Three-dimensional digital core image and attribute feature similarity calculation method
By building a three-dimensional digital core image information and attribute parameters feature similarity evaluation model, and using goodness of fit and Euclidean distance to calculate the core similarity, the problem of poor accuracy of the attribute similarity analysis of three-dimensional digital core attributes in the existing technology is solved, and the automation processing capability and result accuracy are improved.
Patent Information
- Application Number
- CN202311334635.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-16
- Publication Date
- 2025-04-18
AI Technical Summary
The prior art is difficult to effectively combine three-dimensional digital core image features and attribute parameters to conduct accurate similarity analysis, resulting in poor accuracy of the results and low working efficiency.
By obtaining the three-dimensional digital core sample image set, statistical and normalized processing are carried out, an image information similarity evaluation model and attribute parameter feature similarity evaluation model are built, and the similarity between cores is calculated using goodness of fit and Euclidean distance.
The automated processing capability and accuracy of results of three-dimensional digital core attribute characteristics analysis are improved, and manpower investment and time cost are reduced.
Smart Images
Figure SMS_1 
Figure SMS_4 
Figure SMS_6
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of data processing and relates to a three-dimensional digital core image and an attribute feature similarity calculation method. Background Art
[0002] Digital core technology originated from the modeling of reservoir rock microscopic pore structure. Based on the two-dimensional statistical characteristic information of rocks or three-dimensional scanning imaging, it can describe the microscopic structure of rocks at the pore scale, and has become the key to quantitatively analyzing rock physical properties and pore distribution and simulating fluid flow in nanopores. Digital core is the basic platform for pore-level microscopic seepage theory research. It can visualize and quantify core samples and effectively describe the pore space of reservoir rocks and study the fluid migration mechanism inside the reservoir.
[0003] In the process of oil reservoir exploration and development, it is usually necessary to collect real core samples from the target geological area of the oil reservoir to be explored and developed, and perform CT scanning to obtain a three-dimensional simulated digital core model. The core attribute characteristics are compared and analyzed based on the simulated core digital model combined with digital technology at the micro scale, which plays an important role in the evaluation of reservoir characteristics and reserve analysis. In the existing application implementation, the three-dimensional core simulation model is generally obtained by CT scanning for drilling core sampling, which can intuitively find the pore characteristics of the core, and obtain the permeability, porosity, rock electricity and other attribute parameters of the digital core through analysis. However, it takes a lot of tedious and repetitive work to manually mark the image features and perform parameter distribution statistics on a large number of core digital models, and to perform attribute feature similarity analysis on the newly collected digital core simulation models. It is difficult to achieve the correlation and fusion of three-dimensional digital lithology image features and attribute parameters, which not only requires professional and experienced human resources, but also leads to low work efficiency and difficulty in ensuring the accuracy of the results.
[0004] Therefore, how to combine the three-dimensional digital core image features with the attribute parameter information features to accurately perform three-dimensional digital core attribute feature similarity analysis, increase automated processing capabilities, and reduce personnel time costs is one of the key technologies that need to be solved. Summary of the invention
[0005] The purpose of the present invention is to provide a three-dimensional digital core image and attribute feature similarity calculation method, which solves the problem of poor accuracy of the existing three-dimensional digital core attribute feature similarity analysis processing results.
[0006] The technical solution adopted by the present invention is a method for calculating the similarity of three-dimensional digital core images and attribute characteristics. A three-dimensional digital core sample set obtained by scanning and analyzing drilling core sampling is used to process and analyze the digital core image characteristics and the statistical distribution characteristics of attribute parameters, and the similarity between two cores is calculated by goodness of fit and Euclidean distance.
[0007] The present invention is also characterized in that:
[0008] Follow the steps below to implement it:
[0009] Step 1, obtaining a three-dimensional digital core sample image set and performing statistics;
[0010] Step 2: Building a three-dimensional digital core image information similarity evaluation model;
[0011] Step 3: construct a lithology attribute parameter characteristic similarity evaluation model;
[0012] Step 4: According to the results of the three-dimensional digital core image information similarity evaluation model and the results obtained by the attribute parameter feature similarity evaluation model, an average value is calculated to obtain the comprehensive similarity results between different cores.
