A tunnel deformation mode recognition and quantitative analysis method and system

Abstract

The application belongs to the field of engineering surveying and data processing, and discloses a tunnel deformation mode recognition and quantitative analysis method and system. The tunnel deformation mode recognition and quantitative analysis method is based on two-dimensional profile data of laser scanning, and comprises the steps of data acquisition, data preprocessing, profile curve fitting, deformation feature extraction, super-threshold binarization, abnormal area positioning, deformation mode recognition and deformation amount calculation and output. The application uses two-dimensional profile data of laser scanning to obtain tunnel cross section profile information, which can more completely reflect the profile deformation state of the tunnel cross section compared with the method of relying on only a small number of monitoring points. Meanwhile, the application constructs a tunnel deformation feature matrix containing horizontal displacement component, vertical displacement component and curvature change component, so that the spatial position change and local geometric shape change are considered in the deformation analysis, which is beneficial to distinguishing different types of cross section deformation.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of engineering surveying and data processing, and particularly relates to a tunnel deformation mode recognition and quantitative analysis method and system. BACKGROUND

[0002] Under long-term operation and external disturbance, the physical form of a complex spatial structure (such as a tunnel) will produce multi-dimensional state changes (such as overall spatial displacement, local profile deformation, curvature change, and multi-variable evolution). If such multi-variable spatial deformation data cannot be continuously collected, real-time transmitted, and high-precision quantified, it will seriously restrict the accurate digital evaluation of the structure state. Therefore, multi-variable comprehensive measurement of complex cross-section profiles is of great significance.

[0003] Traditional measurement methods rely on level meters, convergence meters and other measurement devices that are dedicated to single variables or based on contact. Such devices can only obtain single-dimensional physical quantities (such as absolute distance or elevation) of discrete nodes in space, and are difficult to fully reflect the multi-variable correlation evolution of the overall profile form of the measured object (such as the full cross-section form); at the same time, manual or contact measurement has obvious deficiencies in data automatic collection, continuous transmission and other capabilities, and is difficult to meet the needs of modern digital measurement.

[0004] With the development of photoelectric sensing and spatial information technology, using non-contact laser scanning devices to obtain high-density spatial profile data points has become the mainstream measurement method. Laser scanning can obtain a large amount of coordinate signals in a short time, which is conducive to realizing panoramic mapping of the measured profile form. However, in actual complex measurement environment, photoelectric receiving signals are easily affected by equipment shielding, dynamic interference and environmental factors, resulting in a large number of outliers and random noise signals mixed in the original measurement data stream. At this time, if only threshold determination of single variable measurement values is used for abnormal triggering, false positives and false negatives of measurement are likely to occur due to system noise interference. Therefore, there is an urgent need for a comprehensive processing method for non-contact two-dimensional spatial measurement data to establish a digital closed-loop measurement framework from original perception signal denoising, multi-dimensional feature parameter extraction, digital mode recognition to comprehensive deformation variable high-precision output. SUMMARY

[0005] The technical problem to be solved by the present application is that the existing deformation monitoring method for complex spatial structures such as tunnels has the problems of insufficient utilization of cross-section profiles, noise points easily leading to misjudgment, difficulty in distinguishing deformation modes, and insufficient calculation accuracy of deformation variables under the condition of incomplete correspondence of point clouds. The present application provides a tunnel deformation mode recognition and quantitative analysis method based on laser scanning two-dimensional profile data.

[0006] The present application relates to tunnel cross-section two-dimensional profile data collection and processing, profile curve reconstruction, deformation feature extraction, abnormal deformation area identification, deformation mode determination and deformation calculation, which can be used for construction period monitoring, operation period inspection, safety early warning and structure health assessment of highway tunnels, railway tunnels, urban rail transit tunnels, hydraulic tunnels and other projects.

[0007] Further, the present application also provides a system corresponding to the above method, so as to realize automatic deployment in an engineering monitoring platform, an edge computing terminal or a server.

[0008] In a first aspect, the present application provides a tunnel deformation mode identification and quantitative analysis method, comprising the following steps: S1, data acquisition: obtaining two-dimensional profile data of the same tunnel cross-section in an initial state and two-dimensional profile data in a monitoring state, and unifying the two sets of two-dimensional profile data to the same cross-section coordinate system, and performing coordinate conversion and registration on the two-dimensional profile data in the initial state and the two-dimensional profile data in the monitoring state according to the position parameters of the scanning device, the cross-section center point, the polar coordinate angle or the preset reference datum; S2, data preprocessing: preprocessing the two-dimensional profile data in the initial state and the two-dimensional profile data in the monitoring state to obtain an initial profile point set and a monitoring profile point set, wherein the preprocessing includes one or more of filtering processing, clustering denoising processing and outlier rejection processing; S3, profile curve fitting: establishing a corresponding relationship between sampling points of the initial profile point set and the monitoring profile point set, and respectively reconstructing an initial profile curve and a monitoring profile curve based on the initial profile point set and the monitoring profile point set; S4, deformation feature extraction: based on the mutually corresponding sampling points and the initial profile curve and the monitoring profile curve, calculating the horizontal displacement component, the vertical displacement component and the curvature change component of the profile curve at the sampling point position of each sampling point, and constructing a tunnel deformation feature matrix; S5, super-threshold binaryzation: comparing the tunnel deformation feature matrix with a preset threshold vector to generate a binary detection matrix for representing whether each feature component of each sampling point is over-standard; S6, abnormal area positioning: calculating the global over-standard proportion of the binary detection matrix, and when the global over-standard proportion exceeds a preset proportion threshold, determining an abnormal deformation area according to the sampling point positions corresponding to the over-standard elements in the binary detection matrix, wherein adjacent or continuously distributed over-standard sampling points are divided into the same abnormal deformation area; S7, deformation mode identification: identifying the tunnel deformation mode according to the spatial distribution, displacement direction and curvature change characteristics of the over-standard elements in the abnormal deformation area; S8, deformation amount calculation and output: taking the initial profile curve as a reference profile, the normal deformation amount of the monitoring profile points in the abnormal deformation area is calculated by using the minimum distance projection method, and the deformation area, deformation mode and deformation amount are output.

