A morphology-based method and system for classifying pores in tight reservoir porous media
By using field emission scanning electron microscopy and machine learning technology, the roundness and aspect ratio of pores are automatically measured, and quantitative classification thresholds are set. This solves the subjective problem of traditional pore morphology classification, realizes the quantification and comparability of pore morphology, and improves the scientificity and reliability of reservoir analysis.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SICHUAN UNIV
- Filing Date
- 2026-03-11
- Publication Date
- 2026-06-02
AI Technical Summary
Existing morphology-based pore classification methods lack rigorous quantitative criteria and broad applicability, resulting in insufficient scientific rigor, repeatability, and cross-sample comparability.
Microstructural images of rock tight reservoir samples were acquired using field emission scanning electron microscopy. A segmentation model was trained using machine learning. By automatically measuring pore roundness and aspect ratio and setting quantitative classification thresholds, the pore morphology was quantified and made comparable.
It achieves a systematic and objective characterization of pore morphology, breaks through the limitations of traditional classification methods, and provides a reliable quantitative tool for fluid migration behavior and reservoir quality evaluation in porous media.
Smart Images
Figure CN122135105A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of geological engineering, oil and gas field development and rock physics, and more specifically to a morphology-based method and system for classifying the pores of porous media in tight reservoirs. Background Technology
[0002] Currently, morphology-based pore classification methods rely heavily on subjective visual judgment and lack rigorous quantitative basis and broad applicability. To date, a robust and quantifiable pore morphology classification system has not been established, resulting in significant deficiencies in the scientific validity, repeatability, and cross-sample comparability of morphology classification results.
[0003] Therefore, how to transform the traditional subjective judgment-based description of pore morphology into a quantifiable and comparable structural index, and realize the pore classification of porous media in tight reservoirs, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0004] In view of the above problems, the present invention aims to provide a morphology-based method and system for classifying the pores of porous media in tight reservoirs that overcomes or at least partially solves the above problems.
[0005] To achieve the above objectives, the present invention adopts the following technical solution:
[0006] A morphology-based method for classifying the pore size of porous media in tight reservoirs, comprising: S1. Obtain microstructure images of rock tight reservoir samples using field emission scanning electron microscopy; S2. Label the obtained scanning electron microscope images with two types of samples, pores and matrix / background, as training sets, and train a segmentation model for rock samples; S3. Apply the trained segmentation model to segment all sample SEM images and generate binarized images that identify the background and pores; S4. Based on the identified binary image, automatically measure the pore morphology parameters to obtain the target parameters pore roundness and aspect ratio; S5. Classify the pore shape according to the preset quantitative classification thresholds for pore roundness and aspect ratio, and output the statistical classification results.
[0007] Preferably, step S2 specifically includes: importing the obtained scanning electron microscope (SEM) image into the software ImageJ, using the TWS plugin, manually marking the two types of samples, pores and matrix / background, on the image by drawing lines / circles as training sets, entering the iterative pore identification stage, using machine learning to optimize the identification accuracy through repeated training and verification to train the classifier, generating a segmentation model for rock samples for identifying pores and matrix / background.
[0008] Preferably, in step S4, the pore morphology parameters are automatically measured using the built-in program of ImageJ to obtain the pore roundness and aspect ratio; The pore roundness is:
[0009] The aspect ratio is:
[0010] in, The area of the pores. Let be the perimeter of the pore. To fit the length of the major axis of the ellipse, This is the length of the minor axis.
[0011] Preferably, the method for determining the quantitative classification threshold in step S5 is as follows: based on the theoretical roundness calculation value of the standard basic geometric shape, combined with the pore morphology presented in the image, image resolution, and the fault tolerance rate of edge recognition, the classification threshold is determined and preset offline.
[0012] Preferably, in step S5, the statistical classification results include: pore type, the specific proportion of each type of pore, and morphological description.
[0013] Preferably, the types of pore shapes include: near-circular, sub-circular, transitional, elongated regular, and complex irregular.
