The Denoising and Trend Extraction Based on the Fusion Method of Wavelet and Emd

Both wavelet algorithms and empirical mode decomposition (EMD) signal analysis method have strong power discriminate the signals from the noise, but the precision of denoise is affected by the end errors of EMD, the effect is more serious especially when the signal carries the information of the trend development. Based on the capability of orthogonal wavelet transform in de-noise, trend extract function of EMD and the characteristics of GPS errors, a new noise filter and trend extraction model is built up. Then, simulated data and real data (i.e., GPS data) are used to test the method separately. The following conclusions are drawn from these tests: (1) Orthogonal wavelet transform and EMD method can better mitigate the random errors which hide in periodic signal; (2) After significantly mitigating the influence of multi-path and others errors by the new model, the accuracy of vertical component position for GPS dynamic deformation can reach the mm level.