A core repositioning method based on digital image intelligent identification

Abstract

This invention discloses a core relocation method based on intelligent digital image recognition, aiming to solve the problem of difficulty in determining the initial attitude of sampled cores. To this end, the core relocation method provided by this invention includes locating the coordinates of a surface borehole, collecting core samples at a target depth from the borehole, generating a columnar image of the borehole wall surface, generating a columnar image of the core surface, performing image matching using OpenCV, determining the depth value and azimuth angle at which the image feature information is most similar based on a similarity value chart, and achieving surface relocation of the unoriented core.

Description

Technical Field

[0001] This invention belongs to the field of geotechnical engineering technology, and in particular relates to a core relocation method based on intelligent digital image recognition. Background Technology

[0002] As depth increases during deep mining, in-situ stress also increases accordingly, making surrounding rock support work more difficult and geological disasters more challenging to prevent. Therefore, measuring in-situ stress becomes increasingly important, and acoustic emission (AE) is currently a primary method for this purpose. However, measuring the magnitude and direction of in-situ stress using AE requires drilling core samples, and the original orientation of the core samples must be determined to indicate the sampling direction. Current drilling core sampling technology cannot determine the original orientation of the core samples.

[0003] To address this, patent 2020109809731 discloses a method for determining the direction of geostress. This method combines wave velocity anisotropy with electrical imaging logging data to evaluate the direction of geostress in the formation, avoiding the shortcomings of electrical imaging images in accurately obtaining the location of collapse fractures and the large errors in determining the direction of geostress using paleomagnetic methods. However, the aforementioned patent determines the orientation of the core by comparing the texture features of the core roll-scan image and the dynamic and static images of the wellbore. It not only fails to provide a specific image matching process, but also determines the target depth by recording the number of core samples taken and the number of core samples contained in each sample. After the core samples are removed from the surface, logging personnel make digital marks on the core samples to confirm the target depth. This target depth determination process is not only cumbersome, but even a slight deviation can cause the core roll-scan image and the dynamic and static images of the wellbore to mismatch, thus making it impossible to determine the relative orientation of the core. Summary of the Invention

[0004] The main objective of this invention is to provide a core repositioning method based on intelligent digital image recognition, which aims to determine the absolute position of the core and solve the problem of difficulty in determining the initial attitude of the sampled core.

[0005] Therefore, the present invention provides a core relocation method based on intelligent digital image recognition, comprising:

[0006] Locate the coordinates of the surface borehole and collect rock cores at the target depth.

[0007] A columnar image of the borehole wall surface was obtained using an imaging device, and the directions of east, south, west, and north were marked on the columnar image of the borehole wall surface.

[0008] A reference line parallel to the core axis is drawn on the core surface, and then the core scanning device is used to scan the core sample surface to generate a columnar image of the core surface.

[0009] The columnar images of the borehole wall surface and the outer surface of the rock core are input into the OpenCV software for image matching.

[0010] During image matching, the reference line on the outer columnar image of the core is aligned with the north direction of the columnar image on the borehole wall surface. Then, the outer columnar image of the core is traversed in the axial and circumferential directions at intervals of set pixels on the columnar image on the borehole wall surface to obtain similarity value maps at different depths.

[0011] Find a normally distributed similarity map among several similarity maps and obtain the depth value h corresponding to the similarity map. Find the azimuth angle θ corresponding to the maximum similarity value in the normally distributed similarity map.

[0012] Based on the obtained depth value h, azimuth angle θ, and reference line on the core surface, the unoriented core is repositioned on the ground.

[0013] Specifically, select a drilling point and locate its coordinates using GPS. Collect rock cores at a depth of h1-h2.

[0014] Specifically, during image matching, the columnar image of the borehole wall surface is first locked within the depth range of h1-h2. Then, the columnar image of the core surface is input, and the core surface columnar image is traversed by sliding downward in the axial direction and rotating counterclockwise in the circumferential direction, starting from h1 of the borehole wall surface columnar image.