[0014] A morphology-based pore classification system for porous media in tight reservoirs, based on the aforementioned morphology-based pore classification method for porous media in tight reservoirs, includes: a field emission scanning electron microscope, a model training module, a pore identification module, an automatic measurement module, and a classification module; Field emission scanning electron microscopy is used to acquire microstructural images of rock tight reservoir samples; The model training module is used to label the obtained scanning electron microscope images with two types of samples, pores and matrix / background, as training sets, and to train a segmentation model for rock samples. The pore recognition module is used to segment all sample SEM images using a trained segmentation model, generating a binarized image that identifies the background and pores. The automatic measurement module is used to automatically measure pore morphology parameters based on the recognized binary image to obtain the target parameters pore roundness and aspect ratio. The classification module is used to classify pore shapes according to preset quantitative classification thresholds for pore roundness and aspect ratio, and output statistical classification results.
[0015] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned morphology-based pore classification method for porous media in tight reservoirs.
[0016] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the morphology-based pore classification method for porous media in tight reservoirs.
[0017] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a morphology-based method and system for classifying pores in porous media of tight reservoirs. It effectively solves the problems of traditional pore morphology classification relying on subjective judgment and lacking unified quantitative standards. By automatically extracting roundness and aspect ratio through machine learning and setting clear mathematical thresholds, it realizes a systematic and objective characterization of pore cross-sectional shape. It transforms the traditional pore morphology description that relies on subjective judgment into a quantifiable and comparable structural index, breaking the limitations of traditional classification methods in cross-sample comparison and dynamic tracking. The present invention can not only systematically reveal the distribution law of pore morphology, but also help to provide a reliable and generalizable quantitative tool system for the study of the mechanism of fluid migration behavior in porous media, reservoir quality evaluation, and structural evolution analysis of the dynamic evolution of pore structure during oil and gas extraction. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0019] Figure 1 This is a schematic diagram of a morphology-based pore classification method for dense reservoir porous media provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the pore classification results of an oil shale reservoir sample provided in an embodiment of the present invention. Detailed Implementation
[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0021] This invention discloses a morphology-based method for classifying the pore size of porous media in tight reservoirs, comprising: S1. Obtain microstructure images of rock tight reservoir samples using field emission scanning electron microscopy; S2. Label the obtained scanning electron microscope images with two types of samples, pores and matrix / background, as training sets, and train a segmentation model for rock samples; S3. Apply the trained segmentation model to segment all sample SEM images and generate binarized images that identify the background and pores; S4. Based on the identified binary image, automatically measure the pore morphology parameters to obtain the target parameters pore roundness and aspect ratio; S5. Classify the pore shape according to the preset quantitative classification thresholds for pore roundness and aspect ratio, and output the statistical classification results.
[0022] To further implement the above technical solution, step S2 specifically includes: importing the obtained scanning electron microscope (SEM) images into the software ImageJ, using the plugin TWS (Trainable Weka Segmentation), manually marking the two types of samples, pores and matrix / background, on the image by drawing lines / circles as training sets, entering the iterative pore identification stage, using machine learning to optimize the identification accuracy through repeated training and verification to train the classifier, generating a segmentation model for rock samples, used to identify pores and matrix / background; finally, using the trained segmentation model to automatically identify pores and matrix / background, completing the accurate segmentation of complex pore structures.
[0023] In this embodiment, the specific process of training the classifier to generate a segmentation model for rock samples is as follows: Representative field emission scanning electron microscope images were selected, and the annotation tools provided by the Trainable Weka Segmentation plugin were used to manually mark the two types of regions, namely pores and matrix, on the images, including pore and matrix regions with different morphologies and grayscale characteristics, in order to ensure the representativeness of the training set. The Trainable Weka Segmentation plugin automatically extracts multidimensional features from the input image, including but not limited to: Gaussian blur and edge detection, Sobel filter, Hessian matrix, statistical features such as mean, variance, and median, as well as texture features, which together constitute a high-dimensional feature space for classification. Based on user-annotated sample regions, the Trainable Weka Segmentation plugin uses a random forest classifier for training. The random forest consists of multiple decision trees, each of which is trained based on a randomly selected subset of features and a subset of samples. Finally, the classification result is output through a voting mechanism. After training, the image to be segmented is input into the model. The Trainable Weka Segmentation plugin classifies each pixel, generates a segmentation probability map of pores and matrix, and further outputs the binarized segmentation result. The training samples or parameters are adjusted according to the segmentation effect to iteratively optimize the model performance.
