A feather-shaped induced slit identification method based on image processing

Abstract

The present application belongs to the technical field of image data processing, and relates to a feather-shaped induced crack identification method based on image processing, comprising the following steps: S1, obtaining a basic gray image; S2, obtaining an ROI region image; S3, based on the ROI region image, performing image enhancement processing to obtain a crack skeleton image; S4, according to the crack skeleton image, performing skeleton extraction to obtain a global skeleton image and a best symmetry axis; S5, according to the global skeleton image, obtaining a plurality of skeleton point clusters through clustering; based on the best symmetry axis, performing sinusoidal curve fitting on each skeleton point cluster to obtain an optimal fitting curve of each skeleton point cluster; and S6, obtaining a feather-shaped induced crack image. The present application selects an optimal channel as a basic gray image, extracts a skeleton after image enhancement of the basic gray image, and performs sinusoidal curve fitting on the skeleton to obtain an accurate feather-shaped induced crack image, thereby solving the problem of insufficient precision of automatic identification of the feather-shaped induced crack.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image data processing, and discloses a feather-shaped induced fracture identification method based on image processing. BACKGROUND

[0002] Under the background of continuous deepening of global oil and gas exploration and development to unconventional reservoirs such as shale and tight sandstone, a high-conductivity artificial fracture network is formed in the reservoir through fracturing reconstruction, and the geometric complexity and spatial distribution range directly determine the final productivity of the oil and gas well. In the artificial fracture network, the feather-shaped induced fracture, as a secondary micro-fracture system with quasi-periodic distribution characteristics along the wall of the main fracture, can significantly expand the reservoir reconstruction volume and effectively communicate the matrix pores, which is crucial for comprehensive evaluation of fracturing effect and prediction of oil and gas production. Due to the inherent characteristics of the feather-shaped induced fracture such as high angle, short interval, incomplete shape, variable topology and strong background interference, the feather-shaped induced fracture often appears as a series of discrete, short and discontinuous textures arranged in a nearly sinusoidal curve in the electrical imaging logging image. Therefore, it is a great challenge to accurately identify the feather-shaped induced fracture through electrical imaging logging technology.

[0003] With the development of computer vision technology, currently, the electrical imaging logging technology usually uses gradient operators such as Sobel operator, Canny algorithm and Laplacian operator to extract the fracture edge, but for the characteristics of the feather-shaped induced fracture such as short interval, multi-branch, edge blur and serious background noise interference, the automatic identification accuracy of the feather-shaped induced fracture is insufficient. SUMMARY

[0004] The purpose of the present application is to provide a feather-shaped induced fracture identification method based on image processing, which solves the problem of insufficient automatic identification accuracy of the feather-shaped induced fracture.

[0005] The specific scheme of the present application is as follows:

[0006] A feather-shaped induced fracture identification method based on image processing, comprising the following steps:

[0007] S1, pre-processing the obtained electrical imaging logging image to obtain a basic gray scale image;

[0008] S2, obtaining an ROI region image through a global configuration system and a YOLOv11s model according to the basic gray scale image;

[0009] S3, performing image enhancement processing based on the ROI region image to obtain a fracture skeleton image;

[0010] S4, performing skeleton extraction according to the fracture skeleton image to obtain a global skeleton image and a best symmetry axis;

[0011] S5, obtaining a plurality of skeleton point clusters by clustering according to the global skeleton image;

[0012] Based on the best symmetry axis, sinusoidal curve fitting is performed on each skeleton point cluster to obtain an optimal fitting curve of each skeleton point cluster;

[0013] S6, based on the optimal fitting curve, the redundant curve is merged to obtain an effective fitting curve, and the effective fitting curve is superimposed on the background of the electrical imaging logging image to obtain a feathered induced fracture image.

[0014] Preferably, obtaining the basic gray image comprises:

[0015] S11, acquiring an electrical imaging logging image;

[0016] S12, according to the electrical imaging logging image, a plurality of channel images are obtained by a channel separation algorithm;

[0017] S13, based on each channel image and the obtained standard gray contrast, each gray contrast improvement rate index is obtained;

[0018] S14, according to each channel image and the obtained standard edge density, each edge density improvement rate index is obtained;

[0019] S15, based on each gray contrast improvement rate index and each edge density improvement rate index, the optimal channel image is selected as the basic gray image.

[0020] Preferably, obtaining each gray contrast improvement rate index comprises:

[0021] Based on each channel image, the average gray of each channel image is calculated to obtain each channel average gray, and the standard deviation of each channel image is calculated according to each channel average gray to obtain each channel gray contrast, and each gray contrast improvement rate index is obtained based on each channel gray contrast and the obtained standard gray contrast.

