I have been trying to export a number of scanned paper maps onto my database, and have been just adding them as raster layers as part of a larger group layer. I georeferenced them, and none of them overlap (they are smaller towns around the county). When I create a mosaic dataset and add the rasters to it, the file ends up being around 40 GB. This is odd, because I also used the same process with about 500 maps of the same file size (which slightly overlapped and filled an entire area) and the final mosaic dataset only took up about 5 GB.
Is there a way to make the file size smaller or store the rasters more efficiently?