Calculate speeds of cars driving by.

[y2kdread]

Hi - Is it possible for Frigate to calculate the speed of cars that are detected by cameras? We are having issues with people speeding on our street and it seems like this would be something that Frigate can do.

[NickM-27]

it's not directly supported but could certainly be done

[y2kdread]

Do you have any ideas on how that could be done?

[NickM-27]

Off the top of my head you could setup two zones and use the MQTT events topic https://docs.frigate.video/integrations/mqtt#frigateevents to get the time between them using frame time

[y2kdread]

Awesome, thanks! I'll have to check that out

[moskovskiy82]

Have set up two zones. They are not that apart (30 metres between them)
Any ideas which data to take?
The frame time?

{
  "before": {
    "id": "1689411553.371775-kvno8z",
    "camera": "roads",
    "frame_time": 1689411559.982616,
    "snapshot_time": 1689411559.782061,
    "label": "car",
    "sub_label": null,
    "top_score": 0.84375,
    "false_positive": false,
    "start_time": 1689411553.371775,
    "end_time": null,
    "score": 0.83984375,
    "box": [
      877,
      93,
      1271,
      423
    ],
    "area": 130020,
    "ratio": 1.1939393939393939,
    "region": [
      348,
      0,
      1280,
      932
    ],
    "stationary": false,
    "motionless_count": 1,
    "position_changes": 1,
    "current_zones": [
      "zone_0"
    ],
    "entered_zones": [
      "zone_1",
      "zone_0"
    ],
    "has_clip": true,
    "has_snapshot": true
  },
  "after": {
    "id": "1689411553.371775-kvno8z",
    "camera": "roads",
    "frame_time": 1689411560.38212,
    "snapshot_time": 1689411559.782061,
    "label": "car",
    "sub_label": null,
    "top_score": 0.84375,
    "false_positive": false,
    "start_time": 1689411553.371775,
    "end_time": 1689411565.581974,
    "score": 0.68359375,
    "box": [
      1084,
      152,
      1277,
      471
    ],
    "area": 61567,
    "ratio": 0.6050156739811913,
    "region": [
      364,
      0,
      1280,
      916
    ],
    "stationary": false,
    "motionless_count": 0,
    "position_changes": 1,
    "current_zones": [
      "zone_0"
    ],
    "entered_zones": [
      "zone_1",
      "zone_0"
    ],
    "has_clip": true,
    "has_snapshot": true
  },
  "type": "end"
}

[consolecwby]

How did this work for you. Looking for a similar solution.

[moskovskiy82]

Was going to ask the same just an hour ago. Any other pointers? Found an app which recognizes the front and rear wheels and does the calculation based on time it takes for them to travel and calculate based on wheelbase. Quite ingenious

[y2kdread]

what is the app that you found that does that? I've never really worked with MQTT before so something that does this out of the box would be much quicker!

[moskovskiy82]

what is the app that you found that does that? I've never really worked with MQTT before so something that does this out of the box would be much quicker!

There is a bunch of projects on github using opencv. How to integrate them to frigate is another question.

[ghzgod]

Wouldn't your cameras FPS be a factor as well? I am not a smart person.

[carlostipton1319]

https://homecleaningblog.com/is-it-possible-to-install-additional-safety-features-on-an-older-model-of-garage-door-opener/if you want to calculate the speeds of cars driving by, you would typically need to measure the distance traveled by the car and the time it takes to cover that distance. The formula for speed is:

Speed = Distance / Time

To calculate the speeds of cars, you would need to have a method of accurately measuring the distance traveled, such as using a measuring tape or a known reference point. You would also need a way to measure the time taken, which could be done using a stopwatch or a timing device.

By dividing the distance traveled by the time taken, you can calculate the speed of the cars driving by.

[moskovskiy82]

A chat GPT spam bot. How nice

[jvijj]

Has anyone found a simple-ish solution for this?

[ctml91]

Would be very cool to have this "Object Speed Detection" natively supported in Friday.

User draws a zone and inputs actual distance between boundaries, then frigate measures the speed and annotates the event.