Method for calculating porosity of buried hill rock core based on high-precision repaired CT (Computed Tomography) image

Abstract

The invention discloses a buried hill core porosity calculation method based on a high-precision repair CT image, and the method comprises the steps: S1, obtaining a micron CT three-dimensional gray image of a buried hill fracture reservoir plunger sample in a research region, and obtaining a three-dimensional digital core model of the buried hill fracture reservoir plunger sample based on the micron CT three-dimensional gray image; s2, obtaining nanometer CT three-dimensional grayscale images of the sub-samples at different positions; s3, establishing a buried hill fractured reservoir core image database under different resolutions, learning and training the image through a residual network model to obtain a high-precision repaired micron CT image, and simulating and calculating the porosity through threshold segmentation; and S4, determining the actually measured porosity of the buried hill crack sample through an indoor experiment, and comparing the actually measured porosity with the simulated calculation porosity. According to the method, the pixel boundary contact information of the micron CT image is recovered through the deep residual network model, the restoration quality of the image is improved, and the simulated porosity is closer to an actual value.

Description

Technical Field

[0001] The invention belongs to the technical field of buried hill core porosity calculation, and in particular relates to a buried hill core porosity calculation method based on high-precision repaired CT images. Background Art

[0002] At present, the exploration and development of offshore oil fields is gradually deepening into complex reservoirs, and higher technical requirements have been put forward in the direction of understanding the genesis of reservoirs, characteristic development mechanism, and research on the law of reservoir development. When carrying out reserve declaration and well site pre-examination in the exploration stage, it is first necessary to obtain reservoir porosity, permeability and other parameters, so as to meet the analysis of exploration reserve potential. At present, the methods for calculating core porosity include conventional rock physics experiments and digital core methods. Now, with the widespread application of digital cores in the fields of core analysis of sandstone, carbonate rock, shale, etc., it can provide an effective technical means for the exploration and development of complex offshore reservoirs. Its basic principle is to use computer image processing technology based on scanned images and complete the reconstruction of digital cores through certain algorithms to truly restore the microscopic structural characteristics of the core. Through the image adaptive threshold segmentation technology, the core porosity can be quickly and quantitatively obtained. The accuracy of the result is affected by the quality and precision of the core image. This is mainly because when scanning the core, its scanning accuracy is inversely proportional to the core size. Taking into account factors such as the representativeness of the core, the acquired image may have problems such as loss of pore information, which directly affects the accuracy of porosity calculation.

[0003] At present, image super-resolution reconstruction technology can restore low-resolution images to corresponding high-resolution images through specific algorithms. This method can effectively overcome the problem of information loss caused by the resolution of micron CT scan core images. Traditional image super-resolution reconstruction methods mainly use neighbor search and other methods to reconstruct each pixel in the low-resolution image by neighbor search or neighbor difference. This method cannot meet the special conditions of each pixel, especially for core images with good feature differences or feature point averages. Therefore, how to restore the information originally lost in the low-resolution image is of great significance for the porosity calculation of buried hill cores based on CT images. Summary of the invention

[0004] The present invention is proposed to solve the problems existing in the prior art, and its purpose is to provide a method for calculating the porosity of buried hill cores based on high-precision repaired CT images.

[0005] The present invention is achieved through the following technical solutions:

[0006] A method for calculating the porosity of buried hill cores based on high-precision repaired CT images comprises the following steps:

[0007] S1. Obtain a micrometer CT three-dimensional grayscale image of a plunger sample of a buried hill fracture reservoir in the study area, and obtain a three-dimensional digital core model of the plunger sample of the buried hill fracture reservoir based on the micrometer CT three-dimensional grayscale image;

[0008] S2, obtaining nano-CT three-dimensional grayscale images of sub-samples at different positions;

[0009] S3. Establish a database of core images of buried hill fracture reservoirs at different resolutions, train the images through residual network model, obtain high-precision repaired micron CT images, and calculate porosity through threshold segmentation simulation;

[0010] S4. Determine the measured porosity of the buried hill fracture samples through indoor experiments and compare the measured porosity with the simulated porosity.