[0009] Preferably, in the step S3, the sampling point correspondence is established in the order of polar angle, profile curve, arc length position or sampling index, and the initial profile curve and the monitoring profile curve are reconstructed by using at least one of the spline interpolation, polynomial fitting or B-spline fitting method to obtain a continuous tunnel section profile curve.

[0010] Preferably, in the step S4, the tunnel deformation feature matrix includes a feature row corresponding to a sampling point, and each feature row includes at least the horizontal displacement component, the vertical displacement component and the curvature change component of the corresponding sampling point; the horizontal displacement component is represented by the coordinate difference between the sampling point in the monitoring state and the corresponding sampling point in the initial state in the horizontal direction; the vertical displacement component is represented by the coordinate difference between the sampling point in the monitoring state and the corresponding sampling point in the initial state in the vertical direction; and the curvature change component is represented by the difference between the curvature of the monitoring profile curve at the corresponding position and the curvature of the initial profile curve at the corresponding position.

[0011] Preferably, generating the binary detection matrix includes: comparing each element in the tunnel deformation feature matrix with a corresponding threshold value in the preset threshold value vector respectively; when the absolute value of the element exceeds the corresponding threshold value, the corresponding element in the binary detection matrix is marked as an out-of-standard state; when the absolute value of the element does not exceed the corresponding threshold value, the corresponding element in the binary detection matrix is marked as an in-standard state; and the global out-of-standard proportion is the ratio of the number of out-of-standard elements in the binary detection matrix to the total number of elements in the binary detection matrix.

[0012] Preferably, the tunnel deformation mode includes local deformation, overall settlement and hole body contraction, wherein, when the out-of-standard elements are concentrated in a local profile section corresponding to continuous sampling points, and there is a curvature change abnormality in the corresponding section, the local deformation is identified; when the vertical displacement components of multiple sampling points are distributed in the same direction, the overall settlement is identified; and when the displacement directions of multiple sampling points point to the tunnel section centroid, the hole body contraction is identified.

[0013] Further preferably, in the step of S8, the minimum distance projection manner comprises: for a to-be-quantified monitoring contour point in the abnormal deformation area, determining a reference curve segment on the initial contour curve closest to the to-be-quantified monitoring contour point; calculating an orthogonal projection point of the to-be-quantified monitoring contour point on the reference curve segment; determining a distance from the to-be-quantified monitoring contour point to the orthogonal projection point as a normal deformation variable of the corresponding position; when the orthogonal projection point is located outside the reference curve segment, selecting an end point or an adjacent reference curve segment closest to the to-be-quantified monitoring contour point to re-determine the projection point.

[0014] Further preferably, when the tunnel deformation mode is overall settlement, an overall settlement deformation variable is output according to a vertical displacement component, a maximum value of vertical displacement or an average value of vertical displacement of the sampling points in the abnormal deformation area; when the tunnel deformation mode is local deformation or hole body contraction, a local deformation variable or a contraction deformation variable is output according to the normal deformation variable obtained by the minimum distance projection manner.

[0015] In a second aspect, the present application provides a tunnel deformation mode recognition and quantification analysis system, comprising: a data acquisition module, which is configured to acquire two-dimensional contour data of a same tunnel section in an initial state and two-dimensional contour data of the same tunnel section in a monitoring state, and to unify the two sets of two-dimensional contour data to a same section coordinate system, and to perform coordinate conversion and registration on the two-dimensional contour data in the initial state and the two-dimensional contour data in the monitoring state according to a position parameter of a scanning device, a section center point, a polar coordinate angle or a preset reference datum; a data preprocessing module, which is configured to perform preprocessing on the two-dimensional contour data in the initial state and the two-dimensional contour data in the monitoring state to obtain an initial contour point set and a monitoring contour point set, and the preprocessing comprises one or more of filtering processing, clustering denoising processing and outlier rejection processing; a contour curve fitting module, which is configured to establish a sampling point correspondence between the initial contour point set and the monitoring contour point set, and to reconstruct an initial contour curve and a monitoring contour curve based on the initial contour point set and the monitoring contour point set, respectively; a deformation feature extraction module, which is configured to calculate a horizontal displacement component, a vertical displacement component and a curvature change component of a contour curve at a sampling point position of each sampling point based on the mutually corresponding sampling points and the initial contour curve and the monitoring contour curve, and to construct a tunnel deformation feature matrix; a super-threshold binaryzation module, which is configured to compare the tunnel deformation feature matrix with a preset threshold vector to generate a binaryzation detection matrix for representing whether each feature component of each sampling point is over-standard. an abnormal area positioning module, configured to calculate a global over-standard proportion of the binarized detection matrix, and determine an abnormal deformation area according to positions of sampling points corresponding to over-standard elements in the binarized detection matrix when the global over-standard proportion exceeds a preset proportion threshold, wherein over-standard sampling points that are adjacent or continuously distributed are divided into the same abnormal deformation area; a deformation mode recognition module, configured to recognize a tunnel deformation mode according to spatial distribution, displacement direction and curvature change characteristics of the over-standard elements in the abnormal deformation area; a deformation variable calculation output module, configured to take the initial profile curve as a reference profile, calculate a normal deformation variable of a monitoring profile point in the abnormal deformation area by using a minimum distance projection method, and output the deformation area, the deformation mode and the deformation variable.

