LPR Parking System: What Actually Happens at the Gate

An LPR parking system is judged at one place: the entry lane. Drivers do not care about model architectures. They care whether the barrier opens when they arrive, and whether it stays closed for the car that should not be there. This guide walks through how a license plate recognition parking setup behaves at the gate, where it usually fails, and how VEZHA® Smart Parking handles those cases.

What an LPR parking system does

At its core, an LPR parking system turns a camera at the entry or exit into an automated checkpoint. The chain is short:

1. Capture. A camera streams video of the lane, over ONVIF or RTSP or from an existing VMS.
2. Recognition. The analytics engine detects the plate and reads the characters. VEZHA LPR also returns vehicle make, model, colour and country identifier.
3. Decision. The plate is checked against lists: staff, visitors, VIP, fleet or a blacklist.
4. Action. The system opens the barrier, sends a notification, starts a timer for the stay, or writes the event to the archive.

Parking-specific software adds the layer that a plain plate reader lacks: who the vehicle belongs to, how long it may stay, and what should happen when the rules are broken.

Lists and tags: where the business logic lives

Most of the work in a parking deployment is not recognition. It is configuring who may enter, when and where. In VEZHA Smart Parking this is done with:

• Custom tags such as management, staff and visitor, so each group gets its own rules.
• Links to the company structure, so a vehicle can be assigned a designated space in a predefined zone.
• Allowed stay intervals, with notifications when a listed vehicle appears, stays too long or is not registered at all.
• A business logic constructor for scenarios that go beyond "open on whitelist".

For a typical office or residential site, start with three lists: permanent users, temporary visitors and blocked vehicles. Add more only when a real scenario needs it. Fewer lists mean fewer surprises for the guard on the night shift.

Search after the fact

Operators spend a lot of time answering questions after an incident. Search by full or partial plate, with filters, saves hours of video review. One example from our product page: a blue SUV in the north section with "5555" in the plate over the last five days. The archive can sit on MySQL, PostgreSQL or Oracle Database, which matters when the customer's IT team already runs one of them.

The re-read problem at the barrier

Here is a situation every parking operator knows. The camera reads the plate and the barrier opens, but the driver hesitates. Or they stop too far back. The site times out, the barrier closes, and the car is still in front of it. The camera keeps reading the same plate, but the system treats it as a duplicate and does nothing. Someone has to press a button or the queue grows.

VEZHA Smart Parking handles this case:

• When the same plate is detected again by the same configured analytics source, the site trigger runs again, so the barrier can re-open.
• For these same-source repeats, the site-level timeout is ignored.
• Re-detections that come from a different analytics source keep the previous logic. This prevents one car from being let in twice through two different lanes.

The distinction between "same source" and "different source" is the point. It removes the false stop at a single lane without opening a loophole across lanes.

Setting up a multi-lane entry

Each entry or exit point is configured as a site. In the site settings you choose which analytics serve it and set an interval (timeout): the time during which a recognition from another camera is ignored. A few rules keep this logic working as intended:

• One camera per lane. A single camera watching two lanes is not a typical installation. Limit each camera's recognition zone to its own lane.
• Direction of travel. Where needed, set the direction of movement in the site settings.
• Separate cameras for entry and exit. The usual setup is one camera facing the entry and another facing the exit. The site interval keeps a read from the other camera from triggering the gate again.

Plates that are hard to read

Real plates are not clean test images. They are dirty, angled, badly lit, and sometimes dark on dark. A recent VEZHA update fixed an issue where dark and chocolate-coloured plates were not detected reliably. If you have sites where such plates are common, re-test them after updating.

Camera placement still decides most of the outcome. Before installation, go through the LPR Camera Install Guidelines in the partner portal documentation: mounting height, angle to the lane and the distance at which the plate is captured.

Regional plate formats

A parking system bought in Mexico City should not be tuned for Kyiv. VEZHA LPR supports region-specific configurations. For Latin America, the region configuration includes OCR Model 2, and the LPR plugin stores longer region codes such as "MX-NLE". Ask for the region profile that matches your country before you test.

Edge or server: choosing the hardware

For most car parks, a central server is the simpler choice to deploy and maintain.

If you prefer analytics close to the camera, VEZHA LPR has an AXIS ACAP integration for Axis cameras with on-board analytics.

You also do not need LPR on every camera. In our 104-camera facility case, four LPR channels at the vehicle entry and exit points were enough for vehicle control, out of 39 analytics channels in total.

Integration with the existing stack

Parking rarely lives alone. The same site has a VMS, an access control system and sometimes a billing platform. Points to check before a project:

• Video sources. VEZHA works with ONVIF and RTSP streams and with major camera brands.
• VMS. VEZHA holds Milestone Verified and Network Optix Verified status, so events can appear in the VMS the operator already uses.
• Cameras with on-board analytics. IncoreSoft is an Axis ADP partner, and VEZHA LPR has an AXIS ACAP integration.
• APIs. REST API and SDK for billing, ticketing or custom dashboards.
• Security. IncoreSoft is certified to ISO/IEC 27001:2022 and ISO 22301:2019.

A short checklist for your next parking project

1. Count lanes and give each lane its own camera, with the recognition zone limited to that lane.
2. Agree the lists and tags with the customer before installation.
3. Check camera placement against the install guidelines.
4. Configure each site: its analytics, the interval for reads from other cameras and, where needed, the direction of travel.
5. Test with the customer's own plates, including dark and dirty ones, and with drivers who stop in the wrong place.
6. Connect events to the VMS and access control the operator already uses.

See the Smart Parking product page and the VEZHA license plate recognition software for details. If you are an integrator, become a partner or request a demo on your own gate footage.

FAQ

What is an LPR parking system?

It is a setup where cameras read license plates at the entry and exit of a car park, and software decides what to do: open the barrier, log the visit, measure the stay or alert staff. It replaces cards and tickets for regular users.

What happens if the barrier closes before the car passes?

In VEZHA Smart Parking, a repeated read of the same plate by the same analytics source triggers the site again and can re-open the barrier, even if the site timeout has passed. Reads from another camera within the site interval are ignored.

How should cameras be set up on a multi-lane entry?

Give each lane its own camera and limit its recognition zone to that lane. Use separate cameras for entry and exit. In the site settings, choose the analytics and set the interval during which reads from another camera are ignored.

Does it work with my VMS?

VEZHA holds Milestone Verified and Network Optix Verified status and works with ONVIF and RTSP cameras. A REST API and SDK are available for other systems.