Bearing fault diagnosis method based on morphological top hat product filtering of optimal scale

The invention discloses a bearing fault diagnosis method based on morphological top hat product filtering of an optimal scale. The diagnosis method comprises: mounting an acceleration sensor near thebearing to acquire bearing vibration signals and determine the initial range of the structural element scale; performing morphological top hat product filtering on the bearing vibration signals at each structural element scale to obtain morphological filtering results of a plurality of scales; calculating the third-order accumulation amount diagonal slice and the diagonal slice spectrum of the morphological filtering result of each scale to obtain third-order accumulative amount diagonal slices and diagonal slice spectra of a plurality of scales; calculating the fault characteristic ratio of the diagonal slice spectrum of each scale and determining a diagonal slice spectrum of the optimal scale according to the maximum fault characteristic ratio criterion; extracting the bearing fault feature information from the