Images Of Sandstone Research Articles

A 3D convolutional neural network model with multiple outputs for simultaneously estimating the reactive transport parameters of sandstone from its CT images

Porosity, tortuosity, specific surface area (SSA), and permeability are four key parameters of reactive transport modeling in sandstone, which are important for understanding solute transport and geochemical reaction processes in sandstone aquifers. These four parameters reflect the characteristics of pore structure of sandstone from different perspectives, and the traditional empirical formulas cannot make accurate predictions of them due to their complexity and heterogeneity. In this paper, eleven types of sandstone CT images were firstly segmented into numerous subsample images, the porosity, tortuosity, SSA, and permeability of the subsamples were calculated, and the dataset was established. The 3D convolutional neural network (CNN) models were subsequently established and trained to predict the key reactive transport parameters based on subsample CT images of sandstones. The results demonstrated that the 3D CNN model with multiple outputs exhibited excellent prediction ability for the four parameters compared to the traditional empirical formulas. In particular, for the prediction of tortuosity and permeability, the 3D CNN model with multiple outputs even showed slightly better prediction ability than its single-output variant model. Additionally, it demonstrated good generalization performance on sandstone CT images not included in the training dataset. The study showed that the 3D CNN model with multiple outputs has the advantages of simplifying operation and saving computational resources, which has the prospect of popularization and application.

Innovative Deep Learning Approaches for High-Precision Segmentation and Characterization of Sandstone Pore Structures in Reservoirs

The detailed characterization of the pore structure in sandstone is pivotal for the assessment of reservoir properties and the efficiency of oil and gas exploration. Traditional fully supervised learning algorithms are limited in performance enhancement and require a substantial amount of accurately annotated data, which can be challenging to obtain. To address this, we introduce a semi-supervised framework with a U-Net backbone network. Our dataset was curated from 295 two-dimensional CT grayscale images, selected at intervals from nine 4 mm sandstone core samples. To augment the dataset, we employed StyleGAN2-ADA to generate a large number of images with a style akin to real sandstone images. This approach allowed us to generate pseudo-labels through semi-supervised learning, with only a small subset of the data being annotated. The accuracy of these pseudo-labels was validated using ensemble learning methods. The experimental results demonstrated a pixel accuracy of 0.9993, with a pore volume discrepancy of just 0.0035 compared to the actual annotated data. Furthermore, by reconstructing the three-dimensional pore structure of the sandstone, we have shown that the synthetic three-dimensional pores can effectively approximate the throat length distribution of the real sandstone pores and exhibit high precision in simulating throat shapes.

Machine Learning Assisted Prediction of Porosity and Related Properties Using Digital Rock Images.

Accurately estimating reservoir rock properties is paramount for modeling the storage and flow of fluids (hydrocarbon, carbon dioxide, and groundwater) in porous media. However, existing laboratory techniques to measure rock properties are usually time-consuming, expensive, and computationally intensive. This work proposes an efficient workflow that uses the machine learning algorithm, based on the convolutional neural network (CNN) framework, to predict rock properties from microcomputed tomography (micro-CT) X-ray images. The workflow involves data preprocessing, label extraction, training, and prediction using the segmented images of the rock to predict porosity, throat area, and pore surface area, which are essential for pore-scale modeling. The model was trained and validated on the Bentheimer sandstone, which was then used to predict properties of other sandstones (Castlegate and Leopard) with different pore structures and flow properties. The model yielded a good prediction for the throat and pore surface area but a significant error for porosity. Subsequently, a new complex model was trained and validated using diverse images from Bentheimer and an additional rock Castlegate, which was then used to predict the properties of Leopard sandstone. The new model improved the prediction of each property, resulting in mean absolute percentage error (MAPE) values of 2.19%, 3.04%, and 6.08% for porosity, pore surface area, and throat area with the binary images, respectively. In addition, we present a novel data-driven method using a simple regression model to predict the absolute permeability of a digital rock sample using the pore network parameters as predictors. The extreme gradient boost (XGBoost), which performed the best among several machine algorithms, was trained and validated using digital rock images from Bentheimer and Castlegate sandstone. The generated model was then used to predict the absolute permeability of the Leopard sandstone with an R 2 of 0.813, which was a significant improvement over the model generated solely by using either the Bentheimer or the Castlegate sandstone images. Furthermore, our analysis showed that the tortuosity had the most significant effect on the absolute permeability prediction of the rock sample. This study showed that we can reliably predict the morphological properties of porous media using computationally efficient models generated from digital rock images, which can be used to build a regression model to predict the crucial petrophysical properties needed to model the flow of fluids in porous media.

