Tunnel surrounding rock classification method and device

Abstract

The application discloses a tunnel surrounding rock grading method and device. The method comprises the following steps: shooting a first tunnel image of a first tunnel face at at least one angle; adopting vibration intervention to the first tunnel face to obtain a second tunnel face; shooting a second tunnel image of the second tunnel face at the angle; obtaining a tunnel face change according to the first tunnel image and the second tunnel image; and grading the tunnel face according to the tunnel face change by using a neural network model. The application firstly adopts active vibration intervention to the tunnel face to obtain geological information change of the tunnel face before and after the intervention; and then evaluates the tunnel face change by using the neural network model, so that different grades of the changeable degree of the current tunnel face of the tunnel surrounding rock are obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of geological exploration, in particular, relates to a tunnel surrounding rock classification method and device. BACKGROUND

[0002] The soil types through which a tunnel passes are various, mainly including rock, silt, cohesive soil, sandy soil, pebble soil, gravel soil, etc. Among them, rock is the hardest soil, with high density and large compressive strength, but also has special structures such as faults, joints and karst caves; silt and cohesive soil are prone to be affected by moisture due to their small particles and low density, so attention needs to be paid in tunnel design and construction.

[0003] The tunnel rock-soil body shows inherent rheological properties, and the deformation of the surrounding rock of the tunnel face also has certain rules. The excavation of the tunnel causes the surrounding rock supporting the tunnel body to be dug out, and the space behind the tunnel face is empty, resulting in the deformation of the surrounding rock towards the tunnel clearance. At the same time, with the tunnel construction, the current tunnel face will show deformation, cracking and other characteristics, and with the progress of the tunnel construction, the deformation and cracking of the tunnel face are evaluated and classified, which is one of the effective measures to analyze the current tunnel face in the construction process to improve the construction safety. SUMMARY

[0004] Therefore, the first aspect of the embodiments of the present application discloses a tunnel surrounding rock classification method.

[0005] The method comprises,

[0006] shooting a first tunnel image of a first tunnel face at at least one angle;

[0007] adopting excitation intervention to the first tunnel face to obtain a second tunnel face;

[0008] shooting a second tunnel image of the second tunnel face at the angle;

[0009] obtaining a tunnel face change according to the first tunnel image and the second tunnel image;

[0010] classifying the tunnel face according to the tunnel face change by using a neural network model.

[0011] The method comprises,

[0012] obtaining a gray scale ratio image of the first tunnel image and the second tunnel image;

[0013] obtaining a clustering result of the gray scale ratio image by using a clustering algorithm;

[0014] obtaining a clustering matrix representing the first tunnel image and the second tunnel image according to the clustering result.

[0015] extracting the number of pixels of the clustering matrix;

[0016] extracting the coordinate relationship of each pixel point of the clustering matrix, and obtaining the number of at least paired or grouped pixel combinations in the clustering matrix according to the adjacent relationship, wherein the number of pixel combinations and the number of pixel points represent the change of the tunnel face.

[0017] wherein the first gradient image of the first tunnel image and the second gradient image of the second tunnel image are obtained by wavelet transform;

[0018] the gray scale ratio image is obtained according to the first gradient image and the second gradient image.

[0019] wherein the first gradient image and the second gradient image at a resolution are obtained by Haar wavelet transform;

[0020] the resolution is selected according to the judgment of the clustering result.

[0021] wherein when the number of pixel points in the clustering matrix deviates too much from the strength and / or frequency of the excitation intervention, the resolution is reduced;

[0022] or when the number of pixel points in the clustering matrix deviates too little from the strength and / or frequency of the excitation intervention, the resolution is increased.

[0023] wherein the clustering result of the gray scale ratio image is obtained by FCMA clustering algorithm.

