Retrieval and Prediction of Three-Dimensional Displacements by Combining the DInSAR and Probability Integral Method in a Mining Area
Chuanguang Zhu, Zhengshuai Wang, Peixian Li, Mahdi Motagh, Liya Zhang, Zongli Jiang, Sichun Long
Abstract
Monitoring ground displacement produced by underground mining is essential to ensure the safety of infrastructure over mining areas. Differential synthetic aperture radar (DInSAR) can only obtain the 1-D [i.e., along the line-of-sight (LOS) direction] displacement component. In this study, we present an improved algorithm for retrieving and predicting 3-D displacement fields induced by underground mining based on the LOS displacement derived from DInSAR and the probability integral method (PIM). Whole parameters included in the standard PIM model are involved in the improved algorithm. In addition, the interaction between multiple working panels is considered and incorporated into the model. Next, a stochastic optimization technique hybridizing the cultural algorithm and random particle swarm optimization has been designed to retrieve model parameters, which can be used to retrieve and predict the 3-D displacement field. Simulated experiments show that the root mean square errors (RMSEs) are 10, 12, and 17 mm in the vertical, east-west, and north-south directions, respectively, by comparing the simulated and retrieved 3-D displacement. Furthermore, the capability of the proposed method is investigated and validated in the Xuehu mining area of China using three ALOS PALSAR acquisitions. Our results agree well with leveling measurements in the vertical direction with an RMSE of 38 mm. Although the retrieved horizontal displacement cannot be validated due to a lack of field surveys, these displacement fields coincide spatially with the evolution of mining excavation.