Frigate Plate Recognizer

Identify license plates via Plate Recognizer or CodeProject.AI and add them as sublabels to blakeblackshear/frigate

Setup

Create a config.yml file in your docker volume with the following contents:

frigate:
  frigate_url: http://127.0.0.1:5000
  mqtt_server: 127.0.0.1
  mqtt_port: 1883 # Optional. Default shown.
  mqtt_username: username
  mqtt_password: password
  main_topic: frigate
  return_topic: plate_recognizer
  frigate_plus: false
  camera:
    - driveway_camera
  objects:
    - car
  min_score: .8
plate_recognizer:
  token: xxxxxxxxxx
  regions:
    - us-ca
logger_level: INFO

Update your frigate url, mqtt server settings. If you are using mqtt authentication, update the username and password. Update the camera name(s) to match the camera name in your frigate config. Add your Plate Recognizer API key and region(s).

You'll need to make an account (free) here and get an API key. You get up to 2,500 lookups per month for free. You will also need to enable car object detection for the cameras you want to use this with. See here on how to locally host Plate Recognizer.

You can specify a custom url for the plate_recognizer api by adding api_url to your config:

plate_recognizer:
  api_url: http://HOST-IP:8080/v1/plate-reader
  token: xxxxxxxxxx
  regions:
    - us-ca

You can also filter by zones and/or cameras. If you want to filter by zones, add zones to your config:

frigate:
  # ...
  zones:
    - front_door
    - back_door

If no objects are speficied in the Frigate options, it will default to [motorcycle, car, bus].

If you have a custom model with Frigate+ then it's able to detect license plates via an event's attributes, you can set frigate_plus to true in your config to activate this feature:

frigate:
  # ...
  frigate_plus: true
  license_plate_min_score: 0 # default is show all but can speficify a min score from 0 - 1 for example 0.8
  max_attempts: 20 # Optional: if set, will limit the number of snapshots sent for recognition for any particular event.

If you're using CodeProject.AI, you'll need to comment out plate_recognizer in your config. Then add and update "api_url" with your CodeProject.AI Service API URL. Your config should look like:

#plate_recognizer:
#  token: xxxxxxxxxx
#  regions:
#    - us-ca
code_project:
  api_url: http://127.0.0.1:32168/v1/image/alpr

Running

docker run -v /path/to/config:/config -e TZ=America/New_York -it --rm --name frigate_plate_recognizer lmerza/frigate_plate_recognizer:latest

or using docker-compose:

services:
  frigate_plate_recognizer:
    image: lmerza/frigate_plate_recognizer:latest
    container_name: frigate_plate_recognizer
    volumes:
      - /path/to/config:/config
    restart: unless-stopped
    environment:
      - TZ=America/New_York

https://hub.docker.com/r/lmerza/frigate_plate_recognizer

Debugging

set logger_level in your config to DEBUG to see more logging information:

logger_level: DEBUG

Logs will be in /config/frigate_plate_recognizer.log

Save Snapshot Images to Path

If you want frigate-plate-recognizer to automatically save snapshots of recognized plates, add the following to your config.yml:

frigate:
  save_snapshots: True # Saves a snapshot called [Camera Name]_[timestamp].png
  draw_box: True # Optional - Draws a box around the plate on the snapshot along with the license plate text (Required Frigate plus setting)
  always_save_snapshot: True # Optional - will save a snapshot of every event sent to frigate_plate_recognizer, even if no plate is detected

Snapshots will be saved into the '/plates' directory within your container - to access them directly, map an additional volume within your docker-compose, e.g.:

services:
  frigate_plate_recognizer:
    image: lmerza/frigate_plate_recognizer:latest
    container_name: frigate_plate_recognizer
    volumes:
      - /path/to/config:/config
      - /path/to/plates:/plates:rw
    restart: unless-stopped
    environment:
      - TZ=America/New_York

Monitor Watched Plates

If you want frigate-plate-recognizer to check recognized plates against a list of watched plates for close matches (including fuzzy recognition), add the following to your config.yml:

frigate:
  watched_plates: #list of plates to watch.
    -  ABC123
    -  DEF456
  fuzzy_match: 0.8 # default is test against plate-recognizer / CP.AI 'candidates' only, but can specify a min score for fuzzy matching if no candidates match watched plates from 0 - 1 for example 0.8

If a watched plate is found in the list of candidates plates returned by plate-recognizer / CP.AI, the response will be updated to use that plate and it's score. The original plate will be added to the MQTT response as an additional original_plate field.

If no candidates match and fuzzy_match is enabled with a value, the recognized plate is compared against each of the watched_plates using fuzzy matching. If a plate is found with a score > fuzzy_match, the response will be updated with that plate. The original plate and the associated fuzzy_score will be added to the MQTT response as additional fields original_plate and fuzzy_score.