A tunnel convergence deformation data acquisition system and method

Abstract

The present application provides a tunnel convergence deformation data acquisition system and method, which relate to the field of optical metrology technology. The system collects laser point cloud data of reflection marking points on the inner wall of the tunnel, performs local modeling on the laser point cloud data, obtains a local point cloud subset, and then performs nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor. The registration error is determined based on the local deformation descriptor and spatial image information of the inner wall of the tunnel. The laser point cloud data is iteratively eliminated and reconstructed based on the registration error to obtain a local deformation matching model. The deformation feature benchmark of the inner wall of the tunnel is determined by the local reflection data obtained by local laser scanning of the reflection marking points on the inner wall of the tunnel. The local deformation matching model determines the deviation outlier value of the inner wall of the tunnel based on the deformation feature benchmark, and the convergence deformation data of the inner wall of the tunnel is obtained by screening the deviation outlier value. The present application can realize nonlinear local small deformation detection to improve the perception ability of irregular tunnel inner wall.

Description

Technical Field

[0001] The present application relates to the field of optical metrology technology, and more specifically, to a tunnel convergence deformation data acquisition system and method. Background Art

[0002] Optical metrology technology refers to the acquisition of target surface structure information through non-contact devices such as visual sensors and laser scanners, combined with image processing, point cloud modeling and geometric calculation methods to achieve high-precision measurement of object deformation. In large-scale curved surface structures such as tunnels, optical methods are gradually replacing traditional contact monitoring methods due to their high efficiency, digitization and automation advantages.

[0003] Tunnel convergence deformation data collection methods primarily rely on manually placed grid markers. These points are then periodically collected using total stations or 3D laser scanners to track tunnel cross-sectional changes. Alternatively, image-based photogrammetry systems are used to identify the spatial displacement of specific markers, combined with geological radar and deformation sensors to capture subtle deformations within or outside the structure. While existing technologies have achieved tunnel convergence deformation monitoring to a certain extent, they still suffer from certain technical flaws and limitations. For example, manually placed grid markers are highly dependent on a regular reference grid, making their deployment difficult and costly in irregular structures or spatially confined scenarios. Furthermore, these points are unable to adapt to nonlinear and localized micro-deformation characteristics, resulting in insufficient deformation recognition capabilities. This limits the accuracy of registration results and makes highly reliable deformation monitoring difficult. Therefore, the industry faces the challenge of detecting nonlinear, localized micro-deformations to improve the perception of irregular tunnel walls. Summary of the Invention

[0004] The present application provides a tunnel convergence deformation data acquisition system and method, which can realize nonlinear local micro-deformation detection to improve the perception ability of irregular tunnel inner walls.

[0005] In a first aspect, the present application provides a tunnel convergence deformation data acquisition system and method, the data acquisition method comprising the following steps:

[0006] Obtain the reflection marking points on the inner wall of the tunnel and determine the spatial image information of the inner wall of the tunnel;

[0007] Collecting laser point cloud data of the reflective marker point, performing local modeling on the laser point cloud data to obtain a local point cloud subset, and then performing nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor;

[0008] Determining a registration error of the reflective marker point based on the local deformation descriptor and the spatial image information, and then iteratively eliminating and reconstructing the laser point cloud data based on the registration error to obtain a local deformation matching model;

[0009] Performing local laser scanning on the reflective marking points on the inner wall of the tunnel to obtain local reflection data, and determining a deformation characteristic benchmark of the inner wall of the tunnel based on the local reflection data;

[0010] The local deformation matching model determines the deviation anomaly value of the tunnel inner wall based on the deformation feature benchmark, and then obtains the convergent deformation data of the tunnel inner wall through screening the deviation anomaly value.

[0011] In this embodiment, the reflective marking points are reflective signs set on the surface of the inner wall of the tunnel, and the spatial image information of the inner wall of the tunnel is determined by a multi-eye vision camera installed on the tunnel detection equipment.

[0012] In this embodiment, the laser point cloud data of the reflective marking points is collected by a laser scanner.

