You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I just saw that Noonlight has opened their API. I was playing with it a bit today, and had some thoughts about other things that could be done in this switch:
I see that you already are thinking about deciding what service (fire/le/medical) to post true on based on what sensor(s) have fired. Maybe do it off of a group (i.e. put sensors in dispatchFire, dispatchLE, dispatchMedical groups).
Add person entities to the alarm based on a config entry. That way any household member or your neighbor can turn off the alarm without knowing your pin. Not sure if noonlight calls them all or not.
It would be really interesting to also call the verification API with video or snapshots from cameras asking if there is a person in the image.
I'm thinking of implementing a two-hit / one-hit verified, dispatch, to reduce false alarms:
a sensor gets tripped
a) notify everyone on the notification list
b) grab some images and send them for verification with a short (30 sec) timeout
c) send everyone on the notification list the images
d) if the verification task comes back as conclusively verified, treat that as a second sensor tripping.
dispatch on two sensor trips
The text was updated successfully, but these errors were encountered:
Greetings,
I just saw that Noonlight has opened their API. I was playing with it a bit today, and had some thoughts about other things that could be done in this switch:
I see that you already are thinking about deciding what service (fire/le/medical) to post true on based on what sensor(s) have fired. Maybe do it off of a group (i.e. put sensors in dispatchFire, dispatchLE, dispatchMedical groups).
Add person entities to the alarm based on a config entry. That way any household member or your neighbor can turn off the alarm without knowing your pin. Not sure if noonlight calls them all or not.
It would be really interesting to also call the verification API with video or snapshots from cameras asking if there is a person in the image.
I'm thinking of implementing a two-hit / one-hit verified, dispatch, to reduce false alarms:
a) notify everyone on the notification list
b) grab some images and send them for verification with a short (30 sec) timeout
c) send everyone on the notification list the images
d) if the verification task comes back as conclusively verified, treat that as a second sensor tripping.
The text was updated successfully, but these errors were encountered: