Image compression–based DS-InSAR method for landslide identification and monitoring of alpine canyon region: a case study of Ahai Reservoir area in Jinsha River Basin

Abstract

Interferometric Synthetic Aperture Radar (InSAR) technology is capable of detecting large areas of potentially unstable slopes. However, traditional time-series InSAR methods yield fewer valid measurement points (MPs) in alpine canyon regions. Distributed Scatterer (DS) Interferometry (DSI) technology serves as a potent tool for monitoring surface deformation in complex land cover areas; nonetheless, it grapples with high computational demands and low efficiency when interpreting deformation across extended time series. This study proposes an image compression–based DSI (ICDSI) method, which, building upon the DSI method, utilizes principal component analysis (PCA) to compress multi-temporal SAR images in the time dimension. It develops a module for compressing long-time sequence SAR images, acquires the compressed image (referred to as a virtual image), and integrates the developed image compression module into the DSI data processing flow to facilitate the inversion of long-time sequence InSAR land surface deformation information. To validate and assess the credibility of the ICDSI method, we processed a total of 78 ascending and 81 descending scenes of Sentinel-1A images spanning the period 2019–2021 using Small Baseline Subset (SBAS), DSI, and the ICDSI method proposed in this paper. Subsequently, these methods were applied to detect landscape displacements on both coasts of the Jinsha River Basin. The investigation reveals that the ICDSI method outperforms SBAS and DSI significantly in monitoring landslide displacements, enabling the detection of more measurement points (MPs) while utilizing less raw data. The accomplishments of this research program carry crucial theoretical implications and practical application value for the detection of surface deformation using long-time series InSAR.