Love the idea of a light sensing trigger for the feeder. Light sensing, wind direction, and strength would be a hot selling add on.
Cheers Don.
Give me one with a rain gauge. My fish do not eat feed well when it's raining, so a feeder with a means to detect precipitation and suspend feeding operations until the shower passes might help me out.
I know absolutely nothing about facial recognition systems, but my feeble guess is that such systems map the face, and compare it against an image stored in the system. Features might be reduced to measurements, which could easily be compared against the "stock" photo already on file. A certain percentage matches, 90-100% ??, and it says ok. Lesser percentage matched, no go.
Building on that, let's have that stock photo be of the pond's surface on a calm day, and compare that image to one taken after an initial, test burst of feed hits the water. Analyze the difference caused by the feeding efforts of the fish. In this case however, if the new surface matches the stock photo, or within a predetermined percentage, then very little feeding is occurring and the feeder shuts down without dispensing any more food.
But if the measurements and the math indicate a greater percentage of change AFTER the test burst hits the water, then open up and feed away.
The problems and issues to overcome would be numerous. Windy days causing surface disturbances, leaves and debris falling on the surface, etc. And I'm sure JKB is right....it wouldn't be cheap. Still, fun to envision.