Research on Monitoring and Warning of Terraced Loess Potential Landslide Based on Data Fusion

Objectives: The frequent occurrence of catastrophic landslide disasters in China seriously threatens the safety of human life and property. Based on data fusion, multi-source heterogeneous landslide hazard monitoring data are analyzed and early warning criteria are proposed, which can effectively avoid risks and reduce economic losses and casualties. Methods: We use the hidden danger of Guamagou terraced loess landslide as an example, the monitoring data of Global Navigation Satellite Systems (GNSS) surface displacement, crack meter displacement and rainfall are obtained. The monitoring data are preprocessed by gross error elimination, data interpolation and data smoothing. Then, on the basis of preprocessing, the data fusion and effect evaluation of data level, feature level and decision level are carried out. Finally, the early warning criterion and classification of hidden danger of terraced loess landslide are put forward. Results: The results show that: Data preprocessing not only significantly improves the quality of monitoring data, but also greatly enhances the accuracy and reliability of the early warning system; The displacement-time curve of the hidden danger of loess landslide in terrace type shows the characteristics of convergence, that is, with the passage of time, the cumulative displacement shows a state of rapid increase first, then slow growth until it tends to be stable, and its deformation rate finally tends to'0'; Data fusion can accurately capture the deformation characteristics of hidden dangers of terraced loess landslides, and the error of prediction and evaluation decreases with the improvement of data fusion level; The tangent angle, cumulative acceleration, rainfall intensity and fracture stage matching characteristics can be used as early warning criteria for hidden dangers of terraced loess landslides. Conclusions: The monitoring and early warning of the hidden danger of terraced loess landslide based on data fusion will provide theoretical and scientific basis for further promoting slope modification projects and protecting the existing terraces, as well as for increasing farmers' income, providing scientific and technological services for "three farmers" and promoting rural revitalization.