[0016] Compared with the prior art, the present application has the following beneficial effects: The present application uses laser scanning two-dimensional profile data to obtain tunnel cross-section profile information, which can more completely reflect the profile deformation state of the tunnel cross-section compared with the method of relying only on a small number of monitoring points. The present application constructs a tunnel deformation feature matrix containing horizontal displacement component, vertical displacement component and curvature change component, so that the spatial position change and local geometric shape change are considered in the deformation analysis, which is beneficial to distinguish different types of cross-section deformation. The present application triggers the abnormality by combining the threshold vector, the binarized detection matrix and the global over-standard proportion, which can reduce the possibility of misjudgment caused by isolated noise points or accidental occlusion points. The present application recognizes modes such as local deformation, overall settlement and hole contraction according to the spatial distribution, displacement direction and curvature change characteristics of the over-standard elements in the abnormal deformation area, which can provide more abundant state information than single displacement value for tunnel deformation measurement. The present application takes the initial profile curve as the reference profile, and calculates the normal deformation variable of the monitoring point in the abnormal area to the reference profile by using the minimum distance projection, which can reduce the quantization deviation caused by directly calculating the distance of the corresponding points when the front and rear point clouds are not completely aligned. The threshold vector of the present application can be set according to the monitoring level, the design allowed deformation amount, the early warning standard or the historical monitoring data, which can adapt to the monitoring requirements in different engineering scenes. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 It is a tunnel deformation mode recognition and quantitative analysis method flow chart.

[0018] Figure 2 It is a tunnel two-dimensional profile data acquisition and coordinate unification schematic diagram.

[0019] Figure 3 It is a tunnel deformation feature matrix construction schematic diagram.

[0020] Figure 4 The tunnel deformation binarization detection matrix construction method is shown in the figure.

[0021] Figure 5 The binarization detection matrix and global over-limit proportion calculation method is shown in the figure.

[0022] Figure 6 The abnormal deformation area and pattern recognition schematic diagram is shown in the figure.

[0023] Figure 7 The minimum distance projection normal deformation variable calculation schematic diagram is shown in the figure.

[0024] Figure 8 The module schematic diagram of the tunnel deformation pattern recognition and quantitative analysis system is shown in the figure.

[0025] Markings in the figure: 100 is a tunnel deformation pattern recognition and quantitative analysis method, S101-S108 are method steps of the tunnel deformation pattern recognition and quantitative analysis method; 201 is a laser scanning device, 202 is an initial tunnel section, 203 is an initial contour point set, 204 is a deformed tunnel section, 205 is a monitoring contour point set, 206 is a unified same section coordinate system; 301 is a horizontal displacement component, 302 is a vertical displacement component, 303 is a curvature change component, and 304 is a tunnel deformation feature matrix; 400 is a tunnel deformation binarization detection matrix construction method, S401-S405 are method steps of the tunnel deformation binarization detection matrix construction method; 500 is a binarization detection matrix and global over-limit proportion calculation method, S501-S503 are method steps of the binarization detection matrix and global over-limit proportion calculation method; 601 is an over-limit sampling point, 602 is an abnormal deformation area, 603 is a local deformation, 604 is an overall settlement, 605 is a hole body shrinkage, and 606 is a displacement direction; 701 is an initial contour curve, 702 is an initial contour point, 703 is a reference curve segment, 704 is a to-be-quantized monitoring contour point, 705 is a normal deformation variable, 800 is a tunnel deformation pattern recognition and quantitative analysis system, 801 is a data acquisition module, 802 is a data preprocessing module, 803 is a contour curve fitting module, 804 is a deformation feature extraction module, 805 is an over-threshold binarization module, 806 is an abnormal area positioning module, 807 is a deformation pattern recognition module, and 808 is a deformation variable calculation output module. DETAILED DESCRIPTION

[0026] The present application will be further described in conjunction with specific embodiments. However, it should not be understood that the above-mentioned subject matter of the present application is limited to the following embodiments, and any technology achieved based on the content of the present application is within the scope of the present application.

[0027] Unless otherwise specified, the terms "upper," "lower," "left," "right," "center," "inner," and "outer," etc., used in the description of specific embodiments of the present invention to indicate orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings, or the orientation or positional relationship in which the product / equipment / device is usually placed during use. These terms are merely for the purpose of facilitating the description of the present invention or simplifying the description in specific embodiments, and for enabling those skilled in the art to quickly understand the solution, and do not indicate or imply that a particular device / component / element must have a specific orientation, or be constructed and operated in a specific positional relationship. Therefore, they should not be construed as limitations on the present invention.

[0028] Furthermore, the use of terms such as "horizontal," "vertical," "suspended," "parallel," and "coaxial" does not imply that the corresponding device / component / element must be absolutely horizontal, vertical, suspended, parallel, or coaxial. Slight tilt or deviation is permissible, as long as it does not affect the normal function of the relevant component. For example, "horizontal" simply means that its direction is more horizontal relative to "vertical," not that the structure must be perfectly horizontal; a slight tilt is acceptable. "Coaxial" means that two components are arranged as coaxially as possible, allowing them to move coaxially or approximately coaxially when their relative positions change. Alternatively, it can be simplified to mean that the corresponding device / component / element, when arranged in "horizontal," "vertical," "suspended," "parallel," or "coaxial" directions, can have an error / deviation of ±10% relative to the corresponding direction, more preferably within ±8%, more preferably within ±6%, more preferably within ±5%, and more preferably within ±4%. For example, the deviation in the "coaxial" direction is controlled within 0.2-1mm, preferably within 0.2-0.5mm. As long as the corresponding device / component / element is within the error / deviation range, it can still achieve its function in the solution of the present invention.

[0029] Furthermore, the use of terms such as "first," "second," and "third" in terminology is merely for distinguishing descriptions of identical or similar components and should not be interpreted as emphasizing or implying the relative importance of a particular component.

