I am having difficulty applying mask files (GML) as footprints to a mosaic dataset that contains Airbus DIMAP data. The mosaic dataset's name field (sourced from the DIM 'dataset' tag) has 1:1 cardinality with the GML 'parent ID'. When adding these products 'as raster' to the mosaic, the footprint is defined by the raster extent. Is there a capability to leverage the underlying mask file programatically?
Unless I'm misunderstanding, you don't want to add the masks as rasters into your MD of imagery. The system will treat them as separate rasters, not a mask to apply to other records in the MD.
Are the mask files rasters? the footprints in a mosaic dataset are a feature class (polygons), so ideally if you have a polygon that defines the valid area of each image, you can use this tool
to define the proper footprint very quickly.
If you don't have appropriate polygons, but those mask rasters are populated with binary values (0 and 1? or 0 and 255?), you can use "build footprints" to generate the footprints. There may be a more elegant approach (one of my colleagues may offer further comments), but what I'm envisioning is:
It's always helpful to let us know your software & version - are you on ArcGIS Pro 2.2?
Yes, you are correct, I am attempting to import the masks via the mosaic dataset geom tool. The issue is that there is not 1:1 cardinality and therefore I cannot reliably use the tool due to this assumptions - This tool matches the feature in the mosaic dataset with the feature in the feature class based on a common attribute field.
Using FME, I can see various unique ID's in the DIM_*.xml and the mask.gml. Unfortunately, none of these ID's are present in the mosaic dataset attribute table and therefore cannot be used. The only attribute available in the mosaic is the DIM dataset name value.
Sorry, I don't understand your data.
No, there are multiple masks per image.
This is all resolved now as I was able to join the mosaic dataset paths (with some string manipulation) to create unique ID's for each mosaic item that corresponds to the mask ID. Once this was established, I then used cursors to transfer the geometries across and imported it back as the footprint.
I have been using ArcGIS Pro for all of this work and have never had the GML mask file automatically applied to define the footprint for each DIMAP image. This was using the raster types as defined in ArcGIS Pro or in MDCS - per each sensor (e.g. Pleiades).
One of my colleagues just confirmed that ArcGIS should automatically apply the GML mask file to define the footprint of each DiMAP image. Are you using the correct raster type when you ingest these images into the mosaic dataset?