So I see there’s a way to setup facial recognition and automobile plate recognition.
Is there a way to configure recognition such that it could measure the percentage of a given hue in the entire image and return a value based upon the calculation.
Or…perhaps a way to establish what a camera scene should normally look like, then report any abnormalities.
Here are two impractical applications of the idea:
A camera mounted over a kitchen sink. The “on” state would be set when the sink is clean with no dishes in it. If anything interrupts what a clean sink should look like, it would be turned “off”. Or, it would be a measure of the percentage of the image that was different (pixel calculation.)
A camera pointed at a play area for kids that’s cleaned up and in order. If anything is out of order, the room isn’t clean yet…boolean on/off. Something like that.
Automatic window washer robot…but he’s as temperamental as his brother, the Botvac. Morons keep getting stuck on wires. I have no cameras at this point…so the whole idea is on hold…