基于GMF-CLEAN的变形监测数据噪声抑制方法

GMF-CLEAN-Based Noise Reduction Method for Deformation Monitoring Data

利用GPS进行建筑物变形监测过程中,白噪声(随机误差)和色噪声(多径效益、测量粗差等)是主要的误差源,影响对观测数据中变形信息的提取。针对这一问题,提出一种基于广义匹配滤波(generalized matched filter,GMF)和CLEAN的噪声抑制方法,首先利用GMF对色噪声进行"白化"处理,将其转换为白噪声,然后利用CLEAN算法从"白化"后的数据中提取变形监测信息,从而实现对色噪声和白噪声的有效滤除。基于仿真数据和实际工程测试数据的实验结果表明,相对于传统的小波方法和经验模态分解(empirical mode composition,EMD)方法,所提方法可以获得更好的噪声抑制性能。

In the process of building deformation monitoring with GPS,white noise(random error)and color noise(multipath benefit,measurement gross error,etc.)are the main error sources,which affect the extraction of deformation information from observation data.To solve this problem,based on the traditional clean algorithm,a noise suppression method based on generalized matched filter(GMF)and clean is proposed.Firstly,the generalized matched filter algorithm is used to"whiten"the color noise and convert it to white noise.Then,the clean algorithm is used to extract the deformation monitoring information from the whitened data Now it can effectively filter out the color noise and white noise.Experimental results based on simulation data and actual engineering test data show that the proposed method can achieve better noise suppression performance than the traditional wavelet method and empirical mode decomposition(EMD)method.