Hi,
I’m trying to create a “virtual” reference image that I can use for flood detection. I wish to combine several Sentinel-1 images to create an “unreal” but true image of the areas normal situation (non-flooded). Do any of you have any experience with creation of such images?
Thanks,
Ine
Well, you could co-register and average for example 5 scenes taken in non-flooded conditions. This reduces speckle considerably so the average image looks much sharper.
Thank you, that created a much sharper reference image. However, when I stack the reference and my flooded image there is a lot of noise caused by my flooded image. Any suggestions to how I can reduce the speckle in just one image?
you can speckle-filter it, which will inevitably degrade the resolution somewhat. Do some experimentation to see which amount of filtering suits you best.