[0011] In the above technical solution, the buried hill fracture reservoir plunger sample is a dual-medium core sample with pores and fractures; the diameter of the buried hill fracture reservoir plunger sample is 25 mm.

[0012] In the above technical solution, the method for acquiring the micron CT three-dimensional grayscale image of the buried hill fracture reservoir plunger sample is a high-resolution micron CT scanning method, and the specific steps of the high-resolution micron CT scanning method are:

[0013] S11. Drill sub-samples at different positions of the plunger samples of the buried hill fracture reservoir in the study area by drilling, and the diameter of the sub-samples is guaranteed to be less than 8 mm;

[0014] S12, performing nanometer-level scanning on the sub-samples at different positions obtained in step S11 to ensure that the spatial resolution of the device scanning is below 500 nm.

[0015] In the above technical solution, the method for establishing the three-dimensional digital core model is: binary segmenting the three-dimensional grayscale image of the buried hill core to obtain the value of each pixel point, and according to the assignment, obtain the three-dimensional digital core model corresponding to the buried hill core sample.

[0016] In the above technical solution, the method for acquiring the nano CT three-dimensional grayscale image is a high-resolution nano CT scanning method.

[0017] In the above technical solution, step S3 specifically includes the following steps:

[0018] S31, using the micron CT 3D grayscale image obtained in step S1 as a low-resolution core image, and using the nano CT 3D grayscale image obtained in step S2 as a high-resolution core image, constructing two databases, HA and LB, and establishing a total of three test sets, SY1, LF2, and ZMS3;

[0019] S32, cutting the corresponding image, and determining that the minimum height and width of each high-resolution core image after cutting are both 1000 pixels;

[0020] S33. Build residual network model training model

[0021] S34, generating an image restoration model based on training samples, for low-resolution micron CT images, such as Figure 5 As shown, batch super-resolution restoration is performed to enlarge the size and improve the resolution details to form high-resolution repaired CT images of different magnifications. The resulting images are compared by introducing the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM).

[0022] The peak signal-to-noise ratio is defined as:

[0023]

[0024] Structural similarity SSIM is defined as:

[0025] SSIM(X,Y)=L(X,Y)*C ( X,Y)*S(X,Y)

[0026] in:

[0027]

[0028] Where: μ X μ Y are the means of X and Y, σ X σ Y are the standard deviations of X and Y, respectively. are the variances of X and Y, σ XY is the covariance of X and Y, C1, C2 and C3 are constants;

[0029] S35, performing adaptive threshold segmentation on the high-resolution repaired CT image obtained in step S34 to form a binary pixel image, in which pores and cracks are assigned a value of 1, and the rock skeleton is assigned a value of 0, and the total number of pore and crack voxels and the total number of rock voxels are counted.

[0030] In the above technical solution, the specific steps of step S32 are: first, select the width and height of the image in the sample data, set the parameter size-min = 1200 pixels for screening, load any filtered image, crop it from a random position to form 96x96 sub-images, and use it as the initial high-resolution image; use 8 times average pooling downsampling to obtain a 12x12 sub-block image, which is used as a low-resolution image, and preprocess the entire data set, and the high-resolution image is converted to [-1,1].

[0031] In the above technical solution, the residual network model training model of step S33 is composed of 4 convolution modules, 32 residual modules and 2 sub-pixel convolution modules; the convolution module defines the intermediate layer channel size as 128, the number of residual modules is 32, the high-resolution image is preprocessed, and the high- and low-resolution images are returned as the i-th pair of training samples.

[0032] In the above technical solution, the calculation method of the simulated porosity is:

[0033]

[0034] Where: is the porosity, dimensionless; n(r) is the total number of pore or fracture voxels, dimensionless; N is the total number of voxels in the core sample, dimensionless.

[0035] A computer-storable computing medium has a computer program, and the computer program can implement the above method.

