Crack identification method and device

Abstract

The invention discloses a crack identification method and device, and the method comprises the steps: obtaining a first image of a stope similar model, the stope similar model being a model of an area needing to be excavated, and the first image being an image of the stope similar model which is not excavated; obtaining a second image of each time step after mining the similar stope model; based on the first image and the second image, obtaining a third image for crack extraction in each time step after mining; and extracting an original skeleton based on the third image, and repairing the original skeleton to obtain a repaired target skeleton. According to the method, the third image which eliminates the defect features of the stope similar model and is used for crack extraction is obtained through the first image and the second image, the original skeleton extracted based on the third image is repaired to obtain the repaired target skeleton, and a large amount of labeled data is not needed, so that the crack recognition result is more accurate, and the recognition efficiency is improved. And the accuracy of crack identification is improved.

Description

Technical Field

[0001] The invention relates to the technical field of stope, and in particular to a crack identification method and device. Background Art

[0002] During the mining process, physical models can be used to study the stability of mine structures (such as roofs, side walls, and floors). Physical models can be used to simulate the mine excavation process under different geological conditions, observe the deformation and failure mode of the surrounding rock, and thus evaluate the safety of the mine structure. In addition, during the test of the physical model, the evolution of the stress field inside the model is usually accompanied by the generation of cracks. Based on this, it is necessary to extract and analyze the cracks on the surface of the model, so that the laws of stress transfer, mine pressure manifestation, and overburden morphology evolution during the collapse of the overburden can be understood through the cracks. At present, a large number of crack identification methods are mainly used to identify the cracks.

[0003] In the prior art, cracks can be identified by image processing or machine learning. Among them, the image processing method may include edge detection, morphological operation, region growing, segmentation algorithm, etc. However, the above method is based on grayscale gradient to identify cracks, and it is difficult to distinguish between new cracks and original defect features of the physical model, making the crack identification result inaccurate.

[0004] In addition, machine learning methods may include the use of deep learning for crack identification. However, the above methods require a large amount of labeled data for training, and the physical model experiment cycle is long, making it difficult to obtain a large amount of valid data, and the accuracy of manual labeling of cracks on the model surface cannot be guaranteed, making machine learning unable to identify all crack features, resulting in inaccurate crack identification results. Summary of the invention

[0005] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0006] To this end, the present invention proposes a crack recognition method, which can obtain a third image for crack extraction that eliminates the defect features of a similar model of a mining area through a first image and a second image, and repair the original skeleton extracted based on the third image to obtain a repaired target skeleton. No large amount of labeled data is required, thereby making the crack recognition result more accurate and improving the accuracy of crack recognition.

[0007] Another object of the present invention is to provide a crack identification device.

[0008] To achieve the above object, the present invention provides a crack identification method, which comprises:

[0009] Acquire a first image of a stope similarity model, wherein the stope similarity model is a model of an area to be mined, and the first image is an image of the stope similarity model before excavation;

[0010] Acquire a second image at each time step after excavation of the similar model of the stope;

[0011] Based on the first image and the second image, obtaining a third image for crack extraction at each time step after mining;

[0012] An original skeleton is extracted based on the third image, and the original skeleton is repaired to obtain a repaired target skeleton.

[0013] The crack identification method of the embodiment of the present invention may also have the following additional technical features:

[0014] In one embodiment of the present invention, obtaining a third image for crack extraction at each time step after mining based on the first image and the second image includes:

[0015] Performing grayscale conversion on the first image and the second image to obtain corresponding first matrices and second matrices;

[0016] Based on the second matrix and the first matrix, interference of features in the stope similarity model is removed to obtain a corresponding third matrix;

[0017] Based on the third matrix, a third image for crack extraction at each time step after mining is obtained.

[0018] In one embodiment of the present invention, obtaining a third image for crack extraction at each time step after mining based on the third matrix includes:

[0019] Calculating the average value of the third matrix to obtain an average value;

[0020] Calculate the standard deviation of the third matrix based on the average value to obtain a corresponding first standard deviation matrix;

[0021] Performing reflection processing and strengthening processing on the first standard deviation matrix to obtain a corresponding fourth matrix;

[0022] Based on the fourth matrix, a third image for crack extraction at each time step after mining is obtained.

[0023] In one embodiment of the present invention, the step of performing reflection processing and enhancement processing on the first standard deviation matrix to obtain a corresponding fourth matrix includes:

[0024] determining whether there is a reflective first target matrix element in the first standard deviation matrix;

[0025] If there is a reflective first target matrix element in the first standard deviation matrix, modify the matrix value corresponding to the first target matrix element to obtain a modified fifth matrix;

[0026] Calculate the standard deviation of the fifth matrix to obtain a corresponding second standard deviation matrix, repeat the above steps until there is no reflective first target matrix element in the second standard deviation matrix, and determine the second standard deviation matrix as the sixth matrix;

[0027] The sixth matrix is ​​enhanced to obtain a corresponding fourth matrix.