[0013] The specific process of step 1 is to count the multi-dimensional pixel feature set {P1, P2, ..., P n}, n is the total number of sample cores, and P is the set of pixel area, number of pores and fractures, total pore and fracture pixel area, and total pore and fracture area pixel ratio.
[0014] In step 1, the length of the core 3D digital model image in the X, Y, and Z directions is the number of pixels in that direction. Each pixel point in one direction has a cross section relative to the 3D model. All feature data of the cross section corresponding to each point in each direction are counted. This set is represented by {P1, P2, ..., P n}, where P is the full feature set of the core image, and n is the number of sample cores; P1 = {P 1X , P 1Y , P 1Z}, where P 1X , P Y1 , P 1Z are the feature sets of core No. 1 in the X, Y, and Z directions respectively; P 1X = {P 1X_xs , P 1X_kf , P 1X_kfs , P 1X_kfsb}, where P 1X_xs , P 1X_kf , P 1X_kfs , P 1X_kfsb are the pixel area, number of pores and fractures, total pixel area of pores and fractures, and pixel ratio of total pixel area of pores and fractures of core 1 in the X direction; P 1X_xs = {P 1X_xs_1 , P 1X_xs_2 ,…,P 1X_xs_x}, where x is the number of pixels in the core in the X direction, and xs_1 represents the pixel area of the section where the first pixel position in the X direction is located.
[0015] Step 2 is implemented according to the following steps:
[0016] Step 2.1, normalize the data of pixel features of different dimensions in the two-dimensional slice images of each core three-dimensional digital model in different directions so that subsequent calculations are performed within the same scale range;
[0017] Taking xs_1 as an example, the normalization formula is:
[0018]
[0019] Among them, P 1X_xs_1 ′ is the normalized result value, P 1X_xs The minimum value in the set, P 1X_xs The maximum value in the collection.
[0020] Step 2.2, by using the goodness of fit method, similarity calculation is performed on the pixel feature data of the same dimension in the same direction of different cores;
[0021] To calculate P 1X_xs and P 2X_xs Taking the similarity of as an example, the calculation formula for goodness of fit is:
[0022]
[0023] Among them, R 2 (P 1X_xs ′,P 2X_xs ′) is P 1X_xs and P 2X_xs The goodness-of-fit result value of the normalized set, For the set P 1X_xs The average value of .
[0024] Step 2.3, averaging the similarity results in different directions and dimensions to obtain the average similarity results of the three-dimensional image information between different cores;
[0025] Taking the calculation of the average similarity results of the first and second cores on the three-dimensional image information as an example, the calculation process is:
[0026]
[0027]
[0028] Among them, R 2(P1, P2) is the average similarity result value of the first and second cores in the three-dimensional image information, R 2 (P 1X ,P 2X ) is the similarity result value of the first and second cores in the X direction.
[0029] The specific process of step 3 is to obtain the attribute parameter set {Q1, Q2, ..., Q n}, Q is the attribute parameter set of the core such as porosity and permeability. The attribute characteristic distance between the cores is calculated by the Euclidean distance method. The larger the distance value, the less similar it is. The inverse of the Euclidean distance result value can be taken as the attribute characteristic similarity result between the cores.
[0030] In step 3, Q1={Q 1_kxd , Q 1_stl}, Q 1_kxd , Q 1_stl are the porosity and permeability values of the first core, respectively, and the similarity results of the attribute characteristics between the first and second cores are calculated as follows:
[0031]
[0032] Among them, A(Q1,Q2) is the attribute characteristic similarity result value between the first and second cores.
[0033] Step 4 calculates the similarity between the image and attribute features of the first and second cores as follows:
[0034]
[0035] Among them, S 1，2 is the attribute characteristic similarity result value between the first and second cores.
[0036] The beneficial effects of the present invention are as follows: the method for calculating the similarity of three-dimensional digital core images and attribute characteristics of the present invention utilizes a three-dimensional digital core sample set obtained through scanning and analysis of drilling core sampling to process and analyze digital core image characteristics and attribute parameter statistical distribution characteristics, and calculates the similarity between cores through goodness of fit and Euclidean distance, thereby further improving the automated processing capability and result accuracy of the three-dimensional digital core attribute characteristic similarity analysis, and providing important technical support for reservoir characteristic evaluation and reserve analysis. DETAILED DESCRIPTION
[0037] The present invention is described in detail below in conjunction with specific implementation modes.
