[Support]:

[cryptedx]

Describe the problem you are having

I have upgraded to the newest version today and I have a setup where Home Assistant pushes me a notification when an object has been detected. The new version has Stationary objects feature. At the moment I don't have any use for it but I think there will be some use cases when I have time to think about it! So, thank you for that but it's making me trouble:

There a two stationary cars from my neighbour at the moment . One of them is causing me headache. It's usually not there all the time but today it is. It's a silver one (side view). Unfortunately I can't post the picture of the car (privacy), but it's 15 meters away from my cam. Another dark blue car (front view) is 5 meters to the right of the silver one This isn't making any trouble.

I have set contour_area: 125 and also tried up to 1000.

The area where the two cars are is also completely masked.

I don't know hot to solve that issue. So I have to deactivate frigate for now. My config worked for more than 6 months now without such an issue.

Thank you for your time and help!

Version

0.10.0-BFECEE9

Frigate config file

detectors:
  coral:
    type: edgetpu
    device: usb

cameras:
  hof:
    ffmpeg:
      inputs:
        # Main stream (1920x1080)
        - path: ***
          roles:
            - record
            - rtmp
        # Sub stream (640x352)
        - path: ***
          roles:
            - detect
    motion:
      contour_area: 125 # reduce the sensitivity for motion detection
    detect:
      width: 640
      height: 352
      fps: 5
    #record:
      # Optional: Enable recording (default: shown below)
      #enabled: True
    #snapshots:
      #enabled: True
      # Optional: Configuration for stationary object tracking
      stationary:
        # Optional: Frequency for running detection on stationary objects (default: shown below)
        # When set to 0, object detection will never be run on stationary objects. If set to 10, 
        # it will be run on every 10th frame.
        interval: 0
        # Optional: Number of frames without a position change for an object to be considered 
        # stationary (default: 10x the frame rate or 10s)
        threshold: 50
        # Optional: Define a maximum number of frames for tracking a stationary object (default: not set, track forever)
        # This can help with false positives for objects that should only be stationary for a limited amount of time.
        # It can also be used to disable stationary object tracking. For example, you may want to set a value for person,
        # but leave car at the default.
        max_frames:
          # Optional: Default for all object types (default: not set, track forever)
          default: 3000
          # Optional: Object specific values
          objects:
            person: 1000

  doorbird:
    ffmpeg:
      inputs:
        # It's not possible to run two seperate streams from a doorbird!
        # Main stream (1920x1080)
        # - path: ***
        #   roles:
        # Sub stream (1280x720)
        - path: ***
          roles:
            - detect
            - record
            - rtmp
    motion:
      mask:
        - 1280,0,1280,315,1280,521,1020,459,740,473,378,454,374,317,437,315,435,289,0,147,0,0,625,0
      # reduce the sensitivity for motion detection
      contour_area: 125 
    detect:
      width: 1280
      height: 720
      fps: 5
      # Optional: Configuration for stationary object tracking
      stationary:
        # Optional: Frequency for running detection on stationary objects (default: shown below)
        # When set to 0, object detection will never be run on stationary objects. If set to 10, 
        # it will be run on every 10th frame.
        interval: 0
        # Optional: Number of frames without a position change for an object to be considered 
        # stationary (default: 10x the frame rate or 10s)
        threshold: 100
        # Optional: Define a maximum number of frames for tracking a stationary object (default: not set, track forever)
        # This can help with false positives for objects that should only be stationary for a limited amount of time.
        # It can also be used to disable stationary object tracking. For example, you may want to set a value for person,
        # but leave car at the default.
        max_frames:
          # Optional: Default for all object types (default: not set, track forever)
          default: 3000
          # Optional: Object specific values
          objects:
            person: 1000

objects:
  track:
    - car
    - person
    - bicycle
  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: Record configuration
# NOTE: Can be overridden at the camera level
record:
  # Optional: Enable recording (default: shown below)
  enabled: True
  # Optional: Number of days to retain recordings regardless of events (default: shown below)
  # NOTE: This should be set to 0 and retention should be defined in events section below
  #       if you only want to retain recordings of events.
  retain:
    days: 7
    mode: motion
  # Optional: Event recording settings
  events:
    # Optional: Maximum length of time to retain video during long events. (default: shown below)
    # NOTE: If an object is being tracked for longer than this amount of time, the retained recordings
    #       will be the last x seconds of the event unless retain_days under record is > 0.
    max_seconds: 300
    # Optional: Number of seconds before the event to include (default: shown below)
    pre_capture: 5
    # Optional: Number of seconds after the event to include (default: shown below)
    post_capture: 5
    # Optional: Objects to save recordings for. (default: all tracked objects)
    # objects:
      # - person
    # Optional: Restrict recordings to objects that entered any of the listed zones (default: no required zones)
    # required_zones: []
    # Optional: Retention settings for recordings of events
    retain:
      # Required: Default retention days (default: shown below)
      default: 10
      mode: active_objects

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: False
  # 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

