You don't have enough money.
Signal transport alone will put you over $1500 between cable, and extenders, plus a
dome and a
power drop at the camera, and while you certainly can find cameras in that price range, they'll have a security camera look.
That's not even to mention adding other sources, a
switcher, or distribution out to the TV's.
My estimation is that if you found a way to do it in the neighborhood of $1500, the final
image quality of the video shown in the restaurant wouldn't be anything anyone would want to watch, negating the entire purpose of investing in the
system in the first place.
Forgive me if I sound critical. I've seen too many people try to do things like this on the cheap only to discover the best thing they could've done was not undertake the project at all.
The stories I keep encountering are people discovering first-hand and to great expense that commercial video applications are not like
whipping out your camera phone to take a video of your kids. Cameras aren't created equally, lenses matter especially in longer distance throws, and cameras may very well put out 1080 but at a noise quality that distorts the
image so much it might as well be 480.
This is all to say that you should manage expectations of stakeholders before you spend a nickel. Maybe security camera quality is sufficient and they'll be okay with that if they know that from the start. You absolutely don't want to get to the end though and have them figure out they got that instead when they thought they were paying for a broadcast quality
system.