[0038] Example 1
[0039] The method for calculating the similarity of three-dimensional digital core images and attribute characteristics of the present invention utilizes a three-dimensional digital core sample set obtained by scanning and analyzing drilling core sampling, processes and analyzes digital core image characteristics and attribute parameter statistical distribution characteristics, and calculates the similarity between cores by goodness of fit and Euclidean distance.
[0040] Example 2
[0041] The three-dimensional digital core image and attribute feature similarity calculation method of the present invention is specifically implemented according to the following steps:
[0042] Step 1, obtaining a three-dimensional digital core sample image set and performing statistics;
[0043] Step 2: Building a three-dimensional digital core image information similarity evaluation model;
[0044] Step 3: construct a lithology attribute parameter characteristic similarity evaluation model;
[0045] Step 4: According to the results of the three-dimensional digital core image information similarity evaluation model and the results obtained by the attribute parameter feature similarity evaluation model, an average value is calculated to obtain the comprehensive similarity results between different cores.
[0046] Example 3
[0047] The three-dimensional digital core image and attribute feature similarity calculation method of the present invention is specifically implemented according to the following steps:
[0048] Step 1: Obtain a set of three-dimensional digital core sample images and perform statistics; Count the multi-dimensional pixel feature sets {P1, P2, ..., P2} corresponding to the cross-section of each pixel position in the three-dimensional digital model image of each core in the X, Y, and Z directions. n}, n is the total number of sample cores, P is the set of pixel area, number of pores and fractures, total pore and fracture pixel area, and total pore and fracture area pixel ratio;
[0049] Step 2: Building a three-dimensional digital core image information similarity evaluation model;
[0050] The data of pixel features of different dimensions in the two-dimensional slice images of each core three-dimensional digital model in different directions are normalized so that subsequent calculations are performed within the same scale range;
[0051] The similarity of pixel feature data in the same direction and dimension of different cores is calculated by goodness of fit method;
[0052] The similarity results in different directions and dimensions are averaged to obtain the average similarity results of the three-dimensional image information between different cores;
[0053] Step 3: construct a lithology attribute parameter characteristic similarity evaluation model;
[0054] Step 4: According to the results of the three-dimensional digital core image information similarity evaluation model and the results obtained by the attribute parameter feature similarity evaluation model, an average value is calculated to obtain the comprehensive similarity results between different cores.
[0055] Example 4
[0056] The three-dimensional digital core image and attribute feature similarity calculation method of the present invention is specifically implemented according to the following steps:
[0057] Step 1: Obtain a set of three-dimensional digital core sample images and perform statistics; Count the multi-dimensional pixel feature sets {P1, P2, ..., P2} corresponding to the cross-section of each pixel position in the three-dimensional digital model image of each core in the X, Y, and Z directions. n}, n is the total number of sample cores, P is the set of pixel area, number of pores and fractures, total pore and fracture pixel area, and total pore and fracture area pixel ratio;
[0058] The length of the core 3D digital model image in the X, Y, and Z directions is the number of pixels in that direction. Each pixel point in one direction has a cross section relative to the 3D model. All feature data of the cross section corresponding to each point in each direction are counted. This set is represented by {P1, P2, ..., P n}, where P is the full feature set of the core image, and n is the number of sample cores; P1 = {P 1X , P 1Y , P 1Z}, where P 1X , P Y1 , P 1Z are the feature sets of core No. 1 in the X, Y, and Z directions respectively; P 1X = {P 1X_xs , P 1X_kf , P 1X_kfs , P 1X_kfsb}, where P 1X_xs , P 1X_kf , P 1X_kfs , P 1X_kfsb are the pixel area, number of pores and fractures, total pixel area of pores and fractures, and pixel ratio of total pixel area of pores and fractures of core 1 in the X direction; P 1X_xs = {P 1X_xs_1 , P 1X_xs_2 ,…,P 1X_xs_x}, where x is the number of pixels (i.e., length) of the core in the X direction, and xs_1 represents the pixel area of the section where the first pixel position is located.