[0030] Furthermore, in the description of the embodiments of the present invention, "several", "more than", and "a number of" represent at least two. The number can be any number, such as two, three, four, five, six, seven, eight, or nine, and can even exceed nine.

[0031] Furthermore, in the description of the technical solution of this invention, unless otherwise explicitly specified / limited / restricted, the terms "set up," "install," "connect," "link," "provided with," "laid out," and "arranged" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to connection methods commonly used in the art, such as welding, riveting, bolting, and threaded connections. Such connections can be mechanical, electrical, or communication connections; they can be direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components.

[0032] Next, combine Figures 1 to 8 This disclosure provides a detailed description of the tunnel deformation pattern recognition and quantification analysis method and system proposed in the embodiments of this disclosure. Among them, Figure 1 The flowchart illustrates the method for tunnel deformation pattern recognition and quantitative analysis. Figures 2 to 7 This is a schematic diagram illustrating each step in a specific implementation of the tunnel deformation pattern recognition and quantification analysis method. Figure 8 This is a schematic diagram of the modules of the tunnel deformation pattern recognition and quantitative analysis system.

[0033] like Figure 1 As shown, this embodiment provides a tunnel deformation pattern recognition and quantification analysis method 100, including steps S101 to S108. Although the entire process of steps S101 to S108 is described in chronological order here, it should be noted that in actual implementation, steps S101 to S108 may not necessarily be performed in strict order.

[0034] Step S101 is the data acquisition step. Specifically, in step S101, two-dimensional contour data of the same tunnel cross-section in the initial state and two-dimensional contour data in the monitoring state are acquired, and the two sets of two-dimensional contour data are unified to the same cross-section coordinate system. The specific process is as follows: Figure 2 As shown. The initial tunnel section 202 can be the baseline state after tunnel construction is completed and the lining has not shown significant deformation, or it can be a pre-determined reference state within a certain monitoring period; the deformed tunnel section 204 is the state after deformation inferred from the initial tunnel section 202, and can be the state after subsequent inspection or loading. The two-dimensional contour data of the tunnel section in the initial state and the two-dimensional contour data in the monitoring state are the reference coordinate point set in the initial state and the deformation coordinate point set in the monitoring state obtained by scanning the same section of the tunnel using laser scanning equipment 201, respectively, defined by S0={(x 0i ,y 0i )} (Initial contour point set 203) and S1={( x 1i , y 1iThe data is represented by the monitoring contour point set 205. In this embodiment, the laser scanning is performed by a laser scanning device 201 that supports cross-sectional scanning. The laser scanning device 201 is preferably a dedicated laser tunnel profiler. Alternatively, the laser scanning device 201 can also be a ground-based 3D laser scanner or a mobile / track-mounted laser scanning system. The two-dimensional contour data is the set of two-dimensional coordinate points obtained by the laser scanning device scanning the same cross-section of the tunnel. Taking a laboratory circular tunnel model as an example, the cross-sectional radius can be 2.0m, the laser scanning device is installed near the center of the tunnel model, and the scanning accuracy can be ±1mm. Two-dimensional coordinate point sets are obtained in the initial state and the monitoring state, respectively. The specific values ​​mentioned above are only examples; in actual engineering, they can be adjusted according to the tunnel cross-sectional dimensions, scanning device accuracy, and sampling density.

[0035] After obtaining the two-dimensional contour data in the initial state (i.e., the initial contour point set 203) and the two-dimensional contour data in the monitoring state (i.e., the monitoring contour point set 205), in step S101, both are unified to the same cross-sectional coordinate system 206. Coordinate unification can be performed based on the position parameters of the scanning equipment, the center point of the cross-section, the polar coordinate angle, the preset reference datum, or the engineering control points. Through coordinate unification, the initial state data and the monitoring state data have a comparable spatial reference.

[0036] Step S102 is a data preprocessing step. Specifically, in step S102, the two-dimensional contour data in the initial state and the two-dimensional contour data in the monitoring state are preprocessed to obtain an initial contour point set and a monitoring contour point set. This preprocessing step includes one or more of the following: filtering, clustering denoising, and outlier removal. In one embodiment, the example preprocessing process includes removing outliers caused by factors such as ventilation ducts, construction trolleys, personnel passing by, obstructions, and random equipment errors to obtain the initial contour point set and the monitoring contour point set. Preprocessing may include wavelet filtering, moving average filtering, median filtering, clustering denoising, and outlier removal. For example, wavelet soft thresholding can be used to reduce random measurement fluctuations, and then DBSCAN clustering or other density clustering methods can be used to identify and remove isolated points. The above processing methods can be used individually or in combination.

[0037] Step S103 is the contour curve fitting step. Specifically, in step S103, a correspondence between sampling points is established between the initial contour point set and the monitored contour point set, and the initial contour curve C0 and the monitored contour curve C1 are reconstructed based on the initial contour point set and the monitored contour point set, respectively. The establishment of corresponding sampling points can be performed according to polar angle order, contour curve order, arc length position, or sampling index order. In a specific embodiment, the point correspondence can be established based on polar angle order. Specifically, with the cross-section center as the pole, the initial contour point set and the monitored contour point set are converted into polar coordinate representations respectively, and sorted according to polar angle; for points with the same or similar polar angle positions, a correspondence is established. For non-circular cross-sections, a correspondence can also be established based on the arc length position of the contour curve, the sampling index order, or interpolated equal arc length sampling points. Contour reconstruction can use cubic spline interpolation, polynomial fitting, B-spline fitting, or other curve fitting methods that can form continuous cross-section boundaries. Continuous contour curves are used for subsequent curvature calculation and minimum distance projection quantization.

