ISIS3 -- USGS Astrogeology ISIS Cube (Version 3)

ISIS3 is a format used by the USGS Planetary Cartography group to store and distribute planetary imagery data. GDAL provides read/creation/update access to ISIS3 formatted imagery data.

ISIS3 files often have the extension .cub, sometimes with an associated .lbl (label) file. When a .lbl file exists it should be used as the dataset name rather than the .cub file. Since GDAL 2.2, the driver also supports imagery stored in a separate GeoTIFF file.

In addition to support for most ISIS3 imagery configurations, this driver also reads georeferencing and coordinate system information as well as selected other header metadata.

Starting with GDAL 2.2, a mask band is attached to each source band. The value of this mask band is 0 when the pixel value is the NULL value or one of the low/high on-intstrument/processed saturation value, or 255 when the pixel value is valid.

Implementation of this driver was supported by the United States Geological Survey.

ISIS3 is part of a family of related formats including PDS and ISIS2.


Starting with GDAL 2.2, the ISIS3 label can be retrieved as JSon-serialized content in the json:ISIS3 metadata domain.

For example:

$ python
from osgeo import gdal
ds = gdal.Open('../autotest/gdrivers/data/isis3_detached.lbl')
$ gdalinfo -json ../autotest/gdrivers/data/isis3_detached.lbl -mdd all

On creation, a source template label can be passed to the SetMetadata() interface in the "json:ISIS3" metadata domain.

Creation support

Starting with GDAL 2.2, the ISIS3 driver supports updating imagery of existing datasets, creating new datasets through the CreateCopy() and Create() interfaces.

When using CreateCopy(), gdal_translate or gdalwarp, an effort is made to preserve as much as possible of the original label when doing ISIS3 to ISIS3 conversions. This can be disabled with the USE_SRC_LABEL=NO creation option.

The available creation options are:


How to create a copy of a source ISIS3 dataset to another ISIS3 dataset while modifying a parameter of IsisCube.Mapping group, by using GDAL Python :
import json
from osgeo import gdal

src_ds = gdal.Open('in.lbl')
# Load source label as JSon
label = json.loads( src_ds.GetMetadata_List('json:ISIS3')[0] )
# Update parameter
label["IsisCube"]["Mapping"]["TargetName"] = "Moon"

# Instanciate new raster
# Note the USE_SRC_MAPPING=YES creation option, since we modified the
# IsisCube.Mapping section, which otherwise is completely rewritten from
# the geotransform and projection attached to the output dataset.
out_ds = gdal.GetDriverByName('ISIS3').Create('out.lbl',
                                              options = ['USE_SRC_MAPPING=YES'])
# Attach the modified label
out_ds.SetMetadata( [json.dumps(label)], 'json:ISIS3' )

# Copy imagery (assumes that each band fits into memory, otherwise a line-by
# line or block-per-block strategy would be more appropriate )
for i in range(src_ds.RasterCount):
    out_ds.GetRasterBand(1).WriteRaster( 0, 0,
                                        src_ds.GetRasterBand(1).ReadRaster() )
out_ds = None
src_ds = None

See Also: