like?
Hardware solution base
Outdoor Box Micro, 2 GB is ideal as a hardware base (more details - https://comboxvideo.ru/). One device can process up to 2 streams from network cameras (RTSP) and, at the same time, act as a local web server and data source for the web interface and mobile applications.
The Outdoor Box Micro comes in an IP66 industrial enclosure and looks like this:
The use of such devices makes it possible to equip objects without a global change in the existing infrastructure
Result and tasks to be solved
As a result, we have the opportunity to implement autonomous perimeter security points within the framework of the house/cottage infrastructure or within the infrastructure of various enterprises (equipped, for example, with several geographically remote zones that require centralized management). Modularity allows implementing solutions with integrated payment and vehicle time tracking in the zone.