[0059] Step 2: Building a three-dimensional digital core image information similarity evaluation model;
[0060] Step 2 is implemented according to the following steps:
[0061] Step 2.1, normalize the data of pixel features of different dimensions in the two-dimensional slice images of each core three-dimensional digital model in different directions so that subsequent calculations are performed within the same scale range;
[0062] Taking xs_1 as an example, the normalization formula is:
[0063]
[0064] Among them, P 1X_xs_1 ′ is the normalized result value, P 1X_xs The minimum value in the set, P 1X_xs The maximum value in the collection.
[0065] Step 2.2, by using the goodness of fit method, similarity calculation is performed on the pixel feature data of the same dimension in the same direction of different cores;
[0066] To calculate P 1X_xs and P 2X_xs Taking the similarity of as an example, the calculation formula for goodness of fit is:
[0067]
[0068] Among them, R 2 (P 1X_xs ′,P 2X_xs ′) is P 1X_xs and P 2X_xs The goodness-of-fit result value of the normalized set, For the set P 1X_xs The average value of .
[0069] Step 2.3, averaging the similarity results in different directions and dimensions to obtain the average similarity results of the three-dimensional image information between different cores;
[0070] Taking the calculation of the average similarity results of the first and second cores on the three-dimensional image information as an example, the calculation process is:
[0071]
[0072]
[0073] Among them, R 2 (P1, P2) is the average similarity result value of the first and second cores in the three-dimensional image information, R 2 (P1X ,P 2X ) is the similarity result value of the first and second cores in the X direction.
[0074] Step 3: construct a lithology attribute parameter characteristic similarity evaluation model;
[0075] The specific process of step 3 is to obtain the attribute parameter set {Q1, Q2, ..., Q n}, Q is the attribute parameter set of the core such as porosity and permeability. The attribute characteristic distance between the cores is calculated by the Euclidean distance method. The larger the distance value, the less similar it is. The reciprocal of the Euclidean distance result value can be taken as the attribute characteristic similarity result between the cores.
[0076] In step 3, Q1={Q 1_kxd , Q 1_stl}, Q 1_kxd , Q 1_stl are the porosity and permeability values of the first core, respectively, and the similarity results of the attribute characteristics between the first and second cores are calculated as follows:
[0077]
[0078] Among them, A(Q1,Q2) is the attribute characteristic similarity result value between the first and second cores.
[0079] Step 4: According to the results of the three-dimensional digital core image information similarity evaluation model and the results obtained by the attribute parameter feature similarity evaluation model, an average value is calculated to obtain the comprehensive similarity results between different cores.
[0080] The similarity between the image and attribute features of the first and second cores is calculated as follows:
[0081]
[0082] Among them, S 1，2 is the attribute characteristic similarity result value between the first and second cores.
[0083] The modeling results of Example 4 are analyzed, and the results show that the porosity of the cross-section of the three-dimensional digital core model of the present invention is highly consistent with the porosity of the cross-section of the original image, and a three-dimensional digital core model closest to the real core is established, which has powerful feature extraction and image generation capabilities. The present invention converts professional experience into a standardized calculation model, which can be easily and quickly operated even by non-professionals. At the same time, it only needs to use an automated calculation model to perform one-click batch operations, which saves more time, effort and money; it can obtain more comprehensive and higher-precision results from more dimensional parameters and standardized calculation models, and has the characteristics of high work efficiency and accurate results.
Claims
1. A method for calculating similarity of three-dimensional digital core images and attribute features, characterized in that: The three-dimensional digital core sample set obtained by scanning and analyzing drilling core sampling is used to process and analyze the digital core image characteristics and the statistical distribution characteristics of attribute parameters, and the similarity between cores is calculated by goodness of fit and Euclidean distance.
2. The three-dimensional digital core image and attribute feature similarity calculation method according to claim 1, characterized in that: Follow the steps below to implement it: Step 1, obtaining a three-dimensional digital core sample image set and performing statistics; Step 2: Building a three-dimensional digital core image information similarity evaluation model; Step 3: construct a lithology attribute parameter characteristic similarity evaluation model; Step 4: According to the results of the three-dimensional digital core image information similarity evaluation model and the results obtained by the attribute parameter feature similarity evaluation model, an average value is calculated to obtain the comprehensive similarity results between different cores.
