Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use

Added by Yang Lei 10 months ago

Dear ISCE users

The dense feature tracking Python modules, autoRIFT and Geogrid, have already been uploaded and documented on GitHub and are ready to use. Below you can find the links and the highlights. Please check it out and forward it to anybody who might feel interested. So far, both optical (Landsat-8) and radar (Sentinel-1A/B) demos are provided over Jakobshavn Glacier of Greenland and the resulting land ice motion velocity maps correspond very well with the coarse velocity maps from the DEM product.

Any feedback is appreciated.

Yang Lei, Alex Gardner and Piyush Agram

autoRIFT (autonomous Repeat Image Feature Tracking)

A Python module of a fast and intelligent algorithm for finding the pixel displacement between two images

autoRIFT can be installed as a standalone Python module (does not support radar coordinates) or with the InSAR Scientific Computing Environment (ISCE: https://github.com/isce-framework/isce2) software that supports handling Cartesian and radar coordinates

Use cases include all dense feature tracking applications, including the measurement of surface displacements occurring between two repeat satellite images as a result of glacier flow, large earthquake displacements, and land slides

autoRIFT can be used for dense feature tracking between two images over a grid defined in an arbitrary geographic-coordinate projection when used in combination with the sister Geogrid Python module (https://github.com/leiyangleon/Geogrid). Example applications include searching radar-coordinate imagery on a polar stereographic grid and searching Universal Transverse Mercator (UTM) imagery at a specified geographic-coordinate grid

Copyright (C) 2019 California Institute of Technology. Government Sponsorship Acknowledged.

Link: https://github.com/leiyangleon/autoRIFT

Geogrid

A Python module for precise mapping between (pixel index, pixel displacement) in imaging coordinates and (geolocation, motion velocity) in geographic coordinates

Geogrid can be installed as a standalone Python module (does not support radar coordinates) or with the InSAR Scientific Computing Environment (ISCE: https://github.com/isce-framework/isce2) software that supports handling Cartesian and radar coordinates

Geogrid can be used for dense feature tracking between two images over a grid defined in an arbitrary geographic-coordinate projection when used in combination with the sister autoRIFT Python module (https://github.com/leiyangleon/autoRIFT). Example applications include searching radar-coordinate imagery on a polar stereographic grid and searching Universal Transverse Mercator (UTM) imagery at a specified geographic-coordinate grid

Copyright (C) 2019 California Institute of Technology. Government Sponsorship Acknowledged.

Link: https://github.com/leiyangleon/Geogrid

Acknowledgement:

This effort was funded by the NASA MEaSUREs program in contribution to the Inter-mission Time Series of Land Ice Velocity and Elevation (ITS_LIVE) project (https://its-live.jpl.nasa.gov/) and through Alex Gardner’s participation in the NASA NISAR Science Team


Replies (21)

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Ravi Bhandari 9 months ago

Hello Piyush
I tried to run testGeogrid_ISCE.py with LC08_L1TP_146038_20161028_20170319_01_T1_B8.TIF and LC08_L1TP_146038_20171031_20171109_01_T1_B8.TIF ad input data and SRTM DEM of 1 arc second as following.
python testGeogrid_ISCE.py -m LC08_L1TP_146038_20161028_20170319_01_T1_B8.TIF -s LC08_L1TP_146038_20171031_20171109_01_T1_B8.TIF -d gang_dem1.tif -fo 1

I got the following error

Optical Image parameters:
X-direction coordinate: 162292.5 15.0
Y-direction coordinate: 3632407.5 -15.0
Dimensions: 15541 15821
Map inputs:
EPSG: 4326
Repeat Time: 864259200.0
XLimits: 30.640816393362254 32.824710403514
YLimits: 77.39490570481183 79.90801730751932
Extent in km: 0.002183894010151743 0.0025131116027074826
DEM: gang_dem1.tif
Velocities:
Slopes:
Outputs:
Window locations: window_location.tif
Window offsets: window_offset.tif
Window rdr_off2vel_x vector: window_rdr_off2vel_x_vec.tif
Window rdr_off2vel_y vector: window_rdr_off2vel_y_vec.tif
Starting processing ....
Xlimits : 77.99986111111112 32.82486111111102
Ylimits : 77.39486111111114 32.00013888888889
Dimensions of geogrid: -162630 x -163421
ERROR 1: Attempt to create -162630x-163421 dataset is illegal,sizes must be larger than zero.
Traceback (most recent call last):
File "abc.py", line 242, in <module>
runGeogridOptical(metadata_m, metadata_s, inps.demfile, inps.dhdxfile, inps.dhdyfile, inps.vxfile, inps.vyfile)
File "abc.py", line 228, in runGeogridOptical
obj.runGeogrid()
File "/home/bhandari/miniconda3/lib/python3.7/site-packages/geogrid/GeogridOptical.py", line 59, in runGeogrid
self.geogrid()
File "/home/bhandari/miniconda3/lib/python3.7/site-packages/geogrid/GeogridOptical.py", line 264, in geogrid
poDstDS.SetGeoTransform( adfGeoTransform )
AttributeError: 'NoneType' object has no attribute 'SetGeoTransform'