[0036] The beneficial effects of the present invention are:

[0037] The present invention provides a method for calculating the porosity of buried hill cores based on high-precision repaired CT images. The high-precision repaired CT images can overcome the limitations of the loss of characteristic pore-fracture information of complex offshore reservoirs and the distortion of image magnification information, solve the core problem of the contradiction between CT scanning sample size and image resolution, realize the reconstruction of low-resolution micron CT images, restore more rock skeleton and pore and fracture lost information, and obtain more accurate porosity. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the method of embodiment 1 of the present invention;

[0039] Figure 2 The hypotonic image training set of Example 1 of the present invention;

[0040] Figure 3 The hidden hill image training set of Example 1 of the present invention;

[0041] Figure 4 is the sandstone image training set of Example 1 of the present invention;

[0042] Figure 5 This is the effect diagram before testing the image in Example 1 of the present invention;

[0043] Figure 6 This is a diagram showing the effect of super-resolution reconstruction of the test image in Example 1 of the present invention.

[0044] For ordinary technicians in this field, other relevant drawings can be obtained based on the above drawings without any creative work. DETAILED DESCRIPTION

[0045] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention is further described below with reference to the accompanying drawings and through specific implementation methods.

[0046] Example 1

[0047] like Figure 1 As shown, a method for calculating the porosity of buried hill cores based on high-precision repaired CT images includes the following steps:

[0048] S1. Obtain a micrometer CT three-dimensional grayscale image of a plunger sample of a buried hill fracture reservoir in the study area, and obtain a three-dimensional digital core model of the plunger sample of the buried hill fracture reservoir based on the micrometer CT three-dimensional grayscale image;

[0049] The buried hill fracture reservoir plunger sample is a dual-medium core sample with pores and fractures;

[0050] The diameter of the plunger sample of the buried hill fracture reservoir is 25 mm;

[0051] The method for acquiring the micron CT three-dimensional grayscale image of the plunger sample of the buried hill fracture reservoir is a high-resolution micron CT scanning method, and the specific steps of the high-resolution micron CT scanning method are:

[0052] S11. Drill sub-samples at different positions of the plunger samples of the buried hill fracture reservoir in the study area by drilling, and the diameter of the sub-samples is guaranteed to be less than 8 mm;

[0053] S12, performing nanometer-level scanning on the sub-samples at different positions obtained in step S11, ensuring that the spatial resolution of the device scanning is below 500nm;

[0054] In this embodiment, the resolution of the micrometer CT three-dimensional grayscale image is 5.0 μm, the voxel size is 2500×2500×2500, and the physical size is 25 mm×25 mm×25 mm.

[0055] The method for establishing the three-dimensional digital core model is as follows: performing binary segmentation on the three-dimensional grayscale image of the buried hill core to obtain a value for each pixel, and obtaining a three-dimensional digital core model corresponding to the buried hill core sample according to the value assignment;

[0056] S2. Obtaining nano-CT three-dimensional grayscale images of sub-samples at different positions;

[0057] The method for acquiring the nano CT three-dimensional grayscale image is a high-resolution nano CT scanning method;

[0058] In this embodiment, the resolution of the nano CT three-dimensional grayscale image is 0.5 μm, the voxel size is 2000×2000×2000, and the physical size is 8 mm×8 mm×8 mm.

[0059] S3, such as Figure 1 As shown in the figure, a buried hill fracture reservoir core image database with different resolutions is established, the image is trained through a residual network model to obtain a high-precision repaired micron CT image, and the porosity is calculated through threshold segmentation simulation, which specifically includes the following steps:

[0060] S31, using the micron CT three-dimensional grayscale image obtained in step S1 as the original low-resolution core image, and using the nano CT three-dimensional grayscale image obtained in step S2 as the original high-resolution core image, constructing two sets of databases, namely, a high-resolution core image database (HA) and a low-resolution core image database (LB), and at the same time establishing three test sets, namely, a low-permeability image training set (SY1), a buried hill image training set (LF2), and a sandstone image training set (ZMS3);

[0061] S32,

[0062] First, the original high-resolution core image is cut to obtain a cut high-resolution core image, wherein the height and width of the cut high-resolution core image are not less than 1000 pixels; the cutting is automatically implemented by code or program, and the image is cut in proportion based on the input image size;

[0063] Then, the cut high-resolution core images are screened to select the cut high-resolution core images with a height and a width of not less than 1200 pixels, any one of the screened cut high-resolution core images is loaded, and cropped from a random position to form 96x96 sub-images, and the 96x96 sub-images are used as the initial high-resolution image;

[0064] Then, the 96x96 sub-images are downsampled by 8 times average pooling to obtain 12x12 sub-block images. The 12x12 sub-block images are used as the initial low-resolution images. The initial high-resolution images and the initial high-resolution images after one-to-one correspondence form a set of learning data. Multiple sets of learning data constitute the overall data set.

