When we see pictures of dinosaurs, they are usually some color.
How do scientists know the skin colour of a creature that’s been extinct for 60 million years? surely no skin has survived anywhere near long enough to be studied? Is it just a guess? How do we know the dinosaurs were not all pink or something?
We are just guessing at the actual colors but color patterns based on fossilized tissue are being revealed. It’s reasonable to assume patterns would use different colors and some reasonable guesses can be made about the range of pigmentation that dinosaurs would have had. But until further evidence is available we are still guessing based on similarities in skin structure among dinosaurs and other animals.
Adding to what TriPolar said - for feathers, at least, some pigmentation has survived, and so the first true colors of some feathered dinosaurs are now known.
But as far as skin is concerned, as far as I know, no colors have been identified.
By analogy with existing animals, mostly. Used to be, all dinosaurs were portrayed in the grey-olive-green colour ranges of Zallinger’s classic Peabody mural. But sometime in, I guess, the 70s, around the same time the dinosaur renaissance started filtering down from the paleontologists to the rest of us, dinosaurs started getting depicted in more varied colours and patterns. Gregory S. Paul, a close associate of Robert Bakker, was at the forefront of this movement.
They were purple. Go ask Barney. 
“Die.” he suggested, laughing merrily.
I love you you love me. 
As has been said, depictions of skin colors in dinosaurs are pretty much speculative, that is to say, made up. There is some color information available in fossil feathers.
Originally dinosaur colors were based on analogy with large modern mammals like elephants, rhinos, and hippos, and large reptiles like crocodiles and tortoises, which are dull colored. When it was realized they were actually more closely related to birds, they began to be depicted with more birdlike colors and patterns. Since both reptiles and birds have good color vision (with four color receptors, unlike our three or the two of most mammals) it’s almost certain dinosaurs did too.
Yes, this.
I think it was also Horner and others’ work with the hadrosaurids like Maiasaura at Egg Mountain, and various trackway studies, which indicated a mammalian-like degree of herding and care of the young - you start seeing stripes and spots and more camouflage markings coming into the art as the popular perception of dinosaur behaviour changes - I’ve seen Lambeosaurus coloured like an okapi, and parasaurolophus done in gazelle colours.
Although I’m always amused by the ceratopsians done with eyespots 
'cos that’ll scare T. Rex…
We can also make some guesses based on what value color might have been to them. Color can serve a number of possible purposes:
Camouflage-- Creatures can try to hide from either their predators or their prey (or both). This usually leads to splotchy or stripy patterns. If your predators or prey have color vision, it also leads to colors which match your environment. In many environments, that means shades of green. This would be irrelevant for the larger dinosaurs (there’s no way you’re going to be able to hide a T. rex, no matter its color, and a brontosaur wouldn’t need to hide even if it could), but it could be relevant for, say, velociraptors.
Mating displays-- Many species, including most modern birds, use colors to attract mates. The details can vary widely, but in general, this leads to bright, distinctive colors, with elaborate patterns. Note that, in most birds, these bright colors are only found in the males, while the females, with more need for camouflage, are usually more drab.
Warning-- Some creatures with unobvious defense mechanisms (usually some sort of poison) have distinctive coloring to warn potential predators that they’re not worth the trouble. These patterns are usually fairly simple but striking, like a skunk’s white stripe. If the potential predators have color vision, they’ll also involve a single bright color, like orange. If any dinosaurs had venom, they probably had coloring like this. Again, probably not for something like a T. rex, which doesn’t need color to make clear that it’s dangerous.
Protection from the sun-- Some creatures will have pigments to protect them from too much sunlight. This is the primary cause of color variation in humans, for instance. Pigments of this sort are usually dark in color.
Incidental-- If none of the above reasons apply, an animal will just be whatever color the functional chemicals it uses happen to be. This usually gives you something dull, in the beige or gray categories.
Colors are built up two ways in animals: pigments, and structural colors. Analysis of any remaining pigments in a fossil can give some indication of coloration. These tend to be browns, greys, blacks, reds, and yellows. Structural colors, as the name implies, rely on the physical structure of something for color, using refractive effects, This is how we get most blues and silvery colors in nature as well as iridescent effects.
So, if a fossil has a feather or scale structure that will refract blue light in combination with a yellow pigment we can say that said animal (at least on that portion of its body) was green. We probably won’t know the exact shade of green, and you have to worry about the environment accidentally contaminating the specimen, and a bunch of other stuff. If a feather (as another example) doesn’t have a refractive effect but does have red pigment that part of the animal was red. If an insect carapace has a certain structure we might be able to say with some (although not perfect) assurance that it was iridescent gold and green.
Which sounds great but most of the time we can’t make such determinations because the fossilization process can chemically alter the specimen, making pigments difficult or impossible to detect, and can also compress it sufficiently to alter the fine structures, or not preserve the fine structures, making it impossible to determine structural coloration. It takes an unusual fossil to provide coloration information.
Some coloration patterns occur again and again in nature, one of the best examples being counter-shading. That’s when animals are dark on top and lighter colored on the bottom. It’s especially common with fish. Counter-shading an illustration of a fossil fish is a pretty good guess. You also see it in snakes, turtles, crocs and their cousins, deer/antelope, deermice, and lots of others.
So it’s possible to make some sort of a directed guess regarding color in extinct animals, but none of the guesses are likely to be perfect representations.
I saw the Thread title, and thought : Dinosaur Street Gangs?