CN114998876A - Sea-land transition phase shale streak layer structure identification method based on rock slice image - Google Patents

Sea-land transition phase shale streak layer structure identification method based on rock slice image Download PDF

Info

Publication number
CN114998876A
CN114998876A CN202210594497.9A CN202210594497A CN114998876A CN 114998876 A CN114998876 A CN 114998876A CN 202210594497 A CN202210594497 A CN 202210594497A CN 114998876 A CN114998876 A CN 114998876A
Authority
CN
China
Prior art keywords
image
frequency domain
distribution
principal component
rock
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210594497.9A
Other languages
Chinese (zh)
Inventor
李佳航
李玮
刘向君
周路
吴丰
吴勇
梁利喜
熊健
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest Petroleum University
Original Assignee
Southwest Petroleum University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southwest Petroleum University filed Critical Southwest Petroleum University
Priority to CN202210594497.9A priority Critical patent/CN114998876A/en
Publication of CN114998876A publication Critical patent/CN114998876A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
    • G06V10/431—Frequency domain transformation; Autocorrelation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a sea-land transition phase shale streak layer structure identification method based on a rock flake image, which mainly comprises the following steps: firstly, performing gray level transformation, Fourier transformation and frequency spectrum characteristic extraction on a rock slice image; then, carrying out principal component analysis on a point set in the frequency domain image, wherein a first principal component corresponds to a long axis in the distribution of the point set, and a second principal component corresponds to a short axis in the distribution of the point set; respectively projecting data points on the frequency domain graph onto the first principal component axis and the second principal component axis by taking the first principal component and the second principal component as coordinate axes to obtain histograms of the frequency domain graph on the two coordinate axes; the ratio of the widths of the two histograms quantifies the morphological characteristics of the data point distribution on the representation frequency domain image; the smaller the ratio, the more obvious the corresponding texture in the original image. The method directly obtains the structural characteristics of the rock core through the frequency domain characteristic analysis of the digital image without image segmentation, and is suitable for rock cores with strong heterogeneity and difficult striae image segmentation, such as sea-land transition phase shale and the like.