[0022] Preferably, obtaining each edge density improvement rate index comprises:

[0023] According to each channel image, the channel edge density of each channel image is calculated by an edge pixel indicator function to obtain each channel edge density;

[0024] Based on each channel edge density and the obtained standard edge density, each edge density improvement rate index is obtained.

[0025] Preferably, obtaining the ROI region image comprises:

[0026] According to the basic gray image, the system automatically calls the corresponding label file of the basic gray image through global configuration, parses and extracts the spatial coordinates and class number of the detection box from the label file; based on the spatial coordinates and class number of the detection box, the YOLOv11s model is used to screen and locate the candidate feather-shaped induced joint area in the basic gray image, and the ROI region image is generated by superimposing the detection box on the basic gray image according to the candidate feather-shaped induced joint area.

[0027] Preferably, obtaining the crack skeleton image comprises:

[0028] S31, obtaining an enhanced image through an adaptive histogram equalization algorithm based on the ROI region image;

[0029] S32, obtaining a filtered image through a median filter algorithm according to the enhanced image;

[0030] S33, obtaining a binary image through an adaptive threshold separation algorithm according to the filtered image, and obtaining a crack skeleton image through a morphological skeletonization algorithm based on the binary image.

[0031] Preferably, obtaining the enhanced image comprises:

[0032] Obtaining a cumulative distribution function through an adaptive histogram equalization algorithm based on the ROI region image, and obtaining the enhanced image through histogram truncation by a contrast limiting factor according to the cumulative distribution function.

[0033] Preferably, obtaining the global skeleton image and the best symmetry axis comprises:

[0034] S41, obtaining a left branch backbone image and a right branch backbone image according to the crack skeleton image through a spatial continuity criterion and a multi-scale length threshold filtering mechanism;

[0035] S42, constructing a global skeleton image through skeleton integration based on the left branch backbone image and the right branch backbone image, and obtaining the best symmetry axis through a mirror mapping mechanism according to the global skeleton image.

[0036] Preferably, obtaining the best symmetry axis comprises:

[0037] Integrating and processing the skeleton point set in the standard coordinate system to construct a global skeleton image by the left branch backbone image and the right branch backbone image;

[0038] Meanwhile, a plurality of groups of symmetry axes are preset, a mirror mapping mechanism is adopted, each group of symmetry axes is traversed through a grid search method, a symmetry index under each group of symmetry axes is calculated, and the symmetry axis corresponding to the maximum symmetry index is selected as the best symmetry axis.

[0039] Preferably, obtaining the optimal fitting curve of each skeleton point cluster comprises:

[0040] According to the spatial continuity and proximity criterion, a skeleton pixel point set in a global skeleton image is clustered to obtain a plurality of skeleton point clusters;

[0041] The position distribution information of each skeleton point cluster relative to the best symmetry axis is estimated according to a sinusoidal curve fitting formula to obtain initial fitting parameters of the sinusoidal curve, and the initial fitting curve of each skeleton point cluster is obtained by initial curve fitting of each skeleton point cluster according to the initial fitting parameters of the sinusoidal curve.

[0042] Based on the initial fitting curve, the optimal fitting parameters of the sinusoidal curve are solved by a nonlinear least squares algorithm for each skeleton point cluster, and the optimal fitting curve of each skeleton point cluster is obtained by fitting according to the optimal fitting parameters of the sinusoidal curve.

[0043] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0044] 1、The present application divides the electrical imaging logging image into multiple channel images through a channel separation algorithm, selects the optimal channel image as the basic gray image by calculating the gray contrast improvement rate index and the edge density improvement rate index of each channel image, ensures that the texture details, gray gradient and edge structure of the feathered induced fracture are retained to the greatest extent, selects the candidate feathered induced fracture area through the YOLOv11s model and superimposes the detection frame to obtain the ROI area image, accurately represents the spatial distribution area of the feathered induced fracture, and enhances the ROI area image through the adaptive histogram equalization algorithm, the value filtering algorithm, the adaptive threshold separation algorithm and the morphological skeletonization algorithm to obtain the fracture skeleton image, enhances the texture features and structural edges of the feathered induced fracture, and improves the expression quality of the fracture features, image coherence, fracture edges, main structure and branch features; the skeleton of the fracture skeleton image is accurately extracted through the spatial continuity criterion, the multi-scale length threshold filtering mechanism and the mirror image mapping mechanism to obtain the global skeleton image and the best symmetry axis, realizes noise suppression and feature retention, accurately obtains the quantitative parameters such as amplitude, angular frequency, phase and translation of the fitting curve through clustering and nonlinear least squares algorithm, so that the distribution rule of the optimal fitting curve in the coordinate space is highly consistent with the geometric shape of the feathered induced fracture, and the accurate feathered induced fracture image is obtained through the merging of the optimal fitting curve, which clearly shows the overall geometric structure of the feathered induced fracture, thereby improving the automatic recognition accuracy of the feathered induced fracture. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 The flowchart of the feathered induced fracture recognition method based on image processing in embodiment 1 of the present application.

