Processing chronological pairs of Sentinel-1 images in a time series for deformation mapping

Microwave Toolbox
1

I would be interested in hearing whether the following broad-brushstroke approach to land deformation mapping is feasible:
Step 1 - download Sentinel-1 SLC using Alaska Satellite Facility baseline tool to ensure appropriate S1 are used. e.g. 8 SLC downloaded over 8 x 12 = 96 days.
Step 2 - gpt to create coherence images for T1 (time 1) & T2 (time 2), T2 & T3, T3 & T4, T4 & T5, … and export to snaphu
Step 3 - unwrap using snaphu
Step 4 - gpt to create displacement images and export to geotif
Step 5 - Import to GIS and create a time series. Coherence could be used as a mask to filter out low coherence. A method for creating a time series could include averaging over an area e.g. 9 x 9 pixels rather than using single pixels to increase the number of valid data points.

The reason I ask is because of a potentially large loss of persistent scatters using PinSAR through temporal degradation of coherence when comparing all slaves to a single master especially in vegetated areas
By comparing two successive times that are only 12 days apart coherence should presumably be higher thereby providing more data for deformation mapping.

Thanks!

1 Like
2

This makes sense to me, good idea. Basically one of the principles of small baseline (SBAS) processing.

1 Like
3

Thanks and much appreciated! I will master snap2stamps first and see what kind of output I get in terms of persistent scatterers. My study area is a mixture of rural settlements and paddy fields.