Coal-rock prediction method and system based on multi-source information fusion and interface reconstruction

Abstract

The present invention provides a coal-rock prediction method and system based on multi-source information fusion and interface reconstruction, which relates to the field of coal-rock interface perception prediction technology, including obtaining a face image and borehole geological information; using the borehole geological information to generate a coal seam trend analysis diagram to determine the start and end positions of the coal-rock layer; inputting the face image into a recognition model to extract two-dimensional coal-rock interface feature points, and establishing a binary tree to match the same feature points in different face images to identify the coal-rock interface coordinates, and using the optimal space circle algorithm to fit the face coal-rock interface three-dimensional model, and reconstructing the coal-rock interface in three dimensions based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface; converting the three-dimensional reconstruction model of the coal-rock interface into a time series and inputting it into an ARIMA model to predict the development trend of the coal-rock interface.

Description

Technical Field

[0001] The present disclosure relates to the technical field of coal-rock interface perception prediction, and in particular to a coal-rock prediction method and system based on multi-source information fusion and interface reconstruction. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] Intelligent tunneling is an important part of intelligent coal mining and intelligent mine construction. Intelligent identification of coal-rock interface, as an important step in intelligent tunneling, can determine the thickness of the coal seam and guide the coal mining machine to automatically and accurately track the coal-rock interface. It can timely adjust the cutting height of the coal mining machine drum to prevent accidental cutting of rocks, reduce the risk of accidents, improve coal quality, and increase the mining rate. It determines whether automated coal mining can be achieved on unmanned working faces and is a technology that must be developed for the construction of intelligent and unmanned coal mining.

[0004] At present, there are two main methods for coal-rock interface identification: (1) using gamma rays, radar, images, etc. to directly detect the coal-rock interface through sensors. This method has the advantages of fast detection speed and high accuracy, but it is greatly affected by interference. For example, the gamma ray detection method is not suitable for rock layers with low radioactive element content or coal seams with more gangue; the radar detection method has serious signal attenuation when the thickness of the coal seam increases, and the detection effect is difficult to guarantee; the image detection method is easily affected by the on-site light source effect, dust concentration, etc., resulting in inaccurate feature extraction, poor prediction effect of the coal-rock interface ahead, and low coal mining efficiency. (2) Indirect identification based on construction equipment operation parameters and information feedback, including infrared, vibration, acoustic spectrum, etc. This method is easy to operate, but it has high requirements for the boundary condition setting of the method model. For example, when infrared, vibration, etc. encounter coal seams with similar hardness to rock, the recognition accuracy will be reduced; the acoustic spectrum detection method is easy to interfere with the judgment when the difference between the coal seam and the rock is not large, and the coal-rock interface can only be measured when the rock is cut and hits the middle groove.

[0005] In addition, the current research on coal-rock interface identification methods is mostly focused on optimizing a single method, such as optimizing drilling equipment, such as probes and thrust rods, or optimizing the way to obtain ground source information. Both are committed to improving the accuracy of information acquisition, but there are few systems and methods that integrate direct and indirect information to accurately predict the coal-rock interface of unmined sections. Summary of the invention

[0006] In order to solve the above problems, the present invention proposes a coal-rock prediction method and system based on multi-source information fusion and interface reconstruction, which integrates the face image information and the borehole geological information, identifies the coal-rock interface and performs three-dimensional reconstruction of the coal-rock interface, and uses the three-dimensional reconstruction of the coal-rock interface and the structural surface to infer the development trend of the coal-rock interface.

[0007] According to some embodiments, the present disclosure adopts the following technical solutions:

[0008] Coal rock prediction method based on multi-source information fusion and interface reconstruction, including:

[0009] Obtain tunnel face images and borehole geological information;

[0010] Use drilling geological information to generate coal seam trend analysis diagrams to determine the start and end positions of coal and rock layers;

[0011] The tunnel face image is input into the recognition model to extract the two-dimensional coal-rock interface feature points, and a binary tree is established to match the same feature points in different tunnel face images to identify the coal-rock interface coordinates. The three-dimensional model of the tunnel face coal-rock interface is fitted using the optimal space circle algorithm, and the coal-rock interface is reconstructed in three dimensions based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface.

[0012] The three-dimensional reconstruction model of the coal-rock interface is converted into a time series and input into the ARIMA model to predict the development trend of the coal-rock interface.

