If you save the image with ImageJ the georeferenced context gets lost in the *.tiff image.
I would simply use the gdal library (with command line tools) to reproject the images.
http://www.gdal.org/
Here is an useful example how you can do that with rgdal (extract the info and reproject):
http://gis.stackexchange.com/questions/65998/how-can-i-use-gdal-to-batch-define-a-projectionSince you are using a IJ macro you can call gdalwarp from within your macro with a native call:
http://rsb.info.nih.gov/ij/macros/ExecExamples.txt(or convert all images with a batch procedure at the end).
I often use rgdal (
http://cran.r-project.org/web/packages/rgdal/index.html) which is an easy to use R wrapper for the gdal library.
If you are interested in here i replied to an answer concerning the use of masks and reprojection with ImageJ and R within my application:
http://bio7.org/?page_id=1918Another possibility is to use the existing java gdal wrapper, GeoTools library or maybe rewrite the *.tiff header (i found this suggestion on the mailing list -
http://imagej.1557.x6.nabble.com/Adding-Text-to-TIFF-Header-td3696853.html).
But since you are using a macro the first suggestions seems to be appropriate.