To get optimal ANPR and recognition results a couple of factors need to be taken into account.

Follow the steps below to get optimal results.

 

 

1.Optimisation starts with the proper positioning.
The chapter positioning of the Module-X is essential for license plate recognition. A good recognition starts with a proper camera projection.
 

2.ANPR inspection tool
The tool visualizes ANPR and is used during installation for camera positioning and adjustment of camera settings to get an optimal ANPR result.
It displays the width and height of the license plate (recommended width between 100-300 pixels), the sharpness and the brightness.
 

3.Increase framerate and select region of interest
Selecting the Region of Interest (ROI) speeds up ANPR and classification tasks enormously.
Region of interest this minimizes the amount pixels and the amount of random background that needs to be analysed. It will improve performance and increase framerate. In the default configuration of the  Module-X, both the color and the ANPR sensor are used to perform license plate recognition. Both sensors should be cropped for optimal performance.
Select ROI ANPR sensor and Select ROI color sensor

 

4.Set the proper threshold for ANPR
Setting the ANPR threshold defines the number of equal recognitions of a lp before it is considered an approved recognition result.
The higher the threshold the more certain the reading is accurate. It is used to capture license plates that are difficult to read even for the human eye.
When the threshold is set to 4, a license plate has to be recognized 4 times before it is approved.
 

5.Setting minimal license plate width
The recommended width of a license plate is between 100-300 pixels. Depending on the mounting position and the cropped region of interest, a license plate can be too small when it enters the scope of a sensor.
Setting the minimal width will prevent the ANPR engine to read license plates that are too small.