[0013] According to some embodiments, the present disclosure adopts the following technical solutions:

[0014] The coal-rock prediction system based on multi-source information fusion and interface reconstruction includes:

[0015] Data acquisition module, used to obtain tunnel face images and borehole geological information;

[0016] Use drilling geological information to generate coal seam trend analysis diagrams to determine the start and end positions of coal and rock layers;

[0017] The reconstruction module is used to input the tunnel face image into the recognition model, extract the two-dimensional coal-rock interface feature points, establish a binary tree to match the same feature points in different tunnel face images, identify the coal-rock interface coordinates, and use the optimal space circle algorithm to fit the three-dimensional model of the tunnel face coal-rock interface. The coal-rock interface is reconstructed in three dimensions based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface;

[0018] The trend prediction module is used to convert the three-dimensional reconstruction model of the coal-rock interface into a time series and input it into the ARIMA model to predict the development trend of the coal-rock interface.

[0019] According to some embodiments, the present disclosure adopts the following technical solutions:

[0020] A non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the coal-rock prediction method based on multi-source information fusion and interface reconstruction is implemented.

[0021] According to some embodiments, the present disclosure adopts the following technical solutions:

[0022] An electronic device comprises: a processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes the coal-rock prediction method based on multi-source information fusion and interface reconstruction.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] The present invention discloses a coal-rock prediction method based on multi-source information fusion and interface reconstruction. By fusing the face image recognition and the information obtained from drilling, the internal coal-rock interface is detected based on the directly detected face coal-rock information and drilling parameters, borehole imaging, apparent resistivity, natural potential and other information. The information of the two is combined to build a model, predict the coal-rock interface information of the unmined section, guide the unmanned intelligent mining of the coal mining machine, and improve the efficiency and accuracy of intelligent coal mining.

[0025] The present invention discloses a coal-rock prediction method based on multi-source information fusion and interface reconstruction. The method inputs a face image into a recognition model, extracts two-dimensional coal-rock interface feature points, establishes a binary tree to match the same feature points in different face images, identifies the coordinates of the coal-rock interface, and uses an optimal spatial circle algorithm to fit a three-dimensional model of the coal-rock interface of the face. The method reconstructs the coal-rock interface in three dimensions based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface. The method can more quickly predict the coal-rock interface of a long unexcavated section ahead, guide the automatic operation of the coal mining machine, and greatly improve the automation and efficiency of coal mining.

[0026] The present invention discloses a coal-rock prediction method based on multi-source information fusion and interface reconstruction, which can accurately predict the coal-rock interface of the unmined section ahead over a long distance. Compared with the traditional method of identifying the coal-rock interface, the method of integrating multi-source information modeling can improve the efficiency of coal excavation and more accurately guide intelligent coal excavation. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure.

[0028] Figure 1is a system flow chart of an embodiment of the present disclosure;

[0029] Figure 2 The present invention is a method flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0030] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0031] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0032] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0033] Example 1

[0034] In one embodiment of the present disclosure, a coal-rock prediction method based on multi-source information fusion and interface reconstruction is provided, comprising the following steps:

[0035] Step 1: Obtain the tunnel face image and drilling geological information;

[0036] Step 2: Generate a coal seam trend analysis diagram using borehole geological information to determine the start and end positions of the coal and rock layers;

[0037] Step 3: Input the tunnel face image into the recognition model, extract the two-dimensional coal-rock interface feature points, and establish a binary tree to match the same feature points in different tunnel face images, identify the coal-rock interface coordinates, and use the optimal space circle algorithm to fit the three-dimensional model of the tunnel face coal-rock interface. Based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface, the coal-rock interface is reconstructed in three dimensions;

[0038] Step 4: Convert the three-dimensional reconstruction model of the coal-rock interface into a time series and input it into the ARIMA model to predict the development trend of the coal-rock interface.

[0039] As an embodiment, the specific implementation process of a coal-rock prediction method based on multi-source information and interface reconstruction disclosed in the present invention is as follows:

[0040] Step 1: Obtaining tunnel face image information

[0041] Specifically, the tunnel face is photographed multiple times by a camera, and the obtained images are stitched together to form a panoramic image of the tunnel face, so that a neural network can be used later to extract coal-rock interface information and obtain prior information of the model tunnel face.

