Except that this problem doesn’t specify distance between horseman, so I think it’s a bit bogus — no need to resolve an individual person to be able to tell that they’re there. And for hair color, if you make assumptions about the clothes being worn, you could perhaps infer color of hair, even if the hair isn’t resolvable (a person being a “single pixel” would have a different hue depending).
With coherent detection I think the separation between eyes would allow for this.