[0028] In one embodiment of the present invention, obtaining a third image for crack extraction at each time step after mining based on the fourth matrix includes:

[0029] Performing gradient calculation and standard deviation calculation on the fourth matrix respectively to obtain a corresponding gradient matrix and a third standard deviation matrix;

[0030] Based on the gradient matrix and the third standard deviation matrix, a seventh matrix is ​​obtained;

[0031] A picture is drawn based on the seventh matrix to obtain a third image for crack extraction at each time step after mining.

[0032] In one embodiment of the present invention, extracting the original skeleton based on the third image and repairing the original skeleton to obtain a repaired target skeleton includes:

[0033] Based on the third image, obtaining a corresponding eighth matrix;

[0034] Identifying the matrix elements in the eighth matrix to obtain the second target matrix elements for extracting the original skeleton;

[0035] determining a target area corresponding to the second target matrix element;

[0036] Based on the length and width of the crack in the target area, extract a third target matrix element in the target area, and repeat the above steps until the original skeleton in the third image is extracted;

[0037] The original skeleton is repaired to obtain a repaired target skeleton.

[0038] In one embodiment of the present invention, the repairing the original skeleton to obtain a repaired target skeleton includes:

[0039] Determining whether the third target matrix element remains connected;

[0040] If it is determined that the third target matrix element does not remain connected, determining whether there are small crack connections between adjacent cracks;

[0041] If it is determined that there are small crack connections between adjacent cracks, the original skeleton is repaired based on the third target matrix element, and the above steps are repeated until the third target matrix element remains connected to obtain a repaired target skeleton.

[0042] Another aspect of the present invention provides a crack identification device, the device comprising:

[0043] A first acquisition module is used to acquire a first image of a stope similarity model, wherein the stope similarity model is a model of an area to be mined, and the first image is an image of the stope similarity model before excavation;

[0044] A second acquisition module is used to acquire a second image at each time step after mining the similar model of the stope;

[0045] A processing module, used for obtaining a third image for crack extraction at each time step after mining based on the first image and the second image;

[0046] The extraction module is used to extract the original skeleton based on the third image, and repair the original skeleton to obtain a repaired target skeleton.

[0047] The crack identification method and device of the embodiment of the present invention can obtain a third image for crack extraction that eliminates the defect features of the similar model of the mining area through the first image and the second image, and repair the original skeleton extracted based on the third image to obtain a repaired target skeleton, without the need for a large amount of labeled data, thereby making the crack identification result more accurate and improving the accuracy of crack identification.

[0048] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0050] Figure 1 is a flow chart of a crack identification method according to an embodiment of the present invention;

[0051] Figure 2 is a schematic diagram of a third image for crack extraction according to an embodiment of the present invention;

[0052] Figure 3is a schematic diagram of determining a midpoint based on a length and a width of a crack in a target region according to an embodiment of the present invention;

[0053] Figure 4 is a schematic diagram of a repaired target skeleton according to an embodiment of the present invention;

[0054] Figure 5 is a structural diagram of a crack identification device according to another embodiment of the present invention. DETAILED DESCRIPTION

[0055] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0056] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0057] The following describes a crack identification method and device according to an embodiment of the present invention with reference to the accompanying drawings.

[0058] Figure 1 4 is a flow chart of a crack identification method according to an embodiment of the present invention.

[0059] like Figure 1 As shown, the method may include the following steps:

[0060] Step 101, obtaining a first image of a stope similarity model.

[0061] In one embodiment of the present invention, the stope similarity model is a model of an area that needs to be excavated.

[0062] In one embodiment of the present invention, the core samples of the area to be excavated can be tested to obtain the thickness and arrangement order of different rock layers, the density and strength of the rock in the area to be excavated. In one embodiment of the present invention, similar materials of rock layers are made according to certain similarity ratios, and the similarity ratios in the material making process can include geometric similarity ratios, density similarity ratios and strength similarity ratios. The selectable materials include river sand, gypsum powder, calcium carbonate, water and silicate cement.

[0063] Furthermore, in one embodiment of the present invention, in the stope similarity model, mica needs to be spread between similar materials simulating different rock layers to weaken the adhesion between the layers, so as to simulate the weak interface between different rock layers and prevent the hydration reaction between the materials. Specifically, in one embodiment of the present invention, the mass G of the spread mica can be calculated by the formula G=[(γ c -γ)ρ m thn m S] / [α(γ c +γ m )], where γ is the surface tension, thnm is the coating thickness, and γ c is the surface tension between different similar materials, γ m is the minimum surface tension between different similar materials, ρ m is the density of mica, α is the spreading coefficient, where the spreading coefficient α>1. And, in one embodiment of the present invention, each layer of similar materials in the above-mentioned stope similarity model needs to be cut vertically multiple times to reduce the bending stiffness of the material. In the cutting process, the cutting spacing of materials at the same height is the same, and the cutting spacing of similar materials corresponding to the lower strength rock layer is shorter.

