小波变换和BP神经网络模型在沉降变形监测中的应用研究  

Application of Wavelet De-noising and BP Neural Network Model in Settlement Deformation Monitoring

作　　者：郑云云 陈姗姗[4] 胡勇[1,2,3] ZHENG Yunyun;CHEN Shanshan;HU Yong(Chongqing Institute of Surveying and Planning for Land Resources and Housing,Chongqing 401120,China;Yuzhong District Land and Resources Management Branch of Chongqing,Chongqing 401120,Chian;Chongqing Engineering Research Center of Land Use and Remote-Sense Monitoring,Chongqing 401120,China;Chongqing Branch of National Remote Sensing Application Engineering Technology Center,Chongqing 401120,China)

机构地区：[1]重庆市国土资源和房屋勘测规划院,重庆401120 [2]重庆市土地利用与遥感监测工程技术研究中心,重庆401120 [3]国家遥感应用工程技术研究中心重庆分中心,重庆401120 [4]重庆市渝中区国土资源管理分局,重庆401120

出　　处：《测绘与空间地理信息》2019年第2期101-103,107,共4页Geomatics & Spatial Information Technology

基　　金：重庆市国土房管局2016年科技计划项目(KJ-2016001);住房和城乡建设部科技项目(2016-k8-054)资助

摘　　要：变形预测在预报工程险情方面起着关键性的作用,针对施工中需及时、准确地预测变形的问题,本文利用小波变换原理对监测数据进行降噪处理,并采用BP神经网络分析不同训练样本下的预测效果和精度水平。实验结果表明:基于小波消噪后的BP网络模型,以连续的近期观测数据作为训练样本,对下期变形预测精度高,效果好,相对误差很小。因此,小波变换和BP神经网络模型在沉降变形监测工程中能作为预测研究与应用的参考。Deformation prediction plays a key role in predicting the danger of engineering.It is necessary to predict the deformation timely and accurately in the construction.In this paper,the wavelet transform principle is used to reduce the noise of monitoring data,and BP neural network model is used to analyze the prediction effect and accuracy for different training samples.The experimental results show that the continuous recent observation data is used as the training sample,the prediction accuracy of the deformation prediction is high,the effect is good and the relative error is small based on the BP network model after wavelet de-noising.Therefore,this method can be used as a reference for prediction research and application in settlement deformation monitoring project.

关 键 词：变形监测 小波消噪 BP网络模型 预测精度 

分 类 号：P25[天文地球—测绘科学与技术] TB22[天文地球—大地测量学与测量工程]