Project

General

Profile

Forums » GAMMA Processing » SAR Data Preprocessing » Sentinel-1 »

Question about mosaicing S1 GRD data after mk_geo_radcal processing

Added by Jack Chang over 6 years ago

Hi all,

I have been processing S1 GRD images these days. I plan to process several
S1 GRD images, perform backscatter correction and terrain geocoded for each
of them, then mosaic them together to obtain image with larger spatial coverage

However, after mosaicing, I notice that along the seams between frames, the backscatter
is unreasonably high which shows as bright line (see figure).

My processing flow is:

par_S1_GRD
multi_look_MLI
mk_geo_radcal (mode 0 to mode 3)
linear_to_dB
multi_mosaic
data2geotiff

(DEM is in EQA projection and w.r.t WGS84)

I am still trying to find out the solution on my own, but also I would like to ask if there
is anyone ever faced this situation before.

Really appreciate your help.

Best,
Jack

Capture.PNG (196 KB) Capture.PNG

Replies (3)

Question about mosaicing S1 GRD data after mk_geo_radcal processing - Added by Redmine Admin over 6 years ago

Dear forum members,

please click the website topic link to see the attached picture.

Best,
Sylvia

Question about mosaicing S1 GRD data after mk_geo_radcal processing - Added by Charles Werner over 6 years ago

Hi Jack,

Perhaps you should first mosaic, then convert to dB... on the border the dB
values would be averaged, and that might not be what you want.
multi_mosiac has 2 different mosaicking modes,

mode         mosaic processing mode
                 0: value of prior image preferred in the case of multiple
valid input values
                 1: average of multiple valid input values calculated

which did you use?

My suggestion is that you concatenate the GRD files first with, MLI_cat, then
terrain geocode.

Charles

RE: Question about mosaicing S1 GRD data after mk_geo_radcal processing - Added by Jack Chang over 6 years ago

Hi Charles,

Thank you very much for your suggestion :)

I used mode 1 for multi_mosaic.

I also realized that in some cases there is gap (NoData) along the seam of
adjacent frames after I mosaic. I was thinking if it could be because of the use of mode 1
when doing multi_mosaic? Since along the seam, if cells from one frame have NoData, while
those from the other frame have data, the average of them would be NoData, which becomes gap
between frames.

Not sure if my guess is the reason, but by concatenating the GRD files before terrain geocode
should should solve this problem. I will try it and see how it works.

Thanks again!

Best,
Jack

    (1-3/3)