Real-time filtering and application of landslide GNSS deformation data based on dynamic frequency conversion

GNSS deformation monitoring devices are crucial ground monitoring tools for capturing multi-directional landslide deformations and their development trends. However, precision and random errors during monitoring often leads to significant fluctuations in computed deformation rates, leading to false alarms in real-time monitoring and warnings. This study addresses the characteristics of typical GNSS deformation monitoring data by analyzing the relationship across three stages between the total filtered data and maximum deviation. It establishes that the filtering of GNSS data should employ 40 data sets corresponding to the initial phase of the gradually decreasing stage. Furthermore, this paper proposes utilizing a data buffer filtering area combined with dynamic frequency monitoring techniques for the prompt elimination of incidental errors. Finally, based on the least squares method and data rejection techniques, the study achieves real-time filtering of GNSS deformation data, verified through monitoring at two typical slope sites. These methods provide a technical references for subsequent real-time landslide analysis and early warning of landslides based on deformation data.