A TS-InSAR clustering approach to detect spatio-temporal changes inground deformation

SAR images can be used to measure changes in the surface of the Earth over time using Time Series Interferometric Synthetic Aperture Radar (TS-InSAR) techniques. TS-InSAR enables the detection and measurement of very small changes in surface deformation, often on the order of millimetres or less. This makes it a powerful tool for monitoring a wide range of natural and man-made phenomena, such as tectonic activity, subsidence, ground water extraction, and the behaviour of engineered structures like buildings and bridges. While TS-InSAR provides deformation measurements, further analysis must be taken to understand the underlying cause of the deformation. In this study, a novel framework has been developed to extract the vast amount of information embedded within the large number of ground deformation Measurement Points (MPs) derived from the Small BAseline Subset (SBAS; Berardino et al., 2002) TS-InSAR technique. The proposed automatic data-mining approach begins with clusterization the TS-InSAR MPs by applying a nonlinear dimensionality-reduction technique, Uniform Manifold Approximation and Projection (UMAP; McInnes et al., 2018), prior to performing clustering with Hierarchical Density based Spatial Clustering of Applications with Noise (HDBSCAN; Campello et al. 2013) in order to group together MPs exhibiting similar deformation behaviour on a large scale. Next, every extracted cluster time series is further investigated by applying a piecewise linear function as a method to detect and quantify accelerations and decelerations of deforming areas.

A test of the method has been conducted over the Bandung Basin (Indonesia) using Sentinel-1 data from October 2015 to December 2020. Application of the method provides an objective way to identify changes in displacement rates over time and provides a wealth of information on the dynamics of surface displacement over a large area. The displacement rates, their spatial variation, and the timing and location of accelerations and decelerations can be used to investigate the physical behaviour of the deforming ground by linking the timing and location of changes in displacement rates to causal and triggering factors.

References
Berardino, P., Fornaro, G., Lanari, R., Sansosti E. A new algorithm for surface deformation monitoring based on small baseline differential SAR interferograms. IEEE Transactions on Geoscience and Remote Sensing, 40 (11) (2002), pp. 2375-2383.
Campello, R.J., Moulavi, D., Sander, J. Density-based Clustering Based on Hierarchical Density Estimates. In Advances in Knowledge Discovery and Data Mining, Proceedings of the Pacific-Asia Conference on Knowledge Discovery and Data Mining; Pei, J., Tseng, V.S., Cao, L., Motoda, H., Xu,
G., Eds.; Springer: Berlin, Germany, 2013; pp. 160–172 McInnes, L. and Healy, J. UMAP: uniform manifold approximation and projection for dimension
reduction. Preprint at https://arxiv.org/abs/1802.03426 (2018).