Frigate Adapter für ioBroker

[bongsao]

Hallo,
zuerst mal Danke für die tolle Arbeit. Der Adapter läuft so weit und macht was er soll. Beim Erstellen einer Vis ist mir aufgefallen, das in den History-Einträgen die Timestrings start_time und end_time fehlerhaft sind. Das sind Kommazahlen. Wenn man rundet, passt es. Das ist halt nur in der Vis ein riesen Aufwand. Ich gehe davon aus, dass das auch so nicht erwünscht ist.
Ansonsten läuft alles super stabil.

Gruß

[tombox]

@bongsao In der Vis kannst du es mit dem widget Timestamp anzeigen
@PrinzEisenherz1 bitte GitHub version testen camera.remote.createEvent

[bongsao]

@tombox
Danke, muss ich ne neue Version installieren. Die stable 1.02 funktioniert ja nicht.

[PrinzEisenherz1]

Hi @tombox erstmal danke das Du das in den Adapter integriert hast und für den tollen Adapter.
Aber irgendwas funktioniert da bei mir nicht so richtig.
Das triggern einer Aufnahme erfolgt wenn ich z.B. das Label ändere, richtig?
Wie ist denn mit den neu erzeugten Datenpunkten für das manuelle Recording verfahren?
Wie befülle ich dann den Body? z.B. duration, usw
Wenn ich den Body wie beim Beispiel von frigate befülle

{
  "sub_label": "some_string", // add sub label to event
  "duration": 30, // predetermined length of event (default: 30 seconds) or can be to null for indeterminate length event
  "include_recording": true, // whether the event should save recordings along with the snapshot that is taken
  "draw": {
    // optional annotations that will be drawn on the snapshot
    "boxes": [
      {
        "box": [0.5, 0.5, 0.25, 0.25], // box consists of x, y, width, height which are on a scale between 0 - 1
        "color": [255, 0, 0], // color of the box, default is red
        "score": 100 // optional score associated with the box
      }
    ]
  }
}

dann bekomme ich in ioBroker folgende Fehlermeldung

frigate.0 2024-02-23 16:33:59.932	error	SyntaxError: Unexpected number in JSON at position 128
frigate.0 2024-02-23 16:33:59.931	error	Cannot parse createEventBody. Please use valid JSON https://docs.frigate.video/integrations/api/#post-apieventscamera_namelabelcreate

Es wird zwar eine Aufnahme gestartet, diese ist aber fehlerhaft.
Die lässt sich nicht abspielen. Da kommt n weißes X.
Wenn ich sie herunterladen möchte dann wird ein clips.json downgeloadet.
Der erzeugte Snapshot wird aber richtig erstellt.
Muss ich da eventuell noch an meiner Config was einstellen?
Hier mal meine config.yml