[0046] Figure 2 The flowchart of obtaining the basic gray image in embodiment 1 of the present application.

[0047] Figure 3 Flow chart for obtaining the fracture skeleton image in Embodiment 1 of the present application. DETAILED DESCRIPTION

[0048] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0049] Embodiment 1

[0050] An image processing-based feathered induced fracture identification method, as shown in FIG. 1, comprises the following steps: Figure 1

[0051] S1, pre-processing the acquired electrical imaging logging image to obtain a basic grayscale image;

[0052] As shown in FIG. 2, obtaining the basic grayscale image comprises: Figure 2

[0053] S11, acquiring the electrical imaging logging image.

[0054] The electrical imaging logging image database is established by a global configuration system, and the electrical imaging logging image is acquired from a preset path based on the electrical imaging logging image database through a batch processing framework.

[0055] The global configuration system comprises a surface instrument, a cable transmission, a downhole instrument, an interpretation workstation and the like; the acquired electrical imaging logging image is a color image and contains a feathered induced fracture.

[0056] S12, obtaining a plurality of channel images through a channel separation algorithm based on the electrical imaging logging image.

[0057] The channel separation algorithm is to use the cv2.split() function of OpenCV to split the multi-channel image into independent single-channel images. represents the number of rows of the electrical imaging logging image, represents the number of columns of the electrical imaging logging image.

[0058] The channel separation algorithm is to use the cv2.split() function of OpenCV to split the multi-channel image into independent single-channel images.

[0059] S13, obtaining each grayscale contrast enhancement rate index based on each channel image and the acquired standard grayscale contrast.

[0060] ​​​The average gray scale of each channel image is calculated based on each channel image, and the standard deviation of each channel image is calculated according to the average gray scale of each channel, to obtain the gray scale contrast of each channel, and the gray scale contrast enhancement rate index of each gray scale is obtained based on the gray scale contrast of each channel and the standard gray scale contrast.

[0061] The calculation formula of the channel average gray scale is:

[0062]

[0063] The channel average gray scale is denoted as , wherein, n represents the number of rows of the electrical imaging logging image, m represents the number of columns of the electrical imaging logging image, The channel gray scale component of the i-th row and the j-th column is denoted as , , , R represents a red channel, G represents a green channel, and B represents a blue channel.

[0064] The calculation formula of the channel gray scale contrast is:

[0065]

[0066] The channel gray scale contrast is denoted as , wherein, n represents the number of rows of the electrical imaging logging image, m represents the number of columns of the electrical imaging logging image, The channel gray scale component of the i-th row and the j-th column is denoted as , and the channel average gray scale is denoted as .

[0067] The channel average gray scale and the channel gray scale contrast are used to reflect the ability of the overall contrast of the electrical imaging logging image, and can be used to judge the potential of the fracture detail retention in the electrical imaging logging image.

[0068] The calculation formula of the gray scale contrast enhancement rate index is:

[0069]

[0070] The gray scale contrast enhancement rate index is denoted as , wherein, the channel gray scale contrast is denoted as , and the standard gray scale contrast is denoted as .

[0071] The gray scale contrast enhancement rate index represents the percentage improvement in contrast relative to the standard gray scale.

[0072] S14, according to each channel image and the obtained standard edge density, the edge density enhancement rate index of each edge is obtained.

[0073] ​​​According to each channel image, each channel edge density is obtained by calculating the channel edge density of each channel image through an edge pixel indication function; and each edge density improvement rate index is obtained based on the channel edge density and the standard edge density.

[0074] The expression of the edge pixel indication function is:

[0075]

[0076] wherein, is the edge pixel indication function.

[0077] The calculation formula of the channel edge density is:

[0078]

[0079] wherein, is the channel edge density, represents the number of rows of the electrical imaging logging image, represents the number of columns of the electrical imaging logging image, is the edge pixel indication function.

[0080] The calculation formula of the edge density improvement rate index is:

[0081]

[0082] wherein, is the edge density improvement rate index, is the channel edge density, is the standard edge density.

[0083] The edge density improvement rate index represents the percentage improvement in texture preservation relative to the standard gray scale image.

[0084] S15, based on each gray contrast improvement rate index and each edge density improvement rate index, an optimal channel image is selected as a basic gray scale image.

[0085] The channel image in which the gray contrast improvement rate index and the edge density improvement rate index of each channel image are both maximum is selected as the basic gray scale image.

[0086] The gray contrast improvement rate index and each edge density improvement rate index are used to quantitatively evaluate the performance of different channels in retaining the characteristics of the fracture edge, so as to ensure that the texture details, gray gradient and edge structure of the feathered induced fracture are retained to the greatest extent, and provide an accurate image basis for subsequent feathered induced fracture identification.

