Georeferenced Files - GeoRef
The georeferencing system produces zip files that contain various files. These zipped files are named using the convention of PhotoID-GeoreferenceMethod-Variations.FileType. The table below explains the Variations-FileType part of these.
File name endingExplanation
warp.tifa geotiff version of the georeferenced astronaut photograph transformed and warped
warp_metadata.txta text file that includes ancillary geotiff metadata such as average RMS error of georegistration
manual-FGDCFederal Geospatial Data Consortium 1998 standard compliant metadata
manual-arcgis_gcp.txtan ArcGIS text file listing tiepoints (column 1 & 2) and the corresponding ground control points (column 3 & 4) used for alignment and alignment fit measures
manual-qgis_gcp.pointsa QGIS points file listing tiepoints and the corresponding ground control points used for alignment and alignment fit measures
uncertainty-warp.tifThis is a warped/aligned version of the uncertainty image. It contains two floating-point channels. The first is the uncertainty as in the non-georeferenced version of the photograph and the second is a "mask" that is 0 where there is no uncertainty data (the warped image does not exist there) or 255 if there is.
no_warp.tif*This is a GeoTIFF version of the photo that is unmodified (unwarped) in an image sense, but contains metadata header fields indicating the list of control/tie/correspondence points found for alignment and some alignment fit uncertainty measures. This version gives an end user all they'd need to create their own aligned image from the embedded control points.
no_warp_metadata.txt*This is a text file containing a formatted dump of the header fields present in the -no_warp.tif version.
uncertainty-no_warp.tif*This is a special synthetic image (single channel floating point) where the number at each pixel represents the uncertainty (standard deviation) in meters we estimate for our fit at that pixel. This provides data to do automated analysis of the relative accuracy of our alignment at each pixel -- it will be more accurate near tie points, and worse further away.
*These files are no longer generated.