[0013] In this embodiment, local modeling is performed on the laser point cloud data to obtain a local point cloud subset, specifically including:

[0014] Performing neighborhood indexing on the laser point cloud data to obtain a neighborhood center point;

[0015] The neighborhood center point is subjected to neighborhood modeling using a preset neighborhood radius to obtain a local point cloud subset.

[0016] In this embodiment, performing nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor specifically includes:

[0017] Determine a fitting coefficient vector based on the point cloud data in the local point cloud subset;

[0018] Performing nonlinear fitting on the point cloud data in the local point cloud subset to obtain a fitting surface;

[0019] Determine the fitting residual of the point cloud data in the local point cloud subset by using the fitting surface and the fitting coefficient vector;

[0020] The point cloud data in the local point cloud subset is fitted based on the fitting residual to obtain a local deformation descriptor.

[0021] In this embodiment, determining the registration error of the reflective marker point according to the local deformation descriptor and the spatial image information specifically includes:

[0022] Performing feature description on the local deformation descriptor to obtain local deformation features of the tunnel inner wall;

[0023] Determine the spatial position of the reflective marker point using the spatial image information, and then determine the surface features and normal vector of the reflective marker point using the spatial position;

[0024] An error evaluation is performed on the curved surface feature and the local deformation feature according to the normal vector to obtain a registration error of the reflective marker point.

[0025] In this embodiment, performing local laser scanning on the reflective marking points on the inner wall of the tunnel to obtain local reflection data specifically includes:

[0026] The inner wall of the tunnel is locally scanned by laser scanning equipment to obtain a local scanning point set;

[0027] Light sensor is used to collect light data reflected by local scanning points in the local scanning point set, and the light data is used as local reflection data.

[0028] In this embodiment, determining the deformation characteristic benchmark of the tunnel inner wall according to the local reflection data specifically includes:

[0029] Performing contour detection on the local reflection data to obtain local contour features of the tunnel inner wall;

[0030] A light fluctuation curve of the inner wall of the tunnel is determined, and then a deformation feature reference of the inner wall of the tunnel is determined based on the light fluctuation curve and the local contour feature.

[0031] In this embodiment, obtaining the convergence deformation data of the tunnel inner wall by screening the deviation abnormal values ​​means performing vector determination on the deviation abnormal values, and then screening to obtain the convergence deformation data of the tunnel inner wall.

[0032] In a second aspect, the present application provides a tunnel convergence deformation data acquisition system for executing a tunnel convergence deformation data acquisition method, the data acquisition system comprising:

[0033] The data acquisition module is used to obtain the reflection marking points on the inner wall of the tunnel and determine the spatial image information of the inner wall of the tunnel;

[0034] a point cloud fitting module, configured to collect laser point cloud data of the reflective marker point, perform local modeling on the laser point cloud data to obtain a local point cloud subset, and then perform nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor;

[0035] a registration optimization module, configured to determine a registration error of the reflective marker points based on the local deformation descriptor and the spatial image information, and then iteratively eliminate and reconstruct the laser point cloud data based on the registration error to obtain a local deformation matching model;

[0036] A feature reference determination module is used to perform local laser scanning on the reflective marking points on the inner wall of the tunnel to obtain local reflection data, and determine the deformation feature reference of the inner wall of the tunnel based on the local reflection data;

[0037] The deformation determination module is used for the local deformation matching model to determine the deviation abnormal value of the tunnel inner wall based on the deformation feature benchmark, and then obtain the convergent deformation data of the tunnel inner wall by screening the deviation abnormal value.

[0038] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0039] The spatial image information of the tunnel inner wall is determined by acquiring the reflective marking points on the tunnel inner wall; the laser point cloud data of the reflective marking points is collected, and the laser point cloud data is locally modeled to obtain a local point cloud subset, and then the local point cloud subset is nonlinearly fitted to obtain a local deformation descriptor; the registration error of the reflective marking points is determined based on the local deformation descriptor and the spatial image information, and then the laser point cloud data is iteratively eliminated and reconstructed based on the registration error to obtain a local deformation matching model; the reflective marking points on the tunnel inner wall are locally laser scanned to obtain local reflection data, and the deformation feature benchmark of the tunnel inner wall is determined based on the local reflection data; the local deformation matching model determines the deviation outlier value of the tunnel inner wall based on the deformation feature benchmark, and then the converged deformation data of the tunnel inner wall is obtained by screening the deviation outlier value.

