Traccar license plate recognition system: a complete guide to ANPR technologies for cars

Automatic License Plate Recognition (ANPR, Automatic Number Plate Recognition) has ceased to be an exotic technology - today it is actively used in parking lots, logistics companies and even by private car owners. Traccar system stands out among its peers due to its open source, flexible configuration and integration with GPS trackers. But how does it work in practice? Is it possible to install it yourself, and what legal restrictions apply in Russia in 2026?

Many people mistakenly think that ANPR systems are the prerogative of large enterprises with budgets in the millions. In fact Traccar allows you to deploy a basic license plate recognition system even on a Raspberry Pi, and cloud solutions reduce costs to a monthly subscription fee. In this article we will analyze the technical nuances, compare equipment and show how to avoid common mistakes when setting up.

It’s worth clarifying right away: Traccar is not only ANPR. This is a full-fledged platform for transport monitoring, where license plate recognition is only one of the modules. However, it is the ANPR functionality that often becomes the key when choosing a system for controlling access to a territory or automating vehicle fleet accounting.

How ANPR works in Traccar: algorithms and technical details

At the heart of any ANPR system is a combination of computer vision and machine learning. Traccar uses open libraries like OpenALPR or Tesseract OCR, adapted to the specifics of license plates of different countries. The recognition process can be divided into 4 key stages:

1. Image Capture — the camera records the vehicle in the coverage area. It is important that the resolution allows you to clearly distinguish the characters on the number (minimum 1920×1080 when shooting from 5–7 meters).

2. License plate detection — the algorithm selects a rectangular area with a number against the background of the body. Proper contrast and lighting settings are critical here.

3. Character segmentation — the program breaks the number into separate characters, removing frames, bolts and other artifacts.

4. Text recognition — OCR engine (for example, Tesseract) converts character images to text format.

Traccar supports plugins for different countries, including Russia, Belarus and Kazakhstan. However the accuracy of recognizing new Russian license plates (from 2023) may drop to 70–80% when shooting at night without IR illumination. This involves changing the font and adding region codes on two lines.

The processing speed of one frame depends on the hardware platform:

• 🖥️ On Raspberry Pi 4 (4GB RAM) - 1-2 seconds at resolution Full HD.
• 💻 On a server with a GPU (for example, NVIDIA Jetson) - up to 10 frames per second.
• ☁️ Cloud solutions (Traccar Cloud) - delay 0.3–0.8 seconds, but requires stable Internet.

⚠️ Attention: If you use Traccar to control access to a parking lot, keep in mind that when the vehicle speed is above 30 km/h, the recognition accuracy drops by 15–20%. For high-speed areas (for example, at the entrance to a logistics complex) specialized cameras with shutter will be required 1/10000s.

Equipment for ANPR: what to choose for Traccar

Traccar is compatible with most IP cameras, but for stable operation of ANPR you need to take into account several criteria. Firstly, resolution: minimum acceptable - 1280×720, but for reliable recognition it is better 1920×1080 or 2560×1440. Secondly, matrix type: CMOS sensors are cheaper but perform worse in low light conditions compared to CCD.

Optimal camera models for Traccar ANPR:

In addition to the cameras, you will need a server to process the data. Options:

• 🖥️ Local server - any PC with Ubuntu 22.04 and 8 GB of RAM. Traccar also works on Windows, but the stability is lower.
• ☁️ cloud — Traccar Cloud (from $15/month) or your own solution on AWS/Yandex Cloud.
• 📱 Mobile solution — Raspberry Pi 4 + camera for small parking lots (up to 20 cars/hour).

⚠️ Attention: If you use cameras with IR illumination, make sure that they comply with GOST R 58401-2019 for electromagnetic radiation levels. Uncertified devices may cause interference with radio equipment (such as walkie-talkies in airport parking lots).

Installing and configuring Traccar for license plate recognition

Deploying Traccar with an ANPR module can be completed in 30–60 minutes if you follow a clear algorithm. Below are step-by-step instructions for installation on Ubuntu 22.04:

1. Installing dependencies:

sudo apt update && sudo apt install -y openjdk-11-jre ffmpeg tesseract-ocr libtesseract-dev

2. Download Traccar:

wget https://github.com/traccar/traccar/releases/download/v5.10/traccar-linux-64-5.10.zip
unzip traccar-linux-64-5.10.zip

3. Setting up a configuration file traccar.xml:

- Specify the path to the camera: <entry key='video.device'>/dev/video0</entry>

- Activate ANPR plugin: <entry key='video.plateReader'>true</entry>

- For Russian numbers, add: <entry key='video.plateReader.country'>ru</entry>

4. Starting the service:

./traccar.run &

After launch, the system will be available at http://localhost:8082. For remote access, configure port forwarding on the router or use NGROK.

