信息通信基于InSAR技术地铁沿线地表形变监测与预测（江西理工大学土木与测绘工程学院，江西赣州341000）摘要：地铁在修建和运行过程中，往往会沿线周边的区域发生沉降。这不仅会使地铁的运行发生危险，还会影响周围工程建筑的稳定性。文章用PS-InSAR技术对研究区37景的Sentinel-1ASAR数据进行处理，提取了南昌地铁沿线400m范围缓冲区内的形变速率场。研究表明地铁沿线大部分区域整体沉降速率大小在-8mm/a~8mm/a之间，个别区域的沉降速率较为严重，最大沉降速率可达-36.7mm/a。最后采用BP神经网络模型和小波神经网络模型对累计形变时间序列数据进行预测实验，结果表明小波神经网络模型在形变数据上的预测精度优于传统的BP神经网络模型，在形变预测方面具有一定的优势和良好的应用前景。关键词：沉降;PS-InSAR;BP神经网络；小波神经网络中图分类号：TP311MonitoringandPredictionofSurfaceDeformationAlonetheMetroLineBasedonInSARTechnology(SchoolofCivilandSurveyingandMappingEngineering,JiangxiUniversityofScienceandtechnology,Ganzhou,341000,China)Abstract:Duringtheconstructionandoperationofasubway,settlementoftenoccursinareasalongtheimmediatesurroundingsoftheline.Itnotonlymakesthemovementofthemetrodangerousbutalsoaffectsthestabilityofthesurroundingengineeringbuildings.Inthispaper,Inthispaper,weprocessedtheSentinel-1ASARdataof37viewsinthestudyareawithPS-InSARtech-niquetoextractthedeformationratefieldwithinthebufferzoneof40OmalongtheNanchangmetro.Thestudyshowsthattheoverallsettlementrateinmostareasalongthesubwaylineisbetween-8mm/aand8mm/a,andthesettlementrateinindividualareasismoreserious,withthemaximumsettlementratereaching-36.7mm/a.Finally,usingBPneuralnetworkmodelandwa-veletneuralnetworkmodeltopredictthecumulativedeformationtimeseriesdata,theresultsshowthatthepredictionaccuracyofwaveletneuralnetworkmodelonthedeformationdataisbetterthanthetraditionalBPneuralnetworkmodel,whichhascer-tainadvantagesandgoodapplicationprospectsindeformationprediction.Keywords:Settlement;PS-InSAR;BPNeuralNetworks;waveletneuralnetwork;settlementtimesequence0引言由于城市的发展，建筑物的规模不断扩大，地下商场、地铁等地下工程设施被修建以及地下水的过度开采导致...