[0042] Step 2: Use special equipment to drill holes and obtain drilling geological information;

[0043] Specifically, a multi-drill bit drilling is performed using a mining resistivity video imaging logging instrument to collect geological information, collect panoramic imaging image information, GR curve, apparent resistivity curve, natural potential curve and other information in the hole, and obtain data such as the borehole orientation, inclination, and borehole depth;

[0044] The mining resistivity video imaging logging instrument consists of a logging instrument host, a logging instrument probe, cables and a push rod. The probe consists of a high-resolution camera, natural gamma, apparent resistivity, natural potential, multiple excitation points, drilling trajectory measurement and control unit.

[0045] The borehole trajectory is obtained through the data of borehole orientation, inclination, and borehole depth. The natural gamma logging curve and video are used for synchronous data analysis at the same depth to interpret the lithology and stratification. The gamma curve, apparent resistivity curve, and natural potential curve are used for stratigraphic analysis. At the coal-rock interface, the gamma curve will show an obvious downward shift (low value), and the apparent resistivity curve will show a jump from low resistivity to high resistivity. The natural potential curve in the coal seam is usually relatively gentle and has small fluctuations. At the same time, the video taken uninterruptedly in the borehole is analyzed frame by frame. Combined with the drilling speed, the start and end positions of the coal (rock) layer are determined, thereby obtaining a comprehensive columnar diagram of the borehole, understanding the coal-rock interface structure inside the unexcavated section, and obtaining the necessary information for coal seam trend analysis;

[0046] By combining the logging database with actual measurements, the curve classification can be used to determine the formation and lithology changes in the survey area.

[0047] Combined with the above information and measurement results, ledger data, etc., Surfer software is used to generate coal thickness contour map, coal seam roof (bottom) plate contour map, profile map, deviation map and other results maps, draw coal seam trend analysis map, and obtain the coordinate information of the coal-rock interface inside the rock mass;

[0048] Step 3: Input the tunnel face image into the recognition model, extract the two-dimensional coal-rock interface feature points, and establish a binary tree to match the same feature points in different tunnel face images, identify the coal-rock interface coordinates, and use the optimal space circle algorithm to fit the three-dimensional model of the tunnel face coal-rock interface. Based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface, the coal-rock interface is reconstructed in three dimensions;

[0049] Furthermore, the existing neural network model is used to perform deep learning on multi-angle coal-rock interfaces, realize the intelligent extraction of two-dimensional feature points of the coal-rock interface, replace the depth direction pixel vector matrix of the fracture segmentation map with the corresponding pixel points of the original image, and realize the recognition and marking of the two-dimensional coal-rock interface feature points.

[0050] Specifically, step 3-1: digitally characterize the geometric parameters (such as length, angle, etc.) of the coal-rock interface, use the feature description function to describe the same feature at different positions of different images in the image set, and establish a local feature set at the position;

[0051] Step 3-2: To obtain stable feature points, Gaussian difference is performed on the same feature point in adjacent images of each group of images to obtain a Gaussian difference pyramid (DOG), which is specifically defined as: using the Gaussian difference pyramid to find the spatial extreme points and determine the exact position of the extreme points through the fitting function.

[0052]

[0053]

[0054]

[0055]

[0056] in, represents the scale space, is the spatial scale factor, is a Gaussian function, Represents the input image.

[0057] Furthermore, the mapping geometry estimation method is used to match features in different images and convert the Euclidean distance , squared sum distance As a similarity measure, the distance of the similar feature point in another coal-rock interface image is calculated, and the Describes the matching degree between two feature points. When the value is the smallest, the matching of feature points is achieved;

[0058]

[0059]

[0060] Step 3-3, to store the matching feature point information, first select one from all the feature points as the root node, and then recursively assign the remaining feature points according to the coordinate values, with the smaller coordinate points as the left child nodes and the larger ones as the right child nodes, until all the feature points are inserted into the tree to form a balanced binary tree structure. The calculation formula calculates the Euclidean distance between the test point and the marked feature point, selects K neighbor feature points, usually 3, and uses the nearest neighbor search algorithm to search the node closest to the feature point to be matched in the binary tree to obtain the multi-angle image feature point matching result.

[0061] Furthermore, based on the above method, the coal-rock interface feature points are matched, and the camera external parameters are solved by using the basic matrix for reconstruction to obtain the three-dimensional information of the matching feature points. The specific steps are as follows;

[0062] Step 1) Select two images with the largest camera baseline to meet the matching ratio requirements of 3D reconstruction. The finite number of rotations can be equivalent to a single rotation of a rotation axis and a rotation angle. The rotation axis is k and the rotation angle is θ. The rotation matrix R is obtained using the following formula. I is the product of the rotation matrix and the transpose of the rotation matrix.