Generation of pore-space images using improved pyramid Wasserstein generative adversarial networks

High-resolution three-dimensional X-ray microscopy can be used to image the pore space of materials. Machine learning algorithms can generate a statistical ensemble of representative images of arbitrary sizes for rock characterization, modeling, and analysis. However, current methods struggle to capture features at different spatial scales observed in many complex rocks which have a wide range of pore size. We use the Improved Pyramid Wasserstein Generative Adversarial Network (IPWGAN) to automatically reproduce multi-scale features in segmented three-dimensional images of porous materials, enabling the reliable generation of large-scale representations of complex porous media. A Laplacian pyramid generator is introduced, which creates pore-space features across a hierarchy of spatial scales. Feature statistics mixing regularization enhances the discriminator’s ability to distinguish between real and generated images by mixing their feature statistics, thereby indirectly enhancing the generator’s ability to capture and reproduce multi-scale pore-space features, leading to increased diversity and realism in the generated images. The method has been tested on five sandstone and carbonate samples. The generated images, which can be of any size – including cm-scale ten-billion-cell images – demonstrate the power of the approach. These images have two-point correlation functions, porosity, permeability, Euler characteristic, curvature, and specific surface area closer to those of the training datasets than existing machine learning techniques. The generated images accurately capture geometric and flow properties, demonstrating a considerable improvement over previously published studies using generative adversarial networks. For instance, the mean relative error in the calculated absolute permeability between the Berea sandstone images generated by IPWGAN and the corresponding real rock images can be reduced by 79%. The work allows representative models of a wide range of porous media to be generated, offering potential benefits in carbon dioxide sequestration, underground hydrogen storage, and enhanced oil recovery.

Controlled ion transport in the subsurface: A coupled advection–diffusion–electromigration system

Ion transport within saturated porous media is an intricate process in which efficient ion delivery is desired in many engineering problems. However, controlling the behavior of ion transport proves challenging, as ion transport is influenced by a variety of driving mechanisms, which requires a systematic understanding. Herein, we study a coupled advection–diffusion–electromigration system for controlled ion transport within porous media using the scaling analysis. Using the Lattice–Boltzmann–Poisson method, we establish a transport regime classification based on an Advection Diffusion Index (ADI) and a novel Electrodiffusivity Index (EDI) for a two-dimensional (2D) microchannel model under various electric potentials, pressure gradients, and concentration conditions. The resulting transport regimes can be well controlled by changing the applied electric potential, the pressure field, and the injected ions concentration. Furthermore, we conduct numerical simulations in a synthetic 2D porous media and an x-ray microcomputed tomography sandstone image to validate the prevailing transport regime. The simulation results highlight that the defined transport regime observed in our simple micromodel domain is also observed in the synthetic two- and three-dimensional domains, but the boundary between each transport regime differs depending on the variation of the pore size within a given domain. Consequently, the proposed ADI and EDI emerge as dimensionless indicators for controlled ion transport. Overall, our proof-of-concept for ion transport control in porous media is demonstrated under advection–diffusion–electromigration transport, demonstrating the richness of transport regimes that can develop and provide future research directions for subsurface engineering applications.

High-Accuracy Image Segmentation Based on Hybrid Attention Mechanism for Sandstone Analysis

Mineral image segmentation based on computer vision is vital to realize automatic mineral analysis. However, current image segmentation methods still cannot effectively solve the problem of sandstone grains that are adjoined and concealed by leaching processes, and the segmentation performance of small and irregular grains still needs to be improved. This investigation explores and designs a Mask R-CNN-based sandstone image segmentation model, including a hybrid attention mechanism, loss function construction, and receptive field enlargement. Simultaneously, we propose a high-quality sandstone dataset with abundant labels named SMISD to facilitate comprehensive training of the model. The experimental results show that the proposed segmentation model has excellent segmentation performance, effectively solving adhesion and overlap between adjacent grains without affecting the classification accuracy. The model has comparable performance to other models on the COCO dataset, and performs better on SMISD than others.

