基于小波抗混叠单子带重构改进算法的测量信号特征提取

On Feature Extraction of Measurement Signal Based on Improved Anti-aliasing Single-Band Reconstruction Algorithm

针对Mallat算法固有的频率混淆特性导致单子带重构算法提取出错误特征频率信息的问题,提出了一种能够消除频率混淆和边界效应的小波抗混叠单子带重构算法。该算法在信号的分解和重构过程中使用不同的纠正滤波器对各子带信号进行处理,同时在每一次卷积后加入去暂态过程点的步骤,可以克服Mallat算法和单子带重构算法中的频率混淆,保证了重构信号与原始信号的长度一致。对仿真信号的计算结果证明了该算法的有效性,为测量信号的特征信息提取分析提供了一种有效手段。

The frequency aliasing which rooted in the Mallat algorithm would lead to a wrong extraction of characteristic frequency in the single sub-band reconstruction algorithm, a completely anti-aliasing single sub-band reconstruction algorithm which could eliminate frequency aliasing and boundary effects is proposed in this paper. The algorithm uses different correcting filters to process each sub-band signal in the process of signal decomposition and reconstruction, at the same time, the steps of removing the transient process points are added after each convolution, which can overcome the frequency aliasing in the Mallat algorithm and the reconstruction algorithm, and the consistency of length between the reconstructed signal and the original signal is guaranteed. The simulation results show the effectiveness of the proposed algorithm, which provides an effective method for the feature information extraction and analysis of monitoring and measurement signals.