Optimizing mosaic datasets
There are often multiple ways in which a mosaic dataset can be optimized for better performance or quality. In some workflows, this optimization does not need to be done at the start, but can be implemented after the images are served. This is sometimes referred to as providing a graded service, where the grade of the service improves over time. The advantage of this approach is that imagery can be quickly served when value is highest and then optimized.
Optimizations can include the following:
- Changing the format of the imagery to a format that is faster to access
- Refining the geometry of the imagery—typically, for both satellite and aerial imagery the initial images are only known with approximate orientation parameters and can later be refined by using better ground control or postprocessing of sensor parameters or refining the terrain model
- Refining the radiometry of the imagery—in some cases, obtaining the most visually appealing imagery can require additional processing of the source imagery or determination of better enhancement parameters
- Changing the properties of the mosaic
For large mosaic datasets it is important to index the fields that are used by the mosaic, such as the fields used in ByAttribute queries.
Mosaic dataset maintenance
Maintenance of mosaic datasets and image services typically involves the addition of new imagery. With the pattern of creating source and derived mosaic datasets, this requires both to be updated. There are three ways in which source mosaic datasets are updated:
Adding the rasters to the original source mosaic dataset—typically, the same raster type used to originally add data is used to add new data.
Running the Synchronize Mosaic Dataset tool on the source mosaic dataset—this reruns the originally used raster types and will add or update parameters of any new sources in the original workspaces. You need to be aware of the parameters you set, because this tool can also overwrite some of the changes you may have made to the items in the mosaic dataset.
- Using a new source mosaic dataset—in this case, a new source mosaic dataset is created to reference the new images.
Once the source mosaic dataset is updated and checked, there are two methods by which the derived mosaic datasets are updated:
- The Synchronize Mosaic Dataset tool can be run on the derived mosaic dataset. This will check all the source mosaic datasets to find new records. If a new source mosaic dataset was created, then this needs to be added with the Add Rasters To Mosaic Dataset tool using the Table raster type.
- Since the creation of a derived mosaic dataset is generally very fast, the derived mosaic dataset is recreated. This is often deemed the quicker option.
Referenced mosaic datasets do not need to be updated (or re-created), as they reference the derived mosaic datasets.