Vidicon works by exposing light on a photoconductive layer, and reading the resistance of that layer using a scanning electron beam. The equivalent circuit is something like an array of FETs, where their drains are connected to a resistor divider, the top resistor being a photoresistor and the bottom one being a current sense resistor. Difference being, instead of a huge array of transistors and the muxes/counters/shifters to drive them, there's only one electron beam, and it's physically scanned over the image.
Of note, this is different from an ordinary CRT, where the electron beam is high energy, and would splat right through any such coating (also ionizing it, making it completely insensitive to light anyway). Even if that current were first absorbed by a thick metal target behind the photoconductive layer, the current would still flow regardless of resistance: with such high voltages behind it, it looks very much like an ideal current source. In the vidicon, a mesh grid near the face, slows the electron beam (at some expense to beam current, because the grid has some cross section) so that the V(I) characteristic is more modest: the target receives less current (more is diverted to the grid) when more voltage is dropped across it. And hence a signal can be read out.
A similar technique was used in classic Tektronix scopes (465 and relatives), this time to accelerate the beam towards the phosphor plate, improving image intensity and deflection sensitivity, at the expense of beam current and focus (because again, the grid steals some current, causes some scattering to the beam; the grid pattern itself is subtly visible on screen, with the right settings).
PbS (lead sulfide) is just another material used for this purpose, apparently for long-ish wave IR? It's a narrow bandgap semiconductor, not usually very useful because it's so conductive at room temperature (galena cat's-whisker diodes being probably the most famous application), but could do a good job here, at limited sensitivity, or better when cooled.
A CCD wouldn't do anything with it (a variable resistance doesn't couple charge), but an array of PbS resistors could be deposited on top of a CMOS array I suppose (much as I introduced the idea above
). It's probably been used at some point; modern MEMS microbolometer arrays perform much better though, and need less active cooling, or none at all.
Tim