[0063]

[0064] Calculate the basic matrix F and the basic matrix E, and decompose the matrix according to the matching points to obtain the camera internal parameter matrix K, the rotation matrix R and the translation vector t.

[0065]

[0066]

[0067] According to the proposed optimal view, the accuracy is selected, the feature points are extracted using the SIFT algorithm, the view with the largest number of feature points is selected, image pairs are continuously added, the camera projection matrix is ​​estimated, and the length and width are both The grid discretizes the coal-rock interface image into several cells. The cells are divided into two different states: empty and full. When there is a coal-rock interface in the cell, it is described as full. The image is divided into several cells according to the weight. Increase the corresponding score , which is empty when it is not visible. Repeat the above process to obtain the pose parameters and sparse 3D point cloud of the reconstructed rock mass image, and obtain the 3D coordinates and normal vectors of the characteristic points of the coal-rock interface based on the 3D reconstruction results;

[0068] Step 2) Fit the coal-rock interface of the tunnel face by the optimal space circle algorithm, and use the six parameters of the space circle (center coordinates (x, y, z) and inclination, inclination and diameter) to characterize the surface coal-rock interface. Assuming that all coal-rock interface feature points fall on the plane of the space circle, the perpendicular bisector between the two feature points must pass through the center of the circle. Assuming that the two feature points , The three-dimensional coordinates are ( , , )and( , , ), the vector connecting the two points It can be expressed as: , , ), the midpoint of the line connecting the two feature points for:

[0069] ( , , )

[0070] Set the center for( , , ), the center With midpoint The vector of the connection direction should satisfy .

[0071] Assuming that all points are on the circle, there is a formula:

[0072]

[0073]

[0074] The radius of the space circle is the average distance from all coal-rock interface feature points to the center of the circle. By using the optimal space circle to extract the coal-rock interface features, the mathematical model of the coal-rock interface can be obtained, and the coordinate information of the feature points of the coal-rock interface at the tunnel face can be obtained.

[0075] Step 4: Convert the three-dimensional reconstruction model of the coal-rock interface into a time series and input it into the ARIMA model to predict the development trend of the coal-rock interface.

[0076] The extension length of the mine tunnel is analogous to a time point, and a spatiotemporal sequence rock mass result prediction model is established through the ARIMA model. By establishing time series data and designing a deep learning prediction network, modeling is performed based on the coordinates of the previously collected feature points, and the mining depth is used as the time point in the spatiotemporal sequence. The three-dimensional reconstruction of the coal-rock interface and the automatic extraction of the structural surface are used, as well as the coordinates of the coal-rock interface development trend are inferred to establish a coal-rock interface spatiotemporal sequence data set, and the spatiotemporal model is trained based on it to further obtain the optimal prediction model. The model establishment mainly includes the following processes:

[0077] Firstly, the unit root (ADF) test method is used to establish a data set for stationarity test of the coordinate sequence of characteristic points of the coal-rock interface;

[0078] Data sets that do not meet the stationarity requirements are differentiated until the data sets meet the stationarity requirements. The subtraction operation between the time series values ​​separated by one time point becomes a first-order difference operation, which is recorded as: Performing another differential operation on the first-order differential operation value is called a second-order differential operation. , and so on, if the time series data set is The subtraction operation between the time series values ​​at time points is called Step difference operation, denoted as: . Let the number of differences be d;

[0079] Determine the parameters , mainly through the autocorrelation plot (ACF) and partial autocorrelation plot (PACF) in the truncation of the judgment, when the ACF and PACF plot correlation characteristics are not obvious. Akaike Information Criterion (AIC) or Bayesian Information Criterion (BIC) is used to determine value.

[0080] The AIC criterion expression is:

[0081]

[0082] The BIC criterion expression is:

[0083]

[0084] is the total number of samples in the dataset, To fit the residual variance, when the AIC value is the smallest, the model is the optimal model. When n is too large and the sample data cannot converge, the BIC criterion is used. When the value is also the smallest, it is the optimal model.

