Hi Valeria,
ok, much better.
Hmm, at first I thought that the angles are too different for it to work, even if you had RGB colors in the scan. But then I noticed that your sample images are from almost the same spot as the scan. My theory is that RC treats each scan as 6 images, into which each station is transformed. The resulting cubes seem to be aligned to the main axes of the coordinate system to which you have geo-referenced the scans. This seems to be at roughly 45° in relation to your facade, which in turn results in an image with a crass vanishing point effect. And this is very bad for photogrammtery to match. So there might be a chance that if you arrange the coordinate system parallel to the facade (if you are able to do that) it might work better. This is really only a guess on my part so don’t put too much hope into it.