Denoising Analysis of Different Data Domains Based on EEMD for Landslide Monitoring

Abstract

There are complex non-modeling errors and random noises that are difficult to effectively separate in the landslide monitoring data, and it is difficult to eliminate the influence by difference method. These noises exist in each satellite separately, and their integrated effects are expressed in the coordinate residuals. However, denoising only in a single data domain has the problem of residual noise. In this paper, the EMD and EEMD methods are used to denoise the measured landslide monitoring data in the double-difference observation domain, the coordinate domain and the integrated data domain of the two. The results show that compared with the EMD method, the EEMD method can effectively reduce the occurrence of modal aliasing, improve the automation level of data processing, and is more suitable for complex monitoring environments; using the EEMD method to simultaneously denoise in the double-difference observation domain and the coordinate domain, the root mean square error is slightly improved, and the standard deviation is, compared with the results of not denoising, increased by 12.3%, 46.9%, and 10.1% in the three directions of E, N, and U, respectively, and the denoising is increased by 8.8%, 9.5%, and 8.7%, respectively, compared with the single data domain. Therefore, using the EEMD method to synthesize the denoising methods in different data domain can effectively reduce the effects of random noise and instantaneous strong noise, and more effectively reflect the true landslide monitoring displacement changes.