[0024] wherein the FCMA clustering algorithm comprises,

[0025]

[0026] wherein v j is the clustering center of the jth class, N is the total number of pixels of the gray scale ratio image, μ is an NxC matrix, μ ij is the membership degree of x i belonging to the jth class, C is the number of clusters, m is the weight, and ||x i -v j || represents the Euclidean distance from the ith data to the jth clustering center, and ‖x i -v k ‖ represents the Euclidean distance from the ith data to the kth clustering center,

[0027] C is 2, which is used to represent the clustering change pixel points and the unchanged pixel points,

[0028] m is a natural number greater than 1.

[0029] Wherein, m is selected according to the judgment of the clustering result.

[0030] Wherein, the obtaining of the change of the tunnel face comprises,

[0031] According to a preset number of adjacency relations, the pixel regions composed of each pixel in the clustering matrix are aggregated to obtain a pixel region;

[0032] The area ratio of the maximum pixel region to the minimum pixel region is calculated.

[0033] The pixel combination number, the pixel number and the area ratio represent the change of the tunnel face.

[0034] The second aspect of the present application discloses a tunnel surrounding rock grading device,

[0035] Wherein, the device comprises,

[0036] The first shooting unit is used for shooting a first tunnel image of a first tunnel face at at least one angle;

[0037] The vibration intervention device is used for vibrating and intervening the first tunnel face to obtain the second tunnel face;

[0038] The second shooting unit is used for shooting a second tunnel image of the second tunnel face at the angle;

[0039] The change analysis unit is used for obtaining the change of the tunnel face according to the first tunnel image and the second tunnel image;

[0040] The change evaluation unit is used for grading the tunnel face according to the change of the tunnel face by using a neural network model.

[0041] For the above-mentioned scheme, the present application will be described in detail below with reference to the accompanying drawings to disclose exemplary embodiments, and other features and advantages of the embodiments of the present application will be clear. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments will be briefly introduced below, and it should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0043] Figure 1 The flowchart of the tunnel surrounding rock grading method of the present embodiment is shown;

[0044] Figure 2 The flowchart of obtaining the change of the tunnel face of the present embodiment is shown;

[0045] Figure 3 A flowchart showing obtaining the change of the tunnel face according to the clustering matrix is shown.

[0046] Figure 4 A structural diagram of a tunnel surrounding rock grading device is shown. DETAILED DESCRIPTION

[0047] For the purpose of facilitating the understanding of the present application, a more complete description of the present application will be made with reference to the accompanying drawings. The embodiments of the present application are shown in the drawings. However, the present application can be realized in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used herein in the specification merely describe the specific embodiments of the present application for the purpose of the description and are not intended to limit the present application.

[0049] The present embodiment discloses a tunnel surrounding rock grading method. The method is used to grade and evaluate the change of the tunnel face in construction, and provides support for tunnel construction safety, support design, etc.

[0050] Figure 1 A flowchart of the tunnel surrounding rock grading method of the present embodiment is shown.

[0051] Figure 1 The method includes the following steps.

[0052] 10. Taking a first tunnel image of the first tunnel face at at least one angle.

[0053] 20. Interfering with the first tunnel face to obtain a second tunnel face by exciting vibration.

[0054] 30. Taking a second tunnel image of the second tunnel face at an angle.

[0055] 40. Obtaining the change of the tunnel face according to the first tunnel image and the second tunnel image.

[0056] 50. Grading the tunnel face according to the change of the tunnel face by using a neural network model.

[0057] Therefore, the method of the present embodiment first interferes with the tunnel face by active exciting vibration to obtain the deformation, crack, etc. of the tunnel face before and after the interference; and then evaluates the change of the tunnel face by using a neural network model to realize different grading of the changeable degree of the current tunnel face.

[0058] In step 10,

[0059] The shooting angle refers to shooting the tunnel face at a fixed angle in the tunnel to ensure that the visible features of the tunnel face can be retained and identified during shooting.

