Abstract: The micro data of rock thin section can help researchers to accurately determine the rock structure, type, void evolution law, fracture development and control, etc. The micrograph of the thin section of rock reflects the characteristics of rock microstructure, grain size and mineral composition. Therefore, this paper takes olivine, augite, hornblende, biotite, plagioclase, andalusite, cordierite, cristobalite, pomegranate, actinolite and oolite as samples, and makes thin slices of rock through 8 steps of sample cutting, coarse grinding, fine grinding, coarse polishing, sticking piece, grinding piece, coarse polishing and fine polishing. The Leica polarizing microscope with dm2700p and 50 magnification was used for image acquisition under the light source environment of single polarized light and orthogonal polarized light. A total of 1446 color rock slice images were obtained, with a storage capacity of 2.15gb. This data set can provide a basic data source for the intelligent classification and recognition of rock thin sections and for geologists to study the genesis of rocks.
Keywords: rock slice; micrograph; image classification; intelligent recognition; rock forming mineral; image database