Sovremennye problemy distantsionnogo zondirovaniya Zemli iz kosmosa, 2020, Vol. 17, No. 2, pp. 49-61
Mass-parallel approach to radar data processing
S.E. Popov
1 , R.Yu. Zamaraev
1 , L.S. Mikov
1 1 Institute of Computational Technologies SB RAS, Novosibirsk, Russia
Accepted: 18.03.2020
DOI: 10.21046/2070-7401-2020-17-2-49-61
The paper describes a modern approach to creating a high-performance computing system for processing satellite radar images based on Apache Spark technology. We consider complete data processing schemes for constructing Earth surface displacement velocities using small baseline methods (SBaS) and constant reflectors (PS). Both methods are implemented in several stages, at each of which the calculation algorithm has its own tuning parameters. Their combination determines the effectiveness of an individual stage and the entire calculation as a whole. Accordingly, the task arises of organizing on massive data a multivariate calculation with user control of intermediate results and selection of parameters. To solve it, adapted schemes for auto running computational tasks in parallel mode in a cluster environment running Apache Spark using executing objects have been developed. A feature of the proposed solutions is the use of custom containers-executors with internal mechanisms of interaction between calculation algorithms and the possibility of combining containers into a single launch scenario to obtain the final solution in the form of surface offsets. The paper provides a general description of the organization of parallel computing and describes the features of the implementation of specific stages of pre-processing in the framework of the proposed approach. Comparative results of testing the computing system on a demonstration cluster are presented. The possibility of significantly reducing the time it takes to perform calculations in the processing of radar data using only open standards and freely distributed software libraries, as well as relatively cheap hardware, is shown.
Keywords: satellite radar data, differential interferometry, calculation of earth surface displacements, mass-parallel computing, Apache Spark
Full textReferences:
- Feoktistov A. A., Zakharov A. I., Gusev M. A., Denisov P. V., Issledovanie vozmozhnostei metoda malykh bazovykh linii na primere modulya SBaS programmnogo paketa SARScape i dannykh RSA ASAR/ENVISat i PALSAR/ALOS. Chast 1. Klyuchevye momenty (Study of the possibilities of the small baseline method using the SBaS module as an example of the SARScape software package and PCA data ASAR/ENVISat and PALSAR/ALOS. Part 1. Key points of the method), Zhurnal radioelektroniki, 2015, No. 9, pp. 1–26.
- Costantini M., Farina A., Zirilli F., A fast phase unwrapping algorithm for SAR interferometry, IEEE Trans. Geoscience and Remote Sensing, 1999, Vol. 37, No. 1, pp. 452–460.
- Crosetto M., Monserrat O., Cuevas-González M., Devanthéry N., Crippa B., Persistent Scatterer Interferometry: A review, ISPRS J. Photogrammetry and Remote Sensing, 2016, Vol. 115, pp. 78–89.
- Ferretti A., Prati C., Rocca F., Wasowski J., Satellite interferometry for monitoring ground deformations in the urban environment, Proc. 10th Congress Intern. Association for Engineering Geology and the Environment (IAEG), 2006, pp. 1–4.
- Gao Sh., Zeng Q., Jiao J., Liang C., Tong Q., Parallel processing of InSAR interferogram filtering with CUDA programming, Science of Surveying and Mapping, 2015, No. 1, pp. 54–68.
- Goldstein R. M., Werner C. L., Radar Interferogram Phase Filtering for Geophysical Applications, Geophysical Research Letters, 1998, Vol. 25, pp. 4035–4038.
- Guerriero A., Anelli V., Pagliara A., Nutricato R., Nitti D., Efficient implementation of InSAR time-consuming algorithm kernels on GPU environment, Proc. IEEE Intern. Geoscience and Remote Sensing Symp. (IGARSS-2015), 2015, pp. 4264–4267.
- Karasev P., Campbell D., Richards M., Obtaining a 35x Speed up in 2D Phase Unwrapping Using Commodity Graphics Processors, Proc. IEEE Radar Conf., 2007, pp. 574–578.
- Lanari R., Casu F., Manzo M., Zeni G., Berardino P., Manunta M., Pepe A., An Overview of the Small BAseline Subset Algorithm: A DInSAR Technique for Surface Deformation Analysis, Pure and Applied Geophysics, 2007, Vol. 164, pp. 637–661.
- Marinkovic P. S., Hanssen R. F., Kampes B. M., Utilization of Parallelization Algorithms in InSAR/PS-InSAR Processing, Proc. Envisat ERS Symp. (ESA SP-572), 2004, pp. 1–7.
- Sousaa J. J., Hooperc J. A., Hanssenc R. F., Bastosd L. C., Ruize A. M., Persistent Scatterer InSAR: A comparison of methodologies based on a model of temporal deformation vs. spatial correlation selection criteria, Remote Sensing of Environment, 2011, Vol. 115, No. 10, pp. 2652–2663.
- Zhang F., Wang B., Xiang M., Accelerating InSAR raw data simulation on GPU using CUDA, Proc. IEEE Intern. Geoscience and Remote Sensing Symp. (IGARSS-2010), 2010, pp. 2932–2935.