摘要
针对高噪声背景下含冲击成分的机械信号的统计特点,提出一种基于盲解卷积的信号特征提取的方法。利用峭度和偏斜度对冲击敏感的特性,应用信号的3阶和4阶矩优化解卷积滤波器,结合小波软阈值消噪方法,提取机械信号中的弱冲击成分。研究表明,此方法能对机械信号中所含的弱冲击成分进行良好的提取。
According to the quality of high-order statistics of the impact, a new method is presented to extract the feature of weak impact component hidden in the mechanical signal. The proposed method is based on blind deconvolution and wavelet soft-thresholding denoising method. Practically, using the presented method, the characteristics of weak impact component can be extracted well.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2002年第6期152-155,共4页
Journal of Mechanical Engineering
关键词
冲击
盲解卷积
阈值消噪
特征提取
机械信号
Impact Soft-thresholding denoisingBlind deconvolution Feature extraction