Permeability prediction of complex porous materials by conjugating generative adversarial and convolutional neural networks

Empirical formulas, established by experimental statistical methods, have been a traditional method for predicting permeability in porous materials, but they suffer from shortcomings in terms of prediction accuracy. In this paper, a framework conjugating generative adversarial nets (GAN) and convolutional neural networks (CNN) is proposed for the rapid prediction of permeability based on 3D Berea sandstone images. This framework aims to surpass the limitations of empirical formulas in achieving accurate predictions. The framework involves three steps: porous material images reconstruction based on a conditional porosity progressive growth of generative adversarial nets (CPPGGAN), permeability label calculation based on the Lattice Boltzmann method (LBM) and deep learning prediction model construction using convolutional neural network (CNN). The CNN predictor trained on 3920 CPPGGAN synthetic images achieved an R2 value of 0.927 and 0.925 on 560 synthetic images and 60 original images, respectively. This indicated the effectiveness of the CPPGGAN synthetic images in the permeability prediction framework. The framework can also be adapted to a wider range of porous material types and shows great potential for exploring materials with exceptional properties in nature.

A multiangle polarised imaging-based method for thin sectionsegmentation.

The most crucial task of petroleum geology is to explore oil and gas reservoirs in the deep underground. As one of the analysis techniques in petroleum geological research, rock thin sectionidentification method includes particle segmentation, which is one of the key steps. A conventional sandstone thin sectionimage typically contains hundreds of mineral particles with blurred boundaries and complex microstructures inside the particles. Moreover, the complex lithology and low porosity of tight sandstone make traditional image segmentation methods unsuitable for solving the complex thin sectionsegmentation problems. This paper combines petrology and image processing technologies. First, polarised sequence images are aligned, and then the images are transformed to the HSV colour space to extract pores. Second, particles are extracted according to their extinction characteristics. Last, a concavity and corner detection matching method is used to process the extracted particles, thereby completing the segmentation of sandstone thin sectionimages. The experimental results show that our proposed method can more accurately fit the boundaries of mineral particles in sandstone images than existing image segmentation methods. Additionally, when applied in actual production scenarios, our method exhibits excellent performance, greatly improving thin sectionidentification efficiency and significantly assisting experts inidentification.

High-precision algorithm for grain segmentation of thin sections by multi-angle optical-microscopic images

ABSTRACT This paper introduces an algorithm for automating the analysis of petrographic thin-section images of sandstones and siltstones. The images of thin sections are obtained in polarized light at magnifications providing good image quality. In addition, the images for each section are obtained at different angles of rotation of the microscope stage. Augmentation is applied to the obtained photographs: the number of images increases due to rotations, shifts, and rescaling of the image. For training the neural network of the Mask R-CNN architecture, transfer learning is used, with initial weights obtained from a huge variety of nongeological images. The results of image segmentation using Mask R-CNN are compared to the Watershed algorithm results and the U-Net network for two metrics. According to the standard Intersection over Union metric, U-Net for high-quality images and Watershed for blurry images show the best results with a slight superiority. However, according to the Grain Size Metric, which evaluates the accuracy of grain-size measurement, the best accuracy (over 95%) is shown by Mask R-CNN. The grain-size analysis is done, and the porosity of the studied petrographic sections is determined. The use of the proposed approaches in the study of thin sections will significantly reduce the time for obtaining the results of grain-size-distribution analysis and porosity determination. This article is the result of multidisciplinary collaboration between geologists and programmers. This has allowed for the merging of profound knowledge in the field of geology with cutting-edge data processing technologies. By employing the presented methodology, geologists can devote more time to interpreting results rather than obtaining them, which in turn enhances the efficiency of research work. The benefits of using this methodology are not limited to just speeding up the process: it also allows for increased accuracy and reliability of the analysis, minimizing human error.

Multiscale fusion of tight sandstone digital rocks using attention-guided generative adversarial network

Tight sandstone exhibits heterogeneity due to its multiple mineral compositions and multiscale pore structures, presenting significant challenges for digital rock modeling. To capture the multiscale information inherent in heterogeneous rocks, we utilized an attention-guided generative adversarial network (AttentionGAN) to combine X-ray micro-computed tomography (micro-CT) and scanning electron microscope (SEM) images of tight sandstone, thereby producing large-scale, high-precision rock images. The reconstructed digital rock images comprehensively showcase the multiscale pore structures, including intergranular and clay micropores, as well as the distribution of minerals. Mineral content, two-point correlation functions, representative volume elements, and pore radius distributions were calculated for both the real and reconstructed images to confirm the efficacy of the proposed AttentionGAN. The results demonstrate that AttentionGAN is capable of generating high-precision rock images with a wide field of view. This offers a beneficial methodology for characterizing heterogeneous rocks and numerically simulating rock physics processes.