Description

Sea-land transition phase shale streak layer structure identification method based on rock slice image
Technical Field
The invention relates to the technical field of oil and gas field exploitation, in particular to a sea-land transition phase shale streak layer structure identification method based on a rock slice image.
Background
The sea-land transition facies shale has large resource amount and sufficient exploration and development potential, but the deposition environment is changeable and the rock structure is complex. If the shale has a striated layer structure, the characteristics of rock physics, mechanics and the like of the shale have obvious anisotropic characteristics, and the shale is obviously different from the shales with other structures. Therefore, before rock physical modeling and reservoir evaluation, identification of the shale with the striated layer structure based on rock slice data is of great significance. Two common methods are available, one is manual identification, and the main disadvantages of the method are that the time consumption is long, the geological level of an interpreter is depended on, and the development degree of a striae is difficult to quantify. In actual engineering projects, the number of samples is often large, manual division not only wastes time and labor, but also is easy to generate wrong selection and wrong selection, and quantitative representation cannot be provided for the texture structure. The other is to identify the striae layer by image segmentation based on digital image processing techniques.
And extracting the rock lamella based on an image segmentation technology. The method specifically comprises the steps of converting a rock slice image into a gray-scale image, carrying out binarization processing on the rock slice through color segmentation, counting the number of white pixel points (namely pixel points with the amplitude value of 1) according to rows or columns, drawing a curve graph, analyzing the growth condition of a grain layer through the fluctuation of a curve, and considering that the grain layer structure exists if the peak value of the curve is locally and suddenly increased. The method has poor recognition effect and limited application range. When the grain layer is approximately parallel to the horizontal axis, the method has a good identification effect, but when an included angle exists between the grain layer structure and the horizontal axis, the identification effect is poor. In the presence of light-colored non-grained structure minerals, misrecognition is likely to occur. For example, as shown in fig. 1, although the left image of fig. 1 has a tilted stripe structure in the row [600,1000], the peak of the curve in the right image is not increased significantly, for example, if the image has a stripe structure that is tilted all the time, the recognition effect is poor. As shown in fig. 2, during threshold segmentation, the threshold value is difficult to determine, white pixel points are generated even when light-colored gravel exists in the rock slice, and the curve peaks at rows 400 and 800 in the right diagram of fig. 2 reach about 550, which is equivalent to the inclined stripe structure peak in the right diagram of fig. 1, but the rock does not have a stripe structure. In the right diagrams of fig. 1 and 2, the ordinate represents the line number of the image, and the abscissa represents the number of white pixels in each line of the binarized image. Therefore, the shale slice image has the characteristic of strong heterogeneity, and the image segmentation difficulty is high by the method.
Disclosure of Invention
The invention aims to provide a sea-land transition facies shale stripe structure identification method established on a rock slice image, aiming at the problems that the existing method for identifying the structure with the stripe layer based on the rock slice data is difficult to accurately identify the structure of the sea-land transition facies shale stripe layer.
The invention provides a sea-land transition phase shale streak layer structure identification method based on a rock flake image, which comprises the following steps:
s1, performing gray scale transformation on the rock slice image; after the gray scale conversion, the data format of the image needs to be converted into a double-precision type, and the gray scale range of the image is changed from original [0,255] to [0,1 ]. The purpose is to prevent data overflow or insufficient precision in the subsequent Fourier transform spectrum obtaining process.
S2, two-dimensional Fourier transform and spectrum feature extraction, which are specifically as follows:
and S21, performing two-dimensional Fourier transform on the rock slice image after gray level transformation to obtain a frequency spectrogram. The two-dimensional fourier transform equation is as follows:
Figure RE-GDA0003728589460000021
wherein F (x, y) is a digital image of size M N, and F (u, v) is a frequency spectrum corresponding to F (x, y); x and y represent coordinate variables of image space domain, u and v are frequency domain discrete variables, u is 0,1,2, …, M-1, upsilon is 0,1,2, …, N-1, j 2 ＝-1。
And S22, performing phase conversion on the frequency spectrogram to move the conversion origin to the center of the frequency spectrogram, then taking logarithm of the amplitude of the image, and finally performing binarization on the image by setting a threshold value to enable the frequency spectrum characteristic to be more obvious. When the image is binarized, the same threshold setting standard is adopted for each rock slice image during processing, and preferably, the threshold is 0.5 time of the maximum amplitude of the frequency spectrum pixel point.
S3, analyzing distribution characteristics of the frequency domain image: the shape of the frequency domain image is quantitatively represented by adopting the long axis and the short axis of the distribution range of the frequency domain image; the method comprises the steps of carrying out principal component analysis on a point set in a frequency domain image, and extracting a first principal component and a second principal component, wherein the first principal component corresponds to a long axis in point set distribution, and the second principal component corresponds to a short axis in point set distribution; the distribution shape of the point set is analyzed by comparing the distribution characteristics of the point set on the major axis and the minor axis.
S4, quantitatively characterizing the distribution characteristics of the frequency domain image, which is as follows:
and S41, projecting the data points on the frequency domain graph to the first main component axis and the second main component axis respectively by taking the extracted first main component and the extracted second main component as abscissa and the extracted frequency as ordinate, so as to obtain histogram distribution of the frequency domain graph on the two coordinate axes.
And S42, removing data points near the distribution boundary of the histogram, wherein the total amount of the removed data points accounts for 10% of the total data point amount, and the effect is to weaken the influence of singular points on the spectral characteristics.
S43, quantitatively characterizing the morphological characteristics of the data point distribution on the frequency domain image by adopting the ratio of the width of the data point distribution histogram on the first principal component axis to the width of the data point distribution histogram on the second principal component axis; the smaller the ratio of the width ratio is, the more the distribution form of the points in the frequency domain image tends to be elliptical, and the more obvious the corresponding texture layer structure in the original image is. Preferably, when the ratio of the widths of the two histograms is less than 0.8, it is determined that the streak layer structure exists on the rock slice image.