Relevant log output

---

FFprobe output from your camera

# ffprobe rtsp://***
ffprobe version 4.3.1 Copyright (c) 2007-2020 the FFmpeg developers
  built with gcc 9 (Ubuntu 9.3.0-17ubuntu1~20.04)
  configuration: --disable-debug --disable-doc --disable-ffplay --enable-shared --enable-avresample --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-gpl --enable-libfreetype --enable-libvidstab --enable-libmfx --enable-libmp3lame --enable-libopus --enable-libtheora --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libxcb --enable-libx265 --enable-libxvid --enable-libx264 --enable-nonfree --enable-openssl --enable-libfdk_aac --enable-postproc --enable-small --enable-version3 --enable-libzmq --extra-libs=-ldl --prefix=/opt/ffmpeg --enable-libopenjpeg --enable-libkvazaar --enable-libaom --extra-libs=-lpthread --enable-vaapi --extra-cflags=-I/opt/ffmpeg/include --extra-ldflags=-L/opt/ffmpeg/lib
  libavutil      56. 51.100 / 56. 51.100
  libavcodec     58. 91.100 / 58. 91.100
  libavformat    58. 45.100 / 58. 45.100
  libavdevice    58. 10.100 / 58. 10.100
  libavfilter     7. 85.100 /  7. 85.100
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  7.100 /  5.  7.100
  libswresample   3.  7.100 /  3.  7.100
  libpostproc    55.  7.100 / 55.  7.100
Input #0, rtsp, from 'rtsp://***':
  Metadata:
    title           : RTSP/RTP stream from DoorBird
    comment         : mpeg/1080p/media.amp
  Duration: N/A, start: 0.094633, bitrate: N/A
    Stream #0:0: Video: h264, yuvj420p(pc, bt709, progressive), 1920x1080, 10 fps, 10 tbr, 90k tbn, 180k tbc

Frigate stats

{"detection_fps":0.0,"detectors":{"coral":{"detection_start":0.0,"inference_speed":9.09,"pid":221}},"doorbird":{"camera_fps":5.0,"capture_pid":234,"detection_fps":0.0,"pid":230,"process_fps":5.0,"skipped_fps":0.0},"hof":{"camera_fps":5.1,"capture_pid":232,"detection_fps":0.0,"pid":228,"process_fps":5.1,"skipped_fps":0.0},"service":{"storage":{"/dev/shm":{"free":5366.0,"mount_type":"tmpfs","total":5368.7,"used":2.7},"/media/frigate/clips":{"free":5101658.4,"mount_type":"fuse.shfs","total":11997243.4,"used":6895584.9},"/media/frigate/recordings":{"free":5101658.4,"mount_type":"fuse.shfs","total":11997243.4,"used":6895584.9},"/tmp/cache":{"free":7337.1,"mount_type":"rootfs","total":8297.5,"used":960.4}},"temperatures":{},"uptime":53,"version":"0.10.0-bfecee9"}}

Operating system

UNRAID

Install method

Docker Compose

Coral version

USB

Network connection

Wired

Camera make and model

Doorbird D1101V

Any other information that may be helpful

I am using this Home Assistant config to push notifications when objects appear:

alias: Notification, Motion, Cam - Frigate object detected
trigger:
  - platform: mqtt
    topic: frigate/events
condition:
  - condition: state
    entity_id: input_boolean.disable_camera_hof_motion_notification_door_open
    state: 'off'
  - condition: state
    entity_id: input_boolean.disable_camera_hof_motion_notification
    state: 'off'
action:
  - service: notify.mobile_app_iphone_12_pro_max
    data_template:
      message: Ein {{trigger.payload_json["after"]["label"]}} wurde erkannt.
      data:
        image: >-
          https://***/api/frigate/notifications/{{trigger.payload_json["after"]["id"]}}/thumbnail.jpg?format=android
        tag: '{{trigger.payload_json["after"]["id"]}}'
        url: /lovelace/kameras
        clickAction: /lovelace/kameras
  - delay:
      hours: 0
      minutes: 0
      seconds: 30
      milliseconds: 0
mode: single

[blakeblackshear]

A couple things:

1. I would remove the stationary block from your detect config. The defaults should work fine in your situation.
2. This isn't a motion issue, so no need to customize the motion settings unless that contour_area setting is better for other reasons.

Under what conditions do you want to be notified of a car? Can you post a picture from your camera without the cars present or with them masked/blurred sufficiently?

I am fairly certain that these cars would have been detected previously when other objects in unmasked areas moved around them. Are you saying that previously they never showed up in the events list or in a notification? Motion masks do not prevent objects from being detected. They just prevent motion in that area from triggering object detection.

[cryptedx]

1. Ok, I have commented it out.
2. I have commented it also out.

Problem still persists. Detection is happening and cars are not moving. Same situation as before. It's Sunday, not much going on ;-) Weather is cloudy. Also nobody or something was moving in front of the cars.

This is my masked cam:

I want to be notified when an object is on my ground. Every time. This area is the main entry to our House.

