[marcovanderkooij]

Hello!

I have been trying to configure APS processing such that the APS results are symmetric spatially. When I use APS option parameter R (and DSF) I can control the spatial scale of the atmospheric residuals (e.g. R is 150). However when we use Sentinel-1 the pixel spacing is approx. 4 m in range and 20 m in azimuth. That means the spatial scale is not symmetric (about 600 m x 3000 m). It would make sense to apply parameters that result in a somewhat symmetric result as atmosphere has similar scales in both direction: e.g. Rx = 750 and Ry = 150). I might have missed a good way of handling this situation. Can you please advise on how to deal with this matter.

Second question, I have tried to use ML Range-az (in InSAR parameters) e.g. 5 x 1 (R-Az). I think it is quite effective for situations where it is ok reduce to about 20 x 20 m (multi-looking 5 in range) for the InSAR process (unless spatial resolution is critical). Can you please comment if the multi-looking is the way to get to spatial symmetry for atmospheric effects.

Thank you,

Marco

[periz]

Hi Marco,
yes, the multi-looking factors helps adjusting the size of pixels with the intent to make them symmetric.
And of course you are right that increasing the size of the filter in one dimension for un-symmetric pixels would make perfect sense. However, after a number of tests, we decided not to implement it, and to keep the size of the filter as small as the smallest resolution allows, based on our experience (or as small as the user likes).
best

[Author]

Posts