Dear @khaldi ,
My post was answering a question asked by hardbasova but it got deleted in the meanwhile. So no need to be offended - i don’t think there are stupid questions in here. Sorry for the misunderstanding.
Have you noticed the values around the boundary areas of @janaya 's created DEM? They are quite different with the values in the core regions of the image.
There was a dividing line along the red arrow. What’s the reasons? And how to solve this problem? @ABraun
Have you noticed the values around the boundary areas of your created DEM? They are quite different with the values in the core regions of the image.
There were two dividing lines in the left and right part of your image, respectively.
What’s the reasons? And how to solve this problem? @ABraun
the strange values at the edge of images result from different coverages of both images.
Please have a look at this post and especially falahfakri’s answer
I have the perfect setting for Sentinel-1 InSAR, desert-like conditions, no vegetation cover, very high coherence (> high temporal coverage (my baseline is 12 days) but keep getting miserable results. My steps:
S1 TOPS Coregistration with ESD
Interferogram generation
Goldstein Phase Filtering
Phase unwrapping with snaphu
Phase to elevation
It seems that already the interferogram is below what it could be. The topography is clearly more distinctive, even in a 30m SRTM. But I cannot get close to that with Sentinel-1 data.
I tried different image pairs to reduce the chance of atmospheric distortions, but the results are the same. Any ideas what else can cause this low quality?
I think you are hitting the limits of the technique, as indeed the coherence is as good as one could ever expect it to be. Perhaps the baseline is quite short? Longer baseline increases sensitivity to elevation, while the atmospheric/ionospheric noise-level stays the same.
ps. great looking coherence-image, are the river-systems current ones or perhaps paleo-ones (covered by sand)?
thank you for the comments. I’ll try the baseline tool of ASF to have a better image pair with higher baseline.
The low coherence areas are wadis which hold coarser sands and are hold water (or at least are saturated) for short times.
the images above had a perpendicular baseline from 13 m, not enough - you are right.
The test below was made with an image pair of 12 days and 138 m, but results are still not very good. I wonder why it worked so well in the Turkey example from above and there are hardly usable results in Mali where conditions are even better.
You can only keep looking at interferograms and hoping to find a pair with the least amount of ionospheric/tropospheric disturbance. What is your use-case BTW? Isn’t SRTM good enough in that area?
Actually I’m just curious about the quality which can be achieved. We use the dem for runoff modeling and groundwater exploration. If some fractures could be enhanced I would be happy.
But I’ll keep testing.
Hi everyone! I have a pair o images with Bper=103,29m and model. Coherence=0,89. I have processed the pair and when it comes to the Histogram of Phase I get this:
I am trying to create a DEM of Nassiryia area. I know that the baseline is not ideal for realizing a DEM (i.e. 53.48), while the coherence is reported in the histogram:
As you can see, they are quite different, but the second one is the real aspect of the topography. Also, I have compared the second output (i.e. bands=“elevation”) with the elevation in Google Earth, but the two data are different. Could is it a problem caused by wrong baseline?
Moreover, I save the dem as GEOtiff, but when I open it on ArcGIS, this is the image:
I know that the “elevation” image is the right while the “elevation_VV” image is wrong. Can save only the “elevation” image as GEOtiff without the “elevation_VV”?
Thank for your reply and I hope I am not hijacking the thread here!
elevation_VV is the one genetated by Sentinel-1, the other is the DEM used for Terrain Correction. The first, although the quality is not perfect, is your result.