Spoiler

database:
    path: /db/frigate.db
mqtt:
  host: 192.168.122.25
  port: 2083
detectors:
    coral:
        type: edgetpu
        device: usb
go2rtc:
  streams:
    haustuer_cam: 
      - rtsp://XXXXXXXXX@192.168.122.15:88/videoMain
      - "ffmpeg:haustuer_cam#audio=opus" # <- copy of the stream which transcodes audio to opus
    haustuer_cam_sub:
      - rtsp://XXXXXXXXX@192.168.122.15:88/videoSub # <- stream which supports video & aac audio. This is only supported for rtsp streams, http must use ffmpeg
      - "ffmpeg:haustuer_cam_sub#audio=opus" # <- copy of the stream which transcodes audio to opus
    garage_cam: 
      - rtsp://XXXXXXXXX@192.168.122.19:554/h264Preview_01_main
      - "ffmpeg:garage_cam#audio=opus" # <- copy of the stream which transcodes audio to opus
    garage_cam_sub:
      - rtsp://XXXXXXXXX@192.168.122.19:554/h264Preview_01_sub # <- stream which supports video & aac audio. This is only supported for rtsp streams, http must use ffmpeg
      - "ffmpeg:garage_cam_sub#audio=opus" # <- copy of the stream which transcodes audio to opus
    garten_cam: 
      - rtsp://XXXXXXXXX@192.168.122.17:88/videoMain
      - "ffmpeg:garten_cam#audio=opus" # <- copy of the stream which transcodes audio to opus
    garten_cam_sub:
      - rtsp://XXXXXXXXX@192.168.122.17:88/videoSub # <- stream which supports video & aac audio. This is only supported for rtsp streams, http must use ffmpeg
      - "ffmpeg:garten_cam_sub#audio=opus" # <- copy of the stream which transcodes audio to opus
    garten_links_cam: 
      - rtsp://XXXXXXXXX@192.168.122.16/11
      - "ffmpeg:garten_links_cam#audio=opus" # <- copy of the stream which transcodes audio to opus
    garten_links_cam_sub:
      - rtsp://XXXXXXXXX@192.168.122.16/12 # <- stream which supports video & aac audio. This is only supported for rtsp streams, http must use ffmpeg
      - "ffmpeg:garten_links_cam_sub#audio=opus" # <- copy of the stream which transcodes audio to opus
    wohnzimmer_cam: 
      - rtsp://XXXXXXXXX@192.168.122.18:88/videoMain
      - "ffmpeg:wohnzimmer_cam#audio=opus" # <- copy of the stream which transcodes audio to opus
    wohnzimmer_cam_sub:
      - rtsp://XXXXXXXXX@192.168.122.18:88/videoSub # <- stream which supports video & aac audio. This is only supported for rtsp streams, http must use ffmpeg
      - "ffmpeg:wohnzimmer_cam_sub#audio=opus" # <- copy of the stream which transcodes audio to opus
ffmpeg:
    output_args:
        record: -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c copy
    hwaccel_args: # This is the section for the H/W Accelaration, you will need to figure out what settings you need to enable - https://docs.frigate.video/configuration/hardware_acceleration
        #- -hwaccel
        #- qsv
        #- -qsv_device
        #- /dev/dri/renderD128
        - -hwaccel
        - vaapi 
        - -hwaccel_device
        - /dev/dri/renderD128 
        #- -hwaccel_output_format
        #- yuv420p
        #- -hwaccel
        #- vaapi
        #- -hwaccel_device
        #- /dev/dri/renderD128
        #- -hwaccel_output_format
        #- yuv420p
record: # how long to record before and after motion
    events:
        pre_capture: 7
        post_capture: 10
cameras:
  haustuer: # <------ Name the camera
    ffmpeg:
      hwaccel_args: preset-vaapi
      output_args:
        record: preset-record-generic-audio-copy
      inputs:
        - path: rtsp://127.0.0.1:8554/haustuer_cam_sub # <----- The stream you want to use for detection
          roles:
            - detect
        - path: rtsp://127.0.0.1:8554/haustuer_cam
          roles:
            - record
            - rtmp
    rtmp:
      enabled: false # <-- RTMP should be disabled if your stream is not H264
    detect:
      enabled: True # <---- disable detection until you have a working camera feed
      width: 1280 # <---- update for your camera's resolution
      height: 720 # <---- update for your camera's resolution
      fps: 5
    objects:
      track:
        - person
        - car
        - dog
        - cat
      filters:
        person:
          # Optional: minimum width*height of the bounding box for the detected object (default: 0)
          #min_area: 5000
          # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
          #max_area: 100000
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          #min_score: 0.5
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.7
          # Optional: mask to prevent this object type from being detected in certain areas (default: no mask)
          # Checks based on the bottom center of the bounding box of the object
          #mask: 0,0,1000,0,1000,200,0,200
    record: # what you want Frigate to record
        enabled: True
        #retain: # uncomment this section if you want Frigate to record this camera 24/7. This can use a lot of space, you have been warned!
          #days: 1
          #mode: all
        events: # this is the default recording mode where events are recorded, the number of days it saves events for is 10.
            retain:
                mode: active_objects
                default: 10
    snapshots:
      # Optional: Enable writing jpg snapshot to /media/frigate/clips (default: shown below)
      # This value can be set via MQTT and will be updated in startup based on retained value
      enabled: true
      # Optional: print a timestamp on the snapshots (default: shown below)
      timestamp: false
      # Optional: draw bounding box on the snapshots (default: shown below)
      bounding_box: true
      # Optional: crop the snapshot (default: shown below)
      crop: false
      # Optional: height to resize the snapshot to (default: original size)
      # height: 175
      # Optional: Restrict snapshots to objects that entered any of the listed zones (default: no required zones)
      required_zones: []
      # Optional: Camera override for retention settings (default: global values)
      retain:
        # Required: Default retention days (default: shown below)
        default: 10
        # Optional: Per object retention days
        #objects:
          #person: 15
    motion:
      mask:
        - 858,215,847,176,746,145,754,115,557,74,528,34,396,37,297,38,349,720,0,720,0,0,1280,0,1280,370