[0024] The random forest classifier consists of: an input layer containing the input image pixels and their neighborhood features, with the feature dimensions adjusted according to the plugin configuration and including multiple filter responses by default; a hidden layer consisting of a random forest composed of multiple decision trees, each tree independently making classification decisions for the pixels; and an output layer that integrates the voting results of all decision trees, outputting the probability that each pixel belongs to the pores or matrix, and generating the final segmentation result accordingly.
[0025] In this embodiment, in step S3, black represents the background and white represents pores in the binarized image. The complex pore structure is automatically and accurately identified and extracted using a trained segmentation model of the rock sample.
[0026] To further implement the above technical solution, in step S4, the pore morphology parameters are automatically measured using the built-in program of ImageJ to obtain the pore roundness and aspect ratio. The pore roundness is:
[0027] The aspect ratio is:
[0028] in, The area of the pores. Let be the perimeter of the pore. To fit the length of the major axis of the ellipse, This is the length of the minor axis.
[0029] To further implement the above technical solution, the method for determining the quantitative classification threshold in step S5 is as follows: based on the theoretical roundness calculation value of the standard basic geometric shape, combined with the pore morphology presented in the image, image resolution, and the fault tolerance rate of edge recognition, the classification threshold is determined and preset offline.
[0030] In this embodiment, the preset classification threshold is specifically as follows: The lower limit for near-circular pores is set to 0.85 to accommodate visually near-circular pores; the boundary between regular and irregular pores is set to 0.6, with values below 0.6 indicating extremely rough edges or highly distorted shapes; the "elongated" boundary is set to aspect ratios AR = 1.5, 2.5, and 5 to distinguish between blocky, short columnar, and fissure-like pores, as detailed in Table 1.
[0031] To further implement the above technical solution, in step S5, the statistical classification results include: pore type, specific proportion of each type of pore, and morphological description.
[0032] To further implement the above technical solutions, the types of pore shapes include: near-circular, sub-circular, transitional, elongated regular, and complex irregular.
[0033] In another embodiment, taking a low-maturity tight oil shale reservoir sample as an example, an oil shale sample that has undergone argon ion polishing was selected. FE-SEM imaging was performed, and the image was imported into ImageJ. The Trainable WekaSegmentation plugin was used for pore and matrix annotation training. After training, the entire image was segmented to obtain a clear binary pore image. Using the software's automatic measurement function, noise was removed, and the roundness (circ.) and aspect ratio (AR) of 3735 independent pores within the field of view were extracted. The extracted data were then classified according to a preset classification threshold to obtain... Figure 2 The classification results show that sub-circular pores account for the highest proportion, reaching 42.9%. Combined with near-circular pores, the total proportion of these two types of regular pores reaches 60.93%, indicating that the pore morphology in this sample is generally regular, and the pore structure is mainly composed of regular intergranular or intercrystalline pores, which have not yet undergone strong structural compression deformation. Transitional pores account for the second highest proportion, while elongated regular pores and complex irregular pores account for a relatively low proportion. Although they are few in number, they may make an important contribution to permeability and further reflect the distribution characteristics of pore morphology. This embodiment no longer relies on the subjective visual judgment of researchers, but classifies pores through objective geometric parameters and clear mathematical thresholds. The classification results not only provide the specific proportion of each type of pore, but also clearly show the morphological characteristics of reservoir pores, providing a solid quantitative data foundation for subsequent analysis of the impact of pore structure on fluid seepage.
