Tops Deburst problem

Hello everyone.

I am processing interferometric information from a state south of Mexico with great seismic activity. I obtained the deformation map where the well-formed fringes can be seen, but as you can notice in the image below, burst merging didn´t go well north of the image. Does anyone knows why?

Steps are: Coregister–>Interferogram formation–>Tops Deburst–>Topografic Phase removal–>Multilooking–>Goldstein Phase filtering–>Geocoding

Images data:

Slave: S1B_IW_SLC__1SDV_20180226T122557_20180226T122620_009794_011B3B_423A

Coherence is 0.35, is this an acceptable value? High values are associated with dense vegetation and water bodies, and low values are in urban areas and zones with scarce vegetation, agricultural fields or bare ground.


In addition, there are zones where decorrelation is evident, corresponding to dense vegetation areas at the sierras and hills, and water bodies at the coast. I think this ok too.

The next step will be to obtain areas with subsequent displacement by the earthquake (landslides, subsidence). I hope!

Enhanced Spectral Diversity ESD. should be applied to tackle the misregistration which appears as jumps between bursts, Please take a look at the following posts,

Source of the post

Source of the post

Source of the post

Coherence is the :


In general coherence value (0 - 1), the highest values are the evidence of the objects stability between the passes,

Basically the urban stable area should has higher coherence values than unstable vegetated area, and in consequently should be more coherent, but in this case the difference is mentioned, leads to ask about the types of Vegetation putting in mind that we have ‘C’ Band,

The height and dense of the vegetation related to the orbit Asc., or Desc. could be an effect as well, in this case “12” days tempor. base line might be not enough to get any change in the vegetation cover,

1 Like

You should try using ESD as @falahfakhri advised. Another issue can be strong azimuthal displacements between the bursts.

The coherence is not very high, but the fringes are still recognizable so it should work. If any trouble during the unwrapping, try multilooking one step further.

Yes, I forgot that. I´m a begginer at this. ESD solved my problem, thanks @qglaude and @falahfakhri.