[0065] Finally, the entire data set is preprocessed and the resolution of the initial high-resolution image is normalized to [-1, 1];

[0066] S33. Build residual network model training model

[0067] The residual network model training model is composed of 4 convolution modules, 32 residual modules and 2 sub-pixel convolution modules; the convolution module defines the intermediate layer channel size as 128, the number of residual modules is 32, the high-resolution image is preprocessed, and a set of learning data formed by the initial high-resolution image and the initial high-resolution image is returned as the i-th pair of training samples; S34, an image restoration model based on the training samples is generated, and the low-resolution micron CT image, such as Figure 5 As shown in Figure 1, batch super-resolution restoration is performed to enlarge the size and improve the resolution details to form high-resolution repaired CT images of different magnifications, such as Figure 6 As shown; the result images are compared by introducing parameter evaluation parameters;

[0068] The parameter evaluation parameters include peak signal-to-noise ratio (PSNR) and structural similarity (SSIM);

[0069] The peak signal-to-noise ratio (PSNR) is expressed as:

[0070]

[0071] The expression of the structural similarity (SSIM) is:

[0072] SSIM(X,Y)=L(X,Y)*C(X,Y)*S(X,Y)

[0073] in:

[0074]

[0075] Where: μ X μ Y are the means of X and Y, σ X σ Y are the standard deviations of X and Y, respectively. are the variances of X and Y, σ XY is the covariance of X and Y, C1, C2 and C3 are constants;

[0076] The evaluation results of the data set parameters in this embodiment are shown in Table 1.

[0077] Table 1 Dataset evaluation parameter results

[0078] Test Set SY1 LF2 ZMS3 PNSR 12.253 13.644 14.060 SSIM 0.338 0.341 0.347 Single sample time 0.541s 0.462s 0.262s

[0079] S35, performing adaptive threshold segmentation on the high-resolution repaired CT image obtained in step S34 to form a binary pixel image, in which pores and cracks are assigned a value of 1, and rock skeletons are assigned a value of 0, and the total number of pore and crack voxels and the total number of rock voxels are counted;

[0080] The calculation method of the simulated porosity is:

[0081]

[0082] Where: is the porosity, dimensionless; n(r) is the total number of pore or fracture voxels, dimensionless; N is the total number of voxels in the core sample, dimensionless;

[0083] S4. Determine the measured porosity of the buried hill fracture samples through indoor experiments and compare the measured porosity with the simulated porosity.

[0084] The applicant declares that the above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Those skilled in the art should understand that any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention shall fall within the protection scope and disclosure scope of the present invention.

Claims

1. A method for calculating the porosity of buried hill cores based on high-precision repaired CT images, characterized in that: The following steps are involved: S1. Obtain a micrometer CT three-dimensional grayscale image of a plunger sample of a buried hill fracture reservoir in the study area, and obtain a three-dimensional digital core model of the plunger sample of the buried hill fracture reservoir based on the micrometer CT three-dimensional grayscale image; S2, obtaining nano-CT three-dimensional grayscale images of sub-samples at different positions; S3. Establish a database of core images of buried hill fracture reservoirs at different resolutions, train the images through residual network model, obtain high-precision repaired micron CT images, and calculate porosity through threshold segmentation simulation; S4. Determine the measured porosity of the buried hill fracture samples through indoor experiments and compare the measured porosity with the simulated porosity.

2. The method for calculating porosity of buried hill core based on high-precision repaired CT images according to claim 1 is characterized in that: The buried hill fracture reservoir plunger sample is a dual-medium core sample with pores and fractures; the diameter of the buried hill fracture reservoir plunger sample is 25 mm.