[0040] It can be seen that in this application, firstly, by randomly deploying reflective markers on the inner wall of the tunnel and combining multi-eye vision equipment to extract spatial image information, it is convenient to enhance the flexibility of deployment in irregular or occluded areas, and provide a reliable reference for subsequent high-precision reconstruction; and by locally modeling the laser point cloud data, extracting a point cloud subset with spatial neighborhood consistency, and introducing a second-order polynomial fitting model to construct a local deformation descriptor, it can effectively capture local curvature changes and normal offsets, which is conducive to the modeling of nonlinear deformation data; secondly, a registration error model is constructed based on image information and point cloud fitting results, and a local deformation matching model is formed through error elimination and reconstruction. , it can dynamically exclude mismatched points or abnormal data, ensuring that the fitting model can truly reflect the actual deformation, which is conducive to enhancing the stability and reliability of the system in harsh environments; then, through local laser scanning and reflection data extraction, combined with contour detection and light fluctuation characteristics, a deformation feature benchmark is established to achieve fusion perception of geometric information and optical information, providing data support for judging whether the current deformation is abnormal; finally, by using the comparative analysis of the local deformation matching model and the deformation feature benchmark, the deviation abnormal values ​​with a direction toward the center of the tunnel inner wall and an amplitude exceeding the threshold are identified, and the data area representing the convergent deformation of the tunnel inner wall is further extracted, which can accurately identify the local deformation trend.

[0041] In summary, the technical solution adopted in this application can realize nonlinear local micro-deformation detection to improve the perception ability of irregular tunnel inner walls. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0043] Figure 1 is a flow chart of a tunnel convergence deformation data acquisition method provided by the present application;

[0044] Figure 2 is an exemplary flow chart for determining a local point cloud subset according to the present application;

[0045] Figure 3 is an exemplary flow chart for determining the registration error of a reflective marker point according to the present application;

[0046] Figure 4 It is a module structure diagram of the data acquisition system provided by this application. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0048] The embodiment of the present application provides a tunnel convergence deformation data acquisition system and method, the core of which is to determine the spatial image information of the tunnel inner wall by acquiring reflective marker points on the tunnel inner wall; collect laser point cloud data of the reflective marker points, perform local modeling on the laser point cloud data to obtain a local point cloud subset, and then perform nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor; determine the alignment error of the reflective marker points based on the local deformation descriptor and the spatial image information, and then iteratively eliminate and reconstruct the laser point cloud data based on the alignment error to obtain a local deformation matching model; perform local laser scanning on the reflective marker points on the tunnel inner wall to obtain local reflection data, and determine the deformation feature benchmark of the tunnel inner wall based on the local reflection data; the local deformation matching model determines the deviation outlier value of the tunnel inner wall based on the deformation feature benchmark, and then obtains the convergence deformation data of the tunnel inner wall through screening the deviation outlier value.

[0049] Example 1: In order to better understand the above technical solution, the following will be combined with the accompanying drawings and specific implementation methods to describe the above technical solution in detail. Figure 1 As shown in FIG. 1 , this figure is an exemplary flow chart of a method for collecting tunnel convergence deformation data according to this embodiment of the present application. The data collection method includes the following steps:

[0050] In step S1, the reflective marking points on the inner wall of the tunnel are acquired to determine the spatial image information of the inner wall of the tunnel.

[0051] In specific implementation, the reflective marking points in this application are the position points of reflective signs set on the surface of the inner wall of the tunnel. The reflective marking points are randomly deployed, and the reflective signs are randomly deployed on the inner wall of the tunnel in advance. The reflective marking points on the inner wall of the tunnel can be obtained by a flash camera. In addition, in this application, the spatial image information of the inner wall of the tunnel can be determined by a multi-eye vision camera installed on the tunnel detection equipment; it should be noted that the spatial image information of the inner wall of the tunnel is the image data of the inner wall of the tunnel obtained by the multi-eye vision camera, and the spatial image information is used to describe the spatial distribution of the reflective marking points and the inner wall of the tunnel.