[0038] Step S104 is the deformation feature extraction step. Specifically, in step S104, based on the corresponding sampling points, the initial contour curve, and the monitored contour curve, the horizontal displacement component, vertical displacement component, and curvature change component of the contour curve at the sampling point position are calculated for each sampling point to construct the tunnel deformation feature matrix. The process is as follows: Figure 3 As shown. Specifically, for the corresponding sampling point, the horizontal displacement component 301 and the vertical displacement component 302 of each sampling point can be obtained by subtracting its x-coordinate and y-coordinate respectively. In addition, the curvature change component 303 can also be calculated based on the curvature of the initial contour curve and the monitored contour curve at the corresponding position, and the tunnel deformation feature matrix M can be constructed from this. Figure 3 The 304 in the matrix represents the tunnel deformation feature matrix. In an optional embodiment, the tunnel deformation feature matrix is ​​represented by the following matrix:

[0039] Overall, the matrix consists of n rows and 3 columns, where n is the total number of sampling points of the tunnel cross-section on which the matrix is ​​based. Each row represents the horizontal displacement component 301, the vertical displacement component 302, and the curvature change component 303 at the i-th sampling point. Let 301 be the spatial horizontal displacement component of the i-th sampling point in the two-dimensional coordinate system, i.e.:

[0040] in Let x be the x-coordinate value of the i-th sampling point under the monitoring state. This represents the x-coordinate value of the sampling point in the initial state corresponding to this point.

[0041] Let 302 be the vertical displacement component of the i-th sampling point in the two-dimensional coordinate system.

[0042] in Let represent the ordinate value of the i-th sampling point under monitoring conditions. This represents the ordinate value of the sampling point in the initial state corresponding to this point.

[0043] Let 303 be the curvature change component of the contour curve at the i-th point. When the contour curve is represented as a function y=f(x), the curvature can be calculated from the first and second derivatives of the curve; when the contour curve is a parametric curve, the curvature can be calculated from the derivative of the parametric curve. For example, by performing cubic spline interpolation or polynomial fitting on the point cloud to obtain the equation f(x), in this case, the curvature is given by the formula... Calculation. Generally, the curvature variation component 303 can be obtained from the following formula:

[0044] in To monitor the curvature of the contour curve at the i-th point under the monitoring state, Let be the curvature of the contour curve in the initial state corresponding to this point.

[0045] Step S105 is the threshold binarization step. Specifically, in step S105, the tunnel deformation feature matrix is ​​compared with a preset threshold vector to generate a binarized detection matrix characterizing whether each feature component at each sampling point exceeds the limit. In an example embodiment, the preset threshold vector includes a horizontal displacement threshold, a vertical displacement threshold, and a curvature change threshold, and is determined based on at least one of the following: tunnel engineering monitoring level, design allowable deformation, early warning standard, or historical monitoring data. For example, in a certain project, the horizontal displacement threshold and the vertical displacement threshold can be determined according to a certain proportion of the total design allowable deformation, and the curvature change threshold can be determined based on historical curvature fluctuations or engineering experience.

[0046] The process of generating the binarized detection matrix can, for example, employ the following methods: Figure 4 The tunnel deformation binarization detection matrix construction method 400 shown is implemented. In step S401, the tunnel deformation feature matrix M calculated as described above is input. In step S402, a preset threshold vector is set. .vector The length of the vector should be equal to the number of columns in the tunnel deformation feature matrix M. In the embodiment shown in step S104, the vector The length should be 3, and its specific value is... The vector contains three element values ​​corresponding to preset thresholds for the horizontal displacement component, the vertical displacement component, and the curvature change component, respectively. In steps S403 and S404, each element m of the tunnel deformation feature matrix is... ij With preset threshold vector j The corresponding elements in the vector are compared in size, and the specific comparison process is shown in step S404. Given a length of 3, each element in the tunnel deformation feature matrix is ​​compared with the corresponding threshold in the preset threshold vector. When the absolute value of each element exceeds the corresponding threshold, the corresponding element in the binarized detection matrix is ​​marked as exceeding the threshold. When the absolute value of each element does not exceed the corresponding threshold, the corresponding element in the binarized detection matrix is ​​marked as not exceeding the threshold. The global exceeding ratio is the ratio of the number of exceeding elements in the binarized detection matrix to the total number of elements in the binarized detection matrix. That is, as shown in step S405, the process of obtaining the binarized detection matrix T from the tunnel deformation feature matrix M involves comparing each element in the tunnel deformation feature matrix M to see if it exceeds the threshold, and recording the result in the binarized detection matrix T. The comparison follows the following formula for judgment:

[0047] in: It is a vector The j-th threshold; It is the j-th eigenvalue of the i-th sampling point in matrix M.

[0048] Depending on the different monitoring levels of the tunnel, the allowable deformation range varies, and the established... The values ​​contained in the value vector should also correspond differently. For example, for vertical displacement... The value is set as shown in Table 1.

[0049] Table 1 Vertical Displacement Monitoring Threshold Table

[0050] For horizontal displacement The values ​​are set as shown in Table 2.

[0051] Table 2 Thresholds for Horizontal Displacement Monitoring

[0052] The binarized detection matrix T obtained based on the preset threshold vector can be used to indicate whether each feature component at each sampling point exceeds the limit. An example of its usage is described below for step S106.

