基于CEEMD和小波包变换的重力数据信噪分离方法

摘要

为了有效地进行海洋重力测量数据的信噪分离，本文提出了基于互补总体经验模式分解(CEEMD)和小波包变换(WPT)的重力数据信噪分离方法。该方法利用CEEMD将海洋重力测量信号分解为从高频到低频的不同固有模式函数(IMF)分量以及趋势项，为进一步提取出各 IMF 分量中的有用重力信号，本文采用小波包变换对各 IMF 分量进行小波包分解降噪，最后将从各分量提取出的有用信号与趋势项进行信号重构，实现重力数据的信噪分离。本文通过仿真数据和实测数据对该方法进行了验证，结果表明本文提出的重力数据信噪分离方法能有效的抑制噪声干扰，保留有用的重力信号，实现较高精度的重力信号提取。%In order to effectively separate the noise from the measurement gravity data, a gravity data signal-noise separation method is proposed based on the theory of complementary ensemble empirical mode decomposition and wavelet packet transform. The measurement gravity data is decomposed by complementary ensemble empirical mode decomposition into different intrinsic mode functions (IMFs) from high frequency to low frequency with the trend item in this method. And each IMF is de-noised by wavelet packet transform for further extracting useful gravity data from IMFs. All the useful gravity data extracted from IMFs and the trend item are reconstructed for signal-noise separation. The method mentioned above is verified by the simulation data and the measured data. The results of the simulation analysis and the engineering experiments indicate that the signal-noise separation method based on the theory of com-plementary ensemble empirical mode decomposition and wavelet packet transform can effectively eliminate the noise of the measurement gravity and reserve the useful gravity signal, and the gravity data extracted by the signal-noise se-paration method proposed in this paper has higher precision.