[0085] After model identification and order determination, the residual sequence is tested to determine whether it can be used for prediction. If the residual sequence presents a random normal distribution, it passes the test. The autocorrelation can be determined by the following formula:

[0086]

[0087] is the total amount of data samples, is the maximum hysteresis number, or ,when When it is extremely large, , and further obtain the detection statistic Q:

[0088]

[0089] When the test statistic , indicating that the model has passed the test. If it does not meet the requirements, return to the previous step for model identification and parameter determination.

[0090] After the characteristic point coordinate data sequence meets the requirements, it is input into the ARIMA model to predict the unexcavated section time point, that is, the unexcavated section coal-rock interface;

[0091] The autoregressive model (AR) is used to predict the coordinates of the coal-rock interface in the unexcavated section;

[0092] The autoregressive model structure is as follows:

[0093]

[0094]

[0095]

[0096]

[0097] is an unknown parameter, called the autoregressive coefficient. is a white noise sequence, is the response data at time t, The variance is Represents the correlation of factors.

[0098] The model prediction steps are as follows:

[0099] (1) Test the stability of the coal-rock interface sequence collected and analyzed. Use differential and logarithmic operations to process the unstable sequence. At the same time, record the number of differentials, recorded as d;

[0100] (2) Perform a stationarity test on the sequence data. When the test result is stationary, identify the model using autocorrelation diagram and partial autocorrelation diagram;

[0101] (3) Parameter estimation of the identified ARIMA model to determine the range of values ​​of the orders p and q;

[0102] (4) Determine the optimal model based on the AIC criterion or the BIC criterion. The model is ARIMA (p, d, q);

[0103] (5) Perform a residual test on the ARIMA (p, d, q) model to determine whether the final applicability of the model can be used for prediction;

[0104] (6) The final model is used to smooth the data of the coal-rock interface sequence, and the AR model in the ARIMA model uses the existing feature point coordinate values ​​and past prediction errors to predict the feature point coordinate values ​​in the unexcavated section.

[0105] Example 2

[0106] In one embodiment of the present disclosure, a coal-rock prediction system based on multi-source information fusion and interface reconstruction is provided, comprising:

[0107] Data acquisition module, used to obtain tunnel face images and borehole geological information;

[0108] Use drilling geological information to generate coal seam trend analysis diagrams to determine the start and end positions of coal and rock layers;

[0109] The reconstruction module is used to input the tunnel face image into the recognition model, extract the two-dimensional coal-rock interface feature points, establish a binary tree to match the same feature points in different tunnel face images, identify the coal-rock interface coordinates, and use the optimal space circle algorithm to fit the three-dimensional model of the tunnel face coal-rock interface. The coal-rock interface is reconstructed in three dimensions based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface;

[0110] The trend prediction module is used to convert the three-dimensional reconstruction model of the coal-rock interface into a time series and input it into the ARIMA model to predict the development trend of the coal-rock interface.

[0111] The data acquisition module collects geological information including the image information of the face and the information obtained from the drilling. The reconstruction module processes the image information and analyzes the drilling information to obtain the coordinates of the coal-rock interface and the trend of the coal seam. Finally, the trend prediction module establishes the ARIMA model, models and predicts the analyzed data, and obtains the coal-rock interface prediction results.

[0112] When collecting information, the camera is rotated around a fixed axis to collect image information of the face; a mining resistivity video imaging logging instrument is used to drill holes to obtain geological information in the hole and drilling data information;

[0113] During information analysis and reconstruction, the neural network model is used to extract the two-dimensional coal-rock interface feature points of the tunnel face image information, and a binary tree is established to match the same feature points in different images, match and identify the coal-rock interface, and then use the optimal space circle algorithm to fit the three-dimensional model of the tunnel face coal-rock interface; the drilling information is input into the Surfer software for analysis, and the coal thickness contour map, coal seam roof (bottom) plate contour map, profile map, deviation map and other result maps are generated, and the coal seam trend analysis map is drawn, and the coordinate information of the coal-rock interface inside the rock mass is obtained;

[0114] When predicting, the excavation depth is analogized to a time series. After verifying the stationarity of the data, the autocorrelation diagram and partial autocorrelation diagram are used for identification. After the optimal model is determined by the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC), a residual test is performed, which is used to guide the intelligent excavation of coal and rock.

[0115] Example 3

[0116] In one embodiment of the present disclosure, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the coal-rock prediction method based on multi-source information fusion and interface reconstruction is implemented.