[0087] S2, according to the basic gray scale image, a ROI region image is obtained through a global configuration system and a YOLOv11s model;

[0088] ​​​According to the basic gray image, the system automatically calls the corresponding label file of the basic gray image through global configuration, and parses and extracts the spatial coordinates and class number of the detection box from the label file; based on the spatial coordinates and class number of the detection box, the YOLOv11s model is used to screen and locate the candidate feather-shaped induced joint area in the basic gray image, and the ROI region image is generated by superimposing the detection box on the basic gray image according to the candidate feather-shaped induced joint area.

[0089] The ROI region image can accurately depict the spatial distribution characteristics of the feather-shaped induced joint and carry the class identifier, providing accurate input basis for subsequent image enhancement and skeleton extraction analysis.

[0090] S3, based on the ROI region image, performing image enhancement processing to obtain a crack skeleton image;

[0091] As shown in Figure 3 , obtaining the crack skeleton image comprises:

[0092] S31, based on the ROI region image, obtaining an enhanced image through an adaptive histogram equalization algorithm.

[0093] Based on the ROI region image, the ROI region image is defined as , wherein represents a pixel coordinate, the cumulative distribution function is obtained through the adaptive histogram equalization algorithm, and the enhanced image is obtained through histogram truncation according to the cumulative distribution function and a contrast limiting factor .

[0094] The expression of the cumulative distribution function is:

[0095] ,

[0096] wherein, is the cumulative distribution function; is a step function, which is 1 when the gray value of the pixel in the ith row and the jth column is less than or equal to the preset gray level , otherwise 0; is the total number of pixels in each local region , and the local region refers to a block region centered on .

[0097] The calculation formula of the histogram truncation is:

[0098] ,

[0099] wherein, is the histogram truncation of the gray level g; is the contrast limiting factor, which can be 2.0; is the cumulative distribution function; is the total number of gray levels, usually = 256, the preset gray level .

[0100] The calculation formula of the enhanced image is:

[0101] ,

[0102] wherein, is the enhanced image, is the ROI region image, is the histogram truncation, is the total number of gray levels.

[0103] The local gray contrast of the ROI region image is effectively improved by the adaptive histogram equalization algorithm, the texture features and structural edges of the feather-shaped induced seam are enhanced, and the expression quality of the seam features is improved.

[0104] S32, a filtered image is obtained by a median filtering algorithm according to the enhanced image.

[0105] The calculation formula of the filtered image is:

[0106] ,

[0107] wherein, is the filtered image; is the enhanced image; represents a filtering window region centered at , that is, the image noise is dynamically adjusted, and .

[0108] By the median filtering algorithm, the local neighborhood median of each pixel point is used as a replacement value, and finally the filtered image obtained can effectively smooth the high-frequency interference, maintain the continuity of the crack main edge, improve the image coherence of the crack main stem and the expression quality of the broken edge, and further suppress the influence of isolated noise points in the image on the skeleton extraction result.

[0109] S33, a binary image is obtained by an adaptive threshold segmentation algorithm according to the filtered image, and a crack skeleton image is obtained by a morphological skeletonization algorithm based on the binary image.

[0110] The adaptive threshold segmentation algorithm is an image processing technology that dynamically calculates the threshold value according to the local region features of the image. It analyzes the local neighborhood information of each pixel in the image, such as gray scale distribution, brightness change, etc., dynamically calculates the best threshold value of the region, and does not use a global uniform threshold value, which can reduce the influence of light changes or shadows on the segmentation result and improve the segmentation accuracy.

[0111] The crack region is represented as a high-contrast black region by the adaptive threshold segmentation algorithm, and is clearly separated from the background.

[0112] The binary image is converted into a single-pixel-wide centerline structure by a morphological skeletonization algorithm to extract the main trunk profile and spatial trend of the feather-shaped induced crack region, and a crack skeleton image is obtained.

[0113] The morphological skeletonization algorithm effectively eliminates isolated noise points, maintains the coherence of the main crack structure, significantly reduces background interference, and clearly presents the main trunk structure and branch characteristics of the feather-shaped induced crack.

[0114] S4, obtaining a global skeleton image and a best symmetry axis according to the crack skeleton image;

[0115] S41, obtaining a left branch skeleton image and a right branch skeleton image according to the crack skeleton image through a spatial continuity criterion and a multi-scale length threshold filtering mechanism.

[0116] The spatial continuity criterion and the multi-scale length threshold filtering mechanism are used to quantitatively analyze and filter the skeleton structure in the crack skeleton image, accurately identify and extract the main branch skeleton, suppress small bifurcations and noise interference, and obtain the left branch skeleton image and the right branch skeleton image.

