论文部分内容阅读
本文叙述了一种可以从地球物理时间或空间序列数据中消除干扰噪声、随机噪声以及线性或非线性漂移的新型自适应滤波程序,为从“原始”信号中消除这些噪声、需要一个(以某种未知方式)与原始信号相关的“参考”信号。参考信号经过自适应滤波并被从原始信号中减去。由此产生的误差被用于最速下降法中以调整滤波器的权重,从而使得最小二乘意义上的均方误差极小。上述过程经过反复迭代,直至取得收敛。滤波器的输出即为原始输入中基本信号的最佳最小二乘估计,而不含噪声的影响。该方法的一个显著特点是它能够分辨一个随时间变化的信号而无需对其中信号和噪声的特征有任何先验的了解,同时也说明了当没有参考信号时,如何从序列数据中消除噪声。该程序已被应用于几种合成信号以表明它的特性;对于该程序中的数学部分做了简要的讨论。文中列出了程序的核心部分,并提出了一些在其它地球科学领域中的可能应用。
This paper describes a new adaptive filtering algorithm that can remove interference noise, random noise and linear or nonlinear drift from geophysical time or space series data. To eliminate these noises from the “original” signal, Unknown method) The “reference” signal associated with the original signal. The reference signal is adaptive filtered and subtracted from the original signal. The resulting error is used in the steepest descent method to adjust the weight of the filter so that the mean square error in the least-squares sense is minimized. The above process is iterative until convergence has been achieved. The output of the filter is the best least-squares estimate of the fundamental signal in the original input without the effects of noise. A notable feature of this method is that it can resolve a signal over time without any priori knowledge of the characteristics of the signal and noise and how to eliminate the noise from the sequence data when there is no reference signal. The program has been applied to several composite signals to demonstrate its characteristics; a brief discussion of the mathematical part of the program is made. The paper lists the core parts of the program and proposes some possible applications in other fields of geosciences.