基于变分贝叶斯学习时序InSAR数据的城市轨道沉降监测方法

Urban Rail Settlement Monitoring Method Based on Variational Bayesian Learning Time Series InSAR Data

为了合理监测城市轨道沉降,保障城市轨道行车安全,提出了基于变分贝叶斯学习时序InSAR数据的城市轨道沉降监测方法。以城市环境中的铁路线路轨道为例,使用永久散射体InSAR技术获取了一系列X段79景较短波长的SAR影像,时间跨度为2016年1月-2023年9月。通过变分贝叶斯学习,我们对SAR图像进行滤波、主辅影像选择与干涉处理,并获得了干涉相位图。然后,通过永久散射体挑选,获取了城市轨道目标点的形变数据。将所获得的形变信息进行地理编码后,将其放入适当的地理空间数据分析平台进行可视化和分析处理。通过转换为垂直方向的形变数据,可以反映研究区域城市轨道的沉降状况。测试结果表明,在研究区域内,能够获取约45280个目标点,包括轨道及其两侧各35 m范围内的数据。基于这些数据,我们可以进行城市轨道沉降观测工作,得到研究区域内城市轨道沉降速率分布,并展示观测期间城市轨道沉降的变化趋势。

In order to monitor urban rail settlement reasonably and ensure the safety of urban rail transit,a method for monitoring urban rail settlement based on variational Bayesian learning time-series InSAR data is proposed.Taking railway tracks in urban environments as an example,a series of X-segment 79 shorter wavelength SAR images were obtained using permanent scatterer InSAR technology,spanning from January 2016 to September 2023.Through variational Bayesian learning,we filtered SAR images,selected primary and secondary images,and processed them with interferometry,obtaining interferometric phase maps.Then,deformation data of urban rail target points were obtained by selecting permanent scatterers.After geocoding the obtained deformation information,it is placed on an appropriate geospatial data analysis platform for visualization and analysis processing.By converting to vertical deformation data,it can reflect the settlement status of urban tracks in the study area.The test results indicate that approximately 45280 target points can be obtained within the study area,including data within a range of 35 meters on each side of the track.Based on these data,we can conduct urban rail settlement observation work,obtain the distribution of urban rail settlement rate in the study area,and display the trend of urban rail settlement during the observation period.