Microstructural and thermal properties of coal measure sandstone subjected to high temperatures

To study the microscopic structure, thermal and mechanical properties of sandstones under the influence of temperature, coal measure sandstones from Southwest China are adopted as the research object to carry out high-temperature tests at 25 °C–1000 °C. The microscopic images of sandstone after thermal treatment are obtained by means of polarizing microscopy and scanning electron microscopy (SEM). Based on thermogravimetric (TG) analysis and differential scanning calorimetric (DSC) analysis, the model function of coal measure sandstone is explored through thermal analysis kinetics (TAK) theory, and the kinetic parameters of thermal decomposition and the thermal decomposition reaction rate of rock are studied. Through the uniaxial compression experiments, the stress‒strain curves and strength characteristics of sandstone under the influence of temperature are obtained. The results show that the temperature has a significant effect on the microstructure, mineral composition and mechanical properties of sandstone. In particular, when the temperature exceeds 400 °C, the thermal fracture phenomenon of rock is obvious, the activity of activated molecules is significantly enhanced, and the kinetic phenomenon of the thermal decomposition reaction of rock appears rapidly. The mechanical properties of rock are weakened under the influence of rock thermal fracture and mineral thermal decomposition. These research results can provide a reference for the analysis of surrounding rock stability and the control of disasters caused by thermal damage in areas such as underground coal gasification (UCG) channels and rock masses subjected to mine fires.

Identification of quartz cement in sandstone through deep learning segmentation of electron microscopy images

Quartz cementation is a crucial factor in controlling the petrophysical properties of sandstone reservoirs. However, reliable identification of sandstone cementation often requires petrographic analysis of rock thin sections using a combination of optical light microscopy, backscattered electron (BSE), and cathodoluminescence (CL) images. This process is currently carried out manually and is both costly and time consuming. In this study, we present the first attempt to automate this process by identifying sandstone cement through convolutional neural networks (CNNs). We used a combination of BSE and CL images acquired from sandstone thin sections sourced from formations in the US, Israel, and the Netherlands. For each image pair we created a labeled mask with 4 classes: (i) quartz grains; (ii) quartz cement; (iii) porosity; and (iv) Other phases. We developed a U-Net with a total of 10 layers: 5 layers for the contracting path and 5 layers for the expansive path. The model is trained to predict the segmentation mask for a given input of paired BSE and CL images. A high level of accuracy is achieved for quartz grains (91%), quartz cement (78%), and porosity (95%). By contrast, the Other phase class was poorly predicted due to a combination of mineral heterogeneity and low representation. Our results indicate that Deep Learning algorithms provide a promising avenue for the automation of quartz cement detection in sandstone images.

Extracting crucial microstructures to characterize the elastic wave velocity and resistivity of Berea sandstone using convolutional neural networks

To reveal the internal structure of the earth, rock physical models are insightful for interpreting geophysical observation data and are often used for estimating subsurface structures. However, as the estimated value of each model significantly depends on the assumed microstructure, the obtained subsurface structure may be biased. Despite the innovations of high-resolution imaging technologies and their application to digital rock physics (DRP) and machine learning (ML), the characteristic microstructure controlling rock physical properties remains unclear. We apply ML and extraction of regression activation mapping (RAM) and learned filters to visualize the characteristic microstructure underlying the elastic and electrical properties of rock (P-wave velocity [Formula: see text], S-wave velocity [Formula: see text], [Formula: see text]/[Formula: see text] ratio, and electrical resistivity). This regression task is implemented using convolutional neural networks (CNNs) that are trained on these properties and calculated with the DRP approach using Berea sandstone images as samples. Although the calculated geophysical properties show perturbations even among digital rocks that have the same porosity, the results predicted using the best-performing CNN model exhibit a relative error of 2%–7% with respect to elastic properties. The visualized RAM revealed that the machine focuses on larger grains for predicting P-wave and S-wave velocities, whereas their ratio ([Formula: see text]/[Formula: see text]) takes into account the pore space. In contrast, electrical properties consider grain edges and surrounding pores, which may be related to the tortuosity of the pore space. Moreover, filter visualizations demonstrated that the geometric characteristics of Berea sandstone could be representative of circular or elliptical shapes for determining elastic properties, whereas tube-like shapes were extracted for electrical resistivity. These findings shed new light on the crucial internal structures used for defining each rock property.