[0015] Specifically, the GPS positioning coordinates use the WGS-84 coordinate system to describe the locations of surface boreholes.

[0016] Specifically, to ensure that the columnar images on the surface of the rock core are not upside down, the dust on the surface of the rock core needs to be cleaned before using the scanning device.

[0017] Specifically, after core sampling is completed, the imaging equipment starts taking pictures from the borehole opening and is lowered at a set speed until it stops taking pictures at the bottom of the hole. The images of the shooting process are stored to obtain a columnar image of the borehole wall surface.

[0018] Specifically, the imaging device is a hole-punch television.

[0019] Specifically, the borehole television includes an imaging probe, a surface control terminal, and a winch. The winch is installed on the ground via a support frame, and the winch's lifting rope extends into the borehole and is connected to the imaging probe. The probe is connected to the surface control terminal via a signal line.

[0020] Specifically, the core scanning device includes a rotating base platform, a drive motor for rotating the rotating base platform, and a high-definition camera. The core is vertically fixed on the rotating base platform. The rotating base platform is also provided with a drive mechanism that drives the high-definition camera to move along the axial direction of the core to capture images of the rotating core. The high-definition camera is connected to the surface control terminal.

[0021] Specifically, the rotating base platform is also equipped with a pressure bar to press the top of the rock core.

[0022] Compared with the prior art, at least one embodiment of the present invention has the following beneficial effects:

[0023] 1. This application uses an image template matching method based on OpenCV to perform axial and circumferential image traversal on the columnar image of the rock core surface and the columnar image of the borehole wall surface at predetermined pixel intervals to obtain similarity value maps at different depths. The operator only needs to find the normally distributed map among all the obtained similarity value maps, read the depth h and azimuth angle θ from the map, and then determine the orientation of the rock core in the borehole based on the depth h and azimuth angle θ. The entire positioning process is intuitive and fast, minimizing the uncertainty of human factors and avoiding errors caused by manual judgment, and has the advantages of high efficiency and accuracy.

[0024] 2. The positioning function is accomplished by using digital image recognition technology, taking into account features such as lithological color, surface crack morphology, impurity color and distribution morphology. This comprehensive approach provides complete information, making the results more reliable and convincing. The positioning is achieved by matching columnar images, resulting in direct and easy-to-understand final results with strong spatial representation, making it convenient for operators to understand and use. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a flowchart illustrating the core relocation method provided in this embodiment of the invention.

[0027] Figure 2 This is a schematic diagram of the reference line for the columnar image of the rock core surface provided in an embodiment of the present invention.

[0028] Figure 3 This is a schematic diagram of intelligent image recognition provided in an embodiment of the present invention;

[0029] Figure 4This is a similarity value map with a normal distribution at depth h provided by an embodiment of the present invention;

[0030] Figure 5 This is a schematic diagram of the borehole television and core scanning device provided in an embodiment of the present invention;

[0031] The components include: 1. High-definition camera; 2. Housing; 3. Rotating base platform; 4. Pressure rod; 5. Rock core; 6. Display screen; 7. Surface control terminal; 8. Signal line; 9. Imaging probe; 10. Transparent protective cover; 11. Attitude sensing module; 12. Drill hole; 13. Support frame; 14. Winch. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] See Figure 1 A core relocation method based on intelligent digital image recognition includes:

[0034] Select the drilling point and locate the coordinates S(X, Y, Z) of the point using GPS. The core samples are collected at a depth of approximately h1 to h2.

[0035] After core sampling is completed, the imaging equipment starts taking pictures from the borehole opening and is lowered at a set speed until it stops taking pictures at the bottom of the hole. The images of the shooting process are stored to obtain a columnar image of the borehole wall surface (hereinafter referred to as the large image), and the east-south-west-north directions are marked on the columnar image of the borehole wall surface. The imaging equipment can be a borehole television or other imaging equipment.

