1: What is the map? Is it a plain raster image? Is it some sort of graphics object in a figure created by some sort of mapping tool that I don't know about?
2: What exactly do you mean by "overlay"? That might mean one of a few different things:
- a binary masking operation -- would that make sense? it would obscure the red dots you mentioned. Is that the point?
- opacity/alpha composition -- that would let the map features show through at least
- any number of image blending methods such as things that are literally called "overlay"
3: I'm assuming that n is just placeholder data and that the actual data is not random. You also mention that the data is a rectangular image. So then the question is what actual transformations are necessary to make it fit the target region. It's unclear what features the actual data has and where they should correpond to the features in the map. All I can assume is that some roughly L-shaped part of the data is kept, some other parts are discarded, and the rest is transformed. Are rigid transformations sufficient?