Pore‐Scale Modeling of Coupled CO2 Flow and Dissolution in 3D Porous Media for Geological Carbon Storage

AbstractDissolution trapping is one of the crucial trapping mechanisms for geological carbon storage in deep saline aquifers. The injected supercritical CO2 (scCO2) flow and dissolution processes are coupled and interact with each other. Therefore, we performed direct numerical simulations in three‐dimensional micro‐CT images of sandstones using the volume of fluid and continuous species transfer method. We investigated the coupled scCO2 flow and dissolution processes at pore‐scale under different rock structures, capillary numbers, and rock wettability conditions. The dynamic evolution of the scCO2/brine phase distribution and scCO2 concentration distribution occurring during the injection period were presented and analyzed. Complicated coupling mechanisms between scCO2‐brine two‐phase flow and interphase mass transfer were also revealed. Our results showed that the scCO2 dissolution was highly dependent on the local distribution of scCO2 clusters. The rock with relatively high porosity and permeability would have more capacity for scCO2 injection resulting in a faster and greater dissolution of scCO2 in brine. The effect of capillary number on the scCO2 dissolution process was related to the range of capillary number. The effective upscaled (macro‐scale) mass transfer coefficient (kA) during scCO2 dissolution was evaluated, and the power‐law relationship between kA and Péclet number was obtained. Rock wettability was found to be another factor controlling the scCO2 dissolution process by affecting the scCO2‐brine interfacial area. Our pore‐scale study provides a deep understanding of the scCO2 dissolution trapping mechanism, which is important to enhance the prediction of sequestration risk and improve sequestration efficiency.

A dataset of gas reservoir microscopic images in Xujiahe Formation in northeast Sichuan Basin

The gas reservoir in Xujiahe Formation in northeast Sichuan Basin holds promising potential for exploration and development. In this paper, we collected and sorted the microscopic image data of the clastic reservoir of Xujiahe Formation gas reservoir in the study area, so as to further study the petrological characteristics, diagenesis, reservoir space and physical properties of the region. The reservoir rocks are mainly lithic sandstone, lithic arkose sandstone and quartz sandstone. There are intergranular solution pores, residual intergranular solution pores, intragranular solution pores, clay mineral intergranular (solution) pores and various microfractures in the reservoir space. According to the observation of the cast thin section, the pores are relatively undeveloped, while some thin sections (micro-cracks) are relatively developed. Based on thin section identification and observation, the sandstone reservoirs of Xujiahe Formation in the study area have undergone processes such as compaction, cementation, dissolution and tectonic fracture. This dataset can provide a direct source for the comparative analysis and reference of sandstone image data. Moreover, it can serve as a benchmark for the intelligent identification of lithology.

Sub-volume analysis of pore-network simulations: Determining the asymptotic longitudinal dispersion coefficient

This work introduces a new method called sub-volume analysis (SVA) to calculate macroscopic model parameters with their uncertainties from pore-scale simulations. We apply the SVA to calculate the longitudinal dispersion coefficient of the advection–dispersion model (ADM) from continuous tracer injection problems simulated with pore-network models. The SVA application presented here allows for verifying if a given network has enough length to reach an asymptotic dispersion regime where ADM applies. We tested the method using statistically constructed pore networks and pore networks built by joining several network units extracted from Berea sandstone images. There is complete agreement within the error margins for the asymptotic longitudinal dispersion coefficients obtained with the SVA and those obtained from breakthrough curve fitting. We also showed that the SVA gives more accurate results than those from the breakthrough curve fitting and much more accurate results than those from the method of spatial moments. Unlike the other methods, the SVA estimates the parameter uncertainty using a single pore-network realization.

Characterization and Degradation of Ancient Architectural Red Sandstone in a Natural Erosion Environment

The properties and appearance of ancient architectural red sandstone will be damaged after being eroded by the natural environment for a long time. In order to investigate the weathering and erosion characteristics of the red sandstone structure of an existing ancient building, ultrasonic testing techniques, combined with scanning electron microscopy (SEM), X-ray diffraction (XRD) and X-ray computed tomography (X-CT), were used to analyze a building in Ganzhou. The variation in chemical substances contained in the red sandstone specimens according to phenology was analyzed by X-ray diffraction (XRD). The characteristic parameters of the CT grayscale images of the red sandstone were extracted and combined with the ultrasonic wave velocity values to comprehensively analyze the degradation characteristics of the red sandstone specimens, and a method to characterize the degradation degree of the red sandstone as a whole plane is proposed. We use the gray model (GM (1, 1)) to predict the surface degradation degree of red sandstone specimens, and gray relation analysis (GRA) to further analyze the correlation between the characteristic parameters of CT grayscale images of red sandstone and its degradation degree. The results show that in the natural erosion environment, dolomite and chlorite are generated on the exposed surface of the red sandstone, which can protect the internal sandstone to a certain extent. The degradation degree of the red sandstone specimens in the horizontal X and Y directions varies, and the proposed method of calculating the overall plane degradation degree of the red sandstone is feasible. The minimum average relative error of the surface degradation degree obtained from the gray prediction GM (1, 1) model is 1.4591%. There is a good correlation between the characteristic parameters of the red sandstone CT grayscale images and the degradation degree.

