In the field of deformation monitoring ,the traditional wavelet de-noising method retains only the low frequency of useful information .It is easy to get rid of intermediate frequency and high frequency useful information .The wavelet packet analysis is a new kind of wavelet analysis method developed in recent years ,which is a more subtle de-noising method for considering useful information of the various bands .The key of the wavelet packet de-noising is to select the appropriate threshold criteria and to process the wavelet packet decomposition coefficients by the threshold ,but the researches using traditional wavelet packet de-noising method are not sufficient .This article is for the lack of traditional wavelet and wavelet packet analysis . According to the distribution of different signals and their noise ,wavelet packet decomposition coefficients are arranged by the frequency order ,and segmented in accordance with information type ,to select the appropriate threshold criteria for each band and to perform threshold processing .It is the method of wavelet packet de-noising with multi-threshold criteria based on frequency order .The results show that this method can effectively remove the noise of each band through theoretical analysis and practical applications .The de-noising ability of this method is better than the other methods such as traditional wavelet de-noising or wavelet packet de-noising . Studies have shown that this method can preserve the useful information from the de-noising signal after de-noising when the sampling frequency is low .Therefore ,it can be widely used in the field of high-precision deformation monitoring .%在形变分析中，传统的小波去噪只保留低频上的有用信息，很容易去掉中频以及高频上的有用信息。小波包分析方法则同时考虑了各个频段上的有用信息，因此是一种更为精细的去噪方法，小波包去噪的关键是对小波包分解系数选取合适的阈值准则并进行阈值处理，但传统的小波包去噪并没有对此进行充分的研究。针对传统小波、小波包分析的不足，本文提出一种基于频率顺序并依据信息类型分段的多阈值准则小波包去噪法。通过理论分析与实际应用，结果表明该方法能够高效剔除各频段的噪声，同时当采样频率较低时能有效保留去噪信号中频率较高的有用信息，其去噪能力优于传统的小波、小波包等其他去噪方法，因此可以广泛应用于高精度变形监测领域中。
展开▼
