All details are explained on the website, circusscientist.com, I have set it out as a tutorial, mainly for my own benefit in case I need to do something similar again in the future. Machine Learning is actually really easy if you use pre-trained models, which in my case proved entirely sufficient for the job.

I will outline the details here.

Outline: https://www.circusscientist.com/monkey-detecting-sprinkler-overview/

    - just the outline with a list of parts

Part one: https://www.circusscientist.com/monkey-detector-part-1-detect-a-stuffed-toy-and-switch-on-an-led/

    - I test out the circuit on a breadboard, using some LED's and try out the Android app on a stuffed toy monkey

Part two: https://www.circusscientist.com/monkey-detector-part-2-object-detection-in-python-using-pytorch/

     - I try out pytorch and python version of the monkey detector. This would work great on a Raspberry Pi

Part three: https://www.circusscientist.com/monkey-detector-part-3-the-monkey-squirter/    

     - the full circuit I am using for the sprinkler, with parts and a bit about how it works with the Android app.

Standalone Android App: https://play.google.com/store/apps/details?id=com.circusscientist.monkeydetector

     - doesn't control the sprinkler, just plays a siren sound, works for monkeys and baboons.

Still to do: A nice video of some Monkeys getting squirted by my sprinkler (triggered by the app of course!)