Tensorflow seems really interesting! But, how do I use the officially supported tensorflow on hass.io running on a RPI3? Will you do a writeup?
I have a dafang with custom fw which I was thinking should provide images.
Thanks!
Writing one now.! Note that Hassio is not currently supported
3 Likes
I’m trying to keep my solution on-prem. But that looks awesome!
Hi! where you able to complete? Or do you know a good tutorial with tensor flow native integration in home assistant?
Tks!
Tks! Great tutorial and I will follow it in detail. Only one question: Tensor can learn to distinguish something between a closed gate and a open gate?
On the example i noticed he can identify a dog. However in my situation for my situation would be important to not only identify the gate, but check if it’s a open gate or clsed.
Is there any good tutorial to teach this to tensorflow?
For binary detection I think you would be better off with classificationbox https://www.hackster.io/robin-cole/motion-activated-image-capture-and-classification-of-birds-6ef1ce
This is great!! Just what I need. I will create a script to do a automated download of all images during one day, to understand what is a “closed gate”, then i can assume that everything else will be an open gate.
Just one question:
Any ideia if it’s possible to set a region, to ensure that I’m only getting the gate on the image? This would highly reduce the false negatives.
Tks!
To crop the region of interest pass the image through https://www.home-assistant.io/components/camera.proxy/