基于小波包变换的信号谱峰检测算法

提出了一种新的基于小波包变换的信号谱峰检测算法,主要思想是利用小波包变换的特点,对信号功率谱进行平滑处理,突出谱峰的特征点（起点、顶点和终点）,然后对其进行三层小波包变换,提取相应细节系数的特征点来估计谱峰的起点、顶点和终点,从而完成谱峰的检测。该方法的特点是无需信号的任何先验信息,是一种盲处理算法。仿真结果表明,信噪比不低于5dB的情况下,信号特征点检测的归一化均方误差（NMME）低于6‰,其性能比传统基于差分的方法有明显的优势。

Signal Peak Identification Using Wavelet Packet Transform

This paper presents a novel method for peak identification of communication signal. Firstly, the spectrum should be smoothed to highlight the feature point （start point, top point and end point）, then by using wavelet packet transform （WPT）, the feature points of peak on frequency domain is detected efficiently. The advantage of this method is that the signal is processed without knowing any information in advance. The simulation results indicate that the normalized mean estimation error （NMME） of feature point position is less than 6‰ when signal-noise-ratio （SNR） is higher than 5 dB. Furthermore. The performance of this method is superior to that of conventional methods based on difference.