>>16841802 (OP)
Absence of color is actually a presence of 3 frequencies - R G and B at the intensities perceived by our eye as equal.
So what I would do here.
1. First we need to split this image into 3 using half-reflecting mirrors or prisms.
2. Then for image G we keep color as it is. For R we need to shift spectrum to red side, by reducing frequencies, for B we need to shift it to blue side, by increasing frequencies.
3. Then adjust intensity of those images using gray filters and unite them into one with prisms and mirrors.
How do we implement step 2?
- It could be done using photoelectric shit, but it's not truly passive.
- It also could be implemented using fluorescent screens. Project picture on such screen, obtain weak emitted light of changed frequency from other side. Is it passive enough?
- Using nonlinear crystals. For that we need to split image into spectrum and use only specific frequencies (like in spectrophotometry). Although true gray color is hard to obtain, I guess we could approximate it closely.