Compared with the prior art, the invention has the advantages that:
(1) compared with a manual identification method and a digital image analysis method based on an image segmentation technology, the method directly obtains the structural characteristics of the rock core through the frequency domain characteristic analysis of the digital image without image segmentation, is particularly suitable for rock cores with strong heterogeneity and difficult striae image segmentation, such as sea-land transition phase shale, and the like, and is also suitable for rocks with incomplete parallel bedding.
(2) The engineering investment is low. The method can effectively reduce the labor and time cost in the process of separating the rock slice grained layer structure, and meets the economic index requirements of actual engineering.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention.
Drawings
Fig. 1 is a conventional identification method.
Fig. 2 is a conventional identification method.
FIG. 3 is a sea-land transition phase shale streak layer identification diagram.
Detailed Description
The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.
The invention provides a sea-land transition phase shale streak layer structure identification method based on a rock flake image, which comprises the following steps of:
(1) gray scale transformation of rock slice images
The rock slice image is a color digital image composed of a finite number of elements, each element having a corresponding position and value, the elements being pixels, each pixel being represented by R, G, B three components, each channel taking a range of values [0,255 ]. In order to process the rock slice image conveniently, gray level conversion is firstly needed to convert the color picture into a gray level picture, so that the rock slice image can display more details, and the contrast (contrast stretching) of the image is improved. Different structures of the rock slices endow each pixel point with different gray levels, the gray level transformation can selectively highlight target features in the image or inhibit unnecessary features in the image, and the distribution of the pixels can be more uniform. After the gray scale conversion, the data format of the image needs to be converted into a double-precision type (double type), and the gray scale range of the image is changed from original [0,255] to [0,1 ]. The purpose is to prevent data overflow or insufficient precision in the subsequent Fourier transform spectrum obtaining process. The grey scale map shown in fig. 3a is obtained after grey scale conversion of the colour rock slice image.
(2) Two-dimensional Fourier transform and spectral feature extraction. The fourier transform may represent any periodic function as a sum of sine and or cosine functions (fourier series) at different frequencies. The frequency of converting the image from the space domain to the frequency domain image is an index for representing the intensity degree of gray level change in the image, and is the gradient of the gray level on a plane space, and the frequency value of the area with faster gray level conversion is higher. Thus, if a texture is present in the sheet, the image frequency is low along the texture direction and high perpendicular to the texture direction. Therefore, if the image has a streak layer, its distribution in the frequency domain will exhibit a certain directionality, as shown in fig. 3 e.
Firstly, performing two-dimensional Fourier transform on a rock slice image (figure 3a) after gray level transformation to obtain a frequency spectrogram (figure 3b), wherein the formula is as follows:
Figure RE-GDA0003728589460000031
wherein F (x, y) is a digital image of size M N, and F (u, v) is a frequency spectrum corresponding to F (x, y); x and y represent coordinate variables of an image space domain, u and v are frequency domain discrete variables, u is 0,1,2, …, M-1, upsilon is 0,1,2, …, N-1, j 2 ＝-1。
Then, the frequency spectrogram is subjected to phase conversion, and the transformation origin is moved to the center of the frequency spectrogram, so that fig. 3c is obtained. The amplitude range after fourier transform is large, which is not favorable for observing the frequency domain image characteristics, the logarithm of the amplitude of the image is required, and fig. 3d is the image after logarithmic transform.
Finally, in order to facilitate analysis of the distribution shape of the data points in the frequency domain image, the image is binarized by setting a threshold value, so that the frequency spectrum characteristic is more obvious, as shown in fig. 3 e. In order to provide reliable contrast for different images, the same threshold setting standard should be adopted for each rock slice image during processing, and the threshold is usually 0.5 times of the maximum amplitude of the frequency spectrum pixel point.
(3) And analyzing the distribution characteristics of the frequency domain image. The shape of the frequency domain image reflects the structural information of the original image, and if the image is circular or approximately circular, the obvious texture layer structure does not exist in the image. If the frequency domain image presents an ellipse, the original image is indicated to have a texture structure. And the long axis and the short axis of the distribution range of the frequency domain image are adopted to quantitatively represent the shape of the frequency domain image. The specific method comprises the following steps: and carrying out principal component analysis on the point set in the frequency domain image, and extracting a first principal component and a second principal component, wherein the first principal component corresponds to a long axis in the distribution of the point set, and the second principal component corresponds to a short axis in the distribution of the point set. The distribution shape of the point set can be effectively analyzed by comparing the distribution characteristics of the point set on the major axis and the minor axis, as shown in fig. 3 f.
(4) And (4) quantitatively characterizing the distribution characteristics of the frequency domain image. The extracted first and second principal components are used as abscissa axes, and the data points on the frequency domain graph are respectively projected onto the first and second principal component axes, so that histogram distribution (fig. 3g) of the frequency domain graph on the two coordinate axes can be obtained, wherein the F1 histogram is the distribution of the data points on the frequency domain graph on the second principal component axis (short axis), and the F2 histogram is the distribution of the data points on the frequency domain graph on the first principal component axis (long axis). In order to reduce the influence of singular points on the spectral characteristics, data points near the distribution boundary are removed, and the total amount of the removed data points accounts for about 10% of the total data points. The new histogram after eliminating the back boundary points is shown in fig. 3 h. And quantitatively characterizing the morphological characteristics of the data point distribution on the frequency domain image by adopting the ratio of the width of the data point distribution histogram on the first principal component axis to the width of the data point distribution histogram on the second principal component axis. The histogram width ratio in fig. 3h is 0.6519. The width ratio range of the histogram is between [0 and 1], and the smaller the ratio is, the more elliptical the distribution form of the points in the frequency domain image tends to be, and the more obvious the corresponding texture layer structure in the original image is. In general, when the histogram width ratio is less than 0.8, it is determined that the streak layer structure exists on the rock slice image.
Although the present invention has been described with reference to a preferred embodiment, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (6)

