Photo and depth from reality capture to material?

I have an ortho projection tiff  with depth  done from Reality capture  . It's a photo with shadows  and a height map from photogrammetry +lidar reconstructed surface mesh.   

Can I use Sampler AI somehow to    a)  get rid of shadows  including the small ones from surface details?   

                                                                b)  make the height/depth  from photogrammetry  more detailed   or just  less blurry normal map ? 

I mean not inventing the normal / height maps on its own but follow the existed height? Just by making less blurry normal map details from it?   Fixing  photogrammetry artifacts at the same time  maybe?