Using Offset-Tracking for landslide event

s1tbx
#1

Dear colleagues,

I had the idea to exploit Offset-Tracking operator to check a landslide event using S-1 GRDH data.

My experiment was based on the method proposed by RUS for glacier velocity - https://rus-copernicus.eu/portal/wp-content/uploads/library/education/training/CRYO02_GlacierVelocity_Greenland_Tutorial.pdf

The results are nice, but after terrain correction I experience a shift eastward and southward

image
image.png1025×853 212 KB

The red polygon is the affected area, the green dotted line is the sea shore line, the green dots are the registered landslides and the red arrows should be the velocities.

image
image.png1037×849 38.4 KB

The red polygon is the affected area, the green dotted line is the sea shore line, the green dots are the registered landslides and the red dost are the points where the velocities were taken.

On the second picture the offset is visible.

Any ideas why is this happening?

#2

I don’t see a difference between both images, to be honest.
What exactly is shifted and what is the source of the green dots?

#3

@ABraun
All green dots and red arrows should be behind the dotted line which is delimiting sea and land.
Actually I expected the crossed arrows from the first figure to coincide with the red polygon.
This is why I decided that there should be a shift.

#4

oh, I see. Were water areas masked out in the SAR images?
What is the distance between the green line and the border of all points?

Did you test against a background image which of the sources is shifted - the green line or the red points.

#5

@ABraun The shift is on the red points. The green dotted line is aligned with the SAR data.
During processing I didn’t mask water areas.

#6

so you expect this area to be here actually? What is the distance?

grafik
grafik.png1004×830 399 KB

Did you overlay the points with the SAR data before the terrain correction (by using the layer manager)?

#7

Dear @ABraun,
sorry for this little :wink: delay.
What I need to do is like this

image
image.png1025×853 212 KB

since the event I’d like to track is located in the red polygon.
I hope I could do this in QGIS 3.X.X
I was curios why this had happened after applying the OffsetTracking operator.

#8

what is the scale of this offset? It would be nice too see this on a basemap.

How did you export the displacements after terrain correction?
Is it an option to retrieve the points’ location directly after offset tracking (without terrain projection) just based on the created xml file? (example)

#9

Do you still have an archive of this file, and if you have it, could you please share it with me?