  garage: # <------ Name the camera
    ffmpeg:
      hwaccel_args: preset-vaapi
      output_args:
        record: preset-record-generic-audio-copy
      inputs:
        - path: rtsp://127.0.0.1:8554/garage_cam_sub # <----- The stream you want to use for detection
          roles:
            - detect
        - path: rtsp://127.0.0.1:8554/garage_cam
          roles:
            - record
            - rtmp
    rtmp:
      enabled: false # <-- RTMP should be disabled if your stream is not H264
    detect:
      enabled: True # <---- disable detection until you have a working camera feed
      width: 640 # <---- update for your camera's resolution
      height: 480 # <---- update for your camera's resolution
      fps: 5
    objects:
      track:
        - person
        - car
        - truck
        - bicycle
        - motorcycle
        - dog
        - cat
      filters:
        person:
          # Optional: minimum width*height of the bounding box for the detected object (default: 0)
          #min_area: 5000
          # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
          #max_area: 100000
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          #min_score: 0.5
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.7
          # Optional: mask to prevent this object type from being detected in certain areas (default: no mask)
          # Checks based on the bottom center of the bounding box of the object
          #mask: 0,0,1000,0,1000,200,0,200
    record: # what you want Frigate to record
        enabled: True
        #retain: # uncomment this section if you want Frigate to record this camera 24/7. This can use a lot of space, you have been warned!
          #days: 1
          #mode: all
        events: # this is the default recording mode where events are recorded, the number of days it saves events for is 10.
            retain:
                mode: active_objects
                default: 10
    snapshots:
      # Optional: Enable writing jpg snapshot to /media/frigate/clips (default: shown below)
      # This value can be set via MQTT and will be updated in startup based on retained value
      enabled: true
      # Optional: print a timestamp on the snapshots (default: shown below)
      timestamp: false
      # Optional: draw bounding box on the snapshots (default: shown below)
      bounding_box: true
      # Optional: crop the snapshot (default: shown below)
      crop: false
      # Optional: height to resize the snapshot to (default: original size)
      # height: 175
      # Optional: Restrict snapshots to objects that entered any of the listed zones (default: no required zones)
      # required_zones: []
      required_zones:
        - hof
        - einfahrt
      # Optional: Camera override for retention settings (default: global values)
      retain:
        # Required: Default retention days (default: shown below)
        default: 10
        # Optional: Per object retention days
        #objects:
          #person: 15
    motion:
      mask:
        #- 0,195,40,158,104,104,158,72,180,21,259,21,393,27,408,62,426,91,448,107,476,100,532,105,582,95,591,0,0,0
        - 0,480,199,480,199,450,0,450
    zones:
      hof:
        coordinates: 640,480,640,204,640,74,556,68,503,92,400,79,388,144,189,136,181,70,73,144,0,207,0,480
        objects:
          - car
          - person
      einfahrt:
        coordinates: 204,128,389,140,383,34,231,26
        objects:
          - car
          - person