I installed autoRIFT and Geo Grid from anaconda cloud.

Please help in this regard.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Yang Lei 9 months ago

Hello,

It was unclear in the instructions that the DEM and the auxiliary files (slope and velocity) have to be geographic Cartesian coordinates (with northing and easting or geographic x and y in m) not geographic coordinates (lat/lon in deg). We will make it clear on the GitHub instructions. Unfortunately, the current version of Geogrid cannot handle the latter. Plus, the above error actually associates with the recent change about WGS84 (EPSG code: 4326) in GDAL 3 versus GDAL 2, where the order of the arguments lon and lat are reversed for the transformation of grid point position. Note even though you could use GDAL 2 to get through the above error, the final result is still wrong because only northing/easting are supported not lat/lon.

We recommend to use geographic Cartesian coordinates for the DEM and auxiliary files, such as UTM, polar sterographic coordinates, etc. You can use GDAL to do the transformation.

BTW, the radar imaging coordinates in the code involve transformation from/to WGS84, and the current version of the code does support both GDAL 2 and GDAL 3 on the reversed order of lon/lat.

Yang

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Ravi Bhandari 9 months ago

Hi
Yang thank you very much for your quick reply.
After changing the projection of DEM to "WGS 84 / NSIDC Sea Ice Polar Stereographic North "
and all data also to the same projection with same Landsat-8 data

python testGeogrid_ISCE.py -m LC08_L1TP_146038_20161028_20170319_01_T1_B8.TIF -s LC08_L1TP_146038_20171031_20171109_01_T1_B8.TIF -d gang_dem2.tif -fo 1

I get the following output

This is the Open Source version of ISCE.
Some of the workflows depend on a separate licensed package.
To obtain the licensed package, please make a request for ISCE
through the website: https://download.jpl.nasa.gov/ops/request/index.cfm.
Alternatively, if you are a member, or can become a member of WinSAR
you may be able to obtain access to a version of the licensed sofware at
https://winsar.unavco.org/software/isce
Optical Image parameters:
X-direction coordinate: 21244769.593898553 59.40198419802484
Y-direction coordinate: 4147487.488470178 59.40198419802484
Dimensions: 2641 2795
Map inputs:
EPSG: 3031
Repeat Time: 864259200.0
XLimits: 21244769.593898553 21401650.234165538
YLimits: 3981458.9426366985 4147487.488470178
Extent in km: 156.8806402669847 166.02854583347963
DEM: gang_dem2.tif
Velocities:
Slopes:
Outputs:
Window locations: window_location.tif
Window offsets: window_offset.tif
Window rdr_off2vel_x vector: window_rdr_off2vel_x_vec.tif
Window rdr_off2vel_y vector: window_rdr_off2vel_y_vec.tif
Starting processing ....
Xlimits : 21244720.146656238 21401736.627533596
Ylimits : 3981426.0196709074 4147597.980482806
Dimensions of geogrid: 1372 x 1452
Traceback (most recent call last):
File "testGeogrid_ISCE.py", line 239, in <module>
runGeogridOptical(metadata_m, metadata_s, inps.demfile, inps.dhdxfile, inps.dhdyfile, inps.vxfile, inps.vyfile)
File "testGeogrid_ISCE.py", line 225, in runGeogridOptical
obj.runGeogrid()
File "/home/bhandari/miniconda3/lib/python3.7/site-packages/geogrid/GeogridOptical.py", line 59, in runGeogrid
self.geogrid()
File "/home/bhandari/miniconda3/lib/python3.7/site-packages/geogrid/GeogridOptical.py", line 403, in geogrid
raster1a[jj] = normal2/(self.repeatTime/self.XSize/365.0/24.0/3600.0)*(normal2*yunit1-normal1*yunit2)/((normal2*xunit0-normal0*xunit2)*(normal2*yunit1-normal1*yunit2)
(normal2*yunit0-normal0*yunit2)*(normal2*xunit1-normal1*xunit2));
UnboundLocalError: local variable 'normal' referenced before assignment

