Frigate Go2RTC Coral

[miralant]

Bonjour,

J’ai actuellement 4 caméras que j’ai connectées à une instance Frigate montée en Docker dans mon NAS Synologie 218++ (10Go de Ram).

Je récupère directement les flux rtsp dans Frigate et j’ai ensuite mis la card HA_frigate pour récupérer les informations.

mqtt:
  host: 192.168.1.30
  port: 1883
  user: mqtt
  password: zzzzzzzz
  stats_interval: 60
  #enabled: false

birdseye:
  enabled: false
rtmp:
  enabled: false

detectors:
  coral:
    type: edgetpu
    device: usb

record:
  enabled: true
  retain:
    days: 3
    mode: motion
  events:
    retain:
      default: 7
      mode: motion

snapshots:
  enabled: true
  retain:
    default: 3

cameras:
  garage: # <------ Name the camera
    enabled: true
    ffmpeg:
      inputs:
      - path: rtsp://yyyyyy:zzzzzzz@192.168.1.14:554  # <----- The stream you want to use for detection
        roles:
        - detect
    onvif:
      host: 192.168.1.14
      port: 8899
      user: yyyyyy
      password: zzzzzzz
    detect:
      enabled: true
    objects:
      track:
      - person
      - dog
      - cat
      filters:
        person:
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          min_score: 0.7
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.8
    record:
      enabled: true
    snapshots:
      enabled: true
  piscine: # <------ Name the camera
    enabled: true
    ffmpeg:
      inputs:
      - path: rtsp://zzzzzzz:yyyyyyyy@192.168.1.63:554  # <----- The stream you want to use for detection
        roles:
        - detect
    onvif:
      host: 192.168.1.63
      port: 8899
      user: zzzzzzz
      password: yyyyyyyy
    detect:
      enabled: true
    objects:
      track:
      - person
      - dog
      - cat
      filters:
        person:
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          min_score: 0.7
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.8
    record:
      enabled: true
    snapshots:
      enabled: true
    motion:
      mask:
      - 0,0,888,0,914,539,0,990
      - 2304,0,2304,597,890,543,886,0
      - 0,1303,2174,1294,2159,1670,0,1831
  salon: # <------ Name the camera
    enabled: true
    ffmpeg:
      inputs:
      - path: rtsp://zzzzzzz:yyyyyyyy@192.168.1.32:554/stream1 # <----- The stream you want to use for detection
        roles:
        - detect
    #onvif:
    #  host: 192.168.1.32
    #  port: 8899
    #  user: zzzzzzz
    #  password: yyyyyyyy
    detect:
      enabled: true
    objects:
      track:
      - person
      - dog
      - cat
      filters:
        person:
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          min_score: 0.7
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.8
    record:
      enabled: true
    snapshots:
      enabled: true
  bureau: # <------ Name the camera
    enabled: true
    ffmpeg:
      inputs:
      - path: rtsp://192.168.1.150:554/ch0_0.h264   # <----- The stream you want to use for detection
        roles:
        - detect
    detect:
      enabled: true
    objects:
      track:
      - person
      - dog
      - cat
      filters:
        person:
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          min_score: 0.7
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.8
    record:
      enabled: true
    snapshots:
      enabled: true

J’ai également installé un CORAL pour l’aider un peu mais ça n’est pas flagrant. Le container consomme allègrement 70% du CPU Comment controler que le CORAL joue bien son rôle ?

Sur quoi pourrai-je jouer pour limiter un peu la consommation de CPU ?

Au hazard de lecture, j’ai vu que l’on pouvait passer via Go2RTC mais je ne comprends pas l’interêt. Pouvez vous m’en dire un peu plus que l’interêt ?

Merci