[0052] In step S2, the laser point cloud data of the reflective marking point is collected, local modeling is performed on the laser point cloud data to obtain a local point cloud subset, and then nonlinear fitting is performed on the local point cloud subset to obtain a local deformation descriptor.

[0053] It should be noted that in this application, the laser point cloud data of the reflective marking points can be collected by a laser scanner. The laser point cloud data refers to a data set obtained by using a laser scanner and contains information such as the three-dimensional coordinates and reflection intensity of the marking points. Among them, the point cloud data collected by the laser scanner contains the three-dimensional spatial coordinate information of the inner wall of the tunnel, which is used to record optical parameters such as the reflection intensity of the reflective marking points on the inner wall of the tunnel, which provides a solid data foundation for subsequent local modeling, nonlinear fitting and error correction; in addition, the laser scanner can use a ground three-dimensional laser scanner to facilitate high-precision data collection of the entire inner wall of the tunnel.

[0054] Preferably, in this embodiment, reference Figure 2 As shown in FIG. 1 , this figure is an exemplary flow chart for determining a local point cloud subset in an embodiment of the present application. In this embodiment, local modeling is performed on the laser point cloud data to obtain a local point cloud subset, which can be specifically achieved by the following steps:

[0055] First, in step S21, the laser point cloud data is subjected to neighborhood indexing to obtain the neighborhood center point;

[0056] Then, in step S22, neighborhood modeling is performed on the neighborhood center point using a preset neighborhood radius to obtain a local point cloud subset.

[0057] In the specific implementation, first, the collected laser point cloud data is denoised by the point cloud denoising technology, and then the neighborhood index algorithm is used to perform neighborhood analysis on the denoised laser point cloud data to obtain the Euclidean distance between all laser point cloud data, and then the laser point cloud data corresponding to the maximum Euclidean distance among all Euclidean distances is used as the neighborhood center point; then, the neighborhood radius is preset, that is: the mean of all Euclidean distances is used as the neighborhood radius, and then the laser point cloud data within the spherical range is selected as the local point cloud subset with the neighborhood center point as the center of the sphere and the neighborhood radius as the radius of the sphere, wherein the laser point cloud data is allowed to be selected repeatedly, that is, the same laser point cloud data can belong to different local point cloud subsets, so that the local point cloud subsets obtained by local modeling have high regional consistency.

[0058] It should be noted that the neighborhood index in this embodiment refers to the use of a neighborhood index algorithm to screen out representative neighborhood center points from the laser point cloud data. Compared with the method of directly processing the global point cloud data, the local modeling method using neighborhood indexing and neighborhood modeling can more accurately capture the subtle deformation characteristics of the local area and improve the local detail depiction capability of the point cloud data. There may be two or more local point cloud subsets, that is, the laser point cloud data corresponding to the maximum Euclidean distance of the neighborhood center point in the Euclidean distance can be two or more.

[0059] In this embodiment, nonlinear fitting is performed on the local point cloud subset to obtain a local deformation descriptor, which can be specifically performed in the following manner, namely:

[0060] Determine a fitting coefficient vector based on the point cloud data in the local point cloud subset;

[0061] Performing nonlinear fitting on the point cloud data in the local point cloud subset to obtain a fitting surface;

[0062] Determine the fitting residual of the point cloud data in the local point cloud subset by using the fitting surface and the fitting coefficient vector;

[0063] The point cloud data in the local point cloud subset is fitted based on the fitting residual to obtain a local deformation descriptor.

