单历元GPS变形信息特征提取的EMD方法

为了准确地对单历元GPS变形信息特征进行提取，提出应用EMD（经验模式分解）方法对变形数据进行去噪处理。通过对实际工程算例进行分析并与小波变换及中值滤波方法进行对比，结果表明：EMD方法的去噪效果优于中值滤波方法，同时EMD方法具有自适应滤波去噪的功能．避免了应用小波去噪时其效果与小渡基函数的选择相关的缺陷。在对变形数据进行有效去噪的基础上，应用EMD方法可以完成对单历元GPS变形信息进行有效识别并提取出微弱的变形特征。

Empirical Mode Decomposition Method for Single Epoch GPS Deformation Information Characteristic Extracting

In order to extract single epoch GPS deformation information accurately,a method using Empirical Mode Decomposition （EMD） for deformation data denosing processing is presented. Through analyzing practical GPS deformation monitoring data, and comparing EMD with wavelet transform and median value filter, the results showed that： denosing effect of EMD is better than median filter method. Meanwhile,data denosing with EMD can avoid the defect that denosing effect of wavelet transform is related with choice of wavelet basis function,because EMD has an advantage to do adaptive filtering noise cancellation. Based on effective denoising for GPS deformation monitoring data, this method can achieve effective recognition for single epoch GPS deformation information. Then, weak de- formation characteristics information can be extracted.