[0064] Furthermore, in one embodiment of the present invention, the surface of the stope-like model is polished and sprayed with white matte paint to cover its original defective features. The thickness of the matte paint should not affect the normal collapse of the model. Based on this, the thickness of the matte paint is Among them, ρ min is the minimum density among different rock formations, g is the gravitational acceleration, w is the width of the similarity model of the stope, hmin is the minimum thickness in the rock formation, σ c is the coating strength.

[0065] Furthermore, in one embodiment of the present invention, after obtaining the mine-similar model through the above steps, a lamp with a power greater than 100W can be used to illuminate the mine-similar model, wherein a soft light box needs to be installed on the lamp to make the brightness of the cracked area on the surface of the mine-similar model uniform.

[0066] And, in one embodiment of the present invention, it is necessary to determine the minimum target pixel required according to the structural parameters of the stope similarity model. Specifically, in one embodiment of the present invention, the minimum target pixel required for the stope similarity model can be determined by a pixel formula, wherein the pixel formula is: Wherein, H is the height of the stope similarity model, L is the length of the stope similarity model, wt is the characteristic width of the crack on the surface of the stope similarity model, and ws is the target width of the crack in the stope similarity model. For example, in one embodiment of the present invention, assuming that the length L and the height H of the stope similarity model are 3.0m and 1.65m respectively, the characteristic width of the crack is wt is 0.5mm, and the target width of the crack is ws is 5 pixels, then the corresponding minimum target pixel calculated by the above pixel formula is 4.95 million pixels, based on which, the pixel of the camera used to shoot the stope similarity model needs to be greater than 4.95 million pixels.

[0067] Furthermore, in one embodiment of the present invention, a first image of the stope similarity model may be acquired by a camera disposed on the stope similarity model, wherein the first image is an image of the stope similarity model before excavation.

[0068] Step 102, obtaining a second image at each time step after mining the similarity model of the stope.

[0069] In one embodiment of the present invention, after mining the similar model of the stope, the similar model of the stope corresponding to each time step after mining can be photographed at least once to obtain at least one second image of each time step, and the camera position and parameters cannot be adjusted during the photography to ensure that the camera parameters corresponding to the images are the same. In one embodiment of the present invention, the interval between different photography can be greater than 5s, and the number of photography corresponding to each time step is greater than 5 times to ensure that the cracks after each time step are captured.

[0070] Furthermore, in one embodiment of the present invention, after obtaining at least one second image of each time step after mining of the mining site similarity model through the above steps, the second images can be numbered, and the corresponding numbering format can be p_n_i, wherein n represents the mining time step, and i represents the i-th second image corresponding to the n-th time step.

[0071] Step 103: Based on the first image and the second image, a third image for crack extraction at each time step after mining is obtained.

[0072] In one embodiment of the present invention, after the first image and the second image are obtained through the above steps, a third image for crack extraction at each time step after mining can be obtained based on the first image and the second image.

[0073] Specifically, in one embodiment of the present invention, a method for obtaining a third image for crack extraction at each time step after mining based on the first image and the second image may include the following steps:

[0074] Step 1031, performing grayscale conversion on the first image and the second image to obtain corresponding first matrix and second matrix;

[0075] Step 1032, removing interference of features in the stope similarity model based on the second matrix and the first matrix to obtain a corresponding third matrix;

[0076] Step 1033: Based on the third matrix, a third image for crack extraction at each time step after mining is obtained.

[0077] Among them, in one embodiment of the present invention, the above-mentioned method of removing interference of features in the mining field similarity model based on the second matrix and the first matrix to obtain the corresponding third matrix may include: subtracting the second matrix from the first matrix to obtain the third matrix, thereby removing interference of original defect features on the surface of the mining field similarity model.

[0078] Furthermore, in one embodiment of the present invention, the method for obtaining the third image for crack extraction at each time step after mining based on the third matrix may include the following steps:

[0079] Step 1, calculating the average value of the third matrix to obtain the average value;

[0080] Step 2: Calculate the standard deviation of the third matrix based on the mean value to obtain the corresponding first standard deviation matrix;

[0081] Step 3, performing reflection processing and strengthening processing on the first standard deviation matrix to obtain a corresponding fourth matrix;

[0082] Step 4: Based on the fourth matrix, obtain the third image used for crack extraction at each time step after mining.

[0083] In one embodiment of the present invention, the third matrix p_n_is is averaged and the average value obtained is n is the number of rows of the third matrix, and m is the number of columns of the third matrix.

