- TensorFlow image_processing component
The tensorflow image processing platform allows you to detect and recognize objects in a camera image using TensorFlow. The state of the entity is the number of objects detected, and recognized objects are listed in the summary attribute.
Note: In the show we mentioned this is done via the cloud, however this is performed locally on your device. Thanks hunterjm on Discord for confirming!
- Z-wave Barrier Class Support
This is a big change for users of Z-wave in Home Assistant. If you’ve got a garage door opener (or anything that uses the Z-wave Barrier classes), you can now link these with Home Assistant.In order to acheive this, Home Assistant has forked Open-zwave, and is now running their own fork of Openzwave.
- BT Smart Hub router Device Tracker
- Météo-france sensor : current weather and 1 hour rain forecast
French weather provider
- WeMo Humidifiers
This includes support for the Holmes Smart Humidifier as well
- GreenEye Monitor
The GreenEye Monitor (GEM) integration for Home Assistant allows you to create sensors for the various data channels of the GEM. Each current transformer (CT) channel, pulse counter, and temperature sensor appears in Home Assistant as a sensor, and can be used in automations.https://www.home-assistant.io/components/greeneye_monitor/
- Mailgun, Twilio, Dialogflow Platforms have been changed to use Webhooks
Home Assistant will now generate unique URL’s which these services will need to call. Before upgrading, check the docs for these components!https://github.com/home-assistant/home-assistant/pull/17464
- Vera Climate Platform will now report different states
The Vera platform will now use standard Home Assistant climate states, instead of those states which were provided by Vera. This will improve other integrations like the Amazon Echo and Google Home reporting the incorrect state.
- Sense Energy Monitor now has binary sensors
Watch out for changes to your entity IDs.
- Pollen.com gets some more accuracy
Previously, Pollen.com calculated historical and forecasted trends by merely looking at two data points and determining the slope – not very accurate. This PR shifts to a numpy-driven method (a moving average) that adds more data points and greatly increases the accuracy.
Breaking Change: Because of the longer statistical view, trend attributes will no longer return a value of “Flat”. All automations relying on that value will need to be updated.
Other Noteworthy Updates
- Lovelace Editing
Lovelace is getting ready for release! There is a super alpha release, allowing you to change your Lovelace UI from the UI. Requires that your card has an ID specified.
- Splitting Lovelace Config has been removed
If you have split your Lovelace YAML files up, similar to how you would split up your configuration.yaml file, this has now been removed.You can still use !secret, and !include has been added. However using these you won’t be able to do much in the new frontend editor.
- Refactor for the Amazon Echo Climate support
A few fixes in this release, which should allow voice control of your climate devices via the Amazon Echo a bit easier now.