Research on Land Subsidence Prediction in Mining Areas Based on SBAS-InSAR and Multi-Model Comparison

Land subsidence in mining areas caused by underground resource exploitation poses a serious threat to surface stability and ecological security. Accurate prediction of land subsidence is crucial for disaster prevention and mitigation in mining areas. This study integrates SBAS-InSAR technology with various predictive models to analyze and forecast surface subsidence in mining areas. First, SBAS-InSAR technology is used to process Sentinel-1 data from 2018 to 2023, extracting time-series deformation data in the study area. Then, based on the key influencing factors of land subsidence, a comparative analysis of multiple models, including SVR, PSO-SVR, and HOLT, is conducted to assess their performance in subsidence prediction. The results show that the PSO-SVR model, optimized with particle swarm optimization, demonstrates superior accuracy and reliability compared to the other models. This provides a robust approach for monitoring and predicting land subsidence in mining areas, offering technical support for risk management and sustainable resource development.