1. A sea-land transition phase shale streak layer structure identification method based on a rock slice image is characterized by comprising the following steps:
s1, performing gray scale transformation on the rock slice image;
s2, two-dimensional Fourier transform and spectrum feature extraction, which are specifically as follows:
s21, performing two-dimensional Fourier transform on the rock slice image after gray level transformation to obtain a frequency spectrogram;
s22, performing phase conversion on the frequency spectrogram to move the conversion origin to the center of the frequency spectrogram, then taking logarithm of the amplitude of the image, and finally performing binarization on the image by setting a threshold value to enable the frequency spectrum characteristic to be more obvious;
s3, analyzing distribution characteristics of the frequency domain image: the shape of the frequency domain image is quantitatively represented by adopting the long axis and the short axis of the distribution range of the frequency domain image; the method comprises the steps of carrying out principal component analysis on a point set in a frequency domain image, and extracting a first principal component and a second principal component, wherein the first principal component corresponds to a long axis in point set distribution, and the second principal component corresponds to a short axis in point set distribution; analyzing the distribution shape of the point set by comparing the distribution characteristics of the point set on the long axis and the short axis;
s4, quantitatively characterizing the distribution characteristics of the frequency domain image, which is as follows:
s41, projecting data points on the frequency domain graph to a first main component axis and a second main component axis respectively by taking the extracted first main component and the extracted second main component as abscissa and the extracted frequency as ordinate, and obtaining histogram distribution of the frequency domain graph on the two coordinate axes;
s42, eliminating data points near the distribution boundary of the histogram, and weakening the influence of singular points on the spectral characteristics;
s43, quantitatively characterizing morphological characteristics of data point distribution on the frequency domain image by adopting the ratio of the width of the data point distribution histogram on the first principal component axis to the width of the data point distribution histogram on the second principal component axis; the smaller the ratio of the width ratio, the more elliptical the distribution form of the points in the frequency domain image tends to be, and the more obvious the corresponding texture structure in the original image.
2. The method for identifying sea-land transition phase shale streak layer structure based on rock slice image according to claim 1, wherein in step S1, the data format of the image needs to be converted into double precision type after gray scale conversion, and the gray scale range of the image is changed from original [0,255] to [0,1 ].
3. The method for identifying a sea-land transition phase shale streak layer structure based on a rock slice image as claimed in claim 1, wherein in step S21, the two-dimensional fourier transform formula is as follows:
Figure FDA0003667236640000011
wherein F (x, y) is a digital image of size M N, and F (u, v) is a frequency spectrum corresponding to F (x, y); x and y represent coordinate variables of an image space domain, u and v are frequency domain discrete variables, u is 0,1,2, …, M-1, upsilon is 0,1,2, …, N-1, j 2 ＝-1。
4. The method for identifying a sea-land transition phase shale streak layer structure based on rock slice images as claimed in claim 3, wherein in step S22, when the images are binarized by setting a threshold, the same threshold setting standard is adopted for each rock slice image during processing, and the threshold is 0.5 times of the maximum amplitude of the frequency spectrum pixel point.
5. The method for identifying a sea-land transition phase shale streak layer structure based on rock slice images as claimed in claim 1, wherein in step S42, the total amount of the removed data points accounts for 10% of the total data points.
6. The method for identifying a sea-land transition phase shale streak layer structure based on a rock flake image according to claim 1, wherein in step S43, when the ratio of the widths of the two histograms is less than 0.8, it is determined that a streak layer structure exists on the rock flake image.
CN202210594497.9A 2022-05-27 2022-05-27 Sea-land transition phase shale streak layer structure identification method based on rock slice image Pending CN114998876A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210594497.9A CN114998876A (en) 2022-05-27 2022-05-27 Sea-land transition phase shale streak layer structure identification method based on rock slice image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210594497.9A CN114998876A (en) 2022-05-27 2022-05-27 Sea-land transition phase shale streak layer structure identification method based on rock slice image

