变形信号与噪声的分离方法及其应用

在对变形体进行监测的时候会得到一系列不同时间的观测结果,比如对滑坡的位移监测,得到一系列的坐标,通过与首期观测数据的比较可获得变形量,然而测量中不可避免的含有误差,如何将变形信息和观测噪声分离是一个需要解决的问题。本文在变形是一个连续过程的假设下,以给定的验前方差为约束条件,通过三次连续逼近样条函数建立变形曲线,达到变形信号与噪声分离的目的,文中通过对某滑坡多年监测数据的分析验证了该方法的可行性。

Distinguish method between deformation and noise information and its application

Observation results can be obtained with time series if conduct a continuous monitoring on a deformation structure.For example,a list of coordinates can be gotten during the monitoring for the displacement of a coast.Actually,the distortion value can be calculated by comparing the current observation with the previous observation.However,the observation error is inevitable because of different measurement circumstances.It is an important issue to distinguish distortion from observation noise.So under a hypothesis that the distortion is accordance with a continuous procedure,the solid distortion curve is established by using cubic spline approximating function in this paper.Then the distortion information can be extracted from error noise by using the given pretest variance error as a restriction condition.Finally,a case study is given by analyzing years of observation from a coast deformation monitoring,and the feasibility of the above method is proven.