[0084] And, in one embodiment of the present invention, based on the above average value a, the standard deviation of the third matrix can be calculated to obtain the corresponding first standard deviation matrix

[0085] Further, in one embodiment of the present invention, after obtaining the first standard deviation matrix through the above steps, it is necessary to perform reflection processing and enhancement processing on the first standard deviation matrix to obtain a corresponding fourth matrix to eliminate noise caused by mica reflection in the image. In one embodiment of the present invention, the standard deviation at the reflection position is different from the standard deviation at other positions. Based on this, the above method of performing reflection processing and enhancement processing on the first standard deviation matrix to obtain the corresponding fourth matrix may include the following steps:

[0086] Step 31, determining whether there is a reflective first target matrix element in the first standard deviation matrix;

[0087] Step 32: if there is a reflective first target matrix element in the first standard deviation matrix, modify the matrix value corresponding to the first target matrix element to obtain a modified fifth matrix;

[0088] Step 33, calculate the standard deviation of the fifth matrix to obtain the corresponding second standard deviation matrix, repeat the above steps until there is no reflective first target matrix element in the second standard deviation matrix, and determine the second standard deviation matrix as the sixth matrix;

[0089] Step 34: Perform enhancement processing on the sixth matrix to obtain the corresponding fourth matrix.

[0090] Wherein, in one embodiment of the present invention, the method for determining whether there is a reflective first target matrix element in the first standard deviation matrix may include: determining whether there is a matrix element exceeding a first threshold in the first standard deviation matrix, and if there is a matrix element exceeding the first threshold, determining that there is a reflective first target matrix element in the first standard deviation matrix; if there is no matrix element exceeding the first threshold, determining that there is no reflective first target matrix element in the first standard deviation matrix. In one embodiment of the present invention, the first threshold can be any value in the range of (10, 20) as needed, for example, the first threshold is 15.

[0091] And, in one embodiment of the present invention, after determining the first target matrix element with reflection in the first standard deviation matrix through the above steps, the matrix value corresponding to the first target matrix element can be modified to obtain a modified fifth matrix. Specifically, in one embodiment of the present invention, the matrix value corresponding to the first target matrix element can be modified through the first formula to obtain a modified fifth matrix, wherein the first formula is in p_n_is(x i ,y j ) is the first target matrix element (x i ,y j )The modified matrix value.

[0092] Furthermore, in one embodiment of the present invention, after obtaining the fifth matrix through the above steps, the standard deviation of the fifth matrix can be calculated through the above steps 1 to 2 to obtain the corresponding second standard deviation matrix, and the above steps 31 to 32 are repeated until there are no reflective first target matrix elements in the second standard deviation matrix, and the second standard deviation matrix is ​​determined as the sixth matrix.

[0093] Furthermore, in one embodiment of the present invention, after the sixth matrix is ​​obtained through the above steps, the sixth matrix can be enhanced to facilitate subsequent crack capture. In one embodiment of the present invention, the straight line characteristics of the surface cracks in the stope similarity model are obvious, and any crack can be fitted by a finite number of horizontal and vertical lines. Based on this, the grayscale of the cracks in the image can be enhanced by scanning the straight line feature matrix to facilitate the identification and extraction of the cracks in the later stage. Specifically, in one embodiment of the present invention, for any a in the sixth matrix i,j After strengthening, the corresponding second-order matrix can be expressed as: The corresponding fourth matrix is ​​obtained through the above steps.

[0094] Furthermore, in one embodiment of the present invention, after the fourth matrix is ​​obtained through the above steps, a third image for crack extraction at each time step after mining can be obtained based on the fourth matrix.

[0095] In one embodiment of the present invention, the grayscale of the crack is obviously different from that of the area without cracks, and the grayscale gradient and standard deviation of the area where the crack is located in the image are significantly greater than those of other areas. Based on this, a third image for crack extraction can be obtained through the gradient matrix and the standard deviation matrix.

[0096] Specifically, in one embodiment of the present invention, the method for obtaining the third image for crack extraction at each time step after mining based on the fourth matrix may include the following steps:

[0097] Step 41, performing gradient calculation and standard deviation calculation on the fourth matrix respectively to obtain a corresponding gradient matrix and a third standard deviation matrix;

[0098] Step 42: obtaining a seventh matrix based on the gradient matrix and the third standard deviation matrix;

[0099] Step 43: Draw a picture based on the seventh matrix to obtain a third image for crack extraction at each time step after mining.

[0100] In one embodiment of the present invention, the fourth matrix is ​​subjected to gradient calculation and standard deviation calculation respectively to obtain a corresponding gradient matrix and a third standard deviation matrix, wherein the matrix elements in the gradient matrix can be

[0101]

[0102] Among them, p_n_iss(x i+1 ,y j ) is the fourth matrix (x i+1 ,y j ) The matrix values ​​corresponding to the matrix elements; the matrix elements in the third standard deviation matrix can be in,

[0103] Further, in one embodiment of the present invention, after the gradient matrix and the third standard deviation matrix are obtained through the above steps, a seventh matrix can be obtained based on the gradient matrix and the third standard deviation matrix. Wherein, in one embodiment of the present invention, the method for obtaining the seventh matrix based on the gradient matrix and the third standard deviation matrix can include: obtaining a ninth matrix based on the gradient matrix and the third standard deviation matrix through the second formula, and obtaining the seventh matrix based on the ninth matrix.

[0104] In one embodiment of the present invention, the second formula is:

[0105] Among them, p_n(x i ,y j ) is the matrix element in the ninth matrix.

