you are required first to identify several pairs of closely matching points on both maps.Well, in all these cases the GCP approach can successfully resolve all your problems in a rather simple way. ( an infamous example surely familiar to many Italian readers is represented by the Cassini-Soldner projection widely adopted by the Italian National Cadastre). Other times it could be a well known CRS unhappily unsupported by the standard EPSG geodetic definitions simply because it's based on some extravagant or not very common projection system completely unfit to match the standard EPSG expectations Just to say, this is the usual condition of many potentially interesting historical maps but in many others cases too it frequently happens to encounter some odd map based on a completely undocumented CRS. In this specific case adopting the usual coordinate transformations based on PROJ.4 and ST_Transform() is completely out of discussion. You could be legitimately interested in merging together both maps, anyway this operation seems to be completely forbidden simply due to the complete lack of critical informations about one of the two maps. Now suppose that only in one case the Coordinate Reference System is well known and fully qualified however (for any good reason) the other map is completely lacking any useful information about its own CRS.

Imagine two different maps obviously representing the same identical geographic area, as the two shown on the side. In this short presentation we'll make any possible effort so to keep the discussion at the more elementary level. Understanding the intended role of GCP is very intuitive, anyway the underlying mathematic and statistic operations are really complex and apparently intimidating. Starting since version 4.3.0 SpatiaLite supports an optional GCP module the GCP acronym simply stands for Ground Control Points.

Ground control points