Application Study of Color Feature Extraction on Ferrographic Image Classifying and Particle Recognition

Color information of ferrographic image is important to particle identification and wear condition analysis. On the basis of studying on image classifying, this paper mainly studied color feature extraction of ferrographic image. The clustering tree, fuzzy clustering technique and statistic analysis were used in quantitative segmentation of background and wear particles of color ferrographic image. Calculated color features of background and different wear particles can be used for farther processing and recognizing of ferrographic image, and can also be used in particle machine automatic recognition and wear condition analysis.