为了去除激光干涉测速系统提取的小波脊的大量高频噪声，提高系统测量精度，采用一种基于经验模态分解的小波脊优化的方法，对高频噪声强度进行了很好的抑制，测量误差为原来未优化误差的42.3%，较大幅度提高了激光干涉测速系统的测量精度。结果表明,该方法对经过时频分析的信号去噪有重要的实用价值。

In order to suppress the high-frequency noise contained in the wavelet ridge obtained from the laser interference velocity measurement system and improve the system accuracy, a method based on empirical mode decomposition was proposed . Theoretical analysis and experimental results show that this method can significantly suppress the high-frequency noise. The measurement error is 42.3% of the original one before optimization. This method improves the system accuracy of the all-fiber displacement interferometer system for any reflector and it is applicable in signal denoising in time-frequency analysis.