论文部分内容阅读
介绍了一种新的信号去噪方法。针对激光雷达信号和 EMD 信号分解的特点,结合统计检测理论,利用中心极限定理和 Grubbs 判据对信号强度起伏的随机性进行检测。检测出的信号强突变部分视为信号的发展趋势保留不参与 EMD 分解,其余部分则使用 EMD 方法分解。对分段处理的信号用半软阈值限幅的办法重构;重构信号可能出现的 pseudo-Gibbs 现象,采用平移不变量去噪原理处理。实测信号的处理结果表明了这种去噪方法的有效性,即能取得较好的去噪效果,又能较好的保留信号的突变信息和强突变信息.
A new method of signal denoising is introduced. According to the characteristics of laser radar signal and EMD signal decomposition, combined with statistical detection theory, the central limit theorem and Grubbs criterion are used to detect the randomness of signal intensity fluctuation. The strong part of the detected signal is regarded as the development trend of the signal. The EMD method is not used for EMD decomposition. The rest is decomposed using the EMD method. The semi-soft threshold clipping method is used to reconstruct the signal of the segmentation process. The possible pseudo-Gibbs phenomenon of the reconstructed signal is treated by the translation invariant denoising principle. The measured signal processing results show the effectiveness of this denoising method, which can achieve better denoising effect, but also better retain the signal mutation information and strong mutation information.