After running a RPi3 for about 5 years I started to get a failing SD card and regular crashes. I finaly took the step to move my setup to an NCU and run my setup on an NCU.
Now with more processing power and memory I have moved my attention to improving my CCTV system.
I found this video (from rob at the hookup) https://www.youtube.com/watch?v=fwoonl5JKgo and tried to follow the setup
Although it claimed free there was bule iris to buy. (maybe I will but not wanting to pay if I don’t need to I set of looking for a free option)
So I installed Agent DVR which with homeassistant integration and lots of features looked like a good option.
So I installed Deepstack on a windows docker (could not make it work using portainer)
Installed the AI tool from https://ipcamtalk.com/threads/tool-tutorial-free-ai-person-detection-for-blue-iris.37330/
and set about trying to integrate Agent DVR insted of Blue Iris
So what works. Cameras setup in Agent DVR and save an image to a folder. Ai tool detects the images and passes it to deepstack.
and gets an analised image back.
This is where it stops. The AI Tool can send a trigger URL when it detects a specified object. Following the video this is fed back to blue iris to trigger the recording of the HD stream until motion stops.
Can this be fed back to Agent DVR to do the same thing and trigger the recording? What would the trigger URL need to be?
If not maybe the trigger URL could be a web hook that can use an automation to triger the record from the Agent DVR service (Theis service does already work)