新阈值二进小波去噪算法在齿轮信号中的应用

Application of new threshold dyadic wavelet de-noising algorithm on gear signal

在齿轮的激光检测过程中,由于各种噪声源的影响,原始信号会出现特征信息弱化的现象,影响了齿轮检测的准确性。为解决该问题,在分析传统软、硬阈值函数不足的基础上,提出了一种新阈值函数和二进小波变换相结合的去噪方法。在该方法中构建了一个可以调节的新阈值函数,该阈值随二进小波分解系数的变化以指数形式动态调整,克服了传统硬阈值函数不连续和传统软阈值函数存在偏差的问题。经仿真实验验证,该方法的信噪比和适用性具有明显的改善。

In the process of gear laser detection, due to the influence of various noise sources, the phenomenon of the orig-inal signal weakening of feature information will appear, which impacts the accurateness of gear detection. To solve this problem, a new de-noising method of combining a new threshold function and dyadic wavelet transform is proposed based on the analysis of the traditional soft and hard threshold function. In this method, adjustable threshold function is built. The threshold adjusts dynamically in the form of index, with the change of binary wavelet coefficients, overcoming the discontinuity of the hard threshold and the offset of the soft threshold. By simulation experiments, the signal to noise ratio and applicability of the method have significantly improved.