Sentinel-1 Back Geocoding

Hi, it’s again me :slight_smile:

I did not really understand how the Sentinel-1 Back Geocoding is working. Is it a kind of “only” geometrical registration? Can you provide some references (to be added also in the help)?

Thanks for your support!

Hello everybody,

regarding the same topic (that has no answer until now) I’ve notice that the back-geocoding now does not allow to co-register products acquired before and after March 2015: Back-Geocoding: Source products cannot be InSAR pair: one with EAP phase correction, one without. Are you planning to implement the solution proposed in: https://sentinel.esa.int/documents/247904/1653440/Sentinel-1-IPF_EAP_Phase_correction and correct the bursts?

Thanks a lot!

Carlo, yes the backgeocoding is based strictly on the geometry. We do plan to implement the correction for the EAP for products prior to March. This should be available in the fall.

But then why the backgeocoding is affected by EAP correction? I would need to do coregistration not for Interferometry but for multitemporal analysis of long backscattering series. For the interferograms generation I will wait :smile:

Thanks!

1 Like

Good point. The check should be in the interferogram. Thanks

I am bringing up this post with another back geocoding error.

I am processing long time series of coherence images and I have observed a bad scenario. I am coregistering S1A/S1B data from orbit 131, common burst from slices 19 and 20 IW1.

Half of back geocoded products are invalid. As I see, if data from S1A is master, back geocoded product contains its data, but layers from slave product are empty.

Maybe someone has experienced this issue and can suggest some good solution?

Also, is there any difference which image should be used as master (newest or oldest) for coherence estimation? As far as I understand, coherence value doesn’t change depending on master/slave temporal order, right?

Really? Noone has problems with S1A/B coregistration?
I was off this processing workflow for a while, now I see that SNAP updates have not made this problem magically disappear. :frowning:

Did you find a solution? I have similar problem some stacks works fine while others with same masters are blank however it has some data and the computation also run without error. Please let me know if you have any idea!

My only solution is to change master/slave image manually. It is ok in graphic interface, but I haven’t tried to specify that in python script.

Anybody know why the S-1 back geocoding don’t proceed beyond a percentage of work. An example is given in the image below

Thankyou

2 Likes

Dear Zane,
I have faced the same problem.do you find a good solution to deal with this problem?

No, fangz070720, I haven’t.
BUT the problem somehow disappeared itself.
I have processed only summer data. In 2019, some bursts were still back-geocoded empty, but in 2020 everything was fine. At least in my AOI and using SNAP 7.0.4.

I am having the exact same problem, 4 pairs of images and cannot pass from 2-3% in hours, happened to me in linux and windows

please give more information - how many bursts did you select and which DEM?

I tried with 2, 4 and 7 bursts, and the areas of Galapagos island, Medellin (Colombia) and Mexico City (search for SRTM). all of them end up stuck in 2, 3 or 5 % depending on the amount of bursts. also I tried in 2 different computers. Also the Galapagos case worked fine for me 2 months ago but now I wanted to replicate it to see if the problem was with the images I am currently using but it got stuck as well. Does it is that it is not being able to download the STRM DEM?

there are currently problems reported with the SRTM, this will be fixed with the next update. But you can select SRTM 1Sec instead of 3Sec or another AutoDownload one to test.

SRTM 1s worked for me, thank you very much

I have the same problem in Back Geocoding step, the percentage doesn’t advance, stops in 3, 4, or even 8% when I use SRTM 3Sec; I have tried with SRTM 1Sec and it almost works because the loading screen finishes but the slave image data disappears in Stack File. I don’t know what to do.

Hi Daniel_gb,

I was erro the same you. I have 12 image, but I each image has only 3 bursts. I run it 5days but still 3%.
Do you have fixed it? Can you help me?
Thanks.
Tan Loi

image

@tanloi10081997 Did you switch to SRTM 1Sec HGT (AutoDownload)?