Viinex offers an affordable solution for building the automation software for parking or a checkpoint. It implements the video acquisition and video storage features, and works with virtually any model of IP camera or IP video encoder.
Besides the video management functions which are covered in other sections, Viinex has an implementation of the module for vehicle license plate recognition.
This module in Viinex 2.0 has several advantages over competitive implementations:
- LPR module in Viinex 2.0 does not require any settings for camera or license plate geometry. This makes a deployment very easy: license plate recognition works immediately after software installation; no extra adjustment steps are needed.
- When installed, an instance of LPR module is not bound to a specific video camera. This makes it possible to recognize car license plates from a several IP cameras with a single instance of LPR.
- The implementation of LPR engine in Viinex 2.0 takes for license plate recognition a single image, not a video sequence. The use of that engine comes down to acquiring an image from the camera, sending that image to the recognizer and getting back the recognition result. The integration of a license plate recognition functionality has never been easier.
- The interaction with LPR engine in Viinex 2.0 is performed via HTTP which is essentially used as a transport for remote procedure call. This allows for issuing a multiple recognition requests to a single LPR instance simultaneously, without the need for external serialization of calls or any other precautions which are usually required for in-process recognition libraries. The number of CPU cores which recognizer implemented in Viinex 2.0 can utilize is configurable, and this allows the LPR implementation to scale as required over multiple CPU cores / multiple checkpoints served by one instance of LPR engine.
LPR engine built into Viinex 2.0 supports templates for vehicle license plates issued by EU states and similar (1-liners with digits and Latin letters of equal height): Belarus, Kazakhstan, Russian Federation and Ukraine. Contact us to request information on recognition of license plates issued in other countries.
To use the LPR engine implementation built into Viinex 2.0, the client application needs to know when a license plate should be recognized, i.e. – when it appears in sight. Generally, this is not always possible without image analysis, but for parking / checkpoint applications it is usually known when a car approached the checkpoint, and this knowledge can be used as a signal for recognition.
In comparison with the case when one engine is bound to every video channel and the recognition is performed continuously on the video stream, the approach for license plate recognition task implemented in Viinex 2.0 saves CPU resources, which makes it possible to use cheaper CPU to process same number of checkpoints.