[0064] In the specific implementation, first, let the spatial coordinates of the point cloud data in the local point cloud subset be ,in, Indicates the size of the i-th point cloud data in the local point cloud subset in the x-axis direction of the spatial coordinate, Indicates the size of the i-th point cloud data in the local point cloud subset in the y-axis direction of the spatial coordinate, It represents the size of the i-th point cloud data in the local point cloud subset in the z-axis direction of the spatial coordinate. i represents the number of the point cloud data in the local point cloud subset, and its value is 1, 2, 3..., N, that is, N is the total number of point cloud data in the local point cloud subset. The actual spatial coordinates of the point cloud data can be obtained by a laser scanner. Then, the second-order polynomial of the point cloud data is used as the fitting model. All the point data in the local point cloud subset are fitted by the least squares method to obtain the fitting coefficient vector, that is, the fitting coefficient vector is obtained by solving the following formula:

[0065]

[0066] in, Represents the function for solving the fitting coefficient vector. By solving this function, the fitting coefficient vector a can be obtained. is the second-order polynomial of the point cloud data, is a constant term, and represents the first-order coefficient, , and Represents the second-order coefficient, which can be obtained by randomly selecting the spatial coordinates of the point cloud data in three local point cloud subsets and substituting them into the second-order polynomial of the point cloud data to solve the equation. 、 、 、 、 and In addition, the process of solving the above function can be implemented by the SciPy library in python; then, the point cloud data in the local point cloud subset is fitted by the surface calculation formula to obtain the fitting surface; then, for each point cloud data in the local point cloud subset, the spatial coordinates of the point cloud data are substituted into the fitting coefficient vector solution formula and the fitting surface respectively, and the calculated fitting coefficient vector and the fitting surface are subjected to Euclidean distance calculation, and the calculated Euclidean distance is used as the fitting residual. Through this step, the fitting residual of the point cloud data in the local point cloud subset can be obtained; finally, the fitting residual of the point cloud data in the local point cloud subset is mapped to the spatial coordinates, and the fitting residual in the spatial coordinates is visualized and analyzed. If the fitting residual distribution in the spatial coordinates is not If the distribution of fitting residuals in the spatial coordinates is uniform, any point cloud data in the local point cloud subset can be substituted into the fitting surface for normal vector offset calculation, and the calculation result is used as the local deformation descriptor. Whether the distribution of fitting residuals in the spatial coordinates is uniform can be analyzed and judged by the residual and fitting value graph in the existing visualization algorithm, which can detect the nonlinear relationship and heteroscedasticity of the fitting residuals. In actual implementation, the residual and fitting value graph can be drawn by the plot_regress_exog function of the statsmodels library in python.

[0067] It should be noted that the modeling modeled by the second-order polynomial in this application can capture the overall deformation trend in the local area. By analyzing the distribution of the fitting residuals, subtle and local abnormal deformations can be identified, thereby improving the monitoring accuracy. The calculated local deformation descriptor is a mathematical description of the local geometric morphology and deformation trend, that is: if the fitting residuals in the spatial coordinates are evenly distributed, the calculated local deformation descriptor can be used to characterize the normal offset of the local area in the three-dimensional spatial coordinate system. If the fitting residuals in the spatial coordinates are unevenly distributed, the calculated local deformation descriptor is used to describe the local curvature changes of the inner wall of the tunnel.

[0068] In step S3, the registration error of the reflective marker point is determined based on the local deformation descriptor and the spatial image information, and then the laser point cloud data is iteratively eliminated and reconstructed based on the registration error to obtain a local deformation matching model.

[0069] Preferably, in this embodiment, reference Figure 3 As shown in FIG, this figure is an exemplary flow chart of determining the registration error of the reflective marker point in an embodiment of the present application. In this embodiment, determining the registration error of the reflective marker point based on the local deformation descriptor and the spatial image information can be specifically implemented by the following steps:

[0070] First, in step S31, the local deformation descriptor is characterized to obtain the local deformation features of the tunnel inner wall;

[0071] Then, in step S32, the spatial position of the reflective marking point is determined by the spatial image information, and the surface features and normal vector of the reflective marking point are determined by the spatial position;

[0072] Finally, in step S33, an error evaluation is performed on the surface feature and the local deformation feature according to the normal vector to obtain the registration error of the reflective marker point.