  garten: # <------ Name the camera
    ffmpeg:
      hwaccel_args: preset-vaapi
      output_args:
        record: preset-record-generic-audio-copy
      inputs:
        - path: rtsp://127.0.0.1:8554/garten_cam_sub # <----- The stream you want to use for detection
          roles:
            - detect
        - path: rtsp://127.0.0.1:8554/garten_cam
          roles:
            - record
            - rtmp
    rtmp:
      enabled: false # <-- RTMP should be disabled if your stream is not H264
    detect:
      enabled: True # <---- disable detection until you have a working camera feed
      width: 1280 # <---- update for your camera's resolution
      height: 720 # <---- update for your camera's resolution
      fps: 5
    objects:
      track:
        - person
        #- car
        - dog
        - cat
      filters:
        person:
          # Optional: minimum width*height of the bounding box for the detected object (default: 0)
          #min_area: 5000
          # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
          #max_area: 100000
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          #min_score: 0.5
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.7
          # Optional: mask to prevent this object type from being detected in certain areas (default: no mask)
          # Checks based on the bottom center of the bounding box of the object
          #mask: 0,0,1000,0,1000,200,0,200
    record: # what you want Frigate to record
        enabled: True
        #retain: # uncomment this section if you want Frigate to record this camera 24/7. This can use a lot of space, you have been warned!
          #days: 1
          #mode: all
        events: # this is the default recording mode where events are recorded, the number of days it saves events for is 10.
            retain:
                mode: active_objects
                default: 10
    snapshots:
      # Optional: Enable writing jpg snapshot to /media/frigate/clips (default: shown below)
      # This value can be set via MQTT and will be updated in startup based on retained value
      enabled: true
      # Optional: print a timestamp on the snapshots (default: shown below)
      timestamp: false
      # Optional: draw bounding box on the snapshots (default: shown below)
      bounding_box: true
      # Optional: crop the snapshot (default: shown below)
      crop: false
      # Optional: height to resize the snapshot to (default: original size)
      # height: 175
      # Optional: Restrict snapshots to objects that entered any of the listed zones (default: no required zones)
      required_zones: []
      # Optional: Camera override for retention settings (default: global values)
      retain:
        # Required: Default retention days (default: shown below)
        default: 10
        # Optional: Per object retention days
        #objects:
          #person: 15
    motion:
      mask:
        - 487,699,485,720,1280,720,1280,373
        - 0,80,0,0,1280,0,1280,256,958,203,953,219,786,235,738,162,687,139,654,118,602,100,552,107,515,59,449,41,350,39,257,42,121,41,129,79

  garten_links: # <------ Name the camera
    ffmpeg:
      hwaccel_args: preset-vaapi
      output_args:
        record: preset-record-generic-audio-copy
      inputs:
        - path: rtsp://127.0.0.1:8554/garten_links_cam_sub # <----- The stream you want to use for detection
          roles:
            - detect
        - path: rtsp://127.0.0.1:8554/garten_links_cam
          roles:
            - record
            - rtmp
    rtmp:
      enabled: false # <-- RTMP should be disabled if your stream is not H264
    detect:
      enabled: True # <---- disable detection until you have a working camera feed
      width: 640 # <---- update for your camera's resolution
      height: 352 # <---- update for your camera's resolution
      fps: 5
    objects:
      track:
        - person
        #- car
        - dog
        - cat
      filters:
        person:
          # Optional: minimum width*height of the bounding box for the detected object (default: 0)
          #min_area: 5000
          # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
          #max_area: 100000
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          #min_score: 0.5
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.7
          # Optional: mask to prevent this object type from being detected in certain areas (default: no mask)
          # Checks based on the bottom center of the bounding box of the object
          #mask: 0,0,1000,0,1000,200,0,200
    record: # what you want Frigate to record
        enabled: True
        #retain: # uncomment this section if you want Frigate to record this camera 24/7. This can use a lot of space, you have been warned!
          #days: 1
          #mode: all
        events: # this is the default recording mode where events are recorded, the number of days it saves events for is 10.
            retain:
                mode: active_objects
                default: 10
    snapshots:
      # Optional: Enable writing jpg snapshot to /media/frigate/clips (default: shown below)
      # This value can be set via MQTT and will be updated in startup based on retained value
      enabled: true
      # Optional: print a timestamp on the snapshots (default: shown below)
      timestamp: false
      # Optional: draw bounding box on the snapshots (default: shown below)
      bounding_box: true
      # Optional: crop the snapshot (default: shown below)
      crop: false
      # Optional: height to resize the snapshot to (default: original size)
      # height: 175
      # Optional: Restrict snapshots to objects that entered any of the listed zones (default: no required zones)
      required_zones: []
      # Optional: Camera override for retention settings (default: global values)
      retain:
        # Required: Default retention days (default: shown below)
        default: 10
        # Optional: Per object retention days
        #objects:
          #person: 15
    motion:
      mask:
        - 237,0,182,21,139,25,99,38,0,79,0,303,0,0
        - 469,0,468,21,640,21,640,0

