A distributed parallel processing framework for sentinel-1 wide-area time series InSAR: Application in Jining City

Abstract

Interferometric Synthetic Aperture Radar plays a critical role in large-scale surface deformation monitoring. Nevertheless, conventional time-series InSAR techniques are often constrained by low computational efficiency and heavy reliance on high-performance hardware when processing massive datasets. To overcome these challenges, we develop a distributed and parallel time-series InSAR processing strategy. Sentinel-1 imagery is segmented into burst-level units, enabling task partitioning and multi-device parallel computation, which substantially reduces the dependence on advanced computing resources. In addition, a mosaicking and correction workflow is established, where quadratic polynomial fitting and distance-weighted blending are applied for burst and swath mosaicking. This ensures the spatial continuity of the deformation field. A case study in Jining City demonstrates that the proposed approach achieves efficient large-area processing on a personal computer, successfully identifying significant subsidence zones in mining areas and characterizing their temporal evolution. Field surveys confirm the occurrence of road fractures and building cracks in regions of severe deformation. Overall, the proposed framework enhances computational efficiency and lowers hardware requirements, thereby providing a cost-effective and practical solution for regional geohazard monitoring and infrastructure safety assessment.