sarproz-group

sentinel-1 coregistration fault

This topic contains 6 replies, has 4 voices, and was last updated by  Yuxiao 1 year, 11 months ago.

  • Author
    Posts
  • #1306

    milan
    Participant

    Hi,

    I finally managed saproz for testing at computer cluster, so far working well.
    But I got problem with coregistering S-1 data. From 30 images, only 1 was coregistered. For the others:
    Reramping front-stitched SLC slave image on Date 20151104

    Buffer size exceeding 200!!! abnormal!!! check your stitching strategy!!!

    Resampling end-stitched SLC slave images on Date 20151104
    WARNING:Index exceeds matrix dimensions.
    WARNING: TOPS Post Processing for date 20151104 FAILED!!! Further analysis including this image might lead to unreliable result!!!

    So I tried to use fitted images from FITTED/FORMER that seemed coregistered on the first sight. Indeed, they show strong vertical phase ramps within some (around half) of their interferograms. It seems it could be partially corrected by APS but I hope I can find a better way? Coregistration refinement failed as well with messages as:
    Error in amp_corr(): wrong number of lines and samples for search image!

    WARNING!!! coregistration of image 20141028 failed!!!!

    image 20141028 will be discarded: check manually!!!

    This may be also connected to SRTM (?) because it automatically downloaded SRTM data for only half of the scene. I tried to apply DEM in Geotiff over the whole area but it didn’t read the file (but the file is not corrupted, I checked it). My area is relorbit 32, swath IW2, size 6280×24260 pixels.

    Thank you for any clues, hope to continue arranging sarproz in connection to sentinel-1 database as a basis of upcoming project possibilities..

    Milan

  • #1307

    periz
    Keymaster

    Yuxiao is going to answer you soon.

    My only comment is: are you using the most recent Sarproz version?

    d

  • #1309

    Yuxiao
    Moderator

    Hi milan,

    can you post the filename of the images you’re using and the area you’re extracting (center, samples and lines) along with your log file and lasterror so I can take a look?

    Thanks!

    Yuxiao

  • #1321

    ivca-cz
    Participant

    Hello,

    it helps to use the previous version (Oct-6), here the interferograms are reasonable. But pay attention to phase jumps. I was lucky that my area does not exceed one burst… I also got this problem with the newest (Oct-28) version for most of the scenes and the interferograms were full of fringes (in the azimuth direction).

    Regards, Ivana

  • #1322

    Yuxiao
    Moderator

    Hi Dear Ivana,

    Could you be more specific on your problems, like, showing a few interferograms, show your area, master and slave images, maybe the log files so we can take a look?

    We have been updating the TOPS code constantly, and there should be new compiled versions solving the problems in Oct-28 version very soon.

    Please contact me on details,

    Thanks!

    Yuxiao

  • #1323

    milan
    Participant

    hi,

    thank you.

    it seems it was some technical problem due to file permissions. I have solved it probably, now everything seems working well – the “stitching” finished with failure only at 2 images now. Only not sure why I still cannot load the DEM.. (does some chmod function is applied for DEM? because chmod will not work..)
    I will probably finish my first s-1 processing during today and will let you know if i succeed or not.. so far it seems it works well and fast, using 23 workers and 96 gb ram it should be quite working 😉

    cheers

  • #1324

    Yuxiao
    Moderator

    Hi Milan,

    After you’ve finished, please send me some screenshots and the logs relating to the process. and the lasterror.mat, if needed.

    We are still working on the code but the goal is to succeed the post-processing of TOPS data for every image. I will need to investigate the failure with the 2 images you’ve mentioned.

    Please let me know!

    Yuxiao

You must be logged in to reply to this topic.