[0036] See Figure 2 A reference line is lightly drawn on the surface of the rock core with a pen, and then the surface of the rock core sample is scanned with a rock core scanning device to generate a columnar image of the rock core surface (hereinafter referred to as the small image). This will facilitate subsequent image matching and help on-site operators to more easily reposition the rock core.

[0037] The columnar images of the borehole wall surface and the outer columnar image of the core were input into the software OpenCV, and the image matching module of OpenCV was used for intelligent matching. The image feature information of intelligent image matching includes lithological color, surface crack morphology, impurity color and distribution morphology, etc.

[0038] Determine the depth value h and azimuth angle θ when the image feature information is most similar based on the similarity value chart;

[0039] Based on the obtained depth value h, azimuth angle θ, and reference line on the core surface, the unoriented core is repositioned on the ground.

[0040] The specific process of determining the depth value h and azimuth angle θ when the image feature information is most similar based on the similarity value chart is as follows: When matching images, ensure that the reference line on the outer columnar image of the core is aligned with the north direction of the columnar image on the borehole wall surface. Then, allow the outer columnar image of the core to traverse the image on the borehole wall surface at intervals of set pixels in the axial and circumferential directions. After the matching is completed, similarity value charts are obtained at different depths (between h1 and h2) and at various azimuth angles.

[0041] Find a normally distributed similarity map among several similarity maps, such as... Figure 3 As shown, the depth value h corresponding to the similarity map is obtained, and the azimuth angle θ corresponding to the maximum similarity value is found in the normally distributed similarity map.

[0042] See Figure 3 The specific process for solving the depth value h and azimuth angle θ is as follows: First, lock the large image within the depth range of h1-h2, then input the small image, ensuring that the reference line of the small image aligns with the north direction of the large image. Then, starting from h1 of the large image, slide the small image downwards in the axial direction and rotate it counterclockwise in the circumferential direction. That is, perform axial and circumferential image traversal along the vertical and rotational traversal directions at predetermined intervals of pixels. After the traversal is complete, some similarity value maps can be obtained. Among the obtained similarity value maps at different depths, only at the unique depth h (i.e., when the small image matches the large image) can we obtain a normally distributed similarity value map, such as... Figure 4 As shown, the remaining similarity maps exhibit irregular variations, and the similarity values ​​are not very high. In the normally distributed similarity map at depth h, the azimuth angle θ corresponding to the maximum similarity value can be found. This value is the angle at which the small map reference line, after being rotated counterclockwise from the north direction of the large map, is most similar to the feature information of the large map. Finally, based on the obtained (h, θ), the operator can reconstruct the initial attitude of the core on-site according to the reference line of the small map.

[0043] This application utilizes an image template matching method based on OpenCV. It performs axial and circumferential image traversal on a columnar image of the core surface against a columnar image of the borehole wall surface at predetermined pixel intervals to obtain similarity value maps at different depths. Operators only need to identify the normally distributed map among all the obtained similarity value maps, read the depth *h* and azimuth angle *θ* from this map, and then determine the core's location within the borehole based on these values. The entire positioning process is intuitive and fast, minimizing the uncertainty of human factors and avoiding errors caused by manual judgment, offering advantages of high efficiency and accuracy. The positioning function employs digital image recognition technology, considering features such as lithological color, surface crack morphology, and impurity color and distribution. This comprehensive approach provides complete information, resulting in more reliable and convincing results. The use of columnar images for matching and positioning yields direct and easily understandable results with strong spatial representation, facilitating operator comprehension and use.

[0044] In this embodiment of the invention, GPS technology is used to locate the coordinates of the borehole points. The description of the surface borehole points adopts the World Geodetic Coordinate System (WGS-84). Compared with other methods, the above technology is more mature and simpler and more convenient to operate.

[0045] Specifically, the columnar image of the rock core (hereinafter referred to as the small image) must be ensured to be inverted. The surface of the rock core is covered with a lot of dust. Before using the scanning device, the dust on the surface of the rock core needs to be cleaned to ensure a smooth scanning process, reduce scanning errors, and improve scanning quality.

