Combined SBAS-InSAR and PSO-BP algorithm for evaluating the risk of geological disasters in alpine valley regions

Geological disasters have posed serious threats to the living, economic security, and local infrastructure, particularly those frequently occurring in the alpine valley regions in recent years. However, the existing evaluation needs to calculate a large number of weighted factors, leading to low timeliness and poor accuracy in the geological disaster data. In this study, a combined SBAS-InSAR (Small Baseline Subsets-Interferometric Synthetic Aperture Radar) and PSO-BP (Particle Swarm Optimization- Back Propagation) algorithm was proposed to evaluate the risk of geological disasters in the alpine valley regions. First, the Small Baseline Subsets-Interferometric Synthetic Aperture Radar (SBAS-InSAR) technology was used to record the deformation rates for the ascending and descending orbits in the study area, where the high-resolution images were captured for the geological disasters data. Next, the Particle Swarm Optimization-Back Propagation (PSO-BP) geological disaster model was trained using the 12 evaluation factors, such as the elevation, slope, ascending and descending deformation rates. Finally, the new model was used to acquire the geological disaster index for the study area. An ArcGIS natural break point grading was used to grade the risk of geological disasters. As such, a risk evaluation was obtained for the validity of the model. The results showed that the combined ascending and descending orbits performed better to identify the geological disasters in the Alpine valley regions, compared with the current single tracks. A more accurate and comprehensive identification of geological disasters was also achieved to avoid the geometric distortion of SAR imaging. Furthermore, the surface shape variables were acquired using the SBAS InSAR technology combined with high-resolution imaging. A rapid and accurate identification was realized to effectively identify the active rock, slippery slope, collapse, and potential geological disasters. A large number of calculations were avoided for the evaluation factors under the PSO-BP method. In addition, informatics and combined empowerment were selected to quantitatively and qualitatively compare the InSAR-ANN model. It was found that the Area Under the Curve (AUC) values of information, combination weighting, and INSAR-ANN were 0.694, 0.721, and 0.785, respectively, and the accuracies were 73.3%, 76.2%, and 79.8%, respectively, indicating the higher efficiency and accuracy of the improved InSAR-ANN model in the risk evaluation on the geological disasters. Consequently, the INSAR-ANN model can be expected to effectively implement a risk assessment of geological disasters in the Alpine canyon areas. This finding can also provide a strong reference for disaster prevention and decision-making mitigation.