What's interesting is that the filter array on the CCD appears to be colored Red, Green and Blue. An efficient red filter, for instance, ought to transmit red light through and reflect or absorb all the others. Perhaps it is the nature of the filter material applied over the individual CCD wells.
In my purpose-built CCD camera for astrophotography, I have a monochrome sensor and a filter wheel that rotates 5 different filters in front of the CCD. The individual filtered images are later combined to produce a color JPEG image or whatever.. My filters are fancy/expensive interference filters, and the RED filter sure doesn't look like red! All those precious red photons are transmitted, and everything else is reflected, so it appears as almost the complement of the color.
IMG_2801 by
lmamakos, on Flickr is an illustration of this. The filters in order from 1 to 5 are: (1) IR Block, (2) Red, (3) Green, (4) Blue, (5) narrowband Hydrogen-Alpha emission line filter (13nm-wide passband). The narrowband filter in position 5 is almost a mirror since it reflects pretty much the entire visual spectrum except for a very narrow band color corresponding to a single ionized Hydrogen emission line.
Thanks for the tear down! Those detectors were sure expensive in those days, and while cheaper now, they are still quite dear, especially for the larger devices.