[0046] Specifically, after core sampling is completed, borehole television begins to capture images from the borehole opening and is lowered at a set speed until the bottom of the borehole is reached. The images of the capturing process are then stored to obtain a columnar image of the borehole wall surface.

[0047] See Figure 5 In some embodiments, the borehole television includes an imaging probe 9, a surface control terminal 7, and a winch 14. The winch 14 is installed on the ground via a support frame 13. The lifting rope of the winch 14 extends into the borehole 12 and is connected to the imaging probe 9. The imaging probe 9 is lowered into the borehole 12 via the lifting rope, and the camera on the imaging probe 9 can then be used to take pictures of the borehole 12 to obtain images of the borehole wall. The imaging probe 9 is connected to the surface control terminal 7 via a signal line 8, and the borehole wall images are transmitted to the surface control terminal 7 via the signal line 8.

[0048] Understandably, the imaging probe 9 is also equipped with an attitude sensing module 11. During the lowering process, the imaging probe 9 will collect image signals of the inner wall of the borehole 12. At the same time, the attitude sensing module 11 inside the imaging probe 9 will collect real-time orientation information, position information and speed information, and then transmit them to the ground control terminal 7 and store them in the data storage module.

[0049] Specifically, the attitude sensing module 11 includes a GPS positioning system, a gyroscope, and a photoelectric pulse encoder. The GPS positioning system and gyroscope can provide the real-time position and orientation of the imaging probe 9 and prevent the imaging probe 9 from shaking. The photoelectric pulse encoder can provide the lifting speed of the imaging probe 9.

[0050] During operation, the imaging probe 9 descends at a constant speed v to ensure that the acquired image achieves an overlap rate of X%, meeting the imaging accuracy requirements. The imaging probe 9 is connected to the ground control terminal 7 via the signal line 8 to transmit image signals in real time. The imaging probe 9 can rotate 360° horizontally and 180° vertically. The bottom of the imaging probe 9 is an opaque baffle, and the surrounding area is a transparent protective cover 10 to protect the imaging probe 9 from damage.

[0051] See Figure 5 The core scanning device includes a rotating base platform 3, a drive motor for rotating the rotating base platform 3, and a high-definition camera 1. The rotating base platform 3 is horizontally installed on the bottom of the core mounting cavity of the outer shell 2. The core 5 (unoriented) is vertically fixed on the rotating base platform 3. The rotating base platform 3 is also equipped with a drive mechanism that drives the high-definition camera 1 to move along the axial direction of the core 5 to capture images of the rotating core 5. The high-definition camera 1 is connected to the surface control terminal 7.

[0052] During operation, the core 5 is fixed on the turntable and rotates counterclockwise at a certain speed. The high-definition camera 1 scans the surface of the core from top to bottom under the drive mechanism and transmits the image data to the surface processing terminal. The surface control terminal 7 will eventually process the transmitted borehole wall surface columnar diagram and non-directional core surface columnar diagram to meet the matching requirements. Then, the core surface repositioning is realized by the image template matching method based on OpenCV.

[0053] See Figure 5 Specifically, a sliding rod is provided on the rotating base platform 3. The high-definition camera 1 moves at a constant speed from top to bottom on the sliding rod to take pictures. At the same time, the rotating base platform 3 drives the rock core to rotate at a constant angular velocity ωrad / s. Parameters such as the camera's movement speed and rotation angular velocity can be adjusted via buttons on the outer casing 2 and displayed on the display screen 6. This design effectively ensures that the scanning device can obtain stable images of the rock core surface and prevents the rock core from being upside down or flipped after removal. To prevent the rock core from shaking during rotation, a pressure rod 4 is also provided on the rotating base platform 3 to press the top of the rock core firmly.