[0053] Step S106 is the abnormal region localization step. Specifically, in step S106, the global out-of-specification ratio is calculated based on the obtained binarized detection matrix. When the global out-of-specification ratio exceeds a preset threshold, the abnormal deformation region is determined based on the sampling point positions corresponding to the out-of-specification elements in the binarized detection matrix. The specific process can be, for example, as follows: Figure 5 The binary detection matrix and global out-of-range percentage calculation method 500 shown are performed. In step S501, the binary detection matrix T obtained after calculation in step S104 is input, and then the number of out-of-range elements in the binary detection matrix T is counted, that is, the number of elements with a value of 1 in the binary detection matrix T. Further, in step S502, the ratio of the number of elements in the binary detection matrix T that are in an out-of-range state to the total number of elements is counted to obtain the global out-of-range percentage R. Considering the horizontal displacement component, the vertical displacement component, and the curvature change component, the value of the global out-of-range percentage R is obtained by the following formula:

[0054] In step S503, it is determined whether the R value is greater than a certain preset threshold (e.g., 2%). If the global excess rate R is greater than the preset threshold (e.g., R > 2%), it is determined as "yes" (significant deformation has occurred), and the system locks the part of the binary detection matrix T with a value of 1 (i.e., T). i,j = 1) The corresponding coordinate region is used to enter the abnormal region identification and pattern recognition step, and adjacent or continuously distributed out-of-standard sampling points are divided into the same abnormal deformation region. If the global out-of-standard percentage R is less than or equal to the preset threshold (for example, R ≤ 2%): it is judged as "No", which is regarded as sensor noise or minor environmental disturbance, and the analysis ends. The preset percentage threshold can be determined according to the on-site noise level, historical false alarm situation or engineering monitoring level; for example, the preset percentage threshold can be 1%-5%, preferably 2%.

[0055] The abnormal deformation region is determined based on the sampling point location corresponding to the element in T that is in an out-of-range state. If multiple out-of-range sampling points are adjacent or continuously distributed on the contour curve, they can be divided into the same abnormal deformation region; if the out-of-range sampling points are distributed in multiple discontinuous segments, multiple abnormal deformation regions can be formed respectively.

[0056] Step S107 is the deformation pattern identification step. Specifically, in step S107, the tunnel deformation pattern is identified based on the spatial distribution, displacement direction, and curvature change characteristics of the elements exceeding the standard within the abnormal deformation area.

[0057] The tunnel deformation modes include local deformation, overall settlement, and tunnel shrinkage, such as... Figure 6As shown in the diagram, 601 represents the sampling point exceeding the standard, 602 represents the abnormal deformation area, and 606 represents the displacement direction. When the exceeding elements are concentrated in the local contour segment corresponding to the continuous sampling points, and the curvature change component of this segment is in an excessive state, it can be identified as local deformation 603. This pattern usually reflects the occurrence of depression, bulge, flattening, or other local contour changes in a certain local arc segment of the tunnel. When the vertical displacement components of multiple sampling points are distributed in the same direction, it can be identified as overall settlement 604. The deformation of overall settlement can be determined based on the vertical displacement components, maximum vertical displacement, or average vertical displacement of the sampling points in the abnormal deformation area. When the displacement direction of multiple sampling points points points towards the centroid of the tunnel cross-section or approximately towards the interior of the cross-section, it can be identified as tunnel shrinkage 605. For different cross-sectional shapes, the centroid can be determined based on the designed cross-sectional contour, the initial contour curve, or the fitted cross-sectional center. In one optional implementation, the displacement direction can be determined based on the angle between the displacement vector of the sampling point and the direction pointing from the sampling point to the centroid of the tunnel cross section; when the angle is less than a preset angle, the displacement direction of the sampling point is considered to point to the centroid of the tunnel cross section.

[0058] Step S108 is the deformation calculation and output step. Specifically, in step S108, using the initial contour curve as a reference contour, the normal deformation of the monitoring contour points within the abnormal deformation region is calculated using the minimum distance projection method, and the deformation region, deformation mode, and deformation are output. For the identified abnormal deformation region, using the initial contour curve C0 as a reference contour, the normal deformation from the monitoring contour point to the reference contour is calculated using the minimum distance projection method. Finally, the abnormal deformation region, deformation mode, and corresponding deformation are output.

[0059] The process of calculating the normal deformation using the minimum distance projection method is as follows: Figure 7 As shown, 701 is the initial contour curve, 702 is the initial contour point, 703 is the reference curve segment, 704 is the contour point to be quantified and monitored, and 705 is the normal deformation. d p .

[0060] Specifically, for one contour point (represented by point P1 with coordinates (x1, y1)) in the quantified monitoring contour points 704 within the abnormal deformation area, a reference curve segment 703 closest to P1 is determined based on the initial contour curve 701; the orthogonal projection point of P1 (represented by point P0 with coordinates (x0, y0), distinct from other contour points in the initial contour points 702 on the initial contour curve 701, such as P0' and P0'') is obtained on this reference curve segment 703; the distance from P1 to P0 is determined as the normal deformation 705 at that location (from...). d p (represented), that is, the formula for calculating deformation is: When the orthogonal projection point P0 is outside the reference curve segment, the endpoint closest to the monitored contour point to be quantified can be selected as the projection point, or an adjacent reference curve segment can be selected to redetermine the projection point. This method reduces distance calculation errors caused by incomplete overlap of sampling points between scans. Optionally, in a specific embodiment, when the deformation mode is identified as overall settlement, the overall settlement deformation can be output based on the vertical displacement components, maximum vertical displacement, or average vertical displacement of the sampling points within the abnormal deformation area; when it is identified as local deformation or cavity shrinkage, the local deformation or shrinkage deformation can be output based on the normal deformation obtained by the minimum distance projection method.

