Elevation value from DEM file and Elevation Value from Google Earth of elevated surface not matching

Microwave Toolbox
1

This is my DEM elevation display:

image.jpg1600Ă—900 528 KB

This is my Snap module display:

image.jpg1600Ă—860 815 KB

This is my DEM Legend:

This is the Google Earth Display

image.jpg1600Ă—860 792 KB

The Work I have done:

  1. Input the two master-slave image.
  2. Perform Coregistration
  3. Form the interferogram
  4. Remove the Topographic Phase
  5. Phase Filter
  6. Export to Snaphu
  7. Import Snaphu
  8. Execute Phase to Elevation

Platform used:

  1. Windows 8.1
  2. SARbian OS (Debian)

Software used:

  1. Snap 5.0
  2. Google Earth

Problem:
After processing the above steps from -

http://sentinel1.s3.amazonaws.com/docs/S1TBX%20Stripmap%20Interferometry%20with%20Sentinel-1%20Tutorial.pdf
with default values and successfully executing snaphu I have obtained the dem file. Now when I opened the Google Earth and entered the coordinates to check of the same point, I found that there is error. For eg, the value in the DEM file is showing 421 m but in the same point in Google Earth its showing 830 m.
Not only that, it was surprising to see that the values are showing same on the base area (For eg. Yellow region). But for the hill area or elevated area its showing different result. A small change in the value is accepted as in Google Earth page its written that the value is not always accurate. But deviation of 400 m is a big deviation. So can anyone tell me whether its a software error or not?

2

My guess is that it is so steep at the sides that there are phase-ambiguities (the radar is “missing” multiples of 2π in InSAR phase) + perhaps some unwrapping-related problems should explain it.

3

Thanks a lot for your reply. So what to do next?

4

If that is what actually happened, you could look for a pair of images with a shorter baseline. What is the baseline in your current pair?

You could also calculate the “altitude of ambiguity” that tells you how much elevation 2π corresponds to, and see whether adding N times that to 421 meters to see if you get close to the correct value. This is to check whether the theory of missing 2π cycles makes sense or not.

1 Like
5

Sir here is my info about the two images…

image.png1600Ă—900 107 KB

6

233 meters is a long baseline, if you could find something with less than 100m you would have a better chance. The height-ambiguity is 34 meters with this one according to SNAP.

On the other hand if there’s very strong foreshortening or layover in the SAR scene at the edge of the rock, then there’s not much you can do even with a short-baseline pair.

1 Like
7

Thank you sir, for having patience and answering me. I will definitely try with shorter pair.