In a single CCD camera all of the colour spectrum is recorded by one sensor. In most cases this is done by means of a bayer type array filter (a grid of green red & blue filters) there is usually twice the number of green pixels than red & blue, the CCD basically records a full resolution greyscale image and the colour is added at the processing stage.
As each pixel has a different coloured filter over it the camera has to interpolate the colour information.
This can result in some minor errors particularly when the sensor signal is amplified (such as using a high ISO rating or increased gain)
In a 3 ccd camera the light is split by a dichroic prism, with light going to a dedicated ccd for each primary colour. This means colour is recorded at full resolution for each pixel with no interpolation.
The result is fewer errors and richer, more accurate colour.
The 3ccd system (or 3cmos) is generally used in video, there has been one 3ccd camera, an early Minolta DSLR called the RD-175, but the ccds were not arrranged in the same way as in a video camera and it was more about increasing contrast and resolution than the usual colour benefits that 3ccd brings.
Some manufacturers have ploughed their own furrow, such as the Sigma Foveon system which uses a single chip but with colour wavelength filtering (much like film) rather than a 3xCCD system.
This has been at the expense of overall resolution, as although each colour has full resolution the luma part (the greyscale) is vastly reduced compared to other manufacturers. At present the foveon chip offers 4mp luma resolution.
Sigma have used their own CCD designs to have extra pixels that record highlights to increase overall dynamic range, but their camera bodys are based on pretty old Nikon designs and lack any processing punch, which has limited their appeal to portraiture where speed isn't as much of an issue.
If it is strictly for video then I would choose a 3ccd type camera every time, though this should not be your sole criteria, look for manual overides and mic inputs etc.
In terms of still photography you are best advised to stick to a single sensor type as technology stands. I'm hopeful Sony may bring the benefits of their professional camcorder experience to their still cameras, perhaps with a 3ccd imaging block, but it is hard to see how this could be done without making the camera bodies much bigger, more expensive, and requiring much more processing clout.
As current cameras have very very high resolutions the inherent colour interpolation of the bayer filter is not as much of a problem as there is far more information to work with, however smaller pixel size has made moire pattern more of an issue, with some manufacturers beefing up their anti-aliasing filters to compensate - at the expense of critical sharpness.
The advice would be to shoot RAW, convert to 16bit psd or tiff files for editing & learn how to use the unsharp mask tool.