Re: TanDEM-X coSSC product processing

[Batuhan Osmanoglu]

Hi Y.S. Rao, 

Not sure what happened there. Sorry about that. Hope you get this email fine. 

The way bistatic is implemented in ADORE is not the "proper way". Normally this should be done in Doris, where the height ambiguity and the Bperp is calculated. Instead what ADORE does is to calculate everything as if it is repeat-pass but then divide the results by two (for comprefpha, comprefdem and slant2h). 

We also noticed that the interferograms were not perfectly flat, but speaking to a german colleague, we decided that it was due to the orbital error + short baseline. To flatten we used the deramp function with an estimate DEM. 

ADORE automatically sets the coherence to be calculated after DEM removal. If you would like to calculate coherence without a DEM please set the coh_method to refphase_only (http://doris.tudelft.nl/usermanual/node111.html). 

Hope this helps, 

batu

On Thu, Jul 17, 2014 at 7:13 PM, Y.S. Rao <ys...@csre.iitb.ac.in> wrote:

Dear Dr. Batuhan,
 I did not get any message with the following email.
Regards,
YS

On Thu, July 17, 2014 6:22 pm, Batuhan Osmanoglu wrote:
> Dear Y.S.
> On Jul 17, 2014 8:41 AM, "Y.S. Rao" <ys...@csre.iitb.ac.in> wrote:
>
>> Dear Dr. Batuhan,
>>    With simple flat-earth task (not used any DEM), *.cint and .srp looks
>> similar. Fringes are not flattened. The test site is Mumbai with
>> average topography. For Himalayan terrain, DORIS does not apply
>> flat-earth correction.  In the calculation of coherence, I get the
>> following message.
>> PROGRESS: Start COHERENCE.
>> PROGRESS: Start COHERENCE.
>> ERROR   : [processor.cc[2968]]: Requested step: 29 (coherence:) seems
>> impossible, because step 32 (comp_refdem:) is not in resultfile.
>> ERROR   : [processor.cc[2968]]: Requested step: 29 (coherence:) seems
>> impossible, because step 32 (comp_refdem:) is not in resultfile.
>> Caught error of EXCEPTION class!
>> It is: input error
>> INFO    : Current time: Thu Jul 17 18:07:35 2014
>>
>> For DEM generation, reference DEM is not required.
>>
>> Regards,
>> y.s.rao
>>
>> On Thu, July 17, 2014 5:12 pm, Y.S. Rao wrote:
>> > I got it.
>> > settings
>> > int_multilook="2 2"
>> > Hope you will comment on baseline.
>> >
>> > regards,
>> > YSRao
>> >
>> >
>> > On Thu, July 17, 2014 4:03 pm, Y.S. Rao wrote:
>> >> Dear Dr. Batuhan,
>> >>    I started working with TanDEM-X. I could read data.
>> >> I created directories for TSX with one date and TDX with another
>> date.
>> >>
>> >> When I gave coarseorb with bistatic=True.
>> >>
>> >> ADORE: coarseorb bistatic=True
>> >> I got the following. DORIS calculated for repeat pass mode. Bperp is
>> >> doubled and also height ambiguity as compared to TanDEM-X
>> specifications
>> >>
>> >> In the results file, I kept coarse_correl = 1 and also fine_coreg=1
>> >> manually. Later, I gave your command s_crop2resample. It worked fine.
>> >> I could run refphase, interfero, etc. commands. I could not
>> understand
>> >> how
>> >> to multilook 2 by 2 for interferogram generation with ADORE-DORIS.
>> >>
>> >> My result file is attached for your reference.
>> >> Regards,
>> >> y.s.rao
>> >> *******************************************************************
>> >> *_Start_coarse_orbits:
>> >> *******************************************************************
>> Some
>> >> info for pixel: 12062, 10677 (not used):