The camera is connected and recognized (check via `v4l2-ctl --list-devices`)|

Video capture drivers installed (`sudo apt install v4l-utils`)|

Illumination configured (IR or white light) for night photography|

Ports 8082 (Traccar) and 554 (RTSP camera stream) are open in the firewall|-->

To increase recognition accuracy, it is recommended:

• 🎯 Calibrating the shooting area — in the camera settings, specify the area where the license plate is expected to appear (for example, x=200,y=100,width=400,height=150).
• 🔍 Model training — if you have specific numbers (for example, corporate numbers), upload 50–100 examples to Traccar for additional training.
• ⚡ Performance optimization - disable unnecessary plugins in traccar.xml, if the server is weak.

Legal aspects of using ANPR in Russia

In 2026, the use of license plate recognition systems in Russia is regulated by several regulations:

• 📜 Federal Law No. 152-FZ (“About personal data”) - the car number is considered personal information if it is linked to the owner.
• 🚗 Government Decree No. 1191 (from 2020) - allows ANPR to control access to private areas without the consent of car owners.
• 📸 GOST R 58401-2019 — establishes requirements for video surveillance, including ANPR systems.

Key points:

✅ Allowed without consent:

- Control of access to private parking or enterprise territory.

- Recording of drivers’ working hours (if they are notified of video surveillance).

❌ Requires consent or notification:

- Storage of data about numbers for longer than 30 days (according to 152-FZ).

- Transfer of data to third parties (for example, to collection agencies).

- Use on public roads (allowed only to traffic police and municipal services).

If you are installing an ANPR at the entrance to an apartment complex, be sure to post a sign notifying you of video surveillance. Example text:

"The territory is subject to video surveillance with license plate recognition for security purposes. Data is stored for 30 days (FZ-152)."

Integration of Traccar ANPR with other systems

One of Traccar's strengths is its ability to interface with external services. For example, you can automatically:

• 🚪 Open the barrier when recognizing the "white list" of numbers (integration with Wiren Board or Raspberry Pi GPIO).
• 📊 Generate reports in 1C or Google Sheets about the time of arrival/departure of transport.
• 🚨 Send notifications in Telegram or Slack when a “black list” of numbers is detected.

Used for integration Traccar API or webhooks. Example code for sending number data to a Telegram bot:

import requests

WEBHOOK_URL = "https://api.telegram.org/botTOKEN/sendMessage"
CHAT_ID = "12345678"
PLATE = "A123BV777" # Data from Traccar

data = {
"chat_id": CHAT_ID,
"text": f"Number recognized: {PLATE}\nTime: {datetime.now()}"
}
requests.post(WEBHOOK_URL, data=data)

To automate the barrier, you can use a script on Python with library RPi.GPIO:

import RPi.GPIO as GPIO
import time

GPIO.setmode(GPIO.BCM)
GPIO.setup(18, GPIO.OUT) # Barrier relay pin

def open_barrier():
GPIO.output(18, GPIO.HIGH)
time.sleep(5) # Keep open for 5 seconds
GPIO.output(18, GPIO.LOW)

Calling a function when recognizing a number from the white list

Common mistakes and how to avoid them

Even experienced administrators encounter problems setting up ANPR in Traccar. Here are the most common mistakes and ways to solve them:

Another common problem is port conflictwhen the camera and Traccar try to use the same port for an RTSP stream. Solution:

1. Check occupied ports: sudo netstat -tulnp | grep 554
2. Change the camera port in settings or reconfigure Traccar to a different port in traccar.xml.

If Traccar does not see the camera, make sure that:

• 🔌 The camera is connected to the same network as the server.
• 📡 RTSP stream available (check via VLC Player at the address rtsp://IP_cameras:554/stream1).
• 🔑 The login/password for the camera is specified in the Traccar config: <entry key='video.url'>rtsp://login:password@ip:554/stream1</entry>

Traccar Alternatives: ANPR System Comparison

Traccar is not the only solution for license plate recognition. Let's look at the alternatives with their pros and cons:

When to choose Traccar?

• 💰 Budget is limited - open source allows you to save on licenses.
• 🔧 Customization is needed - you can modify the recognition logic to suit your tasks.
• 🌐 Integration with GPS trackers is required - Traccar was originally designed for vehicle monitoring.

When to consider alternatives?

• 📈 Requires maximum % recognition - Avigilon or PlateRecognizer more precisely by 3–5%.
• ☁️ I don’t want to mess with servers - cloud solutions are easier to support.
• 🚛 You work with international transport - OpenALPR better supports foreign number formats.

FAQ: Frequently asked questions about Traccar ANPR

import pandas as pd
df = pd.read_csv('traccar_anpr_export.csv')
df.to_excel('report_by_numbers.xlsx', index=False)