[0073] In the specific implementation, first, a local weighted average algorithm is written in Python, and the local deformation descriptor is characterized by using the written local weighted average algorithm. The local curvature and normal offset of the tunnel inner wall can be mapped to a heat map, and then the discrete image features in the heat map are aggregated through Kriging interpolation, and the aggregated features are used as the local deformation features of the tunnel inner wall, wherein the local deformation features are used to describe the local curvature changes and normal offsets of the tunnel inner wall, which can be used as a basis for whether the tunnel inner wall undergoes convergent deformation; then, the multi-eye vision camera that collects spatial image information can be used as the center of the three-dimensional coordinates, and modeled according to the ratio of actual observation distance to pixel scale of 10:1 to obtain the spatial position of the reflection mark point, and three or more reflection mark points are randomly selected for surface Solve, and then use the solution result as the surface composed of the reflection marker points, and then aggregate the surface composed of the reflection marker points through Kriging interpolation to obtain the surface features of the reflection marker points, and then solve the surface composed of the reflection marker points through the solution formula of the normal vector to obtain the normal vector of the reflection marker point, wherein the above calculation process can be implemented by the NumPy library in Python and is not limited here; finally, the normal vector and the local normal offset of the tunnel inner wall are vector-summed, and the direction of the summed vector is used as the convolution direction, and the surface features and the local curvature changes of the tunnel inner wall are convolved and fused based on the convolution direction through a convolutional neural network, and then the convolved surface is used as the local error fusion surface of the tunnel inner wall, and then the normal vector of the error fusion surface is used as the registration error of the reflection marker point.

[0074] It should be noted that the registration error in this application refers to the difference between the actual spatial position and the observed position of the reflection marker point when the laser point cloud and the spatial image information are aligned in the three-dimensional coordinate system. It can be used to describe the degree of deviation between the three-dimensional point cloud data. The smaller the registration error, the higher the data registration accuracy. The larger the registration error, the lower the data registration accuracy, and the data needs to be optimized by data elimination and registration. In addition, the error fusion surface in this embodiment refers to the surface that describes the local error characteristics of the inner wall of the tunnel, which is used to accurately detect the subtle features of the local convergence deformation of the inner wall of the tunnel.

[0075] In this embodiment, the laser point cloud data is iteratively eliminated and reconstructed based on the registration error to obtain a local deformation matching model, which can be specifically obtained in the following manner, namely:

[0076] Acquire a local point cloud subset from the laser point cloud data;

[0077] Iterating and eliminating the point cloud data in the local point cloud subset according to the registration error to obtain registered point cloud data;

[0078] The registered point cloud data is reconstructed to obtain a local deformation matching model.

[0079] In the specific implementation, first, the point cloud data points whose size exceeds the registration error are regarded as abnormal matching data, and statistical techniques are used to eliminate the abnormal matching data. The point cloud data in the local point cloud subset are repeatedly marked and eliminated for abnormal matching data, and then the eliminated point cloud data are subjected to surface fitting (surface equation calculation). When there is no abnormal matching data, all the remaining point cloud data in the local point cloud subset are regarded as the registered point cloud data; then, the second-order polynomial is used as the fitting model to reconstruct the registered point cloud data, and the reconstructed surface equation is used as the local deformation matching model.

[0080] It should be noted that the local deformation matching model in this application refers to the fitting surface of local point cloud data obtained by eliminating abnormal matching data and reconstruction processing. The local deformation matching model can reflect the true geometric shape of the local area in the tunnel. Through iterative optimization, the local deformation matching model can still maintain a high accuracy when facing interference factors such as noise and occlusion, thereby improving the robustness of the overall tunnel inner wall convergence deformation monitoring.

[0081] In step S4, a local laser scan is performed on the reflective marking points on the inner wall of the tunnel to obtain local reflection data, and a deformation characteristic benchmark of the inner wall of the tunnel is determined based on the local reflection data.

