Spatiotemporal deformation trend assessment based on ICOPS and ICA methods using multitemporal InSAR data in industrial development zones
摘要
The expansion of the industrial sector has created a need for comprehensive infrastructure development. Remote sensing technology, particularly synthetic aperture radar (SAR), has shown substantial importance in the analysis of surface deformation in the context of urban development. A comprehensive study was conducted in the Bugok industrial area, South Korea, in which improved combined scatterer interferometry with optimized point scatterer (ICOPS) methodology was implemented. The ICOPS methodology demonstrated the reliability of using machine learning algorithms in refining postprocessing time series. This investigation utilized Sentinel-1 SAR satellite data from ascending and descending tracks, and Cosmo SkyMed (CSK) acquired from 2014 to 2022. Furthermore, the independent component analysis (ICA) method was used to decompose the dominant trend and clustering of deformation data. The implementation of the ICOPS method facilitated the visualization of linear and nonlinear spatiotemporal deformation during the observation period. The time-series analysis indicated subsidence with a magnitude of 0.76–2.83 cm/year, reaching a maximum cumulative deformation of approximately 20 cm. Based on the principal component data on vertical deformation, there are seasonal, linear‒seasonal, and exponential trend characteristics in the study area. The contributing factors include the high density of industrial structures and intensive groundwater extraction, resulting in increased loads on soil structures and consequently continued subsidence in the linear cluster area. Furthermore, this study underlines the potential application of the InSAR methodology in the context of postconstruction infrastructure monitoring, contributing significantly to risk management and estimation.