Automated detection and fissure width quantification of ground fissures using FastICA-enhanced distributed fiber optic sensing

Abstract

The increasing impact of ground fissures on urban infrastructure demands monitoring approaches that not only provide high spatial resolution but also allow robust and automated interpretation across heterogeneous sensing systems. However, interpreting long-term distributed fiber optic sensing (DFOS) data remains challenging, particularly when different cable types respond to the same deformation processes with comparable spatial patterns but substantially different measurement scales. In this study, a DFOS-based monitoring framework integrating fast independent component analysis (FastICA) with a segmented integration strategy is proposed to enable automated detection of fissure-related strain anomalies and quantitative estimation of fissure width. Continuous observations were conducted using simultaneous measurements from a metal-reinforced strain cable (C1) and a micro-anchored fiber optic cable (C2) deployed across three active ground fissures. The proposed method successfully localized all active fissure zones and identified their initiation timing and temporal evolution. Results show that both cable types consistently detect the same fissure locations and temporal deformation features. The C1 cable exhibits reliable sensitivity to fissure occurrence across all monitored sites but yields fissure width estimates with limited quantitative accuracy. In contrast, the C2 cable enables robust reconstruction of the spatiotemporal evolution of the main fissure width, which shows strong agreement with differential settlement derived from interferometric synthetic aperture radar (InSAR) observations, with correlation coefficients reaching up to 0.97. For fissures wider than 10 mm, the temporal responses of the two cable types are highly correlated (0.88–0.94), indicating stable identification of dominant deformation events despite pronounced amplitude discrepancies. Overall, the proposed DFOS–FastICA framework facilitates automated, real-time tracking of ground fissure dynamics while explicitly accounting for cable-dependent response characteristics. The results highlight the complementary roles of different DFOS cable designs and demonstrate the potential of multi-cable integration for reliable urban geological hazard assessment and infrastructure risk management.