在大桥动态位移监测中,为了更好地滤除噪声,提出了一种改进的交叉证认EMD小波滤波方法。即先用EMD对信号进行分解,再用交叉证认方法自适应地算出噪声主导分量,最后用小波阈值对噪声主导分量进行滤波;其中阈值函数选取极为关键,给出了一种新的阈值函数计算方法;最后进行信号重构。实例分析表明,该方法能够更有效地滤除噪声,提取大桥振动信息,是一种高效的去噪方法。
In order to filter out noise, an improved EMD-wavelet filtering method based on cross-validation is proposed in the bridge dynamic displacement monitoring. At first the signal is decomposed with the EMD-wavelet method, and then we calculate the dominant component of the noise with cross-validation method adaptively, and finally the dominant component of noise is filtered with the wavelet threshold; It is very critical selecting the function of thresholds, and so this paper proposes a new threshold function calculation method; Finally we reconstruct the signal. Example analysis shows that the method can more effectively filter out noise and puts forward the bridge vibration information, and it is a highly efficient de-noising method.