[0117] The spatial continuity criterion is one of the core principles of natural zoning, which requires that the natural regions divided in space remain complete and continuous, and cannot be divided or repeated. This principle, together with the principles of uniformity and relative consistency, forms the methodological framework of natural zoning, and the regional division is realized through sequential division and sequential merging.

[0118] The multi-scale length threshold filtering mechanism is a kind of image or point cloud filtering method combining multi-scale analysis and threshold processing, which realizes noise suppression and feature preservation by setting different scale thresholds.

[0119] S42, based on the left branch skeleton image and the right branch skeleton image, the skeleton is integrated to construct a global skeleton image, and the best symmetry axis is obtained through a mirror mapping mechanism according to the global skeleton image.

[0120] The left branch skeleton image and the right branch skeleton image are integrated to construct a global skeleton image in the standard coordinate system. At the same time, a plurality of symmetry axes are preset, and a mirror mapping mechanism is adopted to traverse each symmetry axis by a grid search method, and calculate the symmetry index of each symmetry axis. The symmetry axis corresponding to the maximum symmetry index is selected as the best symmetry axis of the feather-shaped induced crack structure in the global skeleton image.

[0121] The calculation formula of the symmetry index is:

[0122] ,

[0123] wherein, S is a symmetry index, represents the number of skeleton points of the left branch bone stem image coinciding with the right branch bone stem image after mirroring the left branch bone stem image, is the total number of skeleton points of the right branch bone stem image, is the total number of skeleton points of the right branch bone stem image.

[0124] S5, obtaining a plurality of skeleton point clusters from the global skeleton image by clustering; based on the best symmetry axis, performing sinusoidal curve fitting on each skeleton point cluster to obtain an optimal fitting curve of each skeleton point cluster;

[0125] The skeleton pixel point set in the global skeleton image is clustered according to the spatial continuity and proximity criterion to obtain a plurality of skeleton point clusters, each skeleton point cluster approximately representing a section of the main stem of the crack in the circumferential direction; the position distribution information of each skeleton point cluster relative to the best symmetry axis is estimated according to the sinusoidal curve fitting formula to obtain initial fitting parameters of the sinusoidal curve; according to the initial fitting parameters of the sinusoidal curve, initial curve fitting is performed on each skeleton point cluster to obtain an initial fitting curve of each skeleton point cluster; based on the initial fitting curve, the optimal fitting parameters of the sinusoidal curve are solved by a nonlinear least squares algorithm for the objective function for each skeleton point cluster; according to the optimal fitting parameters of the sinusoidal curve, the fitting of each skeleton point cluster is performed to obtain the optimal fitting curve of each skeleton point cluster.

[0126] Clustering is a process of dividing a data set into multiple classes or clusters composed of similar objects, the purpose of which is to maximize the similarity of objects within the same cluster and maximize the difference between different clusters.

[0127] The sinusoidal curve fitting formula is:

[0128] ,

[0129] wherein, y is a sinusoidal curve, A represents the amplitude, represents the angular frequency, is the phase, b represents the translation.

[0130] The initial fitting parameters of the sinusoidal curve are used for subsequent nonlinear optimization fitting, which can improve the convergence efficiency and fitting accuracy.

[0131] The expression of the objective function is:

[0132] ,

[0133] wherein, L is a least squares error function, i.e., the sum of squared errors between all skeleton points and the fitted sinusoidal curve; A represents the amplitude; Indicates angular frequency; For phase; b Represents the translation amount; For the first point cluster in the skeleton The coordinates of n pixels, i = 1, 2, ..., n.

[0134] The optimal fitting curve obtained by fitting the optimal fitting parameters of the sine curve solved by the nonlinear least squares algorithm effectively eliminates local fluctuations, enhances the smoothness and continuity of the fitting curve, initially improves the redundancy of the fitting curve, preserves the spatial distribution characteristics of the skeleton point clusters, and makes the distribution law of the optimal fitting curves of each skeleton point cluster in the coordinate space highly consistent with the geometric shape of the feather-shaped induced slit.

[0135] In solving for the optimal fitting parameters of the sine curve, error convergence constraints are applied to eliminate invalid results that do not converge or have excessively large errors, ensuring the physical rationality and structural interpretability of the fitting expression.

[0136] Error convergence constraints include: setting the maximum number of iterations, the convergence threshold for parameter updates and changes in the objective function during the fitting process. The maximum number of iterations can be 10, and the convergence threshold can be 0.5.

[0137] S6. Based on the optimal fitting curve, redundant curves are merged to obtain the effective fitting curve. The effective fitting curve is then superimposed on the background of the electrical imaging logging image to obtain the feather-shaped induced fracture image.