It creates 4 files with name window_rdr_off2vel_y_vec.tif window_rdr_off2vel_x_vec.tif window_offset.tif window_location.tif
but these files contain Only NoData Values. The I ran autoRIFT but it also produces veloicty.tif but only nan values
I got stuck there.
Please help in sorting out the issue.
Or plase give data used in your Demo

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Yang Lei 9 months ago

Thank you very much for pointing this out. As you might have seen on our GitHub page (https://github.com/leiyangleon/Geogrid), the recommended use of Geogrid (by taking full advantage of it) is as below

CMD #1
testGeogrid_ISCE.py -m image1 -s image2 -d demname -sx dhdxname -sy dhdyname -vx vxname -vy vyname -fo 1

Other combination of the inputs are also desired, like

CMD #2 (the one you used here)
testGeogrid_ISCE.py -m image1 -s image2 -d demname -fo 1

or

CMD #3
testGeogrid_ISCE.py -m image1 -s image2 -d demname -sx dhdxname -sy dhdyname -fo 1

However, these options are not currently supported for optical data, although they are supported for radar data. It is good you pointed this out as other users probably want to use the module based on their own combination of input files. FYI, we will release another version soon (hopefully within 2 weeks) by accommodating all of the above options for both optical and radar data as well as provide some other output format (e.g. netCDF). Once it is done, I will also send you a notice.

As for your question about the outputs, please refer to the GitHub instructions on their descriptions. They can be summarized below:

Pixel geolocations (depends on DEM): window_location.tif
Initial pixel offsets (depends on DEM, slope and velocity): window_offset.tif
Conversion coefficients from imaging x-, y-offset to velocity in geographic X direction (depends on DEM, and slope): window_rdr_off2vel_x_vec.tif
Conversion coefficients from imaging x-, y-offset to velocity in geographic Y direction (depends on DEM, and slope): window_rdr_off2vel_y_vec.tif

For example, if you use CMD #1, all four outputs will be meaningful. But if you use CMD #2, only window_location.tif will be meaningful, window_offset.tif will be all zero's and the other two will be nodata everywhere. This information will also be added to the GitHub page in the upcoming release.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Ravi Bhandari 9 months ago

Hi
Yang Thank You for your response & big Sorry for irking and troubling you continuously.

I tried testGeogrid_ISCE.py with following Sentinel data
S1B_IW_SLC__1SSV_20161228T004217_20161228T004244_003589_006258_725C.SAFE and S1A_IW_SLC__1SDV_20170103T004259_20170103T004327_014660_017D8E_CC99.SAFE

I got the following error

Traceback (most recent call last):
File "testGeogrid_ISCE.py", line 242, in <module>
metadata_m = loadMetadata(inps.indir_m)
File "testGeogrid_ISCE.py", line 115, in loadMetadata
info.sensingStart = min([x.sensingStart for x in frames])
ValueError: min() arg is an empty sequence

When I went through the file

def loadMetadata(indir):
'''
Input file.
'''
import os
import numpy as np

frames = []
for swath in range(1,4):
inxml = os.path.join(indir, 'IW{0}.xml'.format(swath))
if os.path.exists(inxml):
ifg = loadProduct(inxml)
frames.append(ifg)

inxml = os.path.join(indir, 'IW{0}.xml'.format(swath))*
*File IW1.xml, IW2.xml etc. does not exist in Sentinel data folder

s1a-iw1-slc-vh-20191110t125501-20191110t125529-029849-036781-001.xml files exist in sentinel data
*but loadproduct() function is not able to read these file *
<iscesys.Component.Configurable.EmptyFacility object at 0x2aaab23b56a0>

Could you please point out if I am making any mistake in processing or I must wait for the next release.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Yang Lei 9 months ago

I assume you just referred to the command in the Demo section of the GitHub page. Unfortunately, this is not enough. In order to use the command, you need to understand what those input symbols actually represent, and these are clearly defined in the Instructions section with a few important notes. For example, it looks like you did not even run topsApp.py from ISCE, which of course did not generate the master data folder (exactly named "master" not the SAFE data folder). Again, this is clearly written in the Instructions.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Ravi Bhandari 8 months ago

Thank You Yang for your reply.