[0061] This invention utilizes laser scanning of two-dimensional contour data to acquire tunnel cross-sectional contour information, providing a more complete reflection of the tunnel cross-section's contour deformation state compared to methods relying on only a few monitoring points. Furthermore, by constructing a tunnel deformation feature matrix that simultaneously includes horizontal displacement components, vertical displacement components, and curvature change components, this invention allows deformation analysis to consider both spatial positional changes and local geometrical variations, facilitating the differentiation of different types of cross-sectional deformation. In addition, this invention uses a combination of threshold vectors, binarized detection matrices, and global exceedance percentages for anomaly triggering, reducing the possibility of misjudgments caused by isolated noise points or occasional obstructions. Based on the spatial distribution, displacement direction, and curvature change characteristics of exceedance elements in the abnormal deformation region, this invention identifies patterns such as local deformation, overall settlement, and tunnel shrinkage, providing richer state information for tunnel safety assessment than a single displacement value. Using the initial contour curve as a reference contour, this invention calculates the normal deformation from the monitoring points in the abnormal region to the reference contour through minimum distance projection, reducing quantization bias caused by directly calculating the distance between corresponding points when the point clouds of previous and subsequent scans are not perfectly aligned. The threshold vector of this invention can be set according to the monitoring level, the allowable deformation amount, the early warning standard, or historical monitoring data, and can adapt to the monitoring requirements under different engineering scenarios.

[0062] Furthermore, embodiments of the present invention also provide a tunnel deformation pattern recognition and quantification analysis system based on laser scanning two-dimensional contour data. Figure 8 This is a schematic diagram of the modules of the tunnel deformation pattern recognition and quantification analysis system 800. The system includes a data acquisition module 801, a data preprocessing module 802, a contour curve fitting module 803, a deformation feature extraction module 804, a threshold binarization module 805, an anomaly region location module 806, a deformation pattern recognition module 807, and a deformation calculation and output module 808.

[0063] The data acquisition module 801 acquires two-dimensional contour data of the same tunnel cross-section in its initial state and in its monitoring state, and unifies the two sets of two-dimensional contour data into the same cross-section coordinate system. Based on the scanning device's position parameters, the cross-section center point, polar coordinate angles, or a preset reference datum, it performs coordinate transformation and registration on the two-dimensional contour data in the initial state and the two-dimensional contour data in the monitoring state. The data preprocessing module 802 filters, clusters for noise reduction, or removes outliers from the two-dimensional contour data. The contour curve fitting module 803 establishes the correspondence between sampling points and reconstructs the initial contour curve and the monitoring contour curve. The deformation feature extraction module 804 calculates the horizontal displacement component, vertical displacement component, and curvature change component, and constructs the tunnel deformation feature matrix. The over-threshold binarization module 805 compares the tunnel deformation feature matrix with a preset threshold vector, generates a binarized detection matrix characterizing whether each feature component at each sampling point exceeds the limit, and calculates the global over-threshold percentage. The abnormal region location module 806 is used to calculate the global excess ratio based on the binarized detection matrix. When the global excess ratio exceeds a preset threshold, the abnormal deformation region is determined based on the sampling point positions corresponding to the excess elements in the binarized detection matrix, wherein adjacent or continuously distributed excess sampling points are divided into the same abnormal deformation region. The deformation pattern recognition module 807 identifies the tunnel deformation pattern based on the spatial distribution, displacement direction, and curvature change characteristics of the excess elements within the abnormal deformation region. The deformation calculation and output module 808 is used to calculate the normal deformation of the monitoring contour points within the abnormal deformation region using the initial contour curve as a reference contour and the minimum distance projection method, and is used to output the deformation region, deformation pattern, and deformation.

[0064] This embodiment also provides an electronic device. This electronic device can be a server, industrial control computer, edge computing device, monitoring terminal, or other device with data processing capabilities. The electronic device includes a processor, a memory, and a computer program stored in the memory and capable of running on the processor. When the computer program is executed by the processor, it implements the methods described in any of the foregoing embodiments.

[0065] This embodiment also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method described in any of the foregoing embodiments. The computer-readable storage medium may include, but is not limited to, a read-only memory, a random access memory, a magnetic disk, an optical disk, flash memory, or other media capable of storing program instructions.

[0066] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any equivalent substitutions, combinations, or improvements made to data preprocessing methods, contour fitting methods, threshold setting methods, abnormal region segmentation methods, deformation pattern recognition rules, or deformation output methods within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for tunnel deformation pattern recognition and quantitative analysis, comprising the following steps: S1, Data Acquisition: Acquire two-dimensional contour data of the same tunnel cross section in the initial state and in the monitoring state, and unify the two sets of two-dimensional contour data into the same cross section coordinate system. According to the position parameters of the scanning device, the center point of the cross section, the polar coordinate angle or the preset reference benchmark, perform coordinate transformation and registration on the two-dimensional contour data in the initial state and the two-dimensional contour data in the monitoring state. S2, Data preprocessing: The two-dimensional contour data in the initial state and the two-dimensional contour data in the monitoring state are preprocessed to obtain the initial contour point set and the monitoring contour point set. The preprocessing includes one or more of the following: filtering, clustering noise reduction, and outlier removal. S3, Contour curve fitting: Establish the correspondence between the sampling points of the initial contour point set and the monitoring contour point set, and reconstruct the initial contour curve and the monitoring contour curve based on the initial contour point set and the monitoring contour point set respectively; S4, Deformation feature extraction: Based on the corresponding sampling points, the initial contour curve, and the monitored contour curve, calculate the horizontal displacement component, vertical displacement component, and curvature change component of the contour curve at the sampling point position for each sampling point, and construct the tunnel deformation feature matrix. S5, Over-threshold binarization: The tunnel deformation feature matrix is ​​compared with a preset threshold vector to generate a binarized detection matrix that characterizes whether each feature component of each sampling point exceeds the standard; S6, Abnormal region location: Calculate the global over-standard ratio based on the binarized detection matrix, and when the global over-standard ratio exceeds the preset ratio threshold, determine the abnormal deformation region based on the sampling point position corresponding to the over-standard element in the binarized detection matrix, wherein adjacent or continuously distributed over-standard sampling points are divided into the same abnormal deformation region. S7, Deformation pattern recognition: Based on the spatial distribution, displacement direction, and curvature change characteristics of the excessive elements within the abnormal deformation area, the tunnel deformation pattern is identified; S8, Deformation Calculation Output: Using the initial contour curve as a reference contour, the normal deformation of the monitoring contour points in the abnormal deformation area is calculated using the minimum distance projection method, and the deformation area, deformation mode and deformation are output.

