prep_aria: add more metadata #139
Conversation
prep_aria: + comment out skimage.transform.resize() because the updated ariaTSsteup.py produces the same size DEM file as the interferogram stack, this eliminates the need of installing scikit-image module while running prep_aria.py with the `ARIA-tools` python environment. + copy over all metadata from unwrappedPhase for the sake of completeness, except for Dates and bperp + add metadata extraction for: - ANTENNA_SIDE - HEADING - RANGE_PIXEL_SIZE - LON/LAT_REF1/2/3/4 - CENTER_LINE_UTC + write all ifgrams in the sorted order of date1_date2 with date1 < date2 + comment out geometry dataset azimuthAngle and shadowMask for now since both of them are not required by mintpy. Together with waterMask, they can be added afterward when the products are available from ARIA-tools. + more print out message for file IO utils/ptime: remove obsolete ifgram_date_list() to that ptime.py can be independent from h5py module add ARIA format link to README.md Update indentation in smallbaselineApp.cfg
|
Hi @ehavazli, could you check if all the operations you have used before work for you? After 1) this merge and 2) fix the issue with bridging, I will cut a new release. |
|
everything ran without failing for a small test case (@hfattahi San Francisco data set) so it looks good to me. |
mintpy/prep_aria.py
Outdated
| meta["WIDTH"] = ds.RasterXSize | ||
| meta["NUMBER_OF_PAIRS"] = ds.RasterCount | ||
|
|
||
| # ARIA standard products currently don't number of range and | ||
| az_angle = float(meta["azimuthAngle"]) |
There was a problem hiding this comment.
@ehavazli and @dbekaert, this value azimuthAngle = -8.77992802893209 does not seem like the azimuthAngle of the satellite orbit (band2 of los.rdr from ISCE/topsStack); but more like the azimuth angle of the LOS vector (from incLocal.rdr), could you confirm?
As a reference, the band2 of los.rdr from ISCE/topsStack for descending track has a typical value of -102
If we got the HEADING angle of the satellite orbit, the ORBIT_DIRECTION can be drived from here.
There was a problem hiding this comment.
azimuth angle in the VRT metadata is an average value. Azimuth angle information is extracted from the product. I am testing the new ariaTSsetup.py with the orbit direction information being included in the metadata of VRT.
Extracted azimuth angle file can be uploaded into geometryGeo.h5 just like incidence angle by adding the option into prep_aria.py
+ use orbitDirection from ARIA metadata + hardwire HEADING value for now based on asc/desc orbit + add -a/--az-angle option + update example
|
Updated according to the latest dev branch of ARIA-tools. All metadata seems reasonable to me now. If one of you two could approve this pull request, I will go forward and merge it. |
prep_aria:
comment out skimage.transform.resize() because the updated ariaTSsteup.py produces the same size DEM file as the interferogram stack, this eliminates the need of installing scikit-image module while running prep_aria.py within the
ARIA-toolspython environment.copy over all metadata from unwrappedPhase for the sake of completeness, except for Dates and bperp
add metadata extraction for:
write all ifgrams in the sorted order of date1_date2 with date1 < date2
comment out geometry dataset azimuthAngle and shadowMask for now since both of them are not required by mintpy. Together with waterMask, they can be added afterward when the products are available from ARIA-tools.
more print out message for file IO
utils/ptime: remove obsolete ifgram_date_list() to that ptime.py can be independent from h5py module
add ARIA format link to README.md
Update indentation in smallbaselineApp.cfg