I have installed ISCE2 on my system from Anaconda using conda insatll -c conda-forge isce2

as per your previous mail I ran topsApp.py

I put the geo_autoRIFT folder in the /home/bhandari/miniconda3/lib/python3.7/site-packages/isce/components/contrib/

folder as instructed in GitHub
Get The following error
[bhandari@master2 sar]$ python testGeogrid_ISCE.py -m master -s slave -d demLat_N30_N33_Lon_E077_E079.dem.wgs84

This is the Open Source version of ISCE.
Some of the workflows depend on a separate licensed package.
To obtain the licensed package, please make a request for ISCE
through the website: https://download.jpl.nasa.gov/ops/request/index.cfm.
Alternatively, if you are a member, or can become a member of WinSAR
you may be able to obtain access to a version of the licensed sofware at
https://winsar.unavco.org/software/isce
Traceback (most recent call last):
File "testGeogrid_ISCE.py", line 247, in <module>
runGeogrid(metadata_m, metadata_s, inps.demfile, inps.dhdxfile, inps.dhdyfile, inps.vxfile, inps.vyfile)
File "testGeogrid_ISCE.py", line 196, in runGeogrid
obj.geogrid()
File "/home/bhandari/miniconda3/lib/python3.7/site-packages/isce/components/contrib/geo_autoRIFT/geogrid/Geogrid.py", line 50, in geogrid
from components.contrib.geo_autoRIFT.geogrid import geogrid
ImportError: cannot import name 'geogrid' from 'components.contrib.geo_autoRIFT.geogrid' (/home/bhandari/miniconda3/lib/python3.7/site-packages/isce/components/contrib/geo_autoRIFT/geogrid/__init__.py)

If I run in standalone mode by modifying
from components.contrib.geo_autoRIFT.geogrid import Geogrid line to
from geogrid import Geogrid

I get the folliwing error

https://winsar.unavco.org/software/isce
Traceback (most recent call last):
File "testGeogrid_ISCE.py", line 247, in <module>
runGeogrid(metadata_m, metadata_s, inps.demfile, inps.dhdxfile, inps.dhdyfile, inps.vxfile, inps.vyfile)
File "testGeogrid_ISCE.py", line 196, in runGeogrid
obj.geogrid()
File "/home/bhandari/miniconda3/lib/python3.7/site-packages/geogrid/Geogrid.py", line 64, in geogrid
self.setState()
File "/home/bhandari/miniconda3/lib/python3.7/site-packages/geogrid/Geogrid.py", line 198, in setState
self._geogrid = geogrid.createGeoGrid_Py()
*AttributeError: module 'geogrid' has no attribute 'createGeoGrid_Py' *

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Yang Lei 8 months ago

As stated in the GitHub page, Installation section,

1. Did you also put the SConscript file along with the geo_autoRIFT folder in /home/bhandari/miniconda3/lib/python3.7/site-packages/isce/components/contrib/?
2. Were you able to run "scons install" from the source code directory?

Failure to do either of the above will prevent you installing autoRIFT and Geogrid in ISCE. This means, if you want to use Conda install of ISCE, you probably have to wait for the next release of ISCE (autoRIFT will be included in the Conda package). Otherwise, for now, you cannot install these modules in the ISCE Conda package by simply copying the geo_autoRIFT folder, which actually won't install them.

Two other notes:

1. Remember we discussed previously that DEM has to be in geographic Cartesian coordinate system not geographic coordinate (lat and lon)? So you need demLat_N30_N33_Lon_E077_E079.dem.wgs84 to run ISCE but you need to transform it to polar stereographic or UTM to run autoRIFT.
2. Use of standalone for calling ISCE-installed version is not recommended however possible only if you have already installed the modules in ISCE, which is not true for the reasons mentioned above. In short, the recommended use is: if you want to process both radar and optical, just stick with ISCE version; otherwise, if you want to use Conda installed standalone version, only optical is supported. This is clearly stated at the beginning of the GitHub page.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Ravi Bhandari 8 months ago

Hi Yang
Thank you for answering even my trivial queries.

