I would try to create as goog as possible model with photogrametry to obtain camera positions and then swap that model with lidar model and try to manually align it or automatically if metashape can do that. Not sure if the texture result would be satisfactory, because bad model comes from bad camera positions.