Consider that the only thing that makes automation useful is the output devices. Next, consider what devices would be needed to do whatever function you want to automate.
Even in large commercial operations only some parts are automated without very large expenses. It is the output devices that are the challenge. And an arduino is not suitable for this application.
I would probably go for something a little more capable like a Raspberry Pi or even a cheap PLC. You will likely see the limitations of the arduino pretty quickly with something like this unless you implemented some sort of field bus network. Still I think the computing capabilities of the arduino may likely be reached fairly quickly.
Still, it is the motion output devices that will do anything useful. The only simple things to control are the water application and the lighting, if you have an indoor garden. beyond that you need motion devices and suddenly it gets complicated.
My Mom does here own gardening, and we have a harsh climate to any metals left outside in the rain/snow/salty air.
If I was to make this here, it would need need a lot more weather proofing.
A lot more people, especially now that the banksters are imploding their reserve USD it seems.... need to grow their own food.
I wish there was enough sun/wind here to make this thing -solar/wind powered, but I doubt it, IDK what batteries are like, we aren't as cloudy as the UK, but not much better.
This is pretty interesting though. I have to admit I have contemplated such a thing on a bit smaller scale. My initial intent was for a fairly small number of plants indoors. It would take various measurements and control various factors such as light duration and intensity, with dosing pumps for food or nutrients and such. Of course it could be scaled up but i never thought about using robotics. That scales much nicer. Kind of embarrassing considering robotic automation is a major part of what I do for a living.