[0106] And, in one embodiment of the present invention, after the ninth matrix is ​​obtained through the above steps, the seventh matrix can be obtained based on the ninth matrix through the third formula, where the third formula is c=p_n(x i ,y j ), c 0 Any value in the range of (30, 40) can be used as required, for example, c 0 It can be 35.

[0107] Further, in one embodiment of the present invention, after the seventh matrix is ​​obtained through the above steps, a picture can be drawn based on the seventh matrix, and the matrix elements with a matrix value of 1 in the seventh matrix are the locations of the cracks. Based on this, a third image (such as Figure 2 shown).

[0108] Step 104: extracting an original skeleton based on the third image, and repairing the original skeleton to obtain a repaired target skeleton.

[0109] In one embodiment of the present invention, after the third image is obtained through the above steps, the original skeleton can be extracted based on the third image, and the original skeleton can be repaired to obtain a repaired target skeleton.

[0110] Specifically, in one embodiment of the present invention, the method of extracting the original skeleton based on the third image and repairing the original skeleton to obtain a repaired target skeleton may include the following steps:

[0111] Step 1041, obtaining a corresponding eighth matrix based on the third image;

[0112] Step 1042, identifying the matrix elements in the eighth matrix to obtain the second target matrix elements for extracting the original skeleton;

[0113] Step 1043, determining the target area corresponding to the second target matrix element;

[0114] Step 1044, extracting a third target matrix element in the target area based on the length and width of the crack in the target area, and repeating the above steps until the original skeleton in the third image is extracted;

[0115] Step 1045, repair the original skeleton to obtain a repaired target skeleton.

[0116] In one embodiment of the present invention, after the third image is obtained through the above steps, an eighth matrix can be obtained based on the third image, wherein the eighth matrix is ​​the same as the seventh matrix.

[0117] Furthermore, in one embodiment of the present invention, the region where the crack is located is a simply connected region. Based on this, when extracting the crack skeleton, the midline of the crack feature can be selected as the skeleton of the crack. Among them, cracks in different regions can be selected for analysis in turn. If the width of the crack in the region is greater than the length, the midpoint of the region is taken horizontally, otherwise the midpoint is taken longitudinally (such as Figure 3 shown).

[0118] In one embodiment of the present invention, the second target matrix element may be a matrix element having a matrix value of 1 in the eighth matrix.

[0119] And, in one embodiment of the present invention, the above method for determining the target area corresponding to the second target matrix element may include: identifying the subsequent rows of the column to which the second target matrix element belongs until a row target element whose matrix value is a preset value is identified, and identifying the subsequent columns of the row to which the second target matrix element belongs until a column target element whose matrix value is a preset value is identified, and determining the area composed of the second target matrix element to the row target element and the second target matrix element to the column target element as the target area corresponding to the second target matrix element.

[0120] For example, in one embodiment of the present invention, identification is performed from top to bottom and from left to right starting from the element at the (1, 1) coordinate in the eighth matrix until all matrix elements in the eighth matrix are identified. i ,y j ) is 1, then identify (x i+1 ,y j ) to (x i+p ,y j ) up to (x i+p ,y j) If the matrix value at is 1, stop the recognition; simultaneously recognize the matrix elements from (x i , y j+1 ) to (x i , y j+q ), until the matrix value at (x i , y j+q ) is 1, then stop the recognition. At this time, the area composed of (x i , y i ) to (x i+p , y j ) and (x i , y j ) to (x i , y j+q ) can be determined as the target area corresponding to the second target matrix element. Among them, in an embodiment of the present invention, p is the width of the target area, and q is the length of the target area.

[0121] Among them, in an embodiment of the present invention, the method for extracting the third target matrix element in the target area based on the length and width of the crack in the target area may include: if the width of the crack in the target area is less than the length, the longitudinal midpoint is determined as the third target matrix element; if the width of the crack in the target area is greater than the length, the transverse midpoint is determined as the third target matrix element.

[0122] For example, in an embodiment of the present invention, if p < q, the matrix values from (x i , y j ) to (x s-1 , y j ) and from (x s+1 , y j ) to (x i+n , y j ) are 0; if p > q, the matrix values from (x i , y j ) to (x i , y t-1 ) and from (x i , y t+1 ) to (x i , y j+q ) are 0, where s and t are [(i + s) / 2] and [(i + t) / 2] respectively, and [] is the rounding symbol.

[0123] Also, in an embodiment of the present invention, repeat the above steps until all the second target matrix elements identified in the eighth matrix have completed the above steps, and the original skeleton in the third image can be extracted through the processed eighth matrix.

