稀疏分解在动态测试信号处理中的应用

摘要

In dynamic test,the sampling rate is high and the noise is strong,a sparse decomposition method based on Gabor dictionary was put forward. This method decomposed the signals iteratively with the matching pursuit algorithm and took the coherent ratio threshold as iteration terminate condition. Standard matching pursuit algorithm was time-consuming. This paper introduced an adaptive genetic algorithm introduced to matching pursuit method, which made computation speed improved effectively. Experimental results indicate that this method can effectively remove the high frequency noise, and can compress the signal.%针对动态测试过程采样率高,噪声大的情况,提出一种基于Gabor原子库稀疏分解的去噪压缩方法.该方法利用匹配追踪算法将信号在超完备Gabor原子库中迭代分解,并采用相干比阈值作为迭代终止条件,可以根据信号噪声水平自适应调整迭代次数.针对匹配追踪算法计算量大的缺点,引入一种自适应遗传算法加以改进,提高了计算效率.试验结果证明了该算法可以有效去除高频噪声,并且实现信号大幅压缩.