>> >>   Btemp:     [days]:           0        // Temporal baseline
>> >>   Bperp      [m]:          323.1        // Perpendicular baseline
>> >>   Bpar       [m]:          108.5        // Parallel baseline
>> >>   Bh         [m]:            313        // Horizontal baseline
>> >>   Bv         [m]:          134.9        // Vertical baseline
>> >>   B          [m]:          340.8        // Baseline (distance between
>> >> sensors)
>> >>   alpha      [deg]:         23.3        // Baseline orientation
>> >>   theta      [deg]:         41.9        // look angle
>> >>   inc_angle  [deg]:         46.1        // incidence angle
>> >>   orbitconv  [deg]:   0.00236492        // angle between orbits
>> >>   Height_amb [m]:          -24.5        // height = h_amb*phase/2pi
>> >> (approximately)
>> >>   Control point master (line,pixel,hei) = (12062, 10677, 0)
>> >>   Control point slave  (line,pixel,hei) = (11851.8, 10617.7, 0)
>> >> Estimated translation slave w.r.t. master (slave-master):
>> >>   Positive offsetL: slave image is to the bottom
>> >>   Positive offsetP: slave image is to the right
>> >> Coarse_orbits_translation_lines:        -210
>> >> Coarse_orbits_translation_pixels:       -59
>> >> ******************************************************************* *
>> >> End_coarse_orbits:_NORMAL
>> >> *******************************************************************
>> >>
>> >>
>> >>
>> >>
>> >>
>> >> On Mon, July 7, 2014 9:49 pm, Batuhan Osmanoglu wrote:
>> >>> Hi Y.S. Rao,
>> >>>
>> >>> I am looking forward to see the comparison ;)
>> >>>
>> >>> Best,
>> >>> Batu.
>> >>> On Jul 7, 2014 7:18 PM, "Y.S. Rao" <ys...@csre.iitb.ac.in> wrote:
>> >>>
>> >>>> You are following simple InSAR method. But Gamma and SARScape
>> follows
>> >>>> SRTM
>> >>>> dem method and gets DInSAR interferogram. I will see whether we get
>> >>>> any
>> >> difference using the two methods.
>> >>>> Best
>> >>>> Ysrao
>> >>>> On Mon, July 7, 2014 5:09 pm, Batuhan Osmanoglu wrote:
>> >>>> > Hi Y.S. Rao,
>> >>>> >
>> >>>> > No worries, I try to help anyone as long as I have time;)
>> >>>> >
>> >>>> > Doris complains if we do not have the results in resampling step,
>> so
>> >>>> in
>> >>>> > ADORE I created a command to copy the info in s_crop to resample.
>> No
>> >> actual
>> >>>> > resampling is done, just a workaround to get doris to run
>> interfero
>> >>>> step.
>> >>>> >
>> >>>> > For DEM generation, what we did was:
>> >>>> > - Generate interferogram
>> >>>> > - remove bistatic flat earth phase
>> >>>> > - filter (with modified gaussian, set pf_method=modgoldstein,
>> >> pf_overlap=15, pf_blocksize=32)
>> >>>> > - unwrap
>> >>>> > - deramp (if you have SRTM etc, you can give the .crddemlp file
>> as
>> >>>> an
>> >> estimate)
>> >>>> > - slant2h
>> >>>> >
>> >>>> > And you can geocode etc. if you like. At first we were thinking
>> that
>> >>>> we
>> >>>> > did
>> >>>> > not need to use deramp since the orbits are very good with TDX,
>> but
>> >>>> from
>> >>>> > experience I can tell that we need it. Otherwise we were getting
>> >> about
>> >>>> a
>> >>>> > fringe over the entire interferogram (50x50km).
>> >>>> >
>> >>>> > best,
>> >>>> > batu.
>> >>>> >
>> >>>> > On Mon, Jul 7, 2014 at 7:32 AM, Y.S. Rao <ys...@csre.iitb.ac.in>
>> >>>> wrote:
>> >>>> >
>> >>>> >> I forgot to mention one thing. TanDEM-X DEM generation procedure
>> is
