HASS-Deepstack-object

Home Assistant custom components for using Deepstack object detection. Deepstack is a service which runs in a docker container and exposes deep-learning models via a REST API. There is no cost for using Deepstack, although you will need a machine with 8 GB RAM. On your machine with docker, pull the latest image (approx. 2GB):

sudo docker pull deepquestai/deepstack

Recommended OS Deepstack docker containers are optimised for Linux or Windows 10 Pro. Mac and regular windows users my experience performance issues.

GPU users Note that if your machine has an Nvidia GPU you can get a 5 x 20 times performance boost by using the GPU.

Legacy machine users If you are using a machine that doesn't support avx or you are having issues with making requests, Deepstack has a specific build for these systems. Use deepquestai/deepstack:noavx instead of deepquestai/deepstack when you are installing or running Deepstack.

Activating the API

Before you get started, you will need to activate the Deepstack API. First, go to www.deepstack.cc and sign up for an account. Choose the basic plan which will give us unlimited access for one installation. You will then see an activation key in your portal.

On your machine with docker, run Deepstack without any recognition so you can activate the API on port 5000:

sudo docker run -v localstorage:/datastore -p 5000:5000 deepquestai/deepstack

Now go to http://YOUR_SERVER_IP_ADDRESS:5000/ on another computer or the same one running Deepstack. Input your activation key from your portal into the text box below "Enter New Activation Key" and press enter. Now stop your docker container. You are now ready to start using Deepstack! To check Deepstack is ready make a request using cURL:

curl -X POST -F image=@development/test-image3.jpg 'http://localhost:5000/v1/vision/detection'

This should return the predictions for that image.

Home Assistant setup

Place the custom_components folder in your configuration directory (or add its contents to an existing custom_components folder). Then configure object detection. Note that at we use scan_interval to (optionally) limit computation, as described here.

Object Detection

Deepstack object detection can identify 80 different kinds of objects, including people (person) and animals. On you machine with docker, run Deepstack with the object detection service active on port 5000:

sudo docker run -e VISION-DETECTION=True -e API-KEY="Mysecretkey" -v localstorage:/datastore -p 5000:5000 deepquestai/deepstack

The deepstack_object component adds an image_processing entity where the state of the entity is the total number of target objects that are above a confidence threshold which has a default value of 80%. The class and number of objects of each object detected (any confidence) is listed in the entity attributes. An event image_processing.object_detected is fired for each object detected. Optionally the processed image can be saved to disk. If save_file_folder is configured two images are created, one with the filename of format deepstack_latest_{target}.jpg which is over-written on each new detection of the target, and another with a unique filename including the timestamp.

Add to your Home-Assistant config:

image_processing:
  - platform: deepstack_object
    ip_address: localhost
    port: 5000
    api_key: Mysecretkey
    timeout: 5
    scan_interval: 20000
    save_file_folder: /config/www/deepstack_person_images
    target: person
    confidence: 50
    source:
      - entity_id: camera.local_file
        name: person_detector

Configuration variables:

ip_address: the ip address of your deepstack instance.
port: the port of your deepstack instance.
api_key: (Optional) Any API key you have set.
timeout: (Optional, default 10 seconds) The timout for requests to deepstack.
save_file_folder: (Optional) The folder to save processed images to. Note that folder path should be added to whitelist_external_dirs
source: Must be a camera.
target: The target object class, default person.
confidence: (Optional) The confidence (in %) above which detected targets are counted in the sensor state. Default value: 80
name: (Optional) A custom name for the the entity.

Event `image_processing.object_detected`

An event image_processing.object_detected is fired for each object detected above the configured confidence threshold. An example use case for this is incrementing a counter when a person is detected. The image_processing.object_detected event payload includes:

classifier : the classifier (i.e. deepstack_object)
entity_id : the entity id responsible for the event
object : the object detected
confidence : the confidence in detection in the range 0 - 1 where 1 is 100% confidence.

An example automation using the image_processing.object_detected event is given below:

- action:
  - data_template:
      title: "New object detection"
      message: "{{ trigger.event.data.object }} with confidence {{ trigger.event.data.confidence }}"
    service: notify.pushbullet
  alias: Object detection automation
  condition: []
  id: '1120092824622'
  trigger:
  - platform: event
    event_type: image_processing.object_detected
    event_data:
      object: person

Event `image_processing.file_saved`

If save_file_folder is configured, an new image will be saved with bounding boxes of detected target objects, and the filename will include the time of the image capture. On saving this image a image_processing.file_saved event is fired, with a payload that includes:

classifier : the classifier (i.e. deepstack_object)
entity_id : the entity id responsible for the event
file : the full path to the saved file

An example automation using the image_processing.file_saved event is given below, which sends a Telegram message with the saved file:

- action:
  - data_template:
      caption: "Captured {{ trigger.event.data.file }}"
      file: "{{ trigger.event.data.file }}"
    service: telegram_bot.send_photo
  alias: New person alert
  condition: []
  id: '1120092824611'
  trigger:
  - platform: event
    event_type: image_processing.file_saved

The second image created on new detections has a fixed filename to make it easy to display with the local_file camera. An example configuration is:

camera:
  - platform: local_file
    file_path: /Users/robincole/.homeassistant/images/deepstack/deepstack_latest_person.jpg
    name: deepstack_latest_person

Face recognition

For face recognition with Deepstack use https://github.com/robmarkcole/HASS-Deepstack-face

Google Coral USB stick

If you have a Google Coral USB stick you can use it as a drop in replacement for Deepstack object detection by using the coral-pi-rest-server. Note that the predictions may differ from those provided by Deepstack.

Support

For code related issues such as suspected bugs, please open an issue on this repo. For general chat or to discuss Home Assistant specific issues related to configuration or use cases, please use this thread on the Home Assistant forums.

Docker tips

Add the -d flag to run the container in background, thanks @arsaboo.

FAQ

Q1: I get the following warning, is this normal?

2019-01-15 06:37:52 WARNING (MainThread) [homeassistant.loader] You are using a custom component for image_processing.deepstack_face which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you do experience issues with Home Assistant.

A1: Yes this is normal

Q2: Will Deepstack always be free, if so how do these guys make a living?

A2: I'm informed there will always be a basic free version with preloaded models, while there will be an enterprise version with advanced features such as custom models and endpoints, which will be subscription based.

Q3: What are the minimum hardware requirements for running Deepstack?

A3. Based on my experience, I would allow 0.5 GB RAM per model.

Q4: Can object detection be configured to detect car/car colour?

A4: The list of detected object classes is at the end of the page here. There is no support for detecting the colour of an object.

m0rphy/HASS-Deepstack-object