[0124] Furthermore, in one embodiment of the present invention, after the original skeleton in the third image is extracted through the above steps, the above method may cause the originally connected area to be disconnected after the original skeleton is extracted. Based on this, after taking the midpoint through the above steps, it is necessary to determine whether it remains connected. If the skeleton is not connected, it means that the skeleton is broken and needs to be repaired to connect two points within a certain distance, so that the crack identification result is more accurate. Also, in one embodiment of the present invention, there are usually small cracks connecting two adjacent cracks, but these small cracks may be missed during the image recognition process. Based on this, during the skeleton repair process, the distance between the crack ends can be identified to distinguish whether there are small cracks connecting adjacent cracks, so as to perform repairs.

[0125] In one embodiment of the present invention, the method for repairing the original skeleton to obtain the repaired target skeleton may include the following steps:

[0126] Step 10451, determining whether the third target matrix element remains connected;

[0127] Step 10452, if it is determined that the third target matrix element is not connected, determine whether there are small crack connections between adjacent cracks;

[0128] Step 10453, if it is determined that there are small crack connections between adjacent cracks, the original skeleton is repaired based on the third target matrix element, and the above steps are repeated until the third target matrix element remains connected to obtain a repaired target skeleton.

[0129] Wherein, in one embodiment of the present invention, the method for determining whether the third target matrix element remains connected may include: determining the number of matrix elements in the connected region corresponding to the third target matrix element that are preset values, and if the number of matrix elements in the connected region that are preset values ​​is greater than or equal to a second threshold, determining that the third target matrix element remains connected; if the number of matrix elements in the connected region that are preset values ​​is less than the second threshold, determining that the third target matrix element does not remain connected. Wherein, in one embodiment of the present invention, the preset value and the second threshold can be set as needed, for example, the preset value is 1, and the second threshold is 1.

[0130] In one embodiment of the present invention, if the third target matrix element is a midpoint in the vertical direction, the connected area corresponding to the third target matrix element is the previous column, the same column and the next column of the previous row of the third target matrix element; if the third target matrix element is a midpoint in the horizontal direction, the connected area corresponding to the third target matrix element is the previous column, the same row and the next row of the previous column of the third target matrix element. For example, assuming that the third target matrix element is (x s ,y j), then the connected area corresponding to the third target matrix element is (x s+1 ,y i-1 ), (x s-1 ,y i-1 ) and (x s ,y i-1 ).

[0131] In one embodiment of the present invention, if it is determined that the elements of the third target matrix remain connected, there is no need to repair the third target matrix.

[0132] And, in one embodiment of the present invention, the above method for determining whether there are small crack connections between adjacent cracks may include: if the third target matrix element is a horizontal midpoint, then determining a fourth target matrix element with a preset value in the previous row of the target row to which the third target matrix element belongs, determining the first absolute value of the difference between the horizontal coordinate of the fourth target matrix element and the horizontal coordinate of the third target matrix element, if the first absolute value is less than a third threshold, then determining that there are small crack connections between adjacent cracks; if the third target matrix element is a vertical midpoint, then determining a fifth target matrix element with a preset value in the previous column of the target row to which the third target matrix element belongs, determining the second absolute value of the difference between the vertical coordinate of the fifth target matrix element and the vertical coordinate of the third target matrix element, if the second absolute value is less than the third threshold, then determining that there are small crack connections between adjacent cracks. Wherein, the third threshold can be set as needed, for example, the third threshold can be 5.

[0133] For example, in one embodiment of the present invention, it is assumed that the third target matrix element (x s ,y j ) is the horizontal midpoint, then determine the fourth target matrix element that is 1 in the previous row s-1 of the target row s to which the third target matrix element belongs, determine the first absolute value |zs| of the difference between the horizontal coordinate z of the fourth target matrix element and the horizontal coordinate s of the third target matrix element, and if the first absolute value |zs| is less than 5, it is determined that there are small crack connections between adjacent cracks.

[0134] In one embodiment of the present invention, if it is determined that there are no small cracks between adjacent cracks of the third target matrix element, there is no need to repair the third target element.

[0135] Further, in one embodiment of the present invention, the method for repairing the original skeleton based on the third target matrix element may include: if the third target matrix element is a horizontal midpoint, then the matrix values ​​corresponding to the matrix elements between the horizontal coordinate corresponding to the fourth target matrix element in the target column to which the third target matrix element belongs and the horizontal coordinate of the third target matrix element are all modified to preset values; if the third target matrix element is a vertical midpoint, then the matrix values ​​corresponding to the matrix elements between the vertical coordinate of the fifth target matrix element in the target row to which the third target matrix element belongs and the vertical coordinate of the third target matrix element are all modified to preset values.

[0136] For example, in one embodiment of the present invention, it is assumed that the third target matrix element (x s ,y j ) is the horizontal midpoint, then the fourth target matrix element (x z ,y j ) to the third target matrix element (x s ,y j ) are all changed to 1.

[0137] Further, in one embodiment of the present invention, the target skeleton after repair is obtained through the above steps as follows: Figure 4 shown.

[0138] Furthermore, in one embodiment of the present invention, after obtaining the target skeleton through the above steps, the above method may further include: evaluating the target skeleton to obtain an evaluation result, and adjusting the target skeleton based on the evaluation result.

