Filter out sentinel-1 images with sensor artefacts or processing error

Hi Everyone,

Is there a way in SNAP or something in Metadata headers to filter out images such as the following which appear to contain processing artefacts or sensor noise? (see attached image for preview)

S1B_IW_GRDH_1SDV_20180108T174936_20180108T175001_009083_0103D3_457A

Thanks,
Sanjay.

Hi @sanjayrana,

Have a look to Sentinel 1 IW GRD visual feature

The only solution would be to mask them out.

M

Thanks. I wouldn’t mind dropping these images actually if there was a way to automatically identify that the images RFI. But if not, is the only option is to manually shift through the images to identify such cases and then put a mask on the affected areas?
Thanks,
Sanjay.

This is very often only found in the cross polarisation (VH). Have you checked th VV image?

Thanks. I haven’t yet but I would need to use both as I work out ratio using VH and VV for classification. I would have hoped that S1-QC team at ESA would have worked out an approach to at least flag the images with RFI. I wonder if a simple pattern recognition type machine learning code can be trained to find these patches.

2 Likes

Flagging of RFI-affected images would indeed be very useful.