Ferrography Abrasive Wear Color Image Recognition Technology Research

With the rapid development of modern industry, the ferrography analysis technology of oil analysis has been widely applied because of its higher detection efficiency and better early diagnosis of the effect of wear and tear failures. The wear particles which is separated from the lubricating oil is the object on researching and analyzing of ferrography analysis technique. The ferrography analysis technology is very effective to the wear analysis, but not everyone can identify the nature of abrasive from the ferrography images, and only the experienced analysts can understand and identify the mechanical wear information carried by ferrography images. Undoubtedly, the research on the analysis theory and techniques of the colorful ferrography images can provide a better guarantee on technical support to the wider use of ferrography technology. This paper is mainly based on how to effectively identify the types of abrasive from ferrography images in the colorful ferrography images analyzing process.Mainly based on the analysis of mechanical wear and the available research of ferrography image processing, and consulting some methods which have got a lot of achievements on wavelet transform, fractal dimension, BP network,ect., this paper will proposes some recommendations and methods on the colorful ferrography image processing, distinguishes six types of abrasive, which seperately are the iron (steel) abrasive; the copper alloy abrasive, the aluminum alloy abrasive and the layered abrasive; spherical abrasive; cutting abrasive; severely sliding wear abrasive of iron-based metal, and exploits the image software on the wear abrasive identification by using VC + +6.0. The whole paper concludes six parts. The first part is the introduction, which outlined the background, the purpose, the significance of this study, and briefly introduced the methods, the content and the structure of this paper. The second part is the basic theories and basic methods, which expounds the ferrography analysis technology and are adopted on the ferrography image processing. The third part is the research of ferrography image recognition technology, which firstly elaborates the characteristics of the internal combustion engine abrasive in colorful ferrography images and the following four aspects of the job we do in ferrography image processing-------image enhancement, image segmentation, feature extraction and abrasive identification, then, highlighted several key methods in the ferrografic image processing, such as on the wavelet enhancement and the multi-threshold segmentation based on wavelet transformation, and finally proposed some improving suggestions by comparing with the general ferrography image processing methods. The fourth part is the design of colorful ferrography image recognition system, which sets forth the structure and the functions of various components of modules of the colorful ferrography image recognition system. The fifth part is the application in the standard 4100QB internal combustion engine, which describes the application on analyzing its reliability of the 400h and 600h oil sample under the colorful ferrography image recognition system, and meantime, verifies the practicality of the colorful ferrography image recognition system which is developed in this paper. The last part is the generalization and prospects, which sums up this paper and points out the direction for further research.