Motion detection triggered in snowfall/

[210392]

Is it possible to configure the configuration file so that the frigate does not detect movement during snowfall?
My current config file

#yaml-language-server: $schema=http://ccab4aaf-frigate:5000/api/config/schema.json

mqtt:
  host: 888.888.8.888
  port: 1883
  topic_prefix: frigate
  client_id: frigate
  user: frigate
  password: frigate

birdseye:
  # Optional: Enable birdseye view (default: shown below)
  enabled: false

#Global Detect Settings
detectors:
  cpu1:
    type: cpu
    num_threads: 2
  cpu2:
    type: cpu
    num_threads: 2

detect:
  height: 640
  width: 720
  fps: 5
  min_initialized: 10
  max_disappeared: 25
  stationary:
    interval: 50
    threshold: 20
    max_frames:
      default: 300
      objects:
        person: 100
        car: 150

#Global Object Settings
objects:
  track:
    - person
    - car
    - motorcycle
  filters:
    person:
      min_area: 5000
      max_area: 20000
    car:
      min_area: 8000
      max_area: 35000

#Global Record Settings
record:
  enabled: true
  retain:
    days: 7
    mode: active_objects

motion:
  threshold: 60
  lightning_threshold: 0.6
  contour_area: 70
  frame_height: 50
  improve_contrast: true
  mqtt_off_delay: 2

#Cameras Settings
cameras:
#Camera Vorota-------------------------------------------------------------------------------
  cam_Vorota:
    ffmpeg:
      inputs:
        - path: rtsp://888.888.8.888:554/user=8888_password=8888_channel=0_stream=1 
          roles:
            - detect
        - path: rtsp://888.888.8.888:554/user=8888_password=8888_channel=0_stream=0
          roles:
            - record
    #
    zones:
      Vorota:
        coordinates: 
          1,1,0.002,0.997,0,1,0,0.612,0.18,0.565,0.519,0.565,0.723,0.563,0.781,0.598,0.489,0.625,0.495,0.78,0.65,0.844,0.781,0.903,1,0.908,1,0.167,0.828,0.176,0.768,0.185,0.785,0.284,0.707,0.322,0.594,0.314,0.559,0.275,0.605,0.215,0.765,0.202,0.764,0.183,0.616,0.203,0.546,0.225,0.366,0.273,0,0.365,0.12,0.433,0.234,0.499,1,0.495
        loitering_time: 2
        inertia: 3
        objects: person
      Vorota2:
        coordinates: 0.564,0.273,0.607,0.22,0.77,0.207,0.776,0.287,0.71,0.317,0.586,0.311
        inertia: 3
        loitering_time: 2
        objects: person
      Vorota3:
        coordinates: 0.491,0.627,0.492,0.78,0.779,0.9,0.997,0.906,0.994,0.795,0.807,0.728,0.781,0.602
        inertia: 3
        loitering_time: 2
        objects: person
    review:
      alerts: {}
      detections: {}
    motion:
      mask: 
        0,0.363,0,0.609,0,0.362,0.553,0.22,0.598,0.206,0.675,0.198,0.814,0.177,1,0.165,1,0,-0.001,0,-0.001,0,0,0.611,0.179,0.564,0.73,0.558,0.783,0.598,0.811,0.73,1,0.795,1,0.495,0.231,0.5
#Camera Dvor-------------------------------------------------------------------------------
  cam_Dvor:
    ffmpeg:
      inputs:
        - path: rtsp://888.888.8.888:554/user=8888_password=8888_channel=0_stream=1
          roles:
            - detect
        - path: rtsp://888.888.8.888:554/user=8888_password=8888_channel=0_stream=0
          roles:
            - record
#
    zones:
      Dvor_1:
        coordinates: 0.996,0.495,0,0.501,0,0.131,0.311,0.097,0.661,0.077,0.81,0.12,0.994,0.194
        inertia: 3
        loitering_time: 2
        objects: person
      Dvor_1_2:
        coordinates: 
          1,0.881,0.871,0.881,0.867,1,0.401,1,0.22,0.628,0.193,0.631,0.186,1,0,1,0,0.6,0.286,0.568,0.368,0.561,0.414,0.57,0.423,0.643,0.463,0.652,0.472,0.566,0.664,0.578,1,0.622
        inertia: 3
        loitering_time: 2
        objects: person
    review: {}
    motion:
      mask:
        - 0.194,1,0.194,0.637,0.223,0.635,0.394,0.997
        - 0.873,1,0.871,0.885,1,0.885,1,1
        - 1,0.618,0.713,0.581,0.56,0.566,0.467,0.561,0.457,0.652,0.426,0.642,0.424,0.563,0.368,0.56,0.234,0.57,0.071,0.586,0,0.597,0,0.597,0,0.503,1,0.499
        - 1,0.191,0.816,0.12,0.667,0.075,0.5,0.085,0.302,0.096,0,0.13,0,0,1,0
#Camera Dvor 2-------------------------------------------------------------------------------
  cam_Dvor_2:
    ffmpeg:
      inputs:
        - path: rtsp://888.888.8.888:554/user=8888_password=8888_channel=0_stream=1
          roles:
            - detect
        - path: rtsp://888.888.8.888:554/user=8888_password=8888_channel=0_stream=0
          roles:
            - record
#
    zones:
      Dvor_2:
        coordinates: 
          -0.001,0.335,0.503,0.081,0.999,0.186,1,0.5,0.004,0.501,0,0.587,0.019,0.579,0.997,0.872,0.999,0.997,0,0.999,0,0.402,0,0.333,-0.003,0.332
        loitering_time: 2
        objects: person
        inertia: 3
    review:
      alerts: {}
    motion:
      mask: 
        0.5,0.081,0.739,0.127,0.833,0.153,0.796,0.501,1,0.5,1,0,0,0,0,0.168,0,0.337,0,0.505,1,0.5,1,0.872,0.022,0.575,0,0.587,0,0.335

semantic_search:
  enabled: false
  reindex: false
  model_size: small

[hawkeye217]

You could try adjusting the threshold and contour area using the Motion Tuner in Settings, though this kind of motion is not easily filtered out.

However, if it is not causing problems, there's no need to try to filter it.

See the motion detection documentation: https://docs.frigate.video/configuration/motion_detection

[210392]

is not causing problems
This causes a lot of video to be recorded, apparently because the Frigate thinks the cars are moving.

[ofcedwards]

To build on @hawkeye217's suggestion, you can use home assistant to automate the motion settings if you find something that works. I used to have an issue like this, but it was wind causing significant motion detection by making bushes/trees move (regularly causing entire hours to record with motion-only retention).

After using the motion tuner to find settings I was content with, I created an automation in home assistant to automatically change the threshold based on the wind speed. This lets me have the high sensitivity I want during optimal conditions and a tolerably lower sensitivity during worse weather conditions.

[hawkeye217]

That's great.