  wohnzimmer: # <------ Name the camera
    ffmpeg:
      hwaccel_args: preset-vaapi
      output_args:
        record: preset-record-generic-audio-copy
      inputs:
        - path: rtsp://127.0.0.1:8554/wohnzimmer_cam_sub # <----- The stream you want to use for detection
          roles:
            - detect
        - path: rtsp://127.0.0.1:8554/wohnzimmer_cam
          roles:
            - record
            - rtmp
    rtmp:
      enabled: false # <-- RTMP should be disabled if your stream is not H264
    detect:
      enabled: True # <---- disable detection until you have a working camera feed
      width: 1280 # <---- update for your camera's resolution
      height: 720 # <---- update for your camera's resolution
      fps: 5
    objects:
      track:
        - person
        - car
        - truck
        - bicycle
        - motorcycle
        - dog
        - cat
      filters:
        person:
          # Optional: minimum width*height of the bounding box for the detected object (default: 0)
          #min_area: 5000
          # Optional: maximum width*height of the bounding box for the detected object (default: 24000000)
          #max_area: 100000
          # Optional: minimum score for the object to initiate tracking (default: shown below)
          #min_score: 0.5
          # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
          threshold: 0.7
          # Optional: mask to prevent this object type from being detected in certain areas (default: no mask)
          # Checks based on the bottom center of the bounding box of the object
          #mask: 0,0,1000,0,1000,200,0,200
    record: # what you want Frigate to record
        enabled: True
        #retain: # uncomment this section if you want Frigate to record this camera 24/7. This can use a lot of space, you have been warned!
          #days: 1
          #mode: all
        events: # this is the default recording mode where events are recorded, the number of days it saves events for is 10.
            retain:
                mode: active_objects
                default: 10
    snapshots:
      # Optional: Enable writing jpg snapshot to /media/frigate/clips (default: shown below)
      # This value can be set via MQTT and will be updated in startup based on retained value
      enabled: true
      # Optional: print a timestamp on the snapshots (default: shown below)
      timestamp: false
      # Optional: draw bounding box on the snapshots (default: shown below)
      bounding_box: true
      # Optional: crop the snapshot (default: shown below)
      crop: false
      # Optional: height to resize the snapshot to (default: original size)
      # height: 175
      # Optional: Restrict snapshots to objects that entered any of the listed zones (default: no required zones)
      required_zones: []
      # Optional: Camera override for retention settings (default: global values)
      retain:
        # Required: Default retention days (default: shown below)
        default: 10
        # Optional: Per object retention days
        #objects:
          #person: 15
    motion:
      mask:
        - 0,720,369,720,364,120,433,121,503,129,542,147,586,184,630,241,1280,632,1280,0,0,0

Dann hätte ich da noch eine Frage, wo ich aber jetzt nicht weiß ob es vom Adapter kommt oder von der 0.13er Version.
Starte ich nämlich den frigate-Container neu, sind alle Kameras automatisch auf Motion-Erkennung.
Bei der Adapter-Version 0.2.8 und Frigate 0.12.x war das nicht so. Da war nach einem Neustart die Einstellung so wie vor dem Neustart.

Noch eine Frage hätte ich da. Besteht vielleicht die Möglichkeit das man bei der neuen Version den integrierten MQTT-Broker deaktivieren kann? Bei mir läuft schon ein anderer.

