Monitoring and Forecasting of Urban Land Subsidence Based on PS-InSAR and Improved CNN Network
Based on the purpose that urban surface subsidence has caused different degrees of damage to urban facilities in recent years,and the traditional monitoring methods cannot meet the requirements of large-scale and long-time monitoring,the method of data processing of 120-view Sentinel-1 ascending orbit images of a city from January 2019 to December 2022 using PS-InSAR and GACOS products is used to obtain the subsidence rate of-58.18mm/a～22.51mm/a.On this basis,seven typical surface subsidence zones are obtained and their spatial and temporal distribution characteristics are analysed.Finally,1000 subsidence points within the B subsidence zone were selected for predic-tion analysis using CNN and improved CNN prediction models,and the R2 and root mean square error of the predicted values were obtained to be improved.The experimental results show that the combination of PS-InSAR technology and improved CNN network for urban ground settle-ment prediction can provide a valuable reference for future urban ground settlement monitoring and urban disaster warning.
land subsidencePS-InSARSentinel-1CNN networkdisaster warning
万方数据
维普
