I don't know if this question applies to only humans but why can cones see much greater detail than rods? Is it possible to have a rod that can detect light intensity and color?
The spectral sensitivity of photoreceptors expressed is the key to color vision. See figure below for the sensitivity of three-types of cone cells (S, M, L) and rod cell (R, dashed line).
From this figure, one can say rod cells provide information about the "blue-greenness" of vision, however, despite their spectral sensitivity, it seems that in human vision rod cells do not contribute to color vision, because they are highly sensitive to intensity, and thus they are mostly saturated in their response (does not induce firing of downstream bipolar cells) during normal daylight conditions. Rod cells specialize for night vision (scotopic conditions) which is crucial for survival, and under this condition the cone cells are pretty much useless.
Cone cells are each connected to their own neurone. This allows them a great deal of resolution as the brain can interpret the exact position of the cone cell that was stimulated by a light photon. However in order to improve low light vision, multiple rod cells are connected to a single neurone - this is called summation. Whilst it does allow for an action potential to be generated in low light conditions, it greatly reduces resolution as the brain can not know precisely which rod cell was stimulated:
Rods can not detect colour as they only come in one variety - cone cells (in humans) come in a red, green and blue specific form to allow for the perception of colour by the brain due to the relative strength of these signals.
While the answers to date are correct regarding the wiring of rods and cones in the primate (specifically human) eye, they are also fundamentally wrong. Neither rods nor cones perceive color. The brain does. The rods and cones are just the receptors providing signals. The first answer in fact says this in its very last sentence.
As one answer says, during the day the rods are saturated (overstimulated) so the brain ignores them. It uses the components of the cone responses to invent the sensation "color". At night the cones are usually only weakly stimulated, so the brain sees only with the more sensitive rods, and little or no color. This is why brightly colored stars such as Betelgeuse and Rigel still appear only faintly tinged (red and blue respectively).
By the way, some primates have even better color perception than humans with four (or five) kinds of cones. It is speculated that color vision is so good in primates because of the need to judge fruit ripeness to eat. Many mammals have fewer cone types than primates.
All of the above answers are great, and very informative. But they are also technically wrong, in certain conditions. Once you understand them, you'll be able to understand this explanation of why.
The canonical answer is that cones are used for color perception in bright light and rods are used in low light. But rods have a peak color sensitivity that is very distinct from the cones (see the chart posted above). And more importantly, there are light levels at which both rods and cones are equally functional for color perception.
This is known as the "Purkinje effect" or "Purkinje shift". Basically, when light levels dim, your red color perception diminishes first, but your blue color perception is enhanced (or at least doesn't diminish nearly as fast). The specific effect is that red objects get darker much faster than blue ones. But the brain isn't yet just perceiving the blue objects as a brighter gray, so it seems there is some color perception built into the brain based on the rods.
It has been proved that rods do add to color vision at certain conditions, especially at mesopic vision, Purkinje effect. Further testing showed that when only rods and L cones are excited, together they produce identifiable hues, although only two monochromatic lights are being used. One very faint blue light enough for rods to react but too dim for S and M cones to react, and one red strong enough to excite L cones, but too red to saturate rods. Rods may add some color information during phootopic vision as well, but only at low fotopic levels when rods are not being saturated yet. That part is still being investigated.
I have found this article to be very informative... http://journal.frontiersin.org/article/10.3389/fpsyg.2014.01594/full