When I was more on the ag side of research, I know a former pilot/engineer who had decided to ruin his sons life by helping him on his farm.

I kid. The guy was actually a fucking gangster. He had built/ modified a customer variable rate drill/ spray rig. The used it in combination with a hyperspectral sensor they had in a hole drilled into the up auger of the harvester. They worked with us to develop a hyperspectral model that used middle infrared light to predict protein.

So basically, you run your harvest, and when you are doing this you are tracking both yield and protein content (with the sensor in the up auger). Then, when it comes time to plant, you've got a map of exactly how much nitrogen was used the previous season. Using this you now have the option to create an optimal yield map, or you can target a specific protein content for the grain.

Anyways, no I don't. But I've seen some pretty cool diy shit.