[0082] In this embodiment, local laser scanning is performed on the reflective marking points on the inner wall of the tunnel to obtain local reflection data in the following manner, namely:

[0083] The inner wall of the tunnel is locally scanned by laser scanning equipment to obtain a local scanning point set;

[0084] Light sensor is used to collect light data reflected by local scanning points in the local scanning point set, and the light data is used as local reflection data.

[0085] In a specific implementation, first, a laser scanning device is used to perform a local scan of the inner wall of the tunnel. The laser scanning device operates according to a pre-set scanning mode to obtain three-dimensional point cloud data of the area where the reflection mark point is located, and the obtained three-dimensional point cloud data is used as a local scanning point set. The local scanning point set contains the spatial coordinate data of each scanning point in the local area of ​​the inner wall of the tunnel, providing a geometric basis for subsequent reflection data collection, wherein the actual detected local contour area is determined by the scanning area of ​​the laser scanning device; then, the light data reflected by the local scanning points in the local scanning point set can be collected by a light sensor, wherein the light sensor in this embodiment can use a laser ranging sensor, wherein the laser ranging sensor has a horizontally rotating rotating base and a spherical card holder for vertical inclination adjustment, and the control module of the laser ranging sensor can be used to realize the collection of light data reflected by the local scanning points in the local scanning point set, and then the collected light data are all used as local reflection data.

[0086] It should be noted that the local reflection data in this application refers to a set of light intensity, reflection distance and related optical data reflected by each scanning point collected by the laser ranging sensor. By collecting reflected light data through the laser ranging sensor, the local reflection data not only contains three-dimensional point cloud information, but also reflects the optical properties of the surface, which helps to improve the accuracy of subsequent reflection marker point alignment and local deformation matching. In addition, through the design with a rotating base and a spherical mount, the laser ranging sensor can flexibly adjust the scanning angle to ensure effective coverage of the target area in complex environments, reduce blind spots and data loss caused by insufficient scanning angles.

[0087] In this embodiment, the following method can be used to determine the deformation characteristic benchmark of the tunnel inner wall based on the local reflection data, namely:

[0088] Performing contour detection on the local reflection data to obtain local contour features of the tunnel inner wall;

[0089] A light fluctuation curve of the inner wall of the tunnel is determined, and then a deformation feature reference of the inner wall of the tunnel is determined based on the light fluctuation curve and the local contour feature.

[0090] In specific implementation, first, the Canny edge detection algorithm is used to perform contour detection on the local reflection data, and then the detection result is used as the local contour feature of the tunnel inner wall. The noise interference of the local contour feature is reduced by Gaussian filtering, median filtering, etc., and then the local contour feature is enhanced by morphological operations (such as expansion and corrosion) to make the extracted local contour feature clearer; then, the polynomial curve fitting technology can be used to model the local reflection data to obtain a light fluctuation curve, which is used to describe the light intensity change of the tunnel inner wall. The average curvature in the local contour feature and the amplitude of the light fluctuation curve are weighted and summed, and the weighted sum value is used as the deformation feature benchmark of the tunnel inner wall, wherein the weight of the weighted summation can be determined by the ratio of the actual detected local contour area to the spatial image area of ​​the tunnel inner wall, and the average curvature in the local contour feature and the amplitude of the light fluctuation curve can be calculated using the NumPy library.

[0091] It should be noted that the deformation feature benchmark in this application is an indicator reflecting the optical properties of the inner wall of the tunnel. The deformation feature benchmark is obtained by weighted summation of local contour features (such as edge contour and average curvature) and the amplitude of the light fluctuation curve, wherein the local contour features include edge position and morphological information, representing the curvature distribution in the local area. The changes in the shape characteristics and light reflection characteristics of the inner wall of the tunnel are comprehensively considered through the local contour features and deformation feature benchmark of the inner wall of the tunnel, which is conducive to a comprehensive deformation description of the inner wall of the tunnel.

[0092] In step S5, the local deformation matching model determines the deviation anomaly of the tunnel inner wall based on the deformation feature benchmark, and then obtains the convergent deformation data of the tunnel inner wall by screening the deviation anomaly.

