[haonan]

Dear SARPROZ
I’m using SARPPROZ’s PSI to process the Sentinel-1 data. I have a question about APS Processing. After Preliminary analysis and Preliminary geocoding, I go to Multi-Image InSAR processing, I have three questions here:
1. I use Linear Trend=30, Height=100, Thermal expans=0.4 and after connections processing, the connection coherence histogram(Temporal coherence) has two peaks. I don’t know if he has any influence on the subsequent operation. Or how to adjust the parameters to modify it?
2. About Reference point selection, I can not find one reference point to make sure every peak of the histogram is zero. Usually, the most inaccurate one is Integrated Residual Height. (Attached is the best point I can choose)
3.Coherence between each slave date and the master date after removing the estimated APS is soooooooo bad. This the most importantly, I have no idea how to solve it.
Based on my screenshots, I think my entire APS estimate was wrong

Attachments:

You must be logged in to view attached files.

[antig]

It is not to mutch information to tell exactly. if You have sentinel 1 images stack, You may use first Smart 5 option (remove thermal manualy). Then You will have data to analyse how big are the movements in AOI

• This reply was modified 2 years, 10 months ago by antig.
• This reply was modified 2 years, 10 months ago by antig.

[periz]

the result is not bad, actually, it’s pretty ok
do not look at the value of the coherence in the temporal plot, that plot is useful only to look at the trend in time, to check whether there are some left signals that were not well modeled
more suggestions:
* u can take a lower threshold for the amplitude stability index, like e.g. 0.7. in this way you will increase the nr of points and the stability of the spatial graph
* why are you using the thermal expansion at this stage? unless strictly necessary, better avoiding it at the APS estimation stage
best

[haonan]

Okay, I will try again by using 0.7 as the amplitude stability index. Using thermal expansion is because I’m processing Sentinel-1 data from 2014 to 2021, I think over this long time series, the results may be affected by thermal expansion. Of course, I will try your suggestion and don’t use it at this time!
Many thanks!

[haonan]

Dear periz,
Thank you so much for your suggestion, I think my result is pretty good!
On this basis, I would like to ask you another question, I’m processing data from two different sensor in one area, so I’m thinking how can I make sure I’m choosing the same reference point in both processing?
Many thanks!

[Author]

Posts