Adaptive mesh refinement in locally conservative level set methods for multiphase fluid displacements in porous media

Multiphase flow in porous media often occurs with the formation and coalescence of fluid ganglia. Accurate predictions of such mechanisms in complex pore geometries require simulation models with local mass conservation and with the option to improve resolution in areas of interest. In this work, we incorporate patch-based, structured adaptive mesh refinement capabilities into a method for local volume conservation that describes the behaviour of disconnected fluid ganglia during level set simulations of capillary-controlled displacement in porous media. We validate the model against analytical solutions for three-phase fluid configurations in idealized pores containing gas, oil, and water, by modelling the intermediate-wet oil layers as separate domains with their volumes preserved. Both the pressures and volumes of disconnected ganglia converge to analytical values with increased refinement levels of the adaptive mesh. Favourable results from strong and weak scaling tests emphasize that the number of patches per processor and the total number of patches are important parameters for efficient parallel simulations with adaptive mesh refinement. Simulations of two-phase imbibition and three-phase gas invasion on segmented 3D images of water-wet sandstone show that adaptive mesh refinement has the highest impact on three-phase displacements, especially concerning the behaviour of the conserved, intermediate-wet phase.

Representative elementary volumes for various characteristics of two-phase flows in porous media: A statistical approach

This article presents a systematic study of the influence of two-phase flowpatterns on the representative elementary volume for various displacement characteristics. This paper considers both drainage and imbibition regimes and related patterns such as compact displacement, viscous, and capillary fingering. As two-phase flow characteristics, the displacement efficiency and interfacial contact areas between the invaded fluid and the displaced fluid, as well as between the invaded fluid and solid particles, are considered. Not limited to displacement patterns, we study the effect of capillary number on the minimum REV size. The findings of this paper are based on numerical simulations performed in two-dimensional artificially created porous structures using the Monte-Carlo movement algorithm. The main conclusions are validated on 3D X-ray computed tomography images of natural sandstones. Two-phase immiscible displacement in a porous medium is modeled using the lattice Boltzmann equations with the multi-relaxation time collision operator in combination with the color-gradient method describing interfacial phenomena and wetting effects. Minimum REV sizes have been identified using statistical data collected from a large number of numerical simulations in two-dimensional samples. Coefficient of variation and average value of the investigated property have been used to identify REV. The results show that REV for two-phase flow in porous media depends on the displacement pattern. Minimum REV sizes for viscous fingering and compact displacement are close. The accuracy of predicting the displacement efficiency for viscous fingering and compact displacement modes is much higher than for capillary fingering. Minimum REV sizes for interfacial contacts are lower than for displacement efficiency.

Hierarchical 3D reconstruction algorithm of core image based on detailed and structural dictionaries

The microstructure of porous media is of great significance to the study and analysis of their physical properties and mechanical properties. At present, the three-dimensional reconstruction method is widely used to obtain digital models of porous media microstructures that can be used to reconstruct their three-dimensional structures from two-dimensional training images. The super-dimensional algorithm, a learning-based method, makes full use of a priori three-dimensional structural information, and its reconstruction effect is better than that of the aforementioned traditional method. However, most current algorithms are based on 128- or 256-size training image reconstruction, which is still far from the actual application of the image size. To make algorithms more suitable to the actual application, we herein propose a hierarchical super-dimensional reconstruction method based on the idea of a multiple grid to complete the reconstruction of large-size images, such as the three-dimensional reconstruction of 512- or even 1000-size images. Using this method, we designed a template of the detailed information dictionary and developed a corresponding matching algorithm for a hierarchical reconstruction process. To further improve the efficiency of reconstruction, dictionary element porosity characteristics were used for classification. Finally, to verify the effectiveness of the proposed method, four groups of sandstone image reconstruction experiments were carried out, and the reference images with different sizes and porosity were reconstructed several times. Furthermore, to determine the accuracy of the reconstruction results, we compared the real structure with the visual and statistical indicators, and pore and throat parameters.