Migrating an .ISDef to a mosaic dataset

This topic applies to ArcGIS for Desktop Standard and ArcGIS for Desktop Advanced only.

The mosaic dataset is optimized for image data management. It is the recommended data model for working with collections of raster and image data and is designed to replace the image service definition (created with ArcGIS Image Server versions 9.2–10). The mosaic dataset can be viewed as a dataset in ArcCatalog, ArcMap, and ArcGlobe. It can be used as a catalog or like a raster dataset layer. It can also be used with other geoprocessing tools as a data source.

Creating and editing

The table below lists the similarities and differences between creating and editing a mosaic dataset versus an image service definition.

Mosaic dataset

Image service definition (.ISDef)

The mosaic dataset is created using core tools in the Data Management toolbox's Mosaic Dataset toolset, such as the Create Mosaic Dataset tool.

The image service definition is created with the Image Service Definition Editor toolbar in ArcMap, which is installed as a component of ArcGIS Image Server (versions 9.2–10).

The ArcGIS for Desktop Standard or ArcGIS for Desktop Advanced Desktop license is required to create or edit a mosaic dataset. However, it can be viewed or used as a dataset with the ArcGIS for Desktop Basic license.

The image service definition can be created and edited in ArcMap, with an ArcGIS for Desktop Basic (ArcView) license, and the ArcGIS Image Server editor component and license.

It is a single dataset in a geodatabase comprising the following:

  • Boundary—Polygon defining the extent of the images in the mosaic dataset
  • Footprints—Polygon for each image
  • Tables—Attribute, raster type, logs, and so on
  • Seamline—Optional polygons for each image

It is a set of folders containing XML files and shapefiles comprising the following:

  • Boundary—Polygon shapefile defining the extent of the images in the mosaic dataset
  • Footprints—Polygon shapefile for each image; also used to maintain the attributes for each image
  • Seamline—Optional polygons for each image

Overviews—By default, they are stored in a folder, outside the geodatabase, with the same name as the geodatabase and an .Overviews extension. When you are using an ArcSDE geodatabase, they are stored within the geodatabase by default.

Service overviews—By default, they are stored in a folder, outside the .ISDef folder, with the same name as the .ISDef (<name>.ISDef) and the .DerivedImages extension (<name>.DerivedImages).

Functions are added to define the on-the-fly processing, such as NDVI, stretching, and pan-sharpening.

Processes are added to define the on-the-fly processing, such as NDVI, stretching, and pan-sharpening.

Modifications to the mosaic dataset are made using the user interface or through ArcObjects. There are no XML files to edit.

The properties of the images, the processing, and the image service definition are maintained in XML files that can be edited.

Automation can be implemented through the geoprocessing framework or ArcObjects.

Automation can be implemented through ISCommands, which are unique to ArcGIS Image Server.

The mosaic dataset is limited by the size of the geodatabase. For example, a file geodatabase is limited to 2 TB.

This limitation is not related to the total size of all the imagery files unless they are also stored within the same geodatabase.

The image service definition is limited by the size of the footprint shapefile, which is 2 GB.

This limitation is not related to the total size of all the imagery files.

Imagery is added using raster types.

Imagery is added using raster types.

The paths to the imagery files are hard coded.

The paths to the imagery files are hard coded.

Any of the polygons can be edited using the ArcMap editing framework.

Any of the polygons can be edited using the ArcMap editing framework.

Comparison of a mosaic dataset and an .ISDef

Serving

The mosaic dataset is served directly from ArcGIS for Server as an image service, with the ArcGIS Image extension (a license added to ArcGIS for Server, not a separately installed or registered component).

You connect to image services in ArcGIS for Desktop by connecting to ArcGIS for Server. No separate client is required.

Migrating

The following lists the steps to migrate your .ISDef to a mosaic dataset:

Steps:
  1. Create a mosaic dataset.
  2. Import your .ISDef using the Add Rasters To Mosaic Dataset tool using the Image Service Definition raster type.

    Point to the ImageService.ISDef file within the .ISDef folder.

    Optionally, you could specify the Raster Process Definition raster type and point to the folder containing the raster process definition files (.RPDefs).

  3. Optionally, modify the metadata and functions on the mosaic dataset to match those on the image service definition.

    The metadata and processes that exist for each image are imported when the data is added; however, anything applied on the image service definition is not carried over automatically.

When you migrate the image service definition to a mosaic dataset, the information contained in the .RPDefs are imported for each raster. This includes the processes (functions), footprint geometry, and attributes. Any processes or settings that have been made to the whole image service definition will need to be redefined for the mosaic dataset, such as mosaic method settings, metadata, image service definition processes, and transmission compression settings.

Once you add the rasters, as described in step 2 above, you should not need the image service definition any longer. However, the service overviews or any derived images you created are stored next to the image service definition (by default) and are now being accessed by the mosaic dataset; therefore, you need to be sure you do not delete these files.

When you migrate the .ISDef to a mosaic dataset, you might also need to update the paths to the source or derived images. You can do this within ArcMap or by using the geoprocessing tools, as described in Repairing paths in a mosaic dataset.

Related Topics

9/10/2014