Please don’t post in the group, post in the forum!
https://www.sarproz.com/groups/sarproz-group/forum/
Sentinel-1 SLC IW intensity grayed out, corregistration not possible
Tagged: tops
- This topic has 15 replies, 5 voices, and was last updated 8 years, 6 months ago by Yuxiao.
-
AuthorPosts
-
-
January 25, 2016 at 6:46 am #1424
Hi,
I’ve found some strange error while working with SLC IW images from Sentinel-1 usually starting from November, 2015. Images from >Nov,2015 in more than one track have grayed out intensity and corregistration fails everytime I am using them.
Few examples are:
SLC IW from ASCENDING track with relativeorbitnumber:102 over the area centered around 49.000N 21.233E:
S1A_IW_SLC__1SDV_20151124T162600_20151124T162627_008749_00C763_7587.SAFE
S1A_IW_SLC__1SDV_20151206T162605_20151206T162632_008924_00CC60_DC67.SAFES1A_IW_SLC__1SDV_20151218T162559_20151218T162626_009099_00D12B_DF64.SAFE
S1A_IW_SLC__1SDV_20151230T162604_20151230T162631_009274_00D62E_B765.SAFE
S1A_IW_SLC__1SDV_20160111T162558_20160111T162625_009449_00DB29_5847.SAFEOR SLC IW from ASCENDING track with relativeorbitnumber:175 (screenshots attached) S1A_IW_SLC__1SDV_20151117T163423_20151117T163450_008647_00C499_7D8C.SAFE
S1A_IW_SLC__1SDV_20151129T163417_20151129T163444_008822_00C96B_B5D8.SAFE
S1A_IW_SLC__1SDV_20151211T163422_20151211T163449_008997_00CE55_5942.SAFE)Many thanks for your response.
Matus
Attachments:
You must be logged in to view attached files. -
January 25, 2016 at 8:09 am #1427
I do not know what you mean with “grayed out”.
I’d suggest to show you a comparison between a normal image and what you call a “grayed out” one.
Isn’t it that it was raining? Anyway, if problems, I guess they belong to the data, not to the extraction code…
About the examples you attached, the first 2 images look coregistrable but shifted.
The third one simply looks different.
Let’s wait also for a comment from Yuxiao.
bye -
January 25, 2016 at 8:53 am #1428
Hi Dear Matus,
I didn’t see anything particular wrong about the software. It looks like the thing is coming from data.
One thing you can do is to check if there’s some heavy raining on either of the dates.
Meanwhile I will take some time to download and take a look at your data also.
Yuxiao
-
January 26, 2016 at 12:57 pm #1429
Dear all,
As I wrote in my previous mail, I received the same problem in co-registration of sentinel images in 2016 and some month at the end of 2015. it seems, there is some discrepancy between this image and previous image in their orbits and also these new images seem darker than previous image. then coregistration fails for theses images.
thanks
Nader-
January 26, 2016 at 12:59 pm #1430
Hi all,
I have been reported the inconsistency between images acquired after November 2015 and before that from a number of people. I am currently investigating this issue and I will come back to you once I get a conclusion.
Thanks to all,
Yuxiao
-
-
January 27, 2016 at 6:32 pm #1439
Hi all,
The bug that impacts most S1A images acquired after November 2015 has been fixed and I have submitted the code to Dan. The next new version of SARPROZ that should come out in the next days should just work fine.
Please let me know if the new codes work well or not,
Thanks to you all!
Yours,
Yuxiao -
January 27, 2016 at 9:26 pm #1440
The new version is available for download through the “manage version” module (pcodes and compiled for win and lnx).
please try it out and let us know!
d -
January 28, 2016 at 7:14 am #1441
SLC extraction is now correct
Thanks
-
January 28, 2016 at 10:00 am #1442
Hi,
Thank you for supporting. Its OK. it seems co-registration of sentinel images is OK.Nader
-
January 28, 2016 at 11:56 am #1443
Hi Yuxiao,
thank you, it is ok now. And the December 2015 interferograms are the most coherent ones (for my AOI)!
Unfortunately, I can still see few phase ramps within the interferograms (about 3 in 30-40 interferograms). It seems that this did not improve since October. Do you need the test data?
Thank you again, Ivana
-
January 28, 2016 at 12:05 pm #1444
same experience here!
-
January 28, 2016 at 12:35 pm #1446
Hi Ivana and Matus,
The phase ramps within the interferograms is usually a case-by-case issue and we do need to take a look case-by-case, since there are a number of issues that might cause the ramp.
It would be the best for you to send me you master image’s file name (or date, but file name preferred), the image pairs (same as master image) that has ramps, and the range (coordinates, samples and lines) you set for your area.
Thanks to you all!
Yuxiao
-
January 28, 2016 at 12:57 pm #1451
Are there any general advices while processing Sentinel-1 IW SLC? Like extracting a bigger area first? I’ve seen hints for increasing the area size by 1500 pixels, etc.
Many thanks to you Yuxiao,
Cheers
-
February 2, 2016 at 11:14 am #1465
Hi Matus,
Sorry for the late reply..
Since the coregistration will rely on the area and quality of overlapping parts between bursts, it is always generally a good idea to extract a bigger area (especially on range direction) so to ensure there are enough overlapping parts for fine-coregistration.
Then, since the fine-coregistration will rely on phase information it is generally true that with a higher coherence in your AOI there’s a better chance of getting a coregistration and thus no ramps. if your AOI is not so coherent then it is also a good idea to extend your AOI a little bit so that coherent areas will be included, this also helps coregistration better.
At last, always make sure that you’ll geocode well so that DEM phase term can be removed well from the final interferogram.
Yuxiao
-
-
January 28, 2016 at 12:55 pm #1449
Hi Yuxiao,
I am sending all the filenames (zipped) I use. As master, I am using 20150703, IW2, center 50.54, 13.5, radius 20 km.
Using MST as with long temp baselines, it gets heavily decorrelated.
So, these are the pairs with phase ramps (is it possible, that the phase ramps are only in the “upper” part of the interferograms?)
20151124 – 20151206 (one hardly visible, but probably present)
20150329 – 20150410 (also hardly visible)
20150305 – 20150317 (nicely visible)
20141024 – 20141117
20141223 – 20150128and maybe some more but I am not sure.
Thank you very much, Ivana
Attachments:
You must be logged in to view attached files.-
January 29, 2016 at 9:03 am #1458
Hi Ivana,
Thanks for the info! I will take a look into the data and let you know!
Yours,
Yuxiao
-
-
-
AuthorPosts
- You must be logged in to reply to this topic.