基于小波包变换的磁记忆信号特征值的提取

Extraction of Feature Value in Metal Magnetic Memory Testing Based on Wavelet Packet Transform

利用自行研制的信号检测系统采集航空铁磁材料试件的磁记忆信号。对信号进行小波包降噪处理,提高磁记忆信号的信噪比;利用Hilbert变换得到磁记忆信号的包络线,再对包络求取梯度特征值。实验结果证明,小波包降噪效果明显,对包络求梯度特征值比直接对信号求取梯度更准确清楚,可望应用于实际检测。

A self-made magnetic memory test system was used to collect magnetic signals of aeronautic ferromagnetic components.Wavelet packet transform was utilized to reduce noise signals,and meanwhile,Hilbert transform was used to get the envelope curve of magnetic signals.According to experiment results,the reduction of noise was improved,and gradient eigenvalues were more exactly picked.The novel method is expected to have practical application worth.