摘要
压缩感知是一种基于信号稀疏性的信号采集与处理框架,能够在信号采集的同时对信号进行压缩.本文提出一种基于压缩感知的滚动轴承振动信号压缩方法,在信号的变换域内,使用幅值百分比作为阈值对变换系数进行稀疏处理,采用高斯随机矩阵对信号进行降维观测,实现数据的压缩.通过研究稀疏处理对信号压缩比和逼近误差的影响,分析阈值选择与信号稀疏比和逼近误差的关系,分析不同变换基对稀疏处理的影响.实验数据分析表明,相对于未经过稀疏处理的信号来说,该方法能有效地提高信号的压缩效果,且保持较好的逼近误差.
Compressed sensing is a signal acquisition and processing method based on signal sparsity.This paper presents a compression method of vibration signal of the rolling bearing based on compressed sensing.In transform domain,the amplitude as a percentage threshold for sparse treatment is used to process the transform coeffi-cients.The Gauss random matrix is used to measure signal and compress data.The relationship between sparse signal compression ratio and approximation error is studied.Different thresholds are then selected to examine the relationship between sparse signal ratio and the approximation error.The influence of different bases on sparsity is also analyzed.The results show that this method can improve the compression effect,and reduce the approxi-mation error.
出处
《昆明理工大学学报(自然科学版)》
CAS
2015年第4期46-50,共5页
Journal of Kunming University of Science and Technology(Natural Science)
基金
国家自然科学基金项目(51265018)
关键词
压缩感知
滚动轴承
稀疏表示
信号压缩
compressive sensing
rolling bearing
sparse representation
signal compression