GNSS 监测数据预处理在变形预测中的应用

GNSS监测技术被广泛应用于变形监测工作中,但GNSS监测数据中会有缺失值、噪声等误差的存在,对预测结果造成影响。引入改进的小波神经网络模型进行变形预测,并考虑该模型的自适应性和容错性,分别采用三次样条插值法、小波滤波法和拉依达准则对原始监测数据进行缺失值填补、去噪和粗差剔除等预处理。并利用实际监测数据进行短期预报分析,对比原始监测数据和预处理后的监测数据的预测结果,结果表明预处理后的监测数据的预测效果更好。

Application of GNSS monitoring data pretreatment to the prediction of deformation

As GNSS monitoring technology is more and more widely used ,there will be some missing data and noise errors in the GNSS monitoring data ,which will affect the result of prediction .The improved wavelet neural network model is used for deformation prediction .And considering the adaptability and fault tolerance of the model ,this paper uses the methods such as cubic spline interpolation ,w avelet filtering and Pauta criterion to pretreat the monitoring data .The actual monitoring data is used for short‐term forecasting analysis .Then comparing the prediction results of original and pretreated monitoring data ,the result show s that the prediction results of pretreated monitoring data are better .