3. The three-dimensional digital core image and attribute feature similarity calculation method according to claim 2, characterized in that: The specific process of step 1 is to count the multi-dimensional pixel feature set {P1, P2, ... ... n }, n is the total number of sample cores, and P is the set of pixel area, number of pores and fractures, total pore and fracture pixel area, and total pore and fracture area pixel ratio.
4. The three-dimensional digital core image and attribute feature similarity calculation method according to claim 3, characterized in that: In step 1, the length of the core three-dimensional digital model image in the X, Y, and Z directions is the number of pixels in that direction; each pixel point in one direction has a section relative to the three-dimensional model, and all feature data of the section corresponding to each point in each direction are counted, and the set is represented as {P1, P2, ..., P n }, where P is the full feature set of the core image, and n is the number of sample cores; P1 = {P 1X , P 1Y , P 1Z }, where P 1X , P Y1 , P 1Z are the feature sets of core No. 1 in the X, Y, and Z directions respectively; P 1X = {P 1X_xs , P 1X_kf , P 1X_kfs , P 1X_kfsb }, where P 1X_xs , P 1X_kf , P 1X_kfs , P 1X_kfsb are the pixel area, number of pores and fractures, total pixel area of pores and fractures, and pixel ratio of total pixel area of pores and fractures of core 1 in the X direction; P 1X_xs = {P 1X_xs_1 , P 1X_xs_2 ,…,P 1X_xs_x }, where x is the number of pixels in the core in the X direction, and xs_1 represents the pixel area of the section where the first pixel position is located.
5. The three-dimensional digital core image and attribute feature similarity calculation method according to claim 4, characterized in that: The step 2 is specifically implemented according to the following steps: Step 2.1, normalize the data of pixel features of different dimensions in the two-dimensional slice images of each core three-dimensional digital model in different directions so that subsequent calculations are performed within the same scale range; Taking xs_1 as an example, the normalization formula is: Among them, P 1X_xs_1 ′ is the normalized result value, P 1X_xs The minimum value in the set, P 1X_xs The maximum value in a set; Step 2.2, by using the goodness of fit method, similarity calculation is performed on the pixel feature data of the same dimension in the same direction of different cores; To calculate P 1X_xs and P 2X_xs Taking the similarity of as an example, the calculation formula for goodness of fit is: Among them, R 2 (P 1X_xs ′,P 2X_xs ′) is P 1X_xs and P 2X_xs The goodness-of-fit result value of the normalized set, For the set P 1X_xs The average value of Step 2.3, averaging the similarity results in different directions and dimensions to obtain the average similarity results of the three-dimensional image information between different cores; Taking the calculation of the average similarity results of the first and second cores on the three-dimensional image information as an example, the calculation process is: Among them, R 2 (P1, P2) is the average similarity result value of the first and second cores in the three-dimensional image information, R 2 (P 1X ,P 2X ) is the similarity result value of the first and second cores in the X direction.
6. The three-dimensional digital core image and attribute feature similarity calculation method according to claim 5, characterized in that: The specific process of step 3 is to obtain the attribute parameter set {Q1, Q2, ..., Q n }, Q is the attribute parameter set of the core such as porosity and permeability. The attribute characteristic distance between the cores is calculated by the Euclidean distance method. The larger the distance value, the less similar it is. The inverse of the Euclidean distance result value can be taken as the attribute characteristic similarity result between the cores.
7. The three-dimensional digital core image and attribute feature similarity calculation method according to claim 6, characterized in that: In step 3, Q1={Q 1_kxd , Q 1_stl }, Q 1_kxd , Q 1_stl are the porosity and permeability values of the first core, respectively, and the similarity results of the attribute characteristics between the first and second cores are calculated as follows: Among them, A(Q1,Q2) is the attribute characteristic similarity result value between the first and second cores.
8. The three-dimensional digital core image and attribute feature similarity calculation method according to claim 7, characterized in that: The similarity results of the image and attribute features calculated in step 4 between the first and second cores are: Among them, S 1，2 is the attribute characteristic similarity result value between the first and second cores.