Analysis of land surface changes based on time-series data interferometric synthetic aperture radar with the application of improved combined scatterers interferometry with optimized point scatterers

Abstract

InSAR  time-series analysis is a powerful remote sensing technique that can measure ground deformation with high spatial and temporal resolution. It has been widely used in various applications, such as the detection and monitoring of land subsidence, earthquakes, landslides, and other geohazards. In recent years, there has been significant development in InSAR time-series analysis, including the development of new algorithms and techniques, the availability of new and improved satellite data, and the application of InSAR in new and innovative ways. In this research, we propose the use of improved combined scatterers interferometry with optimized point scatterers (ICOPS) time-series analysis for InSAR application in the continuous monitoring of surface deformation. The key features of ICOPS include the optimization of point scatterers, the use of combined scatterers, and the optimization based on machine learning. These features enable ICOPS to effectively handle non-linear deformation, slow rate deformations, and to provide more accurate and reliable deformation estimates. The ICOPS method can be briefly explained by optimizing the measurement points obtained based on combined scatterers interferometry of persistent scatterers (PS) and distributed scatterers (DS) using machine learning and statistical based approaches. Distributed scatterers (DS) are used as a complement to persistent scatterers which provide increased spatial coverage with an assessment of pixel quality. After the measurement points are obtained, then the optimization process begins with support vector regression (SVR) which can handle non-linear cases and is capable of processing large data. Then, optimization results using machine learning will then be maximized using the optimized hot-spot analysis (OHSA) method to obtain spatially clustered deformation maps. In practically, we applied ICOPS to InSAR data of a surface deformation case study in the several regions and compared the results with those obtained using other InSAR methods. For the application, we applied the ICOPS in the Yellowstone National Park, USA, for detecting surface deformation around Yellowstone caldera related with volcanic activity. Then, the application for ICOPS was used in land subsidence in coastal city in Semarang, Indonesia. We also try the ICOPS for measure the surface deformation related with the construction activity in reclaimed area in Dangjin, South Korea. The results showed that ICOPS can accurately and reliably monitor surface deformation, even in complex and challenging scenarios. In summary, our study demonstrates the potential of ICOPS time-series analysis for InSAR application in the continuous monitoring of surface deformation and highlights its advantages over other methods. This work can contribute to the development of more effective and robust InSAR-based monitoring systems for surface deformation and support the sustainable and resilient management of our built and natural environments.