[0138] Based on the optimal fitted curve, the geometric consistency between curve pairs is measured by similarity quantification index. Under the premise of satisfying the fitting error constraint, the original curve segments are merged to obtain the effective fitted curve, avoiding the repetitive expression caused by multiple fittings. The similarity quantification index includes parameter difference degree, fitting residual difference, etc. During the fitting process, the fitting parameters, matching index and error statistics of each fitted curve are stored in a standard structure.

[0139] The fitting error constraint refers to the following: before merging the segments of the same origin curves, the optimal fitting curve is checked for errors. The normalized mean square error must be lower than a first error threshold, which can be 0.15, and the coefficient of determination R0 must be within a certain range. 2 The error exceeds the second error threshold, which can be 0.6. This occurs when the normalized mean square error is below the first error threshold and the coefficient of determination R is also satisfied. 2 Only curves exceeding the second error threshold are considered valid and included in the merging process; otherwise, they are discarded to ensure the accuracy and reliability of the final results.

[0140] The formula for calculating the coefficient of determination is:

[0141] ,

[0142] wherein, R2is a correlation coefficient; the y-coordinate of the i-th actual observation value, i.e. the fracture skeleton point; the predicted value of the fitting curve at the corresponding ; the mean of the observed values.

[0143] Structural screening and merging of redundant curve segments or overlapping curve segments in the optimal fitting curve not only reduces redundancy, but also maintains the complete expression of key geometric features, effectively reduces the redundancy of the fitting curve, improves the accuracy and uniqueness of the expression of the fracture geometry, and clearly shows the overall geometric structure of the plume-induced fracture.

[0144] The electrical imaging logging image is divided into multiple channel images through a channel separation algorithm, the optimal channel image is selected as the basic gray image by calculating the gray contrast improvement rate index and the edge density improvement rate index of each channel image, ensuring that the texture details, gray gradient and edge structure of the plume-induced fracture are retained to the greatest extent; the YOLOv11s model is used to screen the candidate plume-induced fracture area and superimpose the detection frame to obtain the ROI region image, which accurately represents the spatial distribution area of the plume-induced fracture; the adaptive histogram equalization algorithm, value filtering algorithm, adaptive threshold separation algorithm and morphological skeletonization algorithm are used to enhance the image of the ROI region to obtain the fracture skeleton image, enhance the texture features and structural edges of the plume-induced fracture, and improve the expression quality of the fracture features, image coherence, fracture edges, main structure, branch features, etc.; the spatial continuity criterion, multi-scale length threshold filtering mechanism and mirror mapping mechanism are used to accurately extract the skeleton of the fracture skeleton image to obtain the global skeleton image and the best symmetry axis, realizing noise suppression and feature preservation; the clustering and nonlinear least squares algorithm are used to accurately obtain the quantitative parameters such as amplitude, angular frequency, phase and translation of the fitting curve, so that the distribution rule of the optimal fitting curve in the coordinate space is highly consistent with the geometric morphology of the plume-induced fracture; the accurate plume-induced fracture image is obtained by merging the optimal fitting curves, which clearly shows the overall geometric structure of the plume-induced fracture; thereby improving the automatic recognition accuracy of the plume-induced fracture.

[0145] Example 2

[0146] The plume-induced fracture identification method for the X block of a certain oilfield includes the following steps:

[0147] S1, obtaining the electrical imaging logging image of the X block of a certain oilfield.

[0148] The global configuration system can preset key parameters, including image storage path, processing mode, performance statistics control, etc. The global configuration system is used to build an electrical imaging logging image database of an X area in an oilfield; and based on the electrical imaging logging image database, the electrical imaging logging images containing the feather-shaped induced fractures are obtained from the preset path through the batch processing framework.

[0149] S2, a plurality of channel images are obtained from the electrical imaging logging image through a channel separation algorithm, and based on each channel image, a gray contrast improvement rate index and an edge density improvement rate index of each channel image are calculated, and an optimal channel image is selected as a basic gray image;

[0150] The red channel image, the green channel image and the blue channel image are obtained from the electrical imaging logging image through the separation algorithm; the gray contrast improvement rate index and the edge density improvement rate index of the red channel image, the green channel image and the blue channel image are calculated based on the red channel image, the green channel image and the blue channel image respectively, and the gray contrast improvement rate index of the red channel image is 0.9%, the edge density improvement rate index is -2.7%, the gray contrast improvement rate index of the green channel image is 4.1%, the edge density improvement rate index is 4.4%, the gray contrast improvement rate index of the blue channel image is -2.8%, and the edge density improvement rate index is 4.2%. It can be seen that the gray contrast improvement rate index and the edge density improvement rate index of the green channel image are the maximum values, which indicates that the green channel is the optimal channel, so the green channel image is taken as the basic gray image.