2. The method for tunnel deformation pattern recognition and quantitative analysis according to claim 1, characterized in that, In step S3, the correspondence of the sampling points is established according to the polar angle order, the contour curve order, the arc length position, or the sampling index order. The initial contour curve and the monitored contour curve are reconstructed using at least one of spline interpolation, polynomial fitting, or B-spline fitting to obtain a continuous tunnel cross-section contour curve.

3. The method for tunnel deformation pattern recognition and quantitative analysis according to claim 1, characterized in that, In step S4, the tunnel deformation feature matrix includes feature rows corresponding to the sampling points, and each feature row includes at least the horizontal displacement component, the vertical displacement component, and the curvature change component corresponding to the sampling point. The horizontal displacement component is represented by the coordinate difference in the horizontal direction between the sampling point in the monitoring state and the sampling point in the corresponding initial state; the vertical displacement component is represented by the coordinate difference in the vertical direction between the sampling point in the monitoring state and the sampling point in the corresponding initial state. The curvature change component is represented by the difference between the curvature of the monitored profile curve at the corresponding position and the curvature of the initial profile curve at the corresponding position.

4. The method for tunnel deformation pattern recognition and quantitative analysis according to claim 1, characterized in that, Generating the binarized detection matrix includes: comparing each element in the tunnel deformation feature matrix with the corresponding threshold in the preset threshold vector; when the absolute value of each element exceeds the corresponding threshold, marking the corresponding element in the binarized detection matrix as exceeding the threshold; when the absolute value of each element does not exceed the corresponding threshold, marking the corresponding element in the binarized detection matrix as not exceeding the threshold; the global exceeding ratio is the ratio of the number of exceeding elements in the binarized detection matrix to the total number of elements in the binarized detection matrix.

5. The method for tunnel deformation pattern recognition and quantitative analysis according to claim 1, characterized in that, The tunnel deformation modes include local deformation, overall settlement, and tunnel shrinkage. When the excessive elements are concentrated in the local contour section corresponding to the continuous sampling points, and the corresponding section has abnormal curvature changes, it is identified as local deformation. When the vertical displacement components of multiple sampling points are distributed in the same direction, it is identified as overall settlement. And when the displacement direction of multiple sampling points points points towards the centroid of the tunnel cross section, it is identified as tunnel shrinkage.

6. The method for tunnel deformation pattern recognition and quantitative analysis according to claim 5, characterized in that, In step S8, the minimum distance projection method includes: for the monitoring contour point to be quantized within the abnormal deformation area, determining the reference curve segment closest to the monitoring contour point on the initial contour curve; obtaining the orthogonal projection point of the monitoring contour point to be quantized on the reference curve segment; determining the distance from the monitoring contour point to be quantized to the orthogonal projection point as the normal deformation of the corresponding position; when the orthogonal projection point is located outside the reference curve segment, selecting the endpoint or adjacent reference curve segment closest to the monitoring contour point to be quantized to redetermine the projection point.

7. The method for tunnel deformation pattern recognition and quantitative analysis according to claim 6, characterized in that, When the tunnel deformation mode is overall settlement, the overall settlement deformation is output based on the vertical displacement component, maximum vertical displacement, or average vertical displacement of the sampling points within the abnormal deformation area; when the tunnel deformation mode is local deformation or tunnel shrinkage, the local deformation or shrinkage deformation is output based on the normal deformation obtained by the minimum distance projection method.

8. A tunnel deformation pattern recognition and quantification analysis system, comprising: The data acquisition module is used to acquire two-dimensional contour data of the same tunnel cross section in the initial state and in the monitoring state, and unify the two sets of two-dimensional contour data into the same cross section coordinate system. According to the position parameters of the scanning device, the center point of the cross section, the polar coordinate angle or the preset reference benchmark, the two-dimensional contour data in the initial state and the two-dimensional contour data in the monitoring state are subjected to coordinate transformation and registration. A data preprocessing module is used to preprocess the two-dimensional contour data in the initial state and the two-dimensional contour data in the monitoring state to obtain an initial contour point set and a monitoring contour point set. The preprocessing includes one or more of the following: filtering, clustering noise reduction, and outlier removal. The contour curve fitting module is used to establish the correspondence between the sampling points of the initial contour point set and the monitoring contour point set, and to reconstruct the initial contour curve and the monitoring contour curve based on the initial contour point set and the monitoring contour point set, respectively. The deformation feature extraction module is used to calculate the horizontal displacement component, vertical displacement component, and curvature change component of the contour curve at the sampling point position of each sampling point based on the corresponding sampling points, the initial contour curve, and the monitored contour curve, and to construct the tunnel deformation feature matrix. The over-threshold binarization module is used to compare the tunnel deformation feature matrix with a preset threshold vector to generate a binarized detection matrix that characterizes whether each feature component of each sampling point exceeds the standard. An abnormal region localization module is used to calculate the global over-standard ratio based on the binarized detection matrix, and when the global over-standard ratio exceeds a preset ratio threshold, to determine the abnormal deformation region based on the sampling point position corresponding to the over-standard element in the binarized detection matrix, wherein adjacent or continuously distributed over-standard sampling points are divided into the same abnormal deformation region. A deformation pattern recognition module is used to identify tunnel deformation patterns based on the spatial distribution, displacement direction, and curvature change characteristics of the excessive elements within the abnormal deformation area. The deformation calculation and output module is used to calculate the normal deformation of the monitoring contour points in the abnormal deformation area using the initial contour curve as a reference contour, and output the deformation area, deformation mode and deformation.

Images