针对轴承故障特征提取问题,提出一种自适应多尺度形态学方法。该方法采用形态闭、开相减构成的差值形态算子提取信号中的正、负冲击成分,基于信号的局部峰值间隔确定扁平结构元素的尺度,产生由若干不同尺度结构元素组成的集合,对信号进行自适应多尺度形态学分析。仿真实验结果表明,该方法能有效提取信号的冲击成分,且较单尺度形态学方法有更好的效果。将该方法应用于轴承故障信号处理,结果表明该方法对提取轴承故障特征频率有良好效果。 Aiming at the problem of bearing fault feature extraction, an adaptive multi-scale morphological method is proposed. In this method, positive and negative impact components in signal are extracted by morphological operator with open form and open form, and the scale of flat structure element is determined based on the local peak spacing of signal to generate a set composed of several structural elements of different scales. Signal adaptive multi-scale morphological analysis. Simulation results show that this method can effectively extract the impact of the signal components, and the single-scale morphological methods have better results. The method is applied to bearing fault signal processing. The results show that this method has a good effect on extracting bearing fault characteristic frequency.