The project defines new raster types in ArcGIS Pro to enable to read and load Sentinel-2 L2A product tiles available for free at Amazon S3.
Sentinel-2 satellites sense visible and infrared spectrum in 12 bands. Rasters that are processed by the pipeline are available in 2 products L1C - top of the atmosphere and L2A - bottom of the atmosphere. These raster products are called granules available at no cost to download from Amazon S3.[See Open Data on AWS] (https://registry.opendata.aws/sentinel-2/) Only data transfer costs from AWS S3 to you computer have to be payed, few cents per GB.
The granules are also cut to 100x100km tiles according to the military grid (kml) which is very handy and easy to work with.
For Level-1C and Level-2A, the granules, also called tiles, are 100x100 km2 ortho-images in UTM/WGS84 projection. Download the Sentinel-2 tiling grid kml. The UTM (Universal Transverse Mercator) system divides the Earth's surface into 60 zones. Each UTM zone has a vertical width of 6° of longitude and horizontal width of 8° of latitude. Further Details
The raster storage on Amazon S3 is organized to a file system structure with metadata and images for every band in separate files. This Sentinel-2 raster type allows you to read the file structure and load the rasters to mosaic dataset in ArcGIS Pro.
3 new raster products become available after installation
- Sentinel-2-L2A-10mTile which is a 4-band raster (red, green, blue and near infrared) with 10m/px resolution
- Sentinel-2-L2A-20mTile which is a 9-band raster with 20m/px resolution
- Sentinel-2-L2A-20mTileCloud which is a 10-band raster with 20m/px resolution, similar to Sentinel-2-L2A-20mTile with additional cloud mast raster band B00.
These multiband rasters are created from source images in the R10m, R20m folders.
Acquisition Date, Cloud Coverage and Vegetation Percentage attributes are also added to filter resulting rasters.
To install create a new directory under
C:\Program Files\ArcGIS\Pro\Resources\Raster\Types\
named Sentinel-2-Tile and copy the Sentinel-2-Tile.py python file and raster function templates *.rft.xml to the directory.
After restarting ArcGIS Pro new raster types will become available in the Add Rasters to Mosaic Dataset geoprocessing tool. It has been tested under ArcGIS Pro 2.2 and 2.3.
You can use standard geoprocessing tools, e.g. Add Rasters to Mosaic Dataset, to load the rasters to a mosaic dataset. Just use the metadata.xml file in the source tile root directory as the raster file.
However, I also provide python geoprocessing toolbox with tools that help you to create appropriate mosaic dataset and add rasters to it. Three tools are included - Create Mosaic Dataset, Create Cloud Mask FeatureClass, Add Tiles.
If you like to include it to a larger scenario you can use the sample script (SentinelImporter.py) that can help you to create the mosaic dataset and add all tiles from a directory (recursive).
To create a mosaic dataset for your imaginery just call (just take care of the spatial reference)
SentinelImporter.createMosaicDataset(workspace, "mosaic_dataset_name", "10m", arcpy.SpatialReference(32634))
Use 10m, 20m or 20c to specify the raster type Sentinel-2-L2A-10mTile, Sentinel-2-L2A-20mTile or Sentinel-2-L2A-20m CloudMask respectively for the mosaic dataset.
To import all tiles in a directory and its subdirectories, call
SentinelImporter.importTiles("E:/Sentinel_tiles_from_amazonS3/", "E:/Sentinel2.gdb/mosaic_dataset_name", "10m")
Additionaly the script can generate also a cloud mask featureclass by parsing qi/MSK_CLOUDS_B00.gml file. Just have a look on the script code. It's only 3 lines altogether...
# create tile mask featureClass
cloudmask_featureclass = CloudMask.createFeatureClass(workspace, "CloudMask", arcpy.SpatialReference(32634))
# create mosaic dataset
mosaic_dataset = SentinelImporter.createMosaicDataset(workspace, "S2-10m", "10m", arcpy.SpatialReference(32634))
# load rasters
loadedRasters = SentinelImporter.importTiles("E:/Sentinel_tiles_from_amazonS3/", mosaic_dataset, "10m", cloudmask_featureclass)