Actually I had installed isce2 from conda
Using conda install. I installed scons on my system and copied geo_AutoRIFT and Sconscript file to the home/bhandari/miniconda3/lib/python3.7/site-packages/isce/components/contrib
directory
and ran scons installed from

home/bhandari/miniconda3/lib/python3.7/site-packages/isce
but it gives error.
Which directory will be isce install directory in case of conda install?

Could you please guide installation of autoRIFT with isce2 when isce2 is installed with conda
Thanking you

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Yang Lei 8 months ago

As I said in the previous reply, with Conda-installed ISCE2, you probably won't install autoRIFT successfully as scons install is not easy to do. I personally don't have much experience on that. If there is any chance, Piyush would know. Otherwise, there are two ways for you to use autoRIFT successfully:

1. Install ISCE2 using macports manually and install autoRIFT by following the GitHub page (Section 6, "With ISCE")
2. Wait for the next release of ISCE (probably ISCE3), which will have autoRIFT included.

Yang

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Ravi Bhandari 8 months ago

Thank You! Very much Yang for your help and support and answering patiently all of my issues.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Yang Lei 7 months ago

FYI, the autoRIFT on GitHub and Conda-forge has been updated to v1.0.4, which solves the bugs in previous versions. Only if you want to 1) install ISCE2 manually and try autoRIFT or 2) install autoRIFT with Conda (that won't work for radar images), you should be good to go. Otherwise, you can continue waiting for ISCE3 to install both ISCE and autoRIFT via Conda.

Yang

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Ravi Bhandari 7 months ago

Great !!!! Thank You Very much Yang for your reply. I am keen to use it, will try to install & use it.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Ravi Bhandari 7 months ago

Hello Yang I installed autoRIFT from conda-forge and copied new testautoRIFT.py
when I try with following

python testautoRIFT.py -m LC08_L1TP_146038_20161028_20170319_01_T1_B8.TIF -s LC08_L1TP_146038_20171031_20171109_01_T1_B8.TIF -fo 1

I get the following error

This is the Open Source version of ISCE.
Some of the workflows depend on a separate licensed package.
To obtain the licensed package, please make a request for ISCE
through the website: https://download.jpl.nasa.gov/ops/request/index.cfm.
Alternatively, if you are a member, or can become a member of WinSAR
you may be able to obtain access to a version of the licensed sofware at
https://winsar.unavco.org/software/isce
Pre-process Start!!!
Pre-process Done!!!
135.4635875225067
Uniform Data Type Done!!!
152.45157623291016
AutoRIFT Start!!!
/home/bhandari/miniconda3/lib/python3.7/site-packages/autoRIFT/autoRIFT.py:1040: RuntimeWarning: invalid value encountered in less
  • C = np.sum(D<option1,axis=0).reshape(output_size)*
    /home/bhandari/miniconda3/lib/python3.7/site-packages/numpy/lib/nanfunctions.py:1115: RuntimeWarning: All-NaN slice encountered
  • overwrite_input=overwrite_input)*
    /home/bhandari/miniconda3/lib/python3.7/site-packages/autoRIFT/autoRIFT.py:1092: RuntimeWarning: invalid value encountered in less_equal
  • M = (np.abs(Dx - DxM) <= (self.MadScalar * DxMad)) & (np.abs(Dy - DyM) <= (self.MadScalar * DyMad)) & M*
    AutoRIFT Done!!!
    14.316216945648193

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Yang Lei 7 months ago

autoRIFT performed successfully. Those are warnings not error messages.

See below the runtime and start/end of each step from what you pasted above:

Pre-process Start!!!
Pre-process Done!!!
135.4635875225067
Uniform Data Type Done!!!
152.45157623291016
AutoRIFT Start!!!
AutoRIFT Done!!!
14.316216945648193

You should be able to find the output (.mat file in this case) in the test data folder.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Joseph Kennedy 7 months ago

If anyone wants to install autoRIFT v1.0.4 + ISCE v2.3.2 into a conda environment for radar images, you can follow the steps I've outlined in this gist:

https://gist.github.com/jhkennedy/5a22251dd6f1b73db1968517f00a8af6

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by ruitang yang 6 months ago

Hi Joe,
I followed the steps to install autoRIFT v1.0.4 + ISCE v2.3.2 on ubuntu 16.04, using anaconda3 (python3.6), but encountered some troubles. I searched the forums for similar posts, but still cannot resolve them, writing here to get any help.

