How to replace images in to a mosaic dataset with overviews ?
Do I have to remove old rasters from the mosaic dataset, add the new rasters ?
Or, can I simply overwrite the files of the old images with the new files ?
What do I have to do about the overviews ?
Thanks
Solved! Go to Solution.
Giuseppe
it is hard for us to advise if we don't fully understand your configuration. The mosaic dataset can be used for a wide range of configurations, e.g. to mosaic multispectral imagery from multiple satellites across the entire world, or to manage multiple years of preprocessed orthos covering the same region, with most recent appearing on top, or manage and share multiresolution elevation datasets (or many other examples).
I would encourage you to spend some time reading http://esriurl.com/imageryWorkflows and specifically the many options in https://doc.arcgis.com/en/imagery/workflows/resources/using-mosaic-datasets-to-manage-imagery.htm
I would first make a zipped copy of your geodatabase just in case something goes wrong. If you break the MD you can delete it and restore your original.
You cannot "overwrite" images in a mosaic dataset - every image is a unique record, with path and filename to the source image. You can remove images by selecting them in the attribute table and then right click on the MD in Catalog and select "Remove Rasters"
You should also delete the overviews if the images you're removing contribute to them (If you have a nationwide mosaic and you're only removing images for one small region, you may not need to delete all overviews)
Then you can add the new rasters, presumably build footprints for them (unless they're simple tiles aligned with the SRS of the MD), make sure the new images are appearing at the right scales, and then rebuild overviews. But my advice may be too simple depending on what else is in your mosaic dataset, and how you have its properties configured. See the URLs above for more information.
Cody
How about just creating a new mosaic dataset instead of trying to replace images in a mosaic dataset?
Are the replacement images just better resolution of same area?
Giuseppe
it is hard for us to advise if we don't fully understand your configuration. The mosaic dataset can be used for a wide range of configurations, e.g. to mosaic multispectral imagery from multiple satellites across the entire world, or to manage multiple years of preprocessed orthos covering the same region, with most recent appearing on top, or manage and share multiresolution elevation datasets (or many other examples).
I would encourage you to spend some time reading http://esriurl.com/imageryWorkflows and specifically the many options in https://doc.arcgis.com/en/imagery/workflows/resources/using-mosaic-datasets-to-manage-imagery.htm
I would first make a zipped copy of your geodatabase just in case something goes wrong. If you break the MD you can delete it and restore your original.
You cannot "overwrite" images in a mosaic dataset - every image is a unique record, with path and filename to the source image. You can remove images by selecting them in the attribute table and then right click on the MD in Catalog and select "Remove Rasters"
You should also delete the overviews if the images you're removing contribute to them (If you have a nationwide mosaic and you're only removing images for one small region, you may not need to delete all overviews)
Then you can add the new rasters, presumably build footprints for them (unless they're simple tiles aligned with the SRS of the MD), make sure the new images are appearing at the right scales, and then rebuild overviews. But my advice may be too simple depending on what else is in your mosaic dataset, and how you have its properties configured. See the URLs above for more information.
Cody
Hello Cody,
many thanks for your reply that is very clear. I have followed your suggestions and replaced successfully the rasters. For your information the mosaic datasets is composed by 4 bands images that have the same resolution. The reason of the replacement is due to the bad radiometry of few images.
Thank you a lot
Giuseppe