基于谐波小波包变换的信号检测

Signal detection based on harmonic wavelet packet transform

利用二进离散小波变换检测周期性信号时,由于频带间的干扰,微弱成分易被强信号所淹没,为解决此问题,提出一种基于谐波小波包变换的周期性信号检测算法。利用谐波小波函数的盒形频谱结构,根据目标声信号的特点,选择合适的分解层数和分析带宽,由此避免分析频带之间的相互干扰。对实测数据的分析结果表明,谐波小波包变换对于低信噪比下的微弱线谱信号成分有一定的增强效果。

The weak signal can be submerged in the strong signal or ambient noise because of the inter-band interference when the dyadic wavelet is used to detect the line spectrum.To solve this problem,a harmonic wavelet packet transform is proposed to detect the line spectrum.With box-like spectrum of the harmonic wavelet function and according to the nature of the target-radiated noise,the analysed level and bandwidth can be appropriately chosen;therefore the inter-band interference can be reduced.Results show that the harmonic wavelet packet transform can enhance the weak line spectrum components under the low signal-to-noise ratio.