[0054] This application uses an OpenCV-Python image matching algorithm to match the feature information (including lithological color, surface crack morphology, impurity color and distribution morphology, etc.) of the core outer columnar image and the borehole wall surface columnar image, resulting in a series of image matching similarity value maps. Finally, the normally distributed similarity value map is selected from all similarity value maps as the basis, thereby objectively finding the relative orientation of the core sample at the target depth of the sampling well wall. The Python algorithm is used to implement the predetermined function, minimizing the uncertainty of human factors and avoiding errors caused by manual judgment. It has the advantages of high efficiency and accuracy. The cylindrical image is used, which is more three-dimensional than the rolling scan image used in the background patent. When we establish a three-dimensional coordinate system on the cylindrical image, we can directly read the depth h and azimuth angle θ in the coordinate system, which is convenient for operators to judge and understand during operation.

[0055] The above embodiments are merely illustrative examples to clearly illustrate the present invention and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all embodiments here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A core repositioning method based on intelligent recognition of digital images, characterized in that, include: Locate the coordinates of the surface borehole and collect rock cores at the target depth. A columnar image of the borehole wall surface was obtained using an imaging device, and the directions of east, south, west, and north were marked on the columnar image of the borehole wall surface. A reference line parallel to the core axis is drawn on the core surface, and then the core scanning device is used to scan the core sample surface to generate a columnar image of the core surface. The columnar images of the borehole wall surface and the outer surface of the rock core are input into the OpenCV software for image matching, where; During image matching, the reference line on the outer columnar image of the core is aligned with the north orientation of the columnar image on the borehole wall surface. Then, the outer columnar image of the core is traversed along the axial and circumferential directions at intervals of a set number of pixels on the borehole wall surface columnar image to obtain similarity value maps at different depths. The similarity value map refers to the distribution of similarity values ​​corresponding to different azimuth angles at a specific depth. Find the normally distributed similarity map among several similarity maps and obtain the depth value corresponding to that similarity map. h Find the azimuth angle corresponding to the maximum similarity value in a normally distributed similarity map. θ ; Based on the obtained depth value h Azimuth θ And reference lines on the surface of the core, enabling the repositioning of unoriented cores on the ground; Select a borehole location and locate its coordinates using GPS. The borehole depth for data acquisition is within [specific range]. h 1 -h 2 Rock cores from the location; During image matching, first lock the columnar image of the borehole wall surface to... h 1 -h 2 Within the depth range, then input the core surface columnar image, allowing the core surface columnar image to be derived from the columnar image on the borehole wall surface. h 1 Starting from the point, slide downwards in the axial direction and rotate counterclockwise in the circumferential direction to traverse the image; GPS positioning coordinates use the WGS-84 coordinate system to describe the location of surface boreholes; To ensure that the columnar images of the rock core are not upside down, the dust on the rock core surface needs to be cleaned before scanning.

2. The core repositioning method according to claim 1, characterized in that: After core sampling is completed, the imaging equipment starts taking pictures from the borehole opening and is lowered at a set speed until it stops taking pictures at the bottom of the hole. The images of the shooting process are stored to obtain a columnar image of the borehole wall surface.

3. The core repositioning method according to claim 1 or 2, characterized in that: The imaging device is a hole-punch television system.

4. The core repositioning method according to claim 3, characterized in that: The borehole television includes an imaging probe (9), a surface control terminal (7), and a winch (14). The winch (14) is installed on the ground via a support frame (13). The lifting rope of the winch (14) extends into the borehole (12) and is connected to the imaging probe (9). The imaging probe (9) is connected to the surface control terminal (7) via a signal line (8).

5. The core repositioning method according to claim 4, characterized in that: The core scanning device includes a rotating base platform (3), a drive motor for driving the rotating base platform (3) to rotate, and a high-definition camera (1). The core (5) is vertically fixed on the rotating base platform (3). The rotating base platform (3) is also provided with a drive mechanism that drives the high-definition camera (1) to move along the axial direction of the core (5) to capture images of the rotating core (5). The high-definition camera (1) is connected to the surface control terminal (7).

6. The core repositioning method according to claim 5, characterized in that: The rotating base platform (3) is also provided with a pressure bar (4) to press the top of the rock core (5) tightly.

Images