IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (Jan 2024)
Independent Component Analysis (ICA) Based Method for Estimating the Deformation of Highways in Permafrost Region (HPICA)—A Case Study of Maduo Section of Gongyu Highway
Abstract
Highways built in permafrost regions are susceptible to deformations and instability of the roadbed caused by climatic factors. Long-term deformation monitoring is essential to reveal the freeze-thaw-related deformations. When using Interferometric Synthetic Aperture Radar (InSAR) for permafrost highway monitoring, the majority of different physical phase components are usually considered as equally weighted, and the permafrost deformation-related components are mostly modeled with an empirical mathematical model. This may induce uncertainty and difficulties to remove the atmospheric delay and orbital error, which affects both the accuracy and efficiency of deformation estimation. To address these limitations, we propose an independent component analysis (ICA) based method for estimating the deformation of highways in permafrost regions (HPICA). In HPICA, the Fast ICA is utilized to separate the original InSAR unwrapped phases, and then the extracted freeze-thaw deformation-related components are modeled considering the climatic factors. The simulated experiments show that the spatial ICA can more accurately separate the deformation-related signals from the mixed signals than that of temporal ICA. The Maduo section of Gongyu Highway on the Tibetan Plateau was selected as a study area in the real-data experiment. The results showed the maximum cumulative settlement spanning January 2020 to January 2022 was up to –140.8 mm. A comparative analysis indicated that the modeling accuracy of HPICA is with significant improvement. Besides, HPICA could reveal the boundaries of different permafrost regions according to the nature of permafrost, thus assisting in spatial classification of different types of soil regions.
Keywords
