The Study of Feature Extraction Method Based on Multi-scale Morphological Spectrum Entropy
Ge, H.M.
Chen, L.
Liang, J.
Download PDF

How to Cite

Ge H., Chen L., Liang J., 2016, The Study of Feature Extraction Method Based on Multi-scale Morphological Spectrum Entropy, Chemical Engineering Transactions, 51, 559-564.
Download PDF

Abstract

Multi-scale morphological decomposition is a decomposition method being proposed based on the basic theory of multi-scale morphology. Using from small to large scale of structural elements to deal with different “Scale domain”, thus the original signal is decomposed into a complete set of orthogonal In multi-scale mathematical morphology operation is the essence of it. This article combines multi-scale decomposition with the information entropy theory, put forward a kind of feature extraction method based on multi-scale morphological spectrum entropy, with multi-scale morphological spectrum entropy (multi-scale morphological spectrum entropy, multi-scale morphological singular spectrum entropy) as a feature vector can effectively distinguish between different running state of rolling bearings, and through the simulation and experiment validation.
Download PDF