A landslide displacement intelligent prediction method and system fusing spatial neighborhood features and time sequence displacement decomposition
By combining SBAS-InSAR technology and wavelet decomposition, adapting linear autoregressive and gated recurrent neural network models respectively, and introducing a spatial neighborhood optimization algorithm, the problem of unutilized spatial correlation and temporal decomposition features in existing landslide displacement prediction is solved, achieving higher accuracy and interpretability in landslide displacement prediction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XIANGTAN UNIV
- Filing Date
- 2026-04-16
- Publication Date
- 2026-07-17
AI Technical Summary
Existing landslide displacement prediction methods neglect spatial correlation characteristics and temporal decomposition characteristics, resulting in prediction models failing to fully utilize neighborhood information and failing to accurately reflect the spatial propagation mechanism and complex characteristics of landslide deformation.
High-precision time-series deformation data were extracted using SBAS-InSAR technology. The displacement was decomposed into trend displacement and periodic displacement by wavelet decomposition. Linear autoregressive model and gated recurrent neural network model were used for prediction, respectively. A spatial correlation optimization algorithm based on Euclidean distance and orientation similarity was introduced for weighted optimization.
It significantly improves the accuracy and interpretability of landslide displacement prediction, overcomes the limitations of a single model in taking into account both trend and periodic predictions, achieves higher prediction accuracy and computational efficiency, and has clear physical meaning.
Smart Images
Figure CN122410522A_ABST