Given a big set of aerial photos taken. 7000x5000 pixels.
What is the best way to serve these in a WMS. When reading up about different WMSs, I get confused about the complexity and have a hard time finding the simple answer to how we can serve large about of images.
One of the confusing things are raster layers. I assume we need to serve raster layers since our original data is aerial photos from a plane. My readings seeem to lead me on the path that a raster layer is one big picture which the WMS then slices up. But its really not possible to stich all the aerial photos together to one big picture. The other path is to create many layers with one aerial photo in each and therefore have a raster layer per image, which also seems stupid.
Is there any WMS implementations out there where we can hook our images into such when a request comes in for a BBOX, then we tell it what image it should use and then it slices it up and not having to have one big image in memory with the hole world. (An answer to all this could be, how do openstreetmap create its internal map?)