Prediction of Surface Subsidence Based on PSO-BP Neural Network

Abstract

Abstract The traditional settlement monitoring has the problems of small monitoring range and low adaptability of settlement prediction model; Based on SBAS-InSAR technology and PSO-BP neural network algorithm, this paper constructs a PSO-BP settlement prediction model based on Sentinel-1A data. Take Jinan(E116°8’, N36°8’)the regional monitoring data is a data sample set. The algorithm proposed in this paper is compared with typical algorithms such as SVM. The results show that the MAE of 30 points predicted by this algorithm is 0.17, which is better than six typical algorithms such as SVM. This algorithm can be used as an effective means to provide early warning of surface subsidence.