My system info: ubuntu 16.04, anaconda3 is installed with Anaconda3-5.2.0-Linux-x86_64.sh (python 3.6). The error message showed it can't find the library hdf5,fftw3,Xm lib, Xt lib,gdal..., but I checked, all of them have installed with the latest version. Although checked SconfigISCE and config.log, I still couldn't solve it.
Error message:

scons: Reading SConscript files ...
Building with scons from python3
Checking for C header file Python.h... yes
Checking for C header file fftw3.h... yes
Checking for C header file hdf5.h... yes
Checking for C header file X11/Xlib.h... yes
Checking for C header file Xm/Xm.h... yes
Checking for C header file omp.h... yes
Checking for C library hdf5... no
Could not find: hdf5 lib for libhdf5
Error: Install hdf5 or libhdf5-dev
Checking for C library fftw3f... no
Could not find: fftw3f lib for libfftw3f
Error: Install fftw3 or libfftw3-dev
Checking for C library Xm... no
Could not find: Xm lib for libXm
Error: Install Xm or libXm-dev
Checking for C library Xt... no
Could not find: Xt lib for libXt
Error: Install Xt or libXt-dev
Checking for F include fftw3 ... yes
GDAL version: 3.0.3

Checking for C++ header file gdal_priv.h... yes
Checking for C library gdal... no
Could not find: libgdal for gdal
Install gdal or include path to libs to LIBPATH
Not all components of ISCE will be installed and can result in errors.
Press Enter to continue.... Ctrl-C to exit
Checking whether cython3 program exists.../home/yrt/anaconda3/envs/autoRIFT/bin/cython3
User did not request CUDA support. Add ENABLE_CUDA = True to SConfigISCE to enable CUDA support
Building with scons from python3
Could not find: hdf5 lib for libhdf5
Error: Install hdf5 or libhdf5-dev
Could not find: fftw3f lib for libfftw3f
Error: Install fftw3 or libfftw3-dev
Could not find: Xm lib for libXm
Error: Install Xm or libXm-dev
Could not find: Xt lib for libXt
Error: Install Xt or libXt-dev
Checking for F include fftw3 ... yes
GDAL version: 3.0.3

Could not find: libgdal for gdal
Install gdal or include path to libs to LIBPATH
Not all components of ISCE will be installed and can result in errors.
Press Enter to continue.... Ctrl-C to exit
User did not request CUDA support. Add ENABLE_CUDA = True to SConfigISCE to enable CUDA support
No module named 'iscesys'
No module named 'iscesys'
No module named 'isceobj'
module '.' has no attribute 'getFactoriesInfo'
cython3 found.

SConfigISCE (1.3 kB)

config.log (15.1 kB)

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Joseph Kennedy 6 months ago

I sat down with Ruitang and we found a small error in the above install script. It's been updated on github and we were able to use it to install autoRIFT+ISCE into a conda environment on her ubuntu laptop.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Emmy Killett 6 months ago

Thanks for that script! I'm trying to get it working in a docker container right now. One issue I ran into is that I had to call the script using "bash -i install_autoRIFT_ISCE_conda.sh" otherwise conda can't activate environments. Because of that change, it's convenient to change both "rm -r" at the end to "rm -rf".

Also, should line 54 actually say "# The directory in which ISCE will be built"? The original script seems to have missed the "#" to comment that line in the SCONS config file.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Joseph Kennedy 6 months ago

Thanks for the suggestions Emmy. I've added the # to line 54 (woops!) and add the f flag to the rm s at the end of the script -- should be up to date on github.

RE: Dense feature tracking Python Modules autoRIFT and Geogrid on GitHub and Ready to Use - Added by Zhiliang Zhang about 1 month ago

Yang Lei wrote:

autoRIFT performed successfully. Those are warnings not error messages.

See below the runtime and start/end of each step from what you pasted above:

Pre-process Start!!!
Pre-process Done!!!
135.4635875225067
Uniform Data Type Done!!!
152.45157623291016
AutoRIFT Start!!!
AutoRIFT Done!!!
14.316216945648193

You should be able to find the output (.mat file in this case) in the test data folder.

Dear Lei Yang
I have run testautoRIFT_ISCE.py successfully and I have get results(offset.mat) successfully! But I find that the dimension of offset.mat is 72*54, but the dimension of input data(SLC data) is 2350*1800, could you tell me how to solve this problems and could you give me some advice about how to get results correctly? Thank you very much!
Best wishes to you!
Elton

result.png (5.8 kB)

(1-21/21)