[0151] S3, according to the basic gray image, the ROI region image is obtained through the global configuration system and the YOLOv11s model;

[0152] According to the basic gray image, the global configuration system automatically calls the label file corresponding to the green channel image, parses and extracts the spatial coordinates and class numbers of the detection frame from the label file; based on the spatial coordinates and class numbers of the detection frame, the YOLOv11s model is used to screen and locate the candidate feather-shaped induced fracture region in the green channel image, and the ROI region image is generated by superimposing the detection frame on the candidate feather-shaped induced fracture region in the green channel image.

[0153] The ROI region image can accurately depict the spatial distribution characteristics of the feather-shaped induced fracture and carry the class identification, providing an accurate input basis for subsequent image enhancement and skeleton extraction analysis.

[0154] S4, based on the ROI region image, an enhanced image is obtained through an adaptive histogram equalization algorithm, a filtered image is obtained through a median filter algorithm according to the enhanced image, a binary image is obtained through an adaptive threshold segmentation algorithm based on the filtered image, a crack skeleton image is obtained through a morphological skeleton algorithm according to the binary image, and left branch skeleton images and right branch skeleton images are obtained through a spatial continuity criterion and a multi-scale length threshold filtering mechanism according to the crack skeleton image;

[0155] From the left branch skeleton image and the right branch skeleton image, it can be seen that the main path of the crack and its bifurcation relationship are significantly restored, and part of the short branch and non-structural noise line are included; the stray branch line is effectively removed, the structure main stem is continuous and complete in shape, and has a good geometric expression basis.

[0156] S5, based on the left branch skeleton image and the right branch skeleton image, a global skeleton image is constructed by integrating the skeleton, and the best symmetry axis is obtained through a mirror mapping mechanism according to the global skeleton image;

[0157] The left branch skeleton image and the right branch skeleton image are integrated to construct a global skeleton image under the standard coordinate system. At the same time, a plurality of symmetry axes are preset, a mirror mapping mechanism is adopted, each symmetry axis is traversed by a grid search method, and a symmetry index and a symmetry score under each symmetry axis are calculated, as shown in Table 1, the symmetry axis corresponding to the maximum values of the symmetry index and the symmetry score is selected as the best symmetry axis of the pinnate induced crack structure.

[0158] The calculation formula of the symmetry score is:

[0159] ,

[0160] Wherein, W i is the symmetry score of the i th symmetry axis; S i is the symmetry index of the i th symmetry axis, i = 1, 2, …, n; S i is the maximum value of the symmetry index in all symmetry indexes.

[0161] Table 1 symmetry index and symmetry score under each symmetry axis

[0162]

[0163] As can be seen from Table 1, when the symmetry axis x = 188.0, the corresponding symmetry index and symmetry score are both maximum values, indicating that the crack structure on both sides of the global skeleton image has the strongest geometric symmetry under the symmetry axis, and x = 188.0 is the best symmetry axis.

[0164] S6, obtaining a plurality of skeleton point clusters according to the global skeleton image through clustering; and performing sinusoidal curve fitting on each skeleton point cluster to obtain an optimal fitting curve of each skeleton point cluster based on the best symmetry axis;

[0165] The skeleton point clusters are obtained by clustering the skeleton pixel point set in the global skeleton image according to the spatial continuity and proximity criterion, each skeleton point cluster approximately represents a section of the trunk circumference of the crack; the position distribution information of each skeleton point cluster relative to the best symmetry axis is estimated according to the sinusoidal curve fitting formula to obtain initial fitting parameters of the sinusoidal curve; the initial curve fitting of each skeleton point cluster is performed according to the initial fitting parameters of the sinusoidal curve to obtain the initial fitting curve of each skeleton point cluster; the optimal fitting parameters of the sinusoidal curve are solved by the nonlinear least square algorithm based on the initial fitting curve to solve the objective function for each skeleton point cluster; and the fitting of each skeleton point cluster is performed according to the optimal fitting parameters of the sinusoidal curve to obtain the optimal fitting curve of each skeleton point cluster.

[0166] S7, performing redundant curve merging based on the optimal fitting curve to obtain an effective fitting curve, and superimposing the effective fitting curve on the background of the electro-imaging logging image to obtain a feathered induced fracture image.

[0167] Based on the optimal fitting curve, the geometric consistency between the curve pairs is measured by a similarity quantification index, and the homologous curve segments are merged to obtain the effective fitting curve under the premise of meeting the fitting error constraint.

