Rock core FIB-SEM image segmentation method based on convolutional neural network

Abstract

The invention discloses a rock core FIB-SEM image segmentation method based on a convolutional neural network, and mainly relates to an image segmentation technology of rock core sequence images. The method comprises the following steps: (1) establishing a core FIB-SEM image data set; (2) constructing a convolutional neural network: embedding a channel attention module into a coding stage, extracting multi-scale features by using an improved feature pyramid attention module, extracting a fine boundary by using multi-scale space attention in a decoding module, and recovering an original resolution by using a sub-pixel convolution module in an up-sampling stage; (3) performing network training and parameter optimization to obtain a model with the best effect; (4) performing a network segmentation result test by using the test set obtained in the step (1); the FIB-SEM pores of the rock core are extracted by using the convolutional neural network, manual operation is not needed, and the segmentation precision is improved.

Description

technical field [0001] The invention relates to an image segmentation technology of a core FIB-SEM image, in particular to a method for segmenting a core FIB-SEM image by a convolutional neural network. Background technique [0002] In recent years, due to the continuous decline in the world's conventional oil and gas energy production, unconventional oil and gas resources with huge reserves have attracted the attention of countries all over the world. Compared with conventional oil and gas resources, unconventional oil and gas resources are stored in smaller-scale spaces, and micro-nano pores are their important storage spaces. Important content of oil and gas energy research. Focused Ion Beam-Scanning Electron Microscope (FIB-SEM, Focused Ion Beam-Scanning Electron Microscope) is a new method to study the structure of unconventional oil and gas reservoirs. Extracting the pore structure in the core FIB-SEM sequence image has important guiding significance for the developm...

AI Technical Summary

Problems solved by technology

However, there are the following difficulties in extracting pores from core FIB-SEM images: (1) Since the observation surface of FIB-SEM imaging is not perpendicular to the electron beam, the signal at the bottom is weaker than that at the upper part, so the scanning image will become darker from top to bottom (2) In addition to pores, other structures in the rock core, such as organic matter, clay minerals, etc., will also form edges with rocks in the FIB-SEM image. It will cause serious interference; (3) SEM imaging will show its internal details, coupled with the effect of charging, the interior of the pores is generally accompanied by bright features, which increases the difficulty of identifying pores

[0003] At present, there are few image segmentation algorithms for core FIB-SEM images

In 2012, some scholars proposed a two-stage method for extracting FIB-SEM images of porous materials, which uses the highlight effect of pores in FIB-SEM images to extract highlight areas, and then backpropagates, but this is for porous materials. The segmentation method requires that there will be no obvious difference in gray level in other parts of the pores, and it is impossible to extract pores from cores that contain various impurities.

In 2016, scholars conducted experiments on edge detection segmentation method, watershed segmentation method and manual or automatic threshold segmentation method for shale FIB-SEM images, but the effect depends on human adjustment, and the pore extraction effect in the sequence diagram is not good.

In 2018, a core FIB-SEM sequence image pore extraction algorithm based on active contours was proposed, but manual pre-selection of marker points is required, and it is difficult to accurately extract fine pore edges for sequence images with large changes

Images