I'd like a 'potential imaging device', please. It would be a camera which would produce an artificial image of a board in which the brightness and/or colour of each pixel would be determined by the voltage on whatever conductor is there, or some neutral background colour if the pixel corresponds to an insulator.
IC pins would, therefore, be clearly visible as rows of conducting points on a neutral background. The false colour palette would be chosen so as to make power rails and ground clearly identifiable, with a suitably intensity-graded response for analogue levels in between. A conventional visible light camera would be optically aligned and used to aid in orientation and recognition.
If the 'potential camera' were really good, it would have a high enough frame rate to be able to tell the difference between pins that are oscillating and those which are at a dc level. Maybe it could have 'virtual probes', in which some number of individual pixels are repeatedly scanned at a very fast rate, as a sort of non-contact oscilloscope. The rest of the image could be refreshed much less often, a few Hz would be fine.
How could such a device ever work? No idea. Maybe it could be done by putting the board into a vacuum chamber and scanning an e-beam across it? Not quite as handy as a thermal camera, but then again, early ones of those weren't exactly portable and easy to use either.