In rock surface stereophotogrammetry and in digital borehole imaging the small opening of a single fracture trace often stays in subpixel scale, and can therefore not be solved in a straight forward way.
Thank You Richard, I agree. I am a geologist and a mineralogist by my first degree, and we have once solved this puzzle with my colleagues using a pattern recognition approach for digital borehole images. Now I am looking for similar solutions with an photo-realistic point cloud case. Point cloud itself does not give strong enough data, but the RGB-image draped to the point cloud just might be able to do the trick.