Regional land subsidence space-time prediction method

The invention discloses a regional land subsidence space-time prediction method which comprises the following steps: S1. by combing with radar image data, acquiring land subsidence information through radar data processing software, and carrying out precision verification on the land subsidence data through a field measurement GPS (global positioning system) and leveling data, thereby obtaining a high-precision land subsidence information database; S2. carrying out simulation and prediction on land subsidence by an improved grey model; and S3-S4. clustering monitor value/analog value ratio and errors by an SOM clustering algorithm to acquire a Markov interval and calculate a Markov matrix, improving the grey model predicted value through an improved Markov model, and carrying out large-scale land subsidence prediction on InSAR. The large-scale ground prediction is carried out through the InSAR, thereby reducing the dependence on manual work and saving the engineering cost. 本发明公开了种区域地面沉降时空预测方法，包括：S1、结合雷达影像数据，通