>> >> different. One should create DInSAR interferogram and unwarp it. The
>> >> unwrapped DInSAR phase is added to the synthetic interferogram
>> >> phase.
>> >>>> >> Finally, the unwrapped phase is converted to height. Are you
>> >>>> following
>> >>>> >> the
>> >>>> >> same way? Is there any effect on baseline (bistatic) in DEM
>> >>>> generation?
>> >>>> >>
>> >>>> >> best
>> >>>> >> YSR
>> >>>> >>
>> >>>> >> On Mon, July 7, 2014 4:50 pm, Batuhan Osmanoglu wrote:
>> >>>> >> > Hello Y.S. Rao,
>> >>>> >> >
>> >>>> >> > Yes I use the bistatic flag when I process tandemx pairs and
>> it
>> >>>> works
>> >>>> >> > fine.
>> >>>> >> > You can skip coregistration, for resample, you need to use
>> >>>> >> s_crop2resample
>> >>>> >> > command.
>> >>>> >> >
>> >>>> >> > I did not integrate your code, mostly because it needed
>> >>>> compilation.
>> >>>> >> As
>> >>>> >> > you
>> >>>> >> > know for now the adore package does not need any compiling. I
>> am
>> >>>> >> looking
>> >>>> >> > into getting it in, but did not find a good solution yet.
>> Thanks
>> >> so
>> >>>> >> much
>> >>>> >> > for your contribution.
>> >>>> >> >
>> >>>> >> > best,
>> >>>> >> > batu.
>> >>>> >> >
>> >>>> >> >
>> >>>> >> > On Mon, Jul 7, 2014 at 12:39 AM, Y.S. Rao
>> <ys...@csre.iitb.ac.in
>> >
>> >>>> >> wrote:
>> >>>> >> >
>> >>>> >> >> Dear Dr. Batuhan,
>> >>>> >> >>    I got the following email long back. If I use bistatic
>> flag
>> >>>> in
>> >>>> >> ADORE
>> >>>> >> >> as
>> >>>> >> >> suggested in the following email (see below), how are the
>> >> results?
>> >>>> >> Did
>> >>>> >> >> you verify it? In processing, I should avoid coregistration?
>> I
>> >>>> need
>> >>>> >> to
>> >>>> >> >> keep the flag on. Did you integrate our python code for
>> faster
>> >> converting half float TanDEM-X data to float data?
>> >>>> >> >>
>> >>>> >> >> Regards,
>> >>>> >> >> y.s.rao
>> >>>> >> >>
>> >>>> >> >> On Tue, August 13, 2013 10:57 pm, Batuhan Osmanoglu wrote:
>> >>>> >> >> > Hello Y.S.
>> >>>> >> >> >
>> >>>> >> >> > Thanks for the update. I think they are in contact with
>> Mahmut
>> >>>> >> Arikan
>> >>>> >> >> > (delft) but not sure if their code has made it in to the
>> >>>> >> repository.
>> >>>> >> >> >
>> >>>> >> >> > In the meantime you can process TDX bistatic interferograms
>> >>>> using
>> >>>> >> the
>> >>>> >> >> > bistatic flag in Adore.
>> >>>> >> >> >
>> >>>> >> >> > best,
>> >>>> >> >> > batu.
>> >>>> >> >> > On Aug 13, 2013 4:56 AM, "Y.S. Rao" <ys...@csre.iitb.ac.in>
>> >>>> wrote:
>> >>>> >> >> >
>> >>>> >> >> >> Dear Dr. Batuhan,
>> >>>> >> >> >>    Recently, one of my students attended DLR-TanDEM-X
>> >>>> meeting.
>> >>>> It
>> >>>> >> is
>> >>>> >> >> >> interesting to know that DORIS code is being modified to
>> >>>> process
>> >>>> >> >> >> TanDEM-X data for DEM generation. Please see the link for
>> the
>> >>>> pdf.
>> >>>> >> >> The
>> >>>> >> >> >> authors are from China. But the first author name looks
>> >>>> European.
>> >>>> >> >> >> http://www.csre.iitb.ac.in/~ysrao/batuhan
>> >>>> >> >> >>
>> >>>> >> >> >> You can also look at the link for all presentation.
>> >>>> >> >> >> https://tandemx-science.dlr.de/
>> >>>> >> >> >> best regards,
>> >>>> >> >> >> y.s.rao
>> >>>> >> >> >>
>> >>>> >> >> >>
>> >>>> >> >> >
>> >>>> >> >>
>> >>>> >> >>
>> >>>> >> >
>> >>>> >>
>> >>>> >>
>> >>>> >
>> >>>
>> >>
>> >>
>> >>
>> >
>> >
>>
>>
>