Slice Image Particle Segmentation Technology Based on Multi Light Sources and Multi Angles

Rock particle segmentation is the basis of intelligent identification of slices, which can provide quantitative data for mineral composition identification, diagenesis evaluation and pore-throat extraction etc. The main difficulty of particle segmentation in a rock slice is that it contains more particles that are densely distributed, and the particles themselves have the characteristics of irregular shape, inconsistent size, rich texture and so on. In view of this difficulty, this paper proposes a semiautomatic rock slice particle segmentation method. The method uses both single polarized image and orthogonal polarized sequence images of a rock slice. Firstly, the HSV color space algorithm is used to extract the pores in the original single polarized image. Then K-means algorithm and SRM algorithm are used to extract particles from the orthogonal polarized sequence images, and the extracted particles are screened manually. Thirdly, OTSU algorithm is used to segment the selected particles combined with the original single polarized image. Finally, the graph cut algorithm is applied to segment the remaining particles. Experimental result shows that this method greatly improves the efficiency of image segmentation, and can ensure good segmentation results.