An OpenALPR agent running on a next generation Intel i5 CPU for desktop computers can process approximately 20 frames per second of video at a resolution of 720p.
In other words, an agent can handle three cameras for low speed vehicles, two cameras for medium speed vehicles and one camera for high speed vehicles.
The memory and disk requirements are relatively constant. Once OpenALPR is initialized, the process uses approximately 500 MB per analysis thread (CPU core), and the memory usage will not increase. The agent records JPEG images on a disc as the plates are recognized.
To start we must have updated our server, for this we execute the following command:
sudo apt update
later this other:
sudo apt upgrade
Once we have updated our server, we execute the following command to start the installation:
bash <(curl https://deb.openalpr.com/install)
Select the install_agent option by pressing the SPACE key (an asterisk must be placed in that option to confirm that it is selected), then to continue, press ENTER.
Then he will ask us for permission to continue with the installation, here we only press the Y key.
After installing the agent will ask us to register it, if we are installing the agent in the same Ubuntu server where OpenALPR Server is installed, we will ask if we want to register it in localhost, for this we type the letter Y.
Then he will ask us for the username and password admin of our OpenALPR server
After this, our agent will be connected to the server. To verify this, we enter the web interface of the OpenALPR server in Configuration / Agents and it will appear in the list of agents.
Configure Agents and add cameras
First we will click on the agent name to modify it:
Then we click on Configure
We will click on the option Add Stream
We post the URL of the RTSP of our IP / DVR / NVR camera
To know the RTSP addresses, visualize the following article:
To verify that the url is correct, we click on the Test button.
If the information is correct, a message from Test Successful will appear, then click on Save Stream
By clicking on the name of the camera we can edit it:
After this we click on Configure
Here we configure the inclusions and exclusions. This helps us so that the camera does not process the entire image, if not just regions of interest. Then click on Save Mask to save the changes.
In this section we must place North America so that the software recognizes 12 x 6 inch plates.
The number of processing cores will control the amount of CPU allocated to the LPR process. The more processing cores you provide (up to the number of CPU cores in the system), the more frames per second (fps) you can process. Higher fps will generally increase accuracy and allow plates to be detected in vehicles that move faster.
The disk quota will control the amount of space reserved for storing vehicle images and license plates. It will work as a displacement buffer, so when it runs out of space, the oldest images will be deleted.