Research Of Improved Watershed Algorithm And Grey Target In Ferrography Image Analysis

Wear is an inevitable phenomenon in the process of operation of machinery and equipment. If not found and repaired in time, it will continue to loss the material and the energy. With the deepening of the wear, it will lead to the failure of some parts of the machinery, endangering personal safety, and even the occurrence of a major accident. Therefore, it is of great social and economic benefits to establish the wear condition monitoring and improve the fault diagnosis technology. With the development of computer technology, the technology of the ferrography analysis is gradually developing to the intelligent development. As an important link, the automatic segmentation and analysis of the ferrography has important significance for the accurate and rapid diagnosis of mechanical equipment fault.In this paper, not only the principle of the ferrography technology, the emphasis and the existing problems are introduced, but also the calculation method of the characteristics of the image processing technique and the morphology of the wear particles are introduced; As ferrography image background color is relatively stable, Otsu's threshold algorithm is used to ferrography image background subtraction in 3O color space(IBS-3O), it can get the accurate removal of the background image. Besides, this paper is proposed to use the Dark Channel Prior Hazeremoval algorithm(DCPH) to improve the image quality.Secondly, the Improved Watershed Algorithm(IWA) is used to segment the ferrography image for large particles. In this algorithm: firstly, Otsu threshold algorithm is used to original image background subtraction, it can separate the abrasive particles and the background; Secondly, Improved Watershed Algorithm(IWA) is used for segmentation, increasing the threshold of the gray level and the four adjacent immersed in the immersion time of the watershed algorithm, the neighbor immersed in the immersion time is changed into the eight neighborhood, which is used to solve the problem of over segmentation of the abnormal large wear debris; the wear particle recognition based on the gray target theory(GTT) is proposed, which use the digital characteristics of particles' size and the concentration of abrasive particles and establish the analysis of ferrography standard model, quantifies abrasive particles concentration and particle size, etc. The quantitative parameters of grey target transformation and standard model, realize the concentration of the abrasive particles analysis through the bull 's-eye degrees.Finally, according to the analysis and design of content, Visual C++ 6 as the software platform and using OpenCV library to develop a processing and analysis system for ferrography image. it can realize the work that include preprocessing, background subtraction and edge detection, abrasive segmentation, abrasive analysis for ferrography image.