Research on deformation monitoring in mining areas based on dynamic hypothesis testing of confidence interval DS-InSAR technology

The large-scale development of coal resources can lead to a series of geological hazards, including ground subsidence, collapses, and soil erosion. Therefore, the surface monitoring of mining areas is important. In the context of Distributed Scatterers Interferometric Synthetic Aperture Radar （DS-InSAR） technology, the selection of homogeneous points is a crucial step that influences subsequent phase optimization and the choice of distributed scatterers targets. Existing methods for selecting of homogeneous points struggle to balance selection efficiency and accuracy. To address this, proposes a Dynamic Hypothesis Testing of Confidence Intervals （Dynamic HTCI） method for selecting homogeneous points. Using traditional homogeneous points selection methods and the improved method proposed in this paper for homogeneous points selection, and its effectiveness has been validated through Monte Carlo simulations and real-world experiments. Based on a total of 32 scenes of Sentinel-1data from 2017 to 2018, monitors and analyzes surface deformation in a coal mining area in Datong City. The results indicate the following conclusions: by comparing the results of selecting homogeneous points in homogeneous and heterogeneous regions, compared with traditional homogeneous point selection methods, this method is more suitable for detecting homogeneous points in mining areas with unclear land differentiation, improving both detection efficiency and selection accuracy. During the monitoring period, 12 significant subsidence funnels are identified in the mining area, with a maximum cumulative subsidence of −157 mm. The point selection density of the improved DS-InSAR method is 9.47 times that of Permanent Scatterer InSAR （PS-InSAR）, and cross-validation with PS-InSAR results confirms the accuracy of improved DS-InSAR method. This method provides a new perspective for the process of homogeneous point selection and enables non-contact, large-scale monitoring of mining areas, offering technical support for disaster prevention and mitigation efforts in mining regions.