3. The method for calculating porosity of buried hill core based on high-precision repaired CT images according to claim 1, characterized in that: The method for acquiring the micron CT three-dimensional grayscale image of the plunger sample of the buried hill fracture reservoir is a high-resolution micron CT scanning method, and the specific steps of the high-resolution micron CT scanning method are: S11. Drill sub-samples at different positions of the plunger samples of the buried hill fracture reservoir in the study area by drilling, and the diameter of the sub-samples is guaranteed to be less than 8 mm; S12, performing nanometer-level scanning on the sub-samples at different positions obtained in step S11 to ensure that the spatial resolution of the device scanning is below 500 nm.

4. The method for calculating porosity of buried hill core based on high-precision repaired CT images according to claim 1, characterized in that: The method for establishing the three-dimensional digital core model is as follows: binary segmentation is performed on the three-dimensional grayscale image of the buried hill core to obtain the value of each pixel, and the three-dimensional digital core model corresponding to the buried hill core sample is obtained according to the value assignment.

5. The method for calculating porosity of buried hill core based on high-precision repaired CT images according to claim 1, characterized in that: The method for acquiring the nano CT three-dimensional grayscale image is a high-resolution nano CT scanning method.

6. The method for calculating porosity of buried hill core based on high-precision repaired CT images according to claim 1, characterized in that: The step S3 specifically comprises the following steps: S31, using the micron CT 3D grayscale image obtained in step S1 as a low-resolution core image, and using the nano CT 3D grayscale image obtained in step S2 as a high-resolution core image, constructing two databases, HA and LB, and establishing a total of three test sets, SY1, LF2, and ZMS3; S32, cutting the corresponding image, and determining that the minimum height and width of each high-resolution core image after cutting are both 1000 pixels; S33, constructing a residual network model training model; S34, generating an image restoration model based on the training samples, performing batch super-resolution restoration on low-resolution micron CT images, enlarging the size and improving the resolution details, forming high-resolution restoration CT images of different magnifications, and comparing the result images by introducing parameter evaluation parameters; the parameter evaluation parameters include peak signal-to-noise ratio and structural similarity; The peak signal-to-noise ratio expression is: The expression of the structural similarity is: SSIM(X,Y)=L(X,Y)*C(X,Y)*S(X,Y) in: In the above formula: PSNR is the peak signal-to-noise ratio, SSIM is the structural similarity SSIM, μ X , μ Y are the means of X and Y, σ X , σ Y are the standard deviations of X and Y respectively; are the variances of X and Y, σ XY is the covariance of X and Y, C1, C2 and C3 are constants; S35. Perform adaptive threshold segmentation on the high-resolution repaired CT image obtained in step S34 to form a binary pixel image, in which pores and cracks are assigned a value of 1, and the rock skeleton is assigned a value of 0, and the total number of pore and crack voxels and the total number of rock voxels are counted.

7. The method for calculating porosity of buried hill core based on high-precision repaired CT images according to claim 6 is characterized in that: The specific steps of step S32 are: first, select the width and height of the image in the sample data, set the parameter size-min=1200 pixels for screening, load any filtered image, crop it from a random position to form 96x96 sub-images, and use it as the initial high-resolution image; use 8 times average pooling downsampling to obtain a 12x12 sub-block image, use the sub-block image as the low-resolution image, and preprocess the entire data set, and convert the high-resolution image to [-1,1].

8. The method for calculating porosity of buried hill core based on high-precision repaired CT images according to claim 6 is characterized in that: The residual network model training model of step S33 is composed of 4 convolution modules, 32 residual modules and 2 sub-pixel convolution modules; the convolution module defines the intermediate layer channel size as 128, the number of residual modules is 32, the high-resolution image is preprocessed, and the high- and low-resolution images are returned as the i-th pair of training samples.

9. The method for calculating porosity of buried hill core based on high-precision repaired CT images according to claim 1, characterized in that: The calculation method of the simulated porosity is: Where: is the porosity, dimensionless; n(r) is the total number of pore or fracture voxels, dimensionless; N is the total number of voxels in the core sample, dimensionless.

10. A computer-storable computing medium having a computer program, characterized in that: The computer program can implement the method according to any one of claims 1 to 9.

Images