[0060] The first tunnel image of the first tunnel face refers to the visible light image of the tunnel face obtained by shooting through a visual device such as a camera. The “first” designation of the tunnel face is used to distinguish the tunnel face that has not changed before excitation. The “first” designation of the tunnel image is used to distinguish the visible light image of the tunnel face that has not changed before excitation.

[0061] In step 20,

[0062] The excitation intervention refers to directly or indirectly acting on the tunnel face by one or more vibration devices to transmit one or more directional vibration components along the longitudinal direction, transverse direction or depth direction of the tunnel face to affect the entire or partial surface of the tunnel face or affect the tunnel face to a certain depth inward, so that the tunnel face changes, such as deformation, cracking, etc.

[0063] The vibration device can be a vibration hammer that vertically or obliquely strikes the tunnel face one or multiple times. The vibration device can also be a drilling device that vertically or obliquely drills the tunnel face. In summary, the vibration device is used to excite the tunnel face in one or more directions to simulate the tunnel construction as much as possible.

[0064] In step 30,

[0065] The first tunnel image of the second tunnel face refers to the “second” designation of the tunnel face to distinguish the tunnel face that has changed after excitation. The “second” designation of the tunnel image is used to distinguish the visible light image of the tunnel face that has changed after excitation.

[0066] In step 40,

[0067] The tunnel face change refers to the changes such as cracks and deformations that can be directly observed from the visible light image after excitation. The present embodiment realizes different classifications of the current tunnel surrounding rock tunnel face by the tunnel face change.

[0068] Figure 2 The flowchart for obtaining the tunnel face change in the present embodiment is shown.

[0069] Figure 2 The flowchart for obtaining the tunnel face change includes the following steps.

[0070] 41. Obtain the gray scale ratio image of the first tunnel image and the second tunnel image.

[0071] 42Obtaining the clustering result of the gray scale ratio image by using the clustering algorithm.

[0072] 43Obtaining the clustering matrix representing the first tunnel image and the second tunnel image according to the clustering result.

[0073] 44Extracting the pixel point number of the clustering matrix, extracting the coordinate relationship of each pixel point of the clustering matrix, and obtaining the pixel combination number of at least pairs or groups in the clustering matrix according to the adjacent relationship, wherein the pixel combination number and the pixel point number represent the change of the working face.

[0074] Further, in step 41,

[0075] For the pixels that change in the first tunnel image and the second tunnel image, the corresponding pixels in the gray scale ratio image are not 1.

[0076] Preferably, the value of each pixel point in the gray scale ratio image can be taken as a logarithm or divided by a weighted value to compress the gray scale value of each pixel point in the gray scale ratio image.

[0077] The gray scale ratio image can be obtained by first using wavelet transform to obtain the first gradient image of the first tunnel image and the second gradient image of the second tunnel image, and then obtaining the gray scale ratio image according to the first gradient image and the second gradient image.

[0078] Preferably, the Haar wavelet transform is used to obtain the first gradient image and the second gradient image at a resolution.

[0079] In some embodiments, the resolution of the Haar wavelet transform can be selected according to the clustering result of step 42. For example, after the Haar wavelet transform selects the lowest resolution in step 41, it is found through manual observation in step 42 that a large number of cracks and deformation protrusions are classified in the same cluster in the clustering result, which can be considered to increase the resolution or the level of the resolution.

[0080] Further, when the number of pixel points in the clustering matrix of step 43 deviates too much from the intensity and / or frequency of the excitation intervention, the resolution of step 41 can be reduced, or when it deviates too little, the resolution of step 41 can be increased.

[0081] Further, in step 42,

[0082] The clustering result of the gray scale ratio image is obtained by using the FCMA clustering algorithm.

