Add GMT_IS_IMAGE case

[Chuan1937]

Add a GMT_IS_IMAGE branch to the gmtinit_get_region_from_data() function, read the image header information using the correct image type and GMT_CONTAINER_ONLY mode, extract the boundary coordinates, and then release the data structure.

Fixes #8878

[Chuan1937]

@joa-quim
I add a new case for this problem, and add a test for it.
I don't know if it can solve this problem in julia.

[joa-quim]

@Chuan1937 Thanks a lot for looking into this. However, I don't think it solves it.

Adding a GMT_IS_IMAGE branch doesn't help because in gmtinit_set_missing_R_from_grid() the call to gmtinit_get_region_from_data is always done with GMT_IS_GRID

if ((err = gmtinit_get_region_from_data (API, GMT_IS_GRID, exact, options, wesn, &API->GMT->common.R.aspect)))

and in fact this doesn't seem to be a problem because at the end the code flow ends up in the function that read an image. I have debugged this a bit further but still think there is something fundamentally broken in gmtinit_set_missing_R_from_grid. I think I can now force it to follow a path that should work but can't still base it in a robust logic.

And your test script. Yes, I guess it will work but it is not testing this case. Reading a file works fine, but reading one that is already loaded in memory (and you would need GMT.jl to test this) is where this keep failing.