The feature extraction and fuzzy clustering of wax crystal images for the rheological research of waxy crude oils

Wax-crystal morph is one of the prime inner factors that influence the rheological property of waxy crude at low temperature. The normal way to identify wax-crystals is comparing qualitatively the wax-crystals obtained from video-microscopy. The paper studies a quantitative characterization and classification of wax crystals of waxy crude oils. In this paper, a blocked 2D entropy threshold method was used to segment wax crystal images, and the grain size, shape characters of wax crystals and the continuous liquid phase characters are extracted through a statistic method. Finally, the fuzzy clustering method was employed to classify the wax crystal images.