So, ich hoffe ich das alles einigermaßen verständlich geschrieben

[crunchip]

@prinzeisenherz1 beim Sprung von 0.12 auf 0.13 sollte man ggf sämtliche Einstellungen/Parameter in der yaml löschen und neu setzen, steht so zumindest in der Beschreibung, hat also mit dem Adapter eher weniger zu tun.

Du gibst doch in deiner yaml ja oben den mqtt broker an

[tombox]

@prinzeisenherz1
Bitte mal die Kommentare aus dem json entfernen das ist nicht erlaubt
der adapter setzt keine einstellungen
ohne integrierte mqtt wird es schwieriger an die Daten zu kommen er müsste sich auf den externen mqtt server verbinden und auf die frigate topics subscriben

[PrinzEisenherz1]

@tombox wenn ich die Kommentare entferne kommt die gleiche Fehlermeldung.
Also wenn ich folgendes in den Datenpunkt schreibe:

{
  "sub_label": "test",
  "duration": 30,
  "include_recording": true
}

Ich bin erstmal wieder bei frigate 0.12 und Adapter auf 0.2.8.
Muss ich mich mal beschäftigen wenn Frau und Kidds bei Schwiegereltern sind

[tombox]

@prinzeisenherz1 das mit dem json sollte jetzt gehen

[martinschm]

Hi,

ich hab v 1.0.2 installiert und hab jetzt in den Logs immer wieder die Meldung "Snapshot sending active but no snapshot available for event xyz".
Wenn ich in meiner Frigate Installation schaue, sind dort aber Snapshots zu den Events zu sehen.

Woran liegt die Fehlermeldung ?

ciao
Martin

[tombox]

@martinschm manche new Events haben noch kein snapshot tritt das auch bei end Events auf

[Globe]

Hi @tombox!
Meinst du bei den Benachrichtigungen lässt sich noch der E-Mail Adapter integrieren? Ich habe zum Test die Instanz mal mit eingetragen. Es werden auch E-Mails gesendet, doch leider beinhalten sie keine Bilder und Infos zum Zeitpunkt, sondern nur den tmp-Bildpfad (bspw. so: /tmp/14356de3-36b9-46fb-83c2-2c48c9d5da17.jpg).

Falls das gehen sollte wäre es top, wenn man die im Adapter hinterlegte Benachrichtigungs-Textvorlage zum Event in der E-Mail als Betreff (und evtl. zusätzlich als Nachrichtentext über dem Bild?) erhalten könnte.

Danke und viele Grüße!

[Tschaeens]

@tombox
Hallo tombox, wäre es vllt möglich den ptz Befehl mit in den Adapter zu integrieren?
Viele Grüße

[tombox]

@globe mail sollte jetzt gehen
@Tschaeens den GET plz info? oder welchen wie oft soll der gerufen werden?

[Tschaeens]

@tombox
geil wäre dieser Datenpunkt zum steuern der ptz Einstellungen:

frigate/<camera_name>/ptz preset_<preset_name>

Danke dir.

[tombox]

@tschaeens finde dazu kein doku
https://docs.frigate.video/integrations/api/

[Tschaeens]

@tombox
den kann man per mqtt setzen.
https://docs.frigate.video/integrations/mqtt

und dann ziemlich weit unten. ptz Steuerung

[tombox]

@tschaeens Ist hinzugefügt

[Tschaeens]

@tombox
funzt! Sau geil. Danke dir.

[Globe]

@tombox

Mail läuft jetzt auch einwandfrei. Merci!

[martinschm]

Hi @tombox ,

ich habe gerade bemerkt, das meine iobroker Platte voll läuft und hab mal geschaut wo der Speicherplatz am meisten genutzt wird. Dabei ist mir mit 7GB das Verzeichnis /tmp aufgefallen, in dem tausende mp4 und jpg sind.

Hab mir die Dateien mal angeschaut und sie sind von meiner Kamera.
Legt der frigate Adapter die Videoschnippsel die er per Telegram schickt dort temporär ab und kann es sein, das es evtl ein Problem beim Löschen der temporären Datei gibt ?

Es sind knapp über 4600 Dateien, die älteste ist vom 18.2

Nutze die v 1.0.2

Viele Grüße
Martin