一种t检验和特征函数处理位移时序方法

A method of processing GNSS displacement time series with t-test and characteristic function

针对GNSS位移时序数据中常存在由各种因素引起的异常值,使得难以准确提取GNSS站点的变形信号,而现有的GNSS位移时序处理方法主要考虑异常值的探测和变形特征参数的估计,但是对异常值进行修复的研究较少这一问题,该文提出了一种基于t检验和先验变形特征函数的GNSS位移时序处理方法,同时实现异常值的自动识别和修复以及变形特征参数的估计。利用该方法对旧金山湾区和南加州的GNSS位移时序数据进行了处理。结果表明,该方法在两个研究区域均能准确判断并有效修复GNSS位移时序中的异常值,提高GNSS位移时序对异常干扰信息的抵抗能力。将计算得到的E、N、U方向的变形速率与MIDAS算法进行对比,发现大部分GNSS站点的平面速率差异和高程速率差异都很小。对比未采用时序处理方法估计的变形速率值,发现该方法处理后的速率估值更接近MIDAS算法的结果,并且速率估值的方差更小,说明该方法能有效提高变形速率估计的精度和可靠性。

Aiming at the problem that outliers caused by various factors often exist in the global navigation satellite system(GNSS)displacement time series data,which made it difficult to accurately extract the deformation signals of GNSS stations,and the existing GNSS displacement time series processing methods mainly considered the detection of outliers and the estimation of deformation characteristic parameters,but there were few studies on the repair of outliers.In this paper,a GNSS displacement time series processing method based on t-test and prior deformation characteristic function was proposed,which realized automatic identification and repaired of outliers and estimation of deformation characteristic parameters.This method was used to process GNSS displacement time series data in the San Francisco Bay Area and Southern California.The results showed that the method could accurately judge and effectively repaired the outliers in the GNSS displacement time series in the two study areas,and improved the resistance of the GNSS displacement time series to abnormal information.Comparing the calculated deformation velocities in the three directions of E,N and U with the median interannual difference adjusted for skewness(MIDAS)algorithm,it was found that most of the GNSS sites had very small differences in plane direction and elevation direction.Compared with the deformation rate estimation without this time series processing method,we found that the rate estimation processed by this method was closer to the result of the MIDAS algorithm,and the variance of the rate estimation was smaller,indicating that this method could effectively improve the accuracy and reliability of the deformation rate estimation.