[0034] A morphology-based pore classification system for porous media in tight reservoirs, based on a morphology-based pore classification method for porous media in tight reservoirs, includes: field emission scanning electron microscopy, a model training module, a pore identification module, an automatic measurement module, and a classification module; Field emission scanning electron microscopy is used to acquire microstructural images of rock tight reservoir samples; The model training module is used to label the obtained scanning electron microscope images with two types of samples, pores and matrix / background, as training sets, and to train a segmentation model for rock samples. The pore recognition module is used to segment all sample SEM images using a trained segmentation model, generating a binarized image that identifies the background and pores. The automatic measurement module is used to automatically measure pore morphology parameters based on the recognized binary image to obtain the target parameters pore roundness and aspect ratio. The classification module is used to classify pore shapes according to preset quantitative classification thresholds for pore roundness and aspect ratio, and output statistical classification results.
[0035] A computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements a morphology-based pore classification method for porous media in tight reservoirs.
[0036] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements a morphology-based method for classifying the pore size of porous media in tight reservoirs.
[0037] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.
[0038] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A morphology-based method for classifying the pore size of porous media in tight reservoirs, characterized in that, include: S1. Obtain microstructure images of rock tight reservoir samples using field emission scanning electron microscopy; S2. Label the obtained scanning electron microscope images with two types of samples, pores and matrix / background, as training sets, and train a segmentation model for rock samples; S3. Apply the trained segmentation model to segment all sample SEM images and generate binarized images that identify the background and pores; S4. Based on the identified binary image, automatically measure the pore morphology parameters to obtain the target parameters pore roundness and aspect ratio; S5. Classify the pore shape according to the preset quantitative classification thresholds for pore roundness and aspect ratio, and output the statistical classification results.
2. The morphology-based pore classification method for tight reservoir porous media as described in claim 1, characterized in that, Step S2 specifically includes: importing the obtained scanning electron microscope (SEM) images into the software ImageJ, using the TWS plugin, manually marking the two types of samples, pores and matrix / background, on the images by drawing lines / circles as training sets, entering the iterative pore identification stage, using machine learning to optimize the identification accuracy through repeated training and verification to train the classifier, generating a segmentation model for rock samples, which is used to identify pores and matrix / background.
3. The morphology-based pore classification method for tight reservoir porous media as described in claim 1, characterized in that, In step S4, the pore morphology parameters are automatically measured using the built-in ImageJ program to obtain the pore roundness and aspect ratio. The pore roundness is: The aspect ratio is: in, The area of the pores. Let be the perimeter of the pore. To fit the length of the major axis of the ellipse, This is the length of the minor axis.
4. The morphology-based pore classification method for tight reservoir porous media as described in claim 1, characterized in that, The method for determining the quantitative classification threshold in step S5 is as follows: based on the theoretical roundness calculation value of the standard basic geometric shape, combined with the pore morphology presented in the image, image resolution, and the fault tolerance rate of edge recognition, the classification threshold is determined and preset offline.
5. The morphology-based pore classification method for tight reservoir porous media as described in claim 1, characterized in that, In step S5, the statistical classification results include: pore type, the specific proportion of each type of pore, and morphological description.
6. The morphology-based pore classification method for tight reservoir porous media as described in claim 1, characterized in that, The types of pore shapes include: near-circular, sub-circular, transitional, elongated regular, and complex irregular.
7. A morphology-based pore classification system for porous media in tight reservoirs, characterized in that, A morphology-based pore classification method for tight reservoir porous media according to any one of claims 1-6 includes: a field emission scanning electron microscope, a model training module, a pore identification module, an automatic measurement module, and a classification module; Field emission scanning electron microscopy is used to acquire microstructural images of rock tight reservoir samples; The model training module is used to label the obtained scanning electron microscope images with two types of samples, pores and matrix / background, as training sets, and to train a segmentation model for rock samples. The pore recognition module is used to segment all sample SEM images using a trained segmentation model, generating a binarized image that identifies the background and pores. The automatic measurement module is used to automatically measure pore morphology parameters based on the recognized binary image to obtain the target parameters pore roundness and aspect ratio. The classification module is used to classify pore shapes according to preset quantitative classification thresholds for pore roundness and aspect ratio, and output statistical classification results.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements a morphology-based pore classification method for porous media in tight reservoirs as described in any one of claims 1 to 6.
9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements a morphology-based pore classification method for porous media in tight reservoirs as described in any one of claims 1 to 6.