[0139] In one embodiment of the present invention, the method for evaluating the target skeleton to obtain an evaluation result may include the following steps:

[0140] Step a, obtaining a fourth image with a skeleton marked corresponding to the second image, and performing grayscale conversion on the fourth image to obtain a corresponding evaluation matrix;

[0141] Step b, obtaining the target matrix corresponding to the target skeleton;

[0142] Step c, adding the evaluation matrix and the target matrix to obtain a tenth matrix;

[0143] Step d, determining the number of first elements whose matrix values ​​are the first numerical values ​​in the tenth matrix;

[0144] Step e, determining the number of second elements in the target matrix whose matrix values ​​are second numerical values;

[0145] Step f: Based on the number of first elements and the number of second elements, an evaluation result is calculated by a fourth formula, wherein the fourth formula is β=Nu / Mu, wherein Nu is the number of first elements and Mu is the number of second elements.

[0146] In one embodiment of the present disclosure, the first value and the second value can be set as needed, for example, the first value is 2 and the second value is 1.

[0147] And, in one embodiment of the present disclosure, after obtaining the evaluation result β through the above steps, the difference between the evaluation result β and the second value can be calculated. If the difference is less than or equal to the fourth threshold, it means that the extraction effect of the target skeleton is very good and no adjustment is required; otherwise, the target skeleton needs to be adjusted. In one embodiment of the present invention, the target skeleton can be adjusted through manual experience.

[0148] The crack recognition method of the embodiment of the present invention obtains a third image for crack extraction that eliminates the defect features of the similar model of the mining area through the first image and the second image, and repairs the original skeleton extracted based on the third image to obtain a repaired target skeleton. It does not require a large amount of labeled data, thereby making the crack recognition result more accurate and improving the accuracy of crack recognition.

[0149] Figure 5 Schematic diagram of the structure of a crack identification device 10 according to an embodiment of the present invention.

[0150] like Figure 5 As shown, the device may include:

[0151] A first acquisition module 501 is used to acquire a first image of a stope similarity model, wherein the stope similarity model is a model of an area to be mined, and the first image is an image of the stope similarity model before mining;

[0152] The second acquisition module 502 is used to acquire a second image at each time step after mining the similar model of the stope;

[0153] A processing module 503 is used to obtain a third image for crack extraction at each time step after mining based on the first image and the second image;

[0154] The extraction module 504 is used to extract the original skeleton based on the third image, and repair the original skeleton to obtain a repaired target skeleton.

[0155] In one embodiment of the present disclosure, the processing module 503 is specifically used to:

[0156] Performing grayscale conversion on the first image and the second image to obtain corresponding first matrix and second matrix;

[0157] Based on the second matrix and the first matrix, interference of features in the stope similarity model is removed to obtain a corresponding third matrix;

[0158] Based on the third matrix, the third image used for crack extraction at each time step after mining is obtained.

[0159] In one embodiment of the present disclosure, the processing module 503 is further configured to:

[0160] Calculate the average value of the third matrix to obtain the average value;

[0161] Calculate the standard deviation of the third matrix based on the mean value to obtain the corresponding first standard deviation matrix;

[0162] Performing reflection processing and strengthening processing on the first standard deviation matrix to obtain a corresponding fourth matrix;

[0163] Based on the fourth matrix, the third image used for crack extraction at each time step after mining is obtained.

[0164] In one embodiment of the present disclosure, the processing module 503 is further configured to:

[0165] determining whether there is a first target matrix element that is reflective in the first standard deviation matrix;

[0166] If there is a reflective first target matrix element in the first standard deviation matrix, modify the matrix value corresponding to the first target matrix element to obtain a modified fifth matrix;

[0167] Calculate the standard deviation of the fifth matrix to obtain the corresponding second standard deviation matrix, repeat the above steps until there is no reflective first target matrix element in the second standard deviation matrix, and determine the second standard deviation matrix as the sixth matrix;

[0168] The sixth matrix is ​​enhanced to obtain the corresponding fourth matrix.

[0169] In one embodiment of the present disclosure, the processing module 503 is further configured to:

[0170] Performing gradient calculation and standard deviation calculation on the fourth matrix respectively to obtain a corresponding gradient matrix and a third standard deviation matrix;

[0171] Based on the gradient matrix and the third standard deviation matrix, a seventh matrix is ​​obtained;

[0172] The picture is drawn based on the seventh matrix to obtain the third image used for crack extraction at each time step after mining.

[0173] In one embodiment of the present disclosure, the extraction module 504 is specifically used to:

[0174] Based on the third image, a corresponding eighth matrix is ​​obtained;

[0175] Identify the matrix elements in the eighth matrix to obtain the second target matrix elements for extracting the original skeleton;

[0176] determining a target region corresponding to an element of a second target matrix;

[0177] Based on the length and width of the crack in the target area, extract the third target matrix element in the target area, and repeat the above steps until the original skeleton in the third image is extracted;

[0178] The original skeleton is repaired to obtain a repaired target skeleton.

