Video analysis for LPR in terms of processing is more demanding than recording/viewing from a camera. Thinking about flexibility, the system has an innovative processing architecture, which is the distributed processing architecture.
The system allows LPR processing of cameras that are recorded on the camera server to be done on one or more computers that have the LPR Server module. The big advantage is that with this flexibility the recording server is not overloaded and does not need to be a "super-machine".
The LPR server automatically checks the computers with the lowest processing power and performs "load balancing", that is, it distributes the processing of video analysis in order to leave all computers with the lowest possible processing capacity, as long as all servers have sufficient licenses.
What determines the number of LPRs that can run on the same server is the processing capacity of this server. The larger the processor, the greater the possibility of running multiple cameras at the same time. The system processes LPR in fixed and PTZ IP cameras and in fixed and PTZ analog cameras, as long as these are converted through encoders or DVRs integrated into the system.
See the diagram below:
In the diagram above, the VMS Server records images from the cameras and sends them to the LPR Servers, which in turn perform the analyzes and return the metadata (information on recognized license plates and vehicles). Between the LPR Servers there is load balancing, if configured for this. When the metadata returns to the VMS Server, it sends the metadata and events to the VMS Clients (Surveillance Clients).