[0083] Specifically, the FCMA clustering algorithm in the present embodiment includes,

[0084]

[0085]

[0086] wherein v j is the cluster center of the jth class, N is the total pixel number of the gray scale ratio image, μ is an NxC matrix, μ ij is the x i membership of the pixel x i -v j || represents the Euclidean distance of the ith data to the jth cluster center, ‖x i -v k ‖ represents the Euclidean distance of the ith data to the kth cluster center,

[0087] C is 2, which is used to represent the cluster changed pixel points and unchanged pixel points,

[0088] m is a natural number greater than 1.

[0089] Preferably, the value of m can be selected according to the judgment of the clustering result of step 42.

[0090] For example, m is initially taken as a random number, and the value of m is adjusted by observing the clustering result.

[0091] Further, in step 44,

[0092] Figure 3 A flowchart for obtaining the change of the working face according to the clustering matrix is shown.

[0093] Figure 3 The method for obtaining the change of the working face according to the clustering matrix includes the following steps.

[0094] 441 The number of pixel points of the clustering matrix is extracted.

[0095] 442 The coordinate relationship of each pixel point of the clustering matrix is extracted, and the number of pixel combinations in pairs or groups in the clustering matrix is obtained according to the adjacent relationship.

[0096] 443 A pixel region formed by each pixel combination in the clustering matrix is obtained according to a preset number of adjacent relationships; and the area ratio of the maximum pixel region to the minimum pixel region is calculated.

[0097] 444 The change of the working face is represented according to the number of pixel combinations, the number of pixel points, and the area ratio.

[0098] Further, in step 441, the number of pixel points of the clustering matrix is extracted, and the pixel point number ratio of the pixel points of the clustering matrix to the total pixel points in the gray scale ratio image is obtained. In step 444, the change of the working face is represented according to the number of pixel combinations, the pixel point number ratio, and the area ratio.

[0099] wherein in step 50,

[0100] The neural network model refers to a BP neural network model. The BP neural network model is a multi-layer feedforward neural network trained based on error backpropagation. The BP neural network model comprises an input layer, a hidden layer and an output layer. The input layer of the BP neural network model comprises a plurality of input neurons, and the plurality of input neurons correspond to the inputs of the neural network model respectively. The BP neural network model has a plurality of hidden layers, and each hidden layer comprises a plurality of hidden neurons. The output layer of the BP neural network model has one output neuron. The BP neural network model can use a S-type transfer function, and a backpropagation error function, wherein Ti is the expected output and Oi is the calculated output of the network.

[0101] Further, the BP neural network model is a fully connected BP neural network model, the input layer of which can comprise three input neurons corresponding to the pixel combination number, the pixel point number ratio and the area of the region respectively, and the output layer of which can comprise one output neuron corresponding to different grades of the changeable degree of the current tunnel surrounding rock face.

[0102] Of course, the neural network model of the embodiment can use other supervised neural network models and is obtained through a large number of sample training.

[0103] Figure 4 A structure schematic diagram of a tunnel surrounding rock grading device disclosed in the embodiment is shown.

[0104] Figure 4 The device comprises a first shooting unit, a vibration intervention device, a second shooting unit, a change analysis unit and a change evaluation unit. The first shooting unit is configured to shoot a first tunnel image of a first tunnel face at at least one angle. The vibration intervention device is configured to intervene in the first tunnel face by vibration to obtain a second tunnel face. The second shooting unit is configured to shoot a second tunnel image of the second tunnel face at the angle. The change analysis unit is configured to obtain the change of the tunnel face according to the first tunnel image and the second tunnel image. The change evaluation unit is configured to grade the tunnel face according to the change of the tunnel face by the neural network model.

[0105] From the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized by means of software and necessary general hardware, and of course can also be realized by hardware, but in many cases the former is a better embodiment.

[0106] Based on such understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a FLASH, a hard disk, and / or an optical disc, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, and / or a network device, etc.) execute the method of various embodiments of the present application.