Publications (1)

Publication Number Publication Date
CN114998876A true CN114998876A (en) 2022-09-02

Family

ID=83029995

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210594497.9A Pending CN114998876A (en) 2022-05-27 2022-05-27 Sea-land transition phase shale streak layer structure identification method based on rock slice image

Country Status (1)

Country Link
CN (1) CN114998876A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115761518A (en) * 2023-01-10 2023-03-07 云南瀚哲科技有限公司 Crop classification method based on remote sensing image data

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115761518A (en) * 2023-01-10 2023-03-07 云南瀚哲科技有限公司 Crop classification method based on remote sensing image data

Similar Documents

Publication Publication Date Title
CN115861135B (en) Image enhancement and recognition method applied to panoramic detection of box body
CN106651872B (en) Pavement crack identification method and system based on Prewitt operator
Karimpouli et al. Coal cleat/fracture segmentation using convolutional neural networks
CN108564085B (en) Method for automatically reading of pointer type instrument
CN107808138B (en) Communication signal identification method based on FasterR-CNN
CN101587189B (en) Texture elementary feature extraction method for synthetizing aperture radar images
CN109447111B (en) Remote sensing supervision classification method based on subclass training samples
CN110070545B (en) Method for automatically extracting urban built-up area by urban texture feature density
Hwang et al. A practical algorithm for the retrieval of floe size distribution of Arctic sea ice from high-resolution satellite Synthetic Aperture Radar imagery
Ebert et al. Semi-automated detection of annual laminae (varves) in lake sediments using a fuzzy logic algorithm
Bora et al. A new efficient color image segmentation approach based on combination of histogram equalization with watershed algorithm
CN109635733A (en) View-based access control model conspicuousness and the modified parking lot of queue and vehicle target detection method
CN114998876A (en) Sea-land transition phase shale streak layer structure identification method based on rock slice image
Yao et al. Expose noise level inconsistency incorporating the inhomogeneity scoring strategy
Ismailova et al. Automated drill cuttings size estimation
CN112991370B (en) Rock core CT crack identification and segmentation method
CN104637060A (en) Image partition method based on neighbor-hood PCA (Principal Component Analysis)-Laplace
CN110991463A (en) Multi-scale guided filtering feature extraction method under guide of super-pixel map
CN111275687B (en) Fine-grained image stitching detection method based on connected region marks
Khamael et al. Using adapted JSEG algorithm with fuzzy C mean for segmentation and counting of white blood cell and nucleus images
Sidike et al. Automatic building change detection through adaptive local textural features and sequential background removal
CN110310263B (en) SAR image residential area detection method based on significance analysis and background prior
CN117197092A (en) Underground coal mine image quality assessment method
CN109344837B (en) SAR image semantic segmentation method based on deep convolutional network and weak supervised learning
CN115908363A (en) Tumor cell counting method, device, equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination