Abstract:

In this thesis, aimed at eight kinds of metal debris existing in lubricating systems, by introducing the digital image process methods, the debris size, shape, texture, and color parameters are researched in detail. Finally, a debris' recognition method layer by layer is provided, and the fuzzy c-mean clustering recognition method is applied to recognize effectively the eight kinds of debris. Examples show that the method in this paper is simple and effective.

Key words: debris, parameter, pattern recognition, fuzzy clustering