Automatic recognition method for rock particles in sandstone microsection

Abstract

The invention discloses an automatic recognition method for rock particles in a sandstone microsection, and the method comprises the steps: 1), carrying out the smoothing filtering preprocessing of the inputted sandstone microsection, extracting the pixel color and texture features, and forming a characteristic vector of pixel similarity measure; 2), searching an adjacent pixel group of each pixel through an image traversal algorithm and a similarity measurement method; 3), calculating the statistical characteristics of the adjacent pixel group, and enabling the statistical characteristics to serve as the characteristic vector of the pixels; 4), training a logic regression classifier based on a rock particle sample data set; 5), predicting the probability that each pixel belongs to quartz, feldspar or rock debris, and determining a noise point; 6), carrying out the processing of the noise point, and determining the dividing and classification of the rock particles. The method employs the image processing technology and a machine learning method, automatically recognizes the rock particles in the sandstone microsection, can solve the positioning of the boundary of a rock particle region and the recognition of foreign matters in the rock particles, is higher in accuracy, and reduces the recognition time and cost of rock particles. The method has application values for mineral discrimination and oil-gas exploration.

Description

technical field [0001] The present invention relates to the technology of image processing by application of computing methods, specifically an automatic identification method for rock particles in sandstone micro-thin slices. The method uses image processing technology and machine learning methods to realize Identification of rock grain composition in microscopic sandstone sections. Background technique [0002] Sandstone is a sedimentary clastic rock accumulated in the basin after weathering, denudation and transportation of rocks in the source area. It is composed of clastics and interstitials. The main rock particle components include quartz, feldspar and cuttings. Sandstone is not only a common building stone, but also the main reservoir of oil and natural gas. Sandstone particle identification and composition are of great significance in mineral identification, oil and gas exploration and other fields. [0003] The existing sandstone micro-thin section rock particle ...

AI Technical Summary

Problems solved by technology

[0003] The existing sandstone microscopic thin-section rock particle identification often uses manual identification, which is not only time-consuming and laborious, but also has poor repeatability

In addition, in sandstone micro-thin slices, the boundaries of some rock grains are staggered, and some areas infiltrate with the surrounding areas, which makes it difficult to distinguish and identify rock grains.

Finally, there are often impurities in rock particles, and the properties of these impurities are quite different from those in the surrounding area, which may affect the identification results of rock particles

Images