LP Recognizer

What is LicensePlateRecognizer?

LicensePlateRecognizer processes images and video frames independently without temporal tracking. It detects license plates and recognizes text in single images, image batches, or video streams (frame-by-frame via predict_batch()).

When to Use LicensePlateRecognizer

Choose LicensePlateRecognizer when you need to:

• Process collections of images or photos

• Analyze video files frame-by-frame without temporal tracking

• Run batch recognition on multiple images or video streams

• Work without real-time tracking requirements

For video with persistent vehicle tracking and Bayesian multi-frame text fusion, use LicensePlateTracker instead - it maintains vehicle identities across frames and supports real-time alerts.

Core Concepts

Configuration

All settings are controlled through LicensePlateRecognizerConfig:

import degirum_vehicle

config = degirum_vehicle.LicensePlateRecognizerConfig(
    license_plate_detection_model_spec=detection_spec,  # Detection model
    license_plate_ocr_model_spec=ocr_spec,             # OCR model
)

lpr = degirum_vehicle.LicensePlateRecognizer(config)

See Configuration Guide for all options.

Methods

LicensePlateRecognizer provides methods for single-frame and batch processing:

• predict() - Recognize license plates in a single image

• predict_batch() - Process multiple images or video frames

See Methods Reference for complete API documentation with examples.