Skip to content

stuomas/TheNoodleSnoop

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TheNoodleSnoop

The clumsy but occasionally useful sidekick of TheSpaghettiDetective.

Introduction

TheNoodleSnoop borrows the pre-trained neural network from obico-server along with some utility functions, and strips away everything else like the web server and GUI. Instead, it uses MQTT to report any findings.

It is configured with the snapshot URL of your camera, and reports findings and the snapshot with detection overlay via MQTT. The user can then handle the message as they prefer, for example in Home Assistant.

Pre-requisites

  • Docker installation
  • Powerful enough host to run the server

Configuration

Set camera source and MQTT settings in docker-compose.yml.

Recommended to create a .env file in the same directory as docker-compose.yml for secrets to prevent accidentally pushing them to public repositories.

Example .env file:

SNAPSHOT_URL=192.168.1.25:8080/snapshot
MQTT_BROKER_IP=192.168.1.4
MQTT_BROKER_PORT=1883
MQTT_BROKER_USER=my-ha-mqtt-user
MQTT_BROKER_PASSWORD=my-ha-mqtt-password

Installation

docker-compose up -d

Home Assistant Integration

Use the MQTT Camera integration to show the detection overlay.

Example configuration:

mqtt:
  camera:
    - topic: thenoodlesnoop/overlay
      name: Bambu Lab P1S NoodleCam
      image_encoding: b64

Use Notify integration for custom notifications about failures.

Example automation:

TBA

Use ha-bambulab custom component for controlling the printer based on detections

Example automation:

TBA

MQTT Camera does not seem to support value_templates, which is the whole reason there is currently a separate MQTT_OVERLAY_TOPIC which you can use in the configuration.

API Reference

${MQTT_BASE_TOPIC}/detections

payload

{
    'failing': <TheNoodleSnoop's opinion whether print is failing [bool]>,
    'raw_detections': <all detections with confidences and locations [list]>,
    'raw_predictions': <some data about predictions [dict]>,
    'overlay': <utf-8 representation of base64 encoded image [string]>,
    'timestamp': <unix timestamp [float]>
}

${MQTT_BASE_TOPIC}/overlay

Convenience topic until I figure out how to use something like value_template in Home Assistant MQTT Camera integration. Payload is the same as payload['overlay'] in MQTT_REPORT_TOPIC.

payload

<utf-8 representation of base64 encoded image [string]>

${MQTT_BASE_TOPIC}/control

Not implemented. Start, stop, re-configuration, etc.