[0179] In one embodiment of the present disclosure, the extraction module 504 is further used to:

[0180] determining whether the third target matrix elements remain connected;

[0181] If it is determined that the third target matrix element does not remain connected, determining whether there are small crack connections between adjacent cracks;

[0182] If it is determined that there are small crack connections between adjacent cracks, the original skeleton is repaired based on the third target matrix element, and the above steps are repeated until the third target matrix element remains connected to obtain a repaired target skeleton.

[0183] The crack identification device of the embodiment of the present invention obtains a third image for crack extraction that eliminates the defect features of the similar model of the mining area through the first image and the second image, and repairs the original skeleton extracted based on the third image to obtain a repaired target skeleton. It does not require a large amount of labeled data, thereby making the crack identification result more accurate and improving the accuracy of crack identification.

[0184] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0185] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

Claims

1. A crack identification method, characterized in that: The method comprises: Acquire a first image of a stope similarity model, wherein the stope similarity model is a model of an area to be mined, and the first image is an image of the stope similarity model before excavation; Acquire a second image at each time step after excavation of the similar model of the stope; Based on the first image and the second image, obtaining a third image for crack extraction at each time step after mining; An original skeleton is extracted based on the third image, and the original skeleton is repaired to obtain a repaired target skeleton.

2. The method according to claim 1, characterized in that The obtaining, based on the first image and the second image, a third image for crack extraction at each time step after mining, comprises: Performing grayscale conversion on the first image and the second image to obtain corresponding first matrices and second matrices; Based on the second matrix and the first matrix, interference of features in the stope similarity model is removed to obtain a corresponding third matrix; Based on the third matrix, a third image for crack extraction at each time step after mining is obtained.

3. The method according to claim 2, characterized in that The method of obtaining a third image for crack extraction at each time step after mining based on the third matrix includes: Calculating the average value of the third matrix to obtain an average value; Calculate the standard deviation of the third matrix based on the average value to obtain a corresponding first standard deviation matrix; Performing reflection processing and strengthening processing on the first standard deviation matrix to obtain a corresponding fourth matrix; Based on the fourth matrix, a third image for crack extraction at each time step after mining is obtained.

4. The method according to claim 3, characterized in that The performing reflection processing and strengthening processing on the first standard deviation matrix to obtain a corresponding fourth matrix includes: determining whether there is a reflective first target matrix element in the first standard deviation matrix; If there is a reflective first target matrix element in the first standard deviation matrix, modify the matrix value corresponding to the first target matrix element to obtain a modified fifth matrix; Calculate the standard deviation of the fifth matrix to obtain a corresponding second standard deviation matrix, repeat the above steps until there is no reflective first target matrix element in the second standard deviation matrix, and determine the second standard deviation matrix as the sixth matrix; The sixth matrix is ​​enhanced to obtain a corresponding fourth matrix.

5. The method according to claim 3, characterized in that: The method of obtaining a third image for crack extraction at each time step after mining based on the fourth matrix includes: Performing gradient calculation and standard deviation calculation on the fourth matrix respectively to obtain a corresponding gradient matrix and a third standard deviation matrix; Based on the gradient matrix and the third standard deviation matrix, a seventh matrix is ​​obtained; A picture is drawn based on the seventh matrix to obtain a third image for crack extraction at each time step after mining.

6. The method according to claim 1, characterized in that The extracting the original skeleton based on the third image and repairing the original skeleton to obtain a repaired target skeleton includes: Based on the third image, obtaining a corresponding eighth matrix; Identifying the matrix elements in the eighth matrix to obtain the second target matrix elements for extracting the original skeleton; determining a target area corresponding to the second target matrix element; Based on the length and width of the crack in the target area, extract a third target matrix element in the target area, and repeat the above steps until the original skeleton in the third image is extracted; The original skeleton is repaired to obtain a repaired target skeleton.

7. The method according to claim 6, characterized in that The repairing of the original skeleton to obtain a repaired target skeleton includes: Determining whether the third target matrix element remains connected; If it is determined that the third target matrix element does not remain connected, determining whether there are small crack connections between adjacent cracks; If it is determined that there are small crack connections between adjacent cracks, the original skeleton is repaired based on the third target matrix element, and the above steps are repeated until the third target matrix element remains connected to obtain a repaired target skeleton.

8. A crack identification device, characterized in that: The device comprises: A first acquisition module is used to acquire a first image of a stope similarity model, wherein the stope similarity model is a model of an area to be mined, and the first image is an image of the stope similarity model before excavation; A second acquisition module is used to acquire a second image at each time step after mining the similar model of the stope; A processing module, used for obtaining a third image for crack extraction at each time step after mining based on the first image and the second image; The extraction module is used to extract the original skeleton based on the third image, and repair the original skeleton to obtain a repaired target skeleton.

9. A computer storage medium, wherein: The computer storage medium stores computer executable instructions; after the computer executable instructions are executed by the processor, any method described in claims 1-7 can be implemented.

10. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images