I am sorry if the question appears naive but I am new to using Sentinel and SNAP. I have some sentinel-1 images inside a zip archive with the measurements and all the xml and xsd files. I can open the tif files in the measurement folder in a GIS such as ArcGIS and they appear georeferenced. I need to clean-up the boundary. So I am trying to use the GRD Border Noise Removal tool both in the cli and the snap gui. I want to output a geotiff. The tool cleans the border but is messes up the georeferencing and changes the bit depth to 32 floating point (from 16 uint.) I need to preserve the georeferencing and the bit depth to produce the geotiff whist cleaning up the boundary. Any ideas?
First of all, in SNAP you need to open the whole product and not just the images. To do this, make sure you’re opening the zip file or the product metadata file (For S1 it would be manifest.safe).
Processing in SNAP usually produces Float32 to maintain accuracy as you chain different processors together. At the end of your processing you can scale the data using Convert DataType operator found in Raster/Data Conversion/
Your final step should then be to write to GeoTiff.
Yes I use the whole zip file as the input. Fair enough for the uint/float but what about the georeferencing?
How does it mess with the georeferencing? I think when you export as GeoTiff it expects a map projection. GRDs as they are don’t have a map projection until after terrain correcting them.