[0117] Example 4

[0118] In one embodiment of the present disclosure, an electronic device is provided, including: a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory, so that the electronic device executes the coal-rock prediction method based on multi-source information fusion and interface reconstruction.

[0119] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0120] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0121] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A coal-rock prediction method based on multi-source information fusion and interface reconstruction, characterized in that: include: Obtain tunnel face images and borehole geological information; Use drilling geological information to generate coal seam trend analysis diagrams to determine the start and end positions of coal and rock layers; The tunnel face image is input into the recognition model, the two-dimensional coal-rock interface feature points are extracted, and a binary tree is established to match the same feature points in different tunnel face images, the coal-rock interface coordinates are identified, and the three-dimensional model of the tunnel face coal-rock interface is fitted using the optimal space circle algorithm. The coal-rock interface is reconstructed in three dimensions based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface; including: The tunnel face image is input into the neural network model to extract the two-dimensional coal-rock interface feature points. The Gaussian difference pyramid is used to find the spatial extreme points and the exact position of the extreme points is determined by fitting functions. The mapping geometry estimation method is used to match features in different images and the Euclidean distance is used as a similarity measure to obtain matching points. A binary tree is established for the feature points through the coordinate axes to obtain the matching results of the feature points of the multi-angle images, and the matching feature point information is stored; based on the matched coal-rock interface feature points, the basic matrix is ​​used to solve the external parameters of the camera for reconstruction to obtain the three-dimensional information of the matching feature points, and the basic matrix is ​​decomposed according to the matching points to obtain the camera internal parameter matrix, rotation matrix and translation vector, and image pairs are continuously added to estimate the camera projection matrix to generate a sparse point cloud, and the three-dimensional coordinates and normal vector of the coal-rock interface are obtained; The three-dimensional reconstruction model of the coal-rock interface is converted into a time series and input into the ARIMA model to predict the development trend of the coal-rock interface.

2. The coal-rock prediction method based on multi-source information fusion and interface reconstruction according to claim 1, characterized in that: During the multi-drill bit drilling process, panoramic imaging image information, GR curve, apparent resistivity curve and natural potential curve are collected in the hole to obtain the borehole azimuth, inclination and drilling depth data. The drilling trajectory is obtained by analyzing the drilling parameter information. The GR curve, apparent resistivity curve and natural potential curve are used to perform stratigraphic analysis to obtain a comprehensive borehole column chart. By analyzing the multi-hole information, the starting and ending positions of the coal and rock layers are determined.

3. The coal-rock prediction method based on multi-source information fusion and interface reconstruction according to claim 1, characterized in that: The three-dimensional model of the coal-rock interface at the tunnel face is fitted by the optimal space circle algorithm, and the six parameters of the space circle are used to characterize the surface coal-rock interface.

4. The coal-rock prediction method based on multi-source information fusion and interface reconstruction according to claim 1, characterized in that: Coal seam trend analysis diagrams are generated using borehole geological information, including coal thickness contour maps, coal seam roof / floor contour maps, profile maps, and deviation maps, and coal seam trend analysis diagrams are drawn.

5. A coal-rock prediction system based on multi-source information fusion and interface reconstruction, which executes a coal-rock prediction method based on multi-source information fusion and interface reconstruction as claimed in any one of claims 1 to 4, characterized in that: include: Data acquisition module, used to obtain tunnel face images and borehole geological information; Generate coal seam trend analysis diagrams using borehole geological information to determine the start and end positions of coal and rock layers; The reconstruction module is used to input the tunnel face image into the recognition model, extract the two-dimensional coal-rock interface feature points, establish a binary tree to match the same feature points in different tunnel face images, identify the coal-rock interface coordinates, and use the optimal space circle algorithm to fit the three-dimensional model of the tunnel face coal-rock interface. The coal-rock interface is reconstructed in three dimensions based on the start and end positions of the coal-rock layer and the three-dimensional model of the coal-rock interface; The trend prediction module is used to convert the three-dimensional reconstruction model of the coal-rock interface into a time series and input it into the ARIMA model to predict the development trend of the coal-rock interface.

6. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by the processor, the coal-rock prediction method based on multi-source information fusion and interface reconstruction as described in any one of claims 1-4 is implemented.

7. An electronic device, characterized in that: include: A processor, a memory and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory so that the electronic device executes the coal-rock prediction method based on multi-source information fusion and interface reconstruction as described in any one of claims 1-4.

Images