[0107] It is noted that the above merely describes the preferred embodiments of the present application and the principles of the applied technology. Those skilled in the art will understand that the present application is not limited to the specific embodiments herein, and that various obvious changes, re-adjustments and substitutions can be made by those skilled in the art without departing from the scope of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments, and can include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1.A method for tunnel surrounding rock classification, characterized in that, the method comprises, taking a first tunnel image of a first tunnel face at at least one angle; interfering with the first tunnel face by using excitation to form a second tunnel face; taking a second tunnel image of the second tunnel face at the angle; obtaining a gray scale ratio image of the first tunnel image and the second tunnel image; obtaining a clustering result of the gray scale ratio image by using a clustering algorithm; obtaining a clustering matrix representing the first tunnel image and the second tunnel image according to the clustering result; extracting the number of pixel points of the clustering matrix; extracting the coordinate relationship of each pixel point of the clustering matrix, and obtaining the number of at least paired or grouped pixel combinations in the clustering matrix according to the neighborhood relationship; obtaining a pixel area formed by each pixel combination in the clustering matrix according to a preset number of neighborhood relationships, and calculating the area ratio of the maximum pixel area to the minimum pixel area; the number of pixel combinations, the number of pixel points, and the area ratio represent the change of the tunnel face; classifying the tunnel face according to the change of the tunnel face by using a neural network model. 2.The method according to claim 1, characterized in that, obtaining a first gradient image of the first tunnel image and a second gradient image of the second tunnel image by using wavelet transform; obtaining the gray scale ratio image according to the first gradient image and the second gradient image. 3.The method according to claim 2, characterized in that, obtaining the first gradient image and the second gradient image at a resolution by using Haar wavelet transform; the resolution is selected according to the judgment of the clustering result. 4.The method according to claim 3, characterized in that, when the number of pixel points in the clustering matrix deviates too much from the intensity and / or frequency of excitation intervention, the resolution is reduced, or when the number of pixel points in the clustering matrix deviates too little from the intensity and / or frequency of excitation intervention, the resolution is increased. 5.The method according to claim 1, characterized in that, obtaining the clustering result of the gray scale ratio image by using an FCMA clustering algorithm. 6.The method according to claim 5, characterized in that, the FCMA clustering algorithm comprises, Among them, v j Let μ be the cluster center of the j-th category, N be the total number of pixels in the grayscale image, and μ be an NxC matrix. uj For x i The membership degree of category j, where C is the number of clusters, m is the weight, and ||x i -v j || represents the Euclidean distance from the i-th data point to the j-th cluster center, ‖x i -v k ‖ represents the Euclidean distance from the i-th data point to the k-th cluster center. C is 2, which is used to represent the clustering change pixel points and the unchanged pixel points, m is a natural number greater than 1. 7.The method according to claim 6, characterized in that, the value of m is selected according to the judgment of the clustering result. 8.A device for tunnel surrounding rock classification, characterized in that, the device comprises, a first shooting unit for taking a first tunnel image of a first tunnel face at at least one angle; an excitation intervention device for interfering with the first tunnel face by using excitation to form a second tunnel face; a second shooting unit for taking a second tunnel image of the second tunnel face at the angle; a change analysis unit for obtaining a gray scale ratio image of the first tunnel image and the second tunnel image; Obtaining a clustering result of the gray ratio image by using a clustering algorithm; Obtaining a clustering matrix representing the first tunnel image and the second tunnel image according to the clustering result; Extracting the number of pixel points of the clustering matrix; Extracting the coordinate relationship of each pixel point of the clustering matrix, and obtaining the number of at least paired or grouped pixel combinations in the clustering matrix according to the neighborhood relationship; aggregating each pixel combination in the clustering matrix according to a preset number of neighborhood relationships to obtain a pixel region formed by the pixel combination, and calculating the area ratio of the maximum pixel region and the minimum pixel region; the number of pixel combinations, the number of pixel points and the area ratio represent the change of the tunnel face; A change evaluation unit is configured to grade the tunnel face according to the change of the tunnel face by using a neural network model.

Images