How to Fix ALOS-1 InSAR Ramps?

Microwave Toolbox Interferometry
#1

Hello,

I’ve recently started experimenting with ALOS-1 and I’ve been using the suggestions from previous ALOS-1 threads. I’m just experimenting with ALOS-1 data and I want to use it with a timeseries analysis eventually such as SBAS or PSI. To my knowledge, there is no major displacement in my test data.

My output seems to be good however is there anyway to fix the ramps using SNAP or other freely available software? From what I’ve seen there is ISCE but I’m still trying to install the open-source version on Github. The installation process is a bit complicated.

If there is also any other feedback regarding this output please do share. This is my first time working with L-band data. Thank you.

Sharing my analysis details:

Scene:
ALPSRP139530000
ALPSRP132820000
Baseline: 28m
Temporal Difference: 46 days

Processing Chain showing modified parameters:

  1. Band Select
    HH Band

  2. Coregistration
    Resampling: NONE
    Num GCPs: 10,000

  3. ALOS Deskew (Data from ASF)

  4. Interferogram
    Degree Flat Earth Polynomial: 8
    Number Flat Earth Estimation Points: 1001
    Orbit Interpolation Degree: 5
    DEM: SRTM 3 Sec

  5. Topophase Removal
    Orbit Interpolation Degree: 5
    DEM: SRTM 1 Sec

  6. Goldstein Phase Filtering

  7. Terrain Correction (For image preview)

image
image1046×714 432 KB

#2

L-band is more sensitive to ionospheric noise. It is possible that these partly cause these ramps. I made similar experience. We discussed this here: Interferogram from Alos palsar 1.1 data

At the moment, SNAP has no operator to remove ionospheric noise, but if you want to give it a try, you can use GITASAR: SAR image processing for beginners

#3

The ISCE2 software on GitHub has several installation methods, so I hope you can find one that works. The stripmapApp pair-by-pair and stripmapStack stack workflows both include estimation of ionospheric corrections using the range split-spectrum method. This works well for ALOS-1 data. As @ABraun wrote, the L-band wavelength of ALOS is much more sensitive to ionospheric effects than C-band data from Sentinel-1, Envisat, and ERS.

1 Like
#4

I was hoping there was a new update since the threads are a few years old :slightly_smiling_face:

If that’s the case I will keep trying with ISCE since that appears to be the best software to handle ionosphere noise. Thank you for the advice.

#5

Split-spectrum processing is going to be implemented in SNAP9.

#6

That’s great to hear! Looking forward to SNAP 9 :slightly_smiling_face: