[Config Support]: Need help with cat and dog detection

[lilp69]

Describe the problem you are having

Hi,
New to frigate.
I've installed it on HA.
Here is my conf:
Frigate Current version: 0.13.2
Home Assistant Core 2024.6.1
Supervisor 2024.06.0
Operating System 12.3
Frontend 20240605.0

I've a lot of false positive with person, I upload them to Frigate+ to edit them.
But camera, never detect dog or car.

Thanks;

Version

0.13.2

Frigate config file

## Connection à MQTT ##
mqtt:
  host: XxX
  port: XxX
  client_id: frigate
  topic_prefix: frigate
  user: XX
  password: XxX


## Choix des caméras ##
go2rtc:
  streams:
    r4252_smart_outdoor_camera:
    - echo:bash /config/custom_components/expose_camera_stream_source/get_stream.sh
      camera.r4252_smart_outdoor_camera

## Liste des caméras avec leurs adresses de connection ##
cameras:
  r4252_smart_outdoor_camera:
    ffmpeg:
      inputs:
      - path: rtsp:https://127.0.0.1:8554/r4252_smart_outdoor_camera?video
        input_args: preset-rtsp-restream-low-latency
        roles:
        - detect

#    motion: ## Création des zones à ignorer ##
#      mask:
#        - 0,461,3,0,1919,0,1919,843,1699,492,1344,458,1346,336,973,317,869,375,866,432
    motion:
      mask:
      - 262,0,259,36,32,38,31,0

## Choix des options de détection et d'enregistrement ##
record:
  enabled: true
  retain:
    days: 7
    mode: all
  events:
# Seconds d'enregistrement avant et après l'èvenement
    pre_capture: 60
    post_capture: 60
    retain:
#Durée de rétention en J
      default: 15
#      mode: motion ## Conserve les vidéos avec mouvements seulement ## ## active_objects ## Conserve les vidéos avec objets seulement ##
      mode: all ## Conserve les vidéos avec détection d'objets ##
      objects:
        dog: 2 ## Durée de conservation par objets ##
        cat: 2
        person: 7

#Objets à suivre
objects:
  # Optional: list of objects to track from labelmap.txt (default: shown below)
  track:
  - person
  - dog
  - cat
  # Optional: filters to reduce false positives for specific object types
  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 width/height of the bounding box for the detected object (default: 0)
#      min_ratio: 0.5
      # Optional: maximum width/height of the bounding box for the detected object (default: 24000000)
#      max_ratio: 2.0
      # 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
    dog:
      # 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 width/height of the bounding box for the detected object (default: 0)
#      min_ratio: 0.5
      # Optional: maximum width/height of the bounding box for the detected object (default: 24000000)
#      max_ratio: 2.0
      # Optional: minimum score for the object to initiate tracking (default: shown below)
      min_score: 0.4
      # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
      threshold: 0.5
      # 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
    cat:
      # 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 width/height of the bounding box for the detected object (default: 0)
#      min_ratio: 0.5
      # Optional: maximum width/height of the bounding box for the detected object (default: 24000000)
#      max_ratio: 2.0
      # Optional: minimum score for the object to initiate tracking (default: shown below)
      min_score: 0.4
      # Optional: minimum decimal percentage for tracked object's computed score to be considered a true positive (default: shown below)
      threshold: 0.5

motion:
  # Optional: The threshold passed to cv2.threshold to determine if a pixel is different enough to be counted as motion. (default: shown below)
  # Increasing this value will make motion detection less sensitive and decreasing it will make motion detection more sensitive.
  # The value should be between 1 and 255.
  threshold: 30


snapshots:
  enabled: true

Relevant log output

Cat and dog not detect.

Frigate stats

No response

Operating system

HassOS

Install method

HassOS Addon

Coral version

USB

Any other information that may be helpful

No response

[NickM-27]

if you upload to frigate+ it won't change what you see, frigate+ only affects your detections once you generate a model and use that.

You are also not using a coral or any hardware object detection which is likely the issue as well as your detection is falling behind

[lilp69]

Thanks for your reply.
How to generate a model and use that ?
I don't understand the part about Coral or hardware object detection. Do I need to install a Coral card in addition to Frigate on the RIP5? Or add motion detection hardware as well? Like what?

[NickM-27]

RPi is not a recommended platform, especially rpi5 as it lacks an h264 hardware decoder

A coral (or a new platform with a compatible detector hardware) would be needed to more efficiently run object detection

[lilp69]

Can I add a Coral like Google Coral USB Edge TPU ML Accelerator Coprocesseur for Raspberry Pi to my setup ? Directly connect by usb to Rpi5 ?

[mimainka]

Yes connect your Coral USB to your RPi5 and then check if its listed under your hardware-view in HA (Settings - System - Hardware - Complete Hardware).
https://docs.frigate.video/troubleshooting/edgetpu#usb-coral-not-detected
If you find the Coral here, you should be able to include it in your config.
You might need to install the full access Add-On of Frigate and turn of protected mode so the Add-On can access your connected hardware.

detectors:
coral:
type: edgetpu
device: usb

//Edit: Jesus im too dumb to get it formated right, just pay attention to the yaml structure :D

[lilp69]

Thanks, I will see to order a Coral and set up to my config.

[lilp69]

I just received my Google Coral https://coral.ai/products/accelerator/, it's seems to be installed well.
https://i.imgur.com/FO0YmGY.png
The add config part :detectors: coral: type: edgetpu device: usb
Now how to train it to detect well (dog, cat, person)?