G004-0032
New advanced ALOS-2 InSAR processors in ISCE2 and some large-scale processing results
Wednesday, 9 December 2020
Poster
Cunren Liang, California Institute of Technology, Pasadena, CA, United States, Mark Simons, CALTECH, Pasadena, CA, United States, Eric Jameson Fielding, Jet Propulsion Lab Caltech, Pasadena, CA, United States and Zhen Liu, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, United States
Abstract:
We present the new ALOS-2 InSAR processors released in JPL/Caltech ISCE2 software including alos2App.py and alos2burstApp.py with an abundance of algorithms, features and enhanced capabilities to tackle various multi-mode InSAR processing problems and facilitate users. Notable features include: 1) One app to process all ALOS-2 same-mode and possible cross-mode InSAR data. 2) Support making interferograms across all three wavelengths. 3) Both full-aperture and burst-by-burst ScanSAR InSAR workflows implemented. 4) ScanSAR system parameters estimated from azimuth spectrum. 5) Automatic estimation of the start times of raw burst from azimuth spectrum. 6) High precision burst synchronization calculation. 7) MBF filter for removing non-overlap spectra caused by burst misalignment and Doppler centroid frequency difference. 8) High precision estimation of offsets between subswaths and frames. 9) High precision mosaic of subswaths and frames. 10) Automatic ionospheric correction. 11) Burst extraction from full-aperture ScanSAR data. 12) Burst-by-burst ScanSAR InSAR processing. 13) Azimuth or along-track deformation from ScanSAR burst Spectral Diversity (SD) or Multiple Aperture InSAR (MAI). 15) Ionospheric correction for burst-by-burst ScanSAR interferometry.
Considering the huge ALOS-2 data volume (~60G/frame), we implemented a number of techniques to speed up the processing and save space, including 1) Optimized workflow to reduce processing time. 2) Using virtual files to reduce the huge amount of input/output. 3) Using OpenMP to speed up for loops. 4) Using CUDA GPU to speed up some of the programs. We also enhanced robustness for automatic processing, and upgraded InSAR programs considering the properties of multi-mode images to enable robust processing.
As the first SAR satellite routinely operating at L-band with wide-swath coverage, ALOS-2 has acquired a wealth of data for time series analysis in many areas since February 2015. In the next few years, we will enter a new era of L-band wide swath data as a fleet of L-band SAR missions will be launched including NISAR (NASA), ALOS-4 (JAXA), ROSE-L (ESA) and Tandem-L (DLR). We present some first large-scale results processed using ALOS-2 data to explore the new opportunities brought by these missions.
