Google logo

Segmenting Large Imagery Files

Google Earth Enterprise Fusion does not allow you to import raw imagery source files larger than 80 GB. (Raw size = number of pixels width * number of pixels height * 3.) Therefore, if you have an imagery source file that is larger than 80 GB, you can split it into two or more source files using the gesplitkhvr command.

The gesplitkhvr tool produces a grid of image files designated as rows and columns. For example, if you specify two rows by two columns, the image is split into four smaller files. If you create a separate resource for each resulting source file, it would look like this in the Preview pane:

Split Image

However, you can reconnect the split images by adding them all to a single resource, which would look like this in the Preview pane:

Split Images Reconnected Into a Single Resource

To split a large source file into multiple smaller files:

  1. At the command prompt, change to the folder that contains the tutorial imagery source files by entering:

    cd /opt/google/share/tutorials/fusion/Imagery

  2. Convert a .tif file to a .khvr file by entering:

    gevirtualraster -o path/virtual_raster.khvr usgsSFHiRes.tif

    where path is the path to the folder where you want to save the resulting .khvr file. (See Creating Imagery Mosaics for more information about the gevirtualraster tool.)

    Note

    The path where you save the resulting .khvr file must be on a known volume. (See “geconfigureassetroot --editvolumes” in the Command reference.)

  3. Change to the folder where you stored the virtual_raster.khvr (the value of path in step 2) by entering:

    cd path

  4. Split the .khvr file into a 4x4 grid (16 image files) with a 10-pixel overlap among all of the files by entering:

    gesplitkhvr --rows 4 --cols 4 --overlap 10 virtual_raster.khvr

    The names of the resulting files appear on the screen as they are created. The file names are constructed from the name of the original file (virtual_raster), a row and column designation for each file (starting with -R1C3), and the .khvr extension.

Now you can import the 16 resulting image files into a resource.