[0168] The above is only a preferred embodiment of the present application and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An image processing-based feathered induced slit identification method, characterized in that, The method comprises the following steps: S1, preprocessing the obtained electrical imaging logging image to obtain a basic gray image; S2, obtaining an ROI region image according to the basic gray image through a global configuration system and a YOLOv11s model; S3, performing image enhancement processing based on the ROI region image to obtain a fracture skeleton image; S4, performing skeleton extraction according to the fracture skeleton image to obtain a global skeleton image and an optimal symmetry axis; The method for obtaining the global skeleton image and the optimal symmetry axis comprises: S41, obtaining left branch skeleton images and right branch skeleton images according to the fracture skeleton image through a spatial continuity criterion and a multi-scale length threshold filtering mechanism; S42, performing skeleton integration to construct a global skeleton image based on the left branch skeleton images and the right branch skeleton images, and obtaining an optimal symmetry axis through a mirror mapping mechanism according to the global skeleton image; S5, obtaining a plurality of skeleton point clusters through clustering according to the global skeleton image; and performing sinusoidal curve fitting on each skeleton point cluster based on the optimal symmetry axis to obtain an optimal fitting curve of each skeleton point cluster; S6, performing redundant curve merging based on the optimal fitting curve to obtain an effective fitting curve, and superimposing the effective fitting curve on the background of the electrical imaging logging image to obtain a feathered induced fracture image.

2. The image processing-based feather joint identification method according to claim 1, wherein The method for obtaining the basic gray image comprises: S11, obtaining an electrical imaging logging image; S12, obtaining a plurality of channel images through a channel separation algorithm according to the electrical imaging logging image; S13, obtaining each gray contrast improvement rate index based on each channel image and an obtained standard gray contrast; S14, obtaining each edge density improvement rate index according to each channel image and an obtained standard edge density; S15, selecting an optimal channel image as the basic gray image based on each gray contrast improvement rate index and each edge density improvement rate index.

3. The image processing-based feather joint identification method according to claim 2, wherein, The method for obtaining each gray contrast improvement rate index comprises: calculating the average gray of each channel image based on each channel image to obtain each channel average gray, calculating the standard deviation of each channel image according to each channel average gray to obtain each channel gray contrast, and obtaining each gray contrast improvement rate index based on each channel gray contrast and an obtained standard gray contrast.

4. The image processing-based feather joint identification method according to claim 2, wherein The method for obtaining each edge density improvement rate index comprises: calculating the channel edge density of each channel image through an edge pixel indicator function according to each channel image to obtain each channel edge density, and obtaining each edge density improvement rate index based on each channel edge density and an obtained standard edge density.

5. The image processing based feather joint identification method of claim 1, wherein, The method for obtaining the ROI region image comprises: automatically calling a label file corresponding to the basic gray image through a global configuration system according to the basic gray image, parsing and extracting the spatial coordinates and the category number of the detection frame from the label file, screening and positioning a candidate feathered induced fracture region in the basic gray image through a YOLOv11s model based on the spatial coordinates and the category number of the detection frame, and generating an ROI region image according to the superposition of the detection frame on the basic gray image.

6. The image processing-based feather joint identification method according to claim 1, wherein, The method for obtaining the fracture skeleton image comprises: S31, obtaining an enhanced image through an adaptive histogram equalization algorithm based on the ROI region image; S32, obtaining a filtered image through a median filter algorithm according to the enhanced image; S33, obtaining a binary image according to the filtered image through an adaptive threshold segmentation algorithm, and obtaining a crack skeleton image through a morphological skeletonization algorithm based on the binary image.

7. The image processing-based feather joint identification method according to claim 6, wherein, The obtaining of the enhanced image includes: Obtaining a cumulative distribution function through an adaptive histogram equalization algorithm based on the ROI region image, and obtaining the enhanced image through histogram truncation according to a contrast limiting factor based on the cumulative distribution function.

8. The image processing-based feather joint identification method of claim 1, wherein, The obtaining of the optimal symmetry axis includes: Integrating the left branch bone image and the right branch bone image to process the skeleton point set in a standard coordinate system to construct a global skeleton image; Meanwhile, a plurality of groups of symmetry axes are preset, a mirror mapping mechanism is adopted, each group of symmetry axes is traversed through a grid search method, a symmetry degree index under each group of symmetry axes is calculated, and a symmetry axis corresponding to a maximum value of the symmetry degree index is selected as the optimal symmetry axis.

9. The image processing-based feather joint identification method according to claim 1, wherein, The obtaining of the optimal fitting curve of each skeleton point cluster includes: Clustering skeleton pixel points in the global skeleton image according to spatial continuity and proximity criteria to obtain a plurality of skeleton point clusters; According to a sinusoidal curve fitting formula, initial fitting parameters of a sinusoidal curve are estimated according to position distribution information of each skeleton point cluster relative to the optimal symmetry axis, initial curve fitting of each skeleton point cluster is performed according to the initial fitting parameters of the sinusoidal curve to obtain an initial fitting curve of each skeleton point cluster; Based on the initial fitting curve, optimal fitting parameters of the sinusoidal curve are solved for a target function through a nonlinear least square algorithm for each skeleton point cluster, and fitting of each skeleton point cluster is performed according to the optimal fitting parameters of the sinusoidal curve to obtain an optimal fitting curve of each skeleton point cluster.

Images