Are you saying that previously they never showed up in the events list or in a notification? --> Yes, never. Just when a car passes from left to right or vice versa I get a notification. This is also an open topic. Since this is a dead-end street there is not much traffic, so it wasn't an urgent one for me, yet. But as you can see the complete area is masked. This is the original image.

[blakeblackshear]

It is working as designed and finding cars since you have that listed in your config. You are getting notified of those cars because you haven't defined any restrictions on when you want to be notified. The recommended way to handle this scenario is to define a zone for your ground and require that an object enter that zone in your notifications, events, etc.

I don't know the exact coordinates of your zone, but here is how you would specify a zone:

cameras:
  ...
  doorbird:
    zones:
      ground:
        coordinates: ... # <---- update with coordinates where bottom center of bounding boxes would be

record:
  ...
  events:
    ...
    required_zones:
      - ground

snapshots:
  ...
  required_zones:
    - ground

Your notification would look like this with the condition added to require the zone:

alias: Notification, Motion, Cam - Frigate object detected
trigger:
  - platform: mqtt
    topic: frigate/events
condition:
  - condition: state
    entity_id: input_boolean.disable_camera_hof_motion_notification_door_open
    state: 'off'
  - condition: state
    entity_id: input_boolean.disable_camera_hof_motion_notification
    state: 'off'
  - "{{ 'ground' in trigger.payload_json['after']['entered_zones'] }}"
action:
  - service: notify.mobile_app_iphone_12_pro_max
    data_template:
      message: Ein {{trigger.payload_json["after"]["label"]}} wurde erkannt.
      data:
        image: >-
          https://***/api/frigate/notifications/{{trigger.payload_json["after"]["id"]}}/thumbnail.jpg?format=android
        tag: '{{trigger.payload_json["after"]["id"]}}'
        url: /lovelace/kameras
        clickAction: /lovelace/kameras
  - delay:
      hours: 0
      minutes: 0
      seconds: 30
      milliseconds: 0
mode: single

This is an example of what I would consider too much masking. With proper use of zones, you should be able to reduce the size of your motion mask and get faster and more reliable detections. Frigate detects and tracks objects as they move in and out of the areas you care about. Ideally, you want Frigate to detect and track person and car objects on the street and only notify you if they enter your ground. With your current mask, when a person steps on your ground, Frigate will only see motion for their foot and may not detect it as a person until they move further into your ground. Also, Frigate requires multiple frames of detections to confirm it as a true positive which will further delay your notifications and may miss the object. Without such a large mask, the person would be detected, confirmed as a true positive, and tracked while they are still across the street. Then, the very first step they take into your ground would trigger a notification.

[blakeblackshear]

You may also want to look at the updated notification blueprint for 0.10.0: https://github.com/blakeblackshear/frigate/discussions/2748#discussioncomment-2152948

The old blueprint ran in single mode and only tracked a single object. With stationary objects in frame this means you were probably missing notifications. This version will send a notification for every tracked object. When they are stationary and start moving again it will send another notification

[blakeblackshear]

Lastly, this note from the docs is important to know when creating zones:

Presence in a zone is evaluated based on the bottom center of the bounding box for the object. It does not matter how much of the bounding box overlaps with the zone.

[cryptedx]

Thank you very much. That was a really great help. I will try to implement everything.

[Coolie1101]

@blakeblackshear

    required_zones:
      - ground

Is this documented, I'm experiencing the same issue since I updated yesterday, but don't see this config option in the docs?

My current global record config =

record:
  enabled: true
  retain_days: 0
  events:
    post_capture: 10
    retain:
      default: 30
      objects:
        person: 60

[NickM-27]

@Coolie1101

It's documented here: https://docs.frigate.video/configuration/masks/#further-clarification

and here: https://docs.frigate.video/guides/stationary_objects/

[cryptedx]

Hi @blakeblackshear,

I didn't had much time but it is working perfect now. Thank you very much so far :-)

But one problem is left. I am getting push notifications outside of my zones.

If a car drives on the street and does not enter my zone I get a notification. I have reduced the mask as you can see but I got this new problem. So I don't know what to do now.

[blakeblackshear]

You need to add a condition to your notification that requires the zone. The required_zones for events and snapshots don't apply to notifications. See here for some examples.

[cryptedx]

I am using the blueprint and have added the condition:

[blakeblackshear]

does not enter my zone

Can you confirm that you don't see the event you were notified about on the events page?

Can you post the yaml the blueprint generated for the automation?

[cryptedx]

Yes, I can confirm that I don't see the event were I have been notified about.

alias: Notification, Motion, Cam - Frigate Camera Doorbird object detected
description: ''
use_blueprint:
  path: hunterjm/frigate_0.10_notification.yaml
  input:
    camera: doorbird
    update_thumbnail: false
    zone_filter: true
    notify_group: notify_group_parents
    zones:
      - zone_hof
    silence_timer: 30
    cooldown: 30

[blakeblackshear]

I'm not sure why you are getting notifications there. You should look at the automation trace tools to see what is happening.

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.