An improved pixel-tracking method for monitoring mining subsidence

Underground mining always induces large vertical displacements in the ground surface and, because of the large subsidence gradient, the phase-unwrapping methods of interferometric synthetic aperture radar (InSAR) are unable to give accurate results. Pixel tracking based on normalized cross-correlation maximization overcomes the limitations imposed by the subsidence gradient and can be used to monitor large displacements. This paper introduces and analyses each component of the offset field and then describes the improved method used to efficiently remove orbital error and topography-related offset caused by rugged terrain with the help of a digital elevation model (DEM). Eleven TerraSAR-X images in spotlight mode and Shuttle Radar Tomography Mission (SRTM) DEM data are used to monitor mining subsidence by the improved pixel-tracking method in Shenmu County of Yulin City. The root mean square error (RMSE) between the improved method and Global Positioning System (GPS) data in the strike and dip directions are 0.143 and 0.108m, respectively. The approach presented here is shown to be appropriate for monitoring large vertical displacements in mining subsidence.