分数阶小波包时频域的信号去噪新方法.pdfVIP

学兔兔 第32卷 第7期 仪 器 仪 表 学 报 Vol_32 No．7 2011年7月 Chinese Journal of Scientific Instrument Ju1．20l1 分数阶小波包时频域的信号去噪新方法 黄雨青，王友仁，罗 慧，孔德明 (南京航空航天大学自动化学院 南京 210016) 摘 要：为了提高信号去噪的效果，提出了一种基于分数阶小波包变换(FRWPT)的信号去噪新方法。该方法根据输出信号信 噪比的大小，用迭代法寻找分数阶小波包变换的最优分数阶P值，通过分数阶小波包变换将带噪信号映射到最优分数阶小波包 时频域内，对变换后的信号进行窄带通滤波，最后通过分数阶小波包逆变换对信号进行重构，实现分数阶小波包时频域内的信 号去噪。以带噪Bumps信号和语音信号为例的去噪实验结果表明，采用该方法去噪后的信号信噪比明显提高，在抑制噪声的 同时可以有效保持细节信息。 关键词：信号去噪；分数阶小波包变换；小波变换；分数阶Fourier变换 中图分类号：TN206 文献标识码：A 国家标准学科分类代码：510．40 New signal de-noising method based on fractional wavelet packet transform in time-frequency domain Huang Yuqing，Wang Youren，Luo Hui，Kong Deming (College ofAutomation Engineering，Na ng University ofAeronautics and Astronautics，Nanjing 210016，China) Abstract：In order to improve the de—noising effect in signal de—noising，a novel signal de—noising method based on fractional wavelet packet transform(FRWPT)is proposed in this paper．The optimal fractional order of FRWPT is obtained with an iterative algorithm according to the SNR of output signals．Noisy signal is transformed into optimal factional wavelet packet time-frequency domain using an optimal FRWPT．Then．the transformed signal is filtered with a narrow band—pass filter．Finally．the signal is reconstructed with an inverse FRWPT．De—noising results of noisy Bumps signal and speech signal show that the SNR of the output signals can be effectively improved．The re- suits also show that the proposed method can preserve detail information effectively and reduce noise at the same time． Key words：s