[0093] In this embodiment, the local deformation matching model may determine the deviation abnormal value of the tunnel inner wall based on the deformation feature benchmark in the following manner, namely:

[0094] Converting the deformation feature datum into three-dimensional coordinates;

[0095] The three-dimensional coordinates of the deformation feature benchmark are substituted into the local deformation matching model for solution to obtain the deviation abnormality value of the tunnel inner wall.

[0096] In specific implementation, the deformation feature benchmark can be converted into three-dimensional coordinates through the polar coordinate conversion equation, and then the three-dimensional coordinates converted from the deformation feature benchmark can be substituted into the local deformation matching model for solution, and then the solved constant can be used as the deviation abnormal value of the tunnel inner wall; it should be noted that the local deformation matching model refers to the one obtained by iteratively eliminating abnormal data and surface reconstruction, which can be used to reflect the deformation state of the local area in the tunnel. The local deformation matching model is obtained by second-order polynomial fitting and is optimized through error elimination, so that the deformation trend it describes is accurate and the data noise is low.

[0097] In this embodiment, obtaining the convergence deformation data of the tunnel inner wall by screening the deviation abnormal values ​​means performing vector determination on the deviation abnormal values, and then screening to obtain the convergence deformation data of the tunnel inner wall.

[0098] It should be noted that by determining the direction of the deformation vector, the tunnel convergence deformation data that deviates from the normal state of the tunnel inner wall can be identified, thereby reducing misjudgment and improving the robustness and accuracy of the overall monitoring system. In specific implementation, the deviation outlier is vector-determined, and then the convergence deformation data of the tunnel inner wall is screened out, that is, the deviation outlier is converted into a vector through unique-hot encoding, and then the vector angle calculation formula is used to calculate the angle between the vector and the tunnel center vector. If the angle exceeds the preset threshold, the deviation outlier is used as the convergence deformation data of the tunnel inner wall. The tunnel center vector and the preset threshold can both be determined by the tunnel construction project, which will not be elaborated here.

[0099] It can be seen that in this application, firstly, by randomly deploying reflective markers on the inner wall of the tunnel and combining multi-eye vision equipment to extract spatial image information, it is convenient to enhance the flexibility of deployment in irregular or occluded areas, and provide a reliable reference for subsequent high-precision reconstruction; and by locally modeling the laser point cloud data, extracting a point cloud subset with spatial neighborhood consistency, and introducing a second-order polynomial fitting model to construct a local deformation descriptor, it can effectively capture local curvature changes and normal offsets, which is conducive to the modeling of nonlinear deformation data; secondly, a registration error model is constructed based on image information and point cloud fitting results, and a local deformation matching model is formed through error elimination and reconstruction. , it can dynamically exclude mismatched points or abnormal data, ensuring that the fitting model can truly reflect the actual deformation, which is conducive to enhancing the stability and reliability of the system in harsh environments; then, through local laser scanning and reflection data extraction, combined with contour detection and light fluctuation characteristics, a deformation feature benchmark is established to achieve fusion perception of geometric information and optical information, providing data support for judging whether the current deformation is abnormal; finally, by using the comparative analysis of the local deformation matching model and the deformation feature benchmark, the deviation abnormal values ​​with a direction toward the center of the tunnel inner wall and an amplitude exceeding the threshold are identified, and the data area representing the convergent deformation of the tunnel inner wall is further extracted, which can accurately identify the local deformation trend.

[0100] In summary, the technical solution adopted in this application can realize nonlinear local micro-deformation detection to improve the perception ability of irregular tunnel inner walls.

[0101] In the second embodiment, the present application provides a tunnel convergence deformation data acquisition system, referring to Figure 4 As shown in FIG, this figure is a module structure diagram of the data acquisition system shown in this embodiment of the present application, and the data acquisition system includes:

[0102] The data acquisition module 100 is used to obtain the reflection marking points on the inner wall of the tunnel and determine the spatial image information of the inner wall of the tunnel;

[0103] A point cloud fitting module 200 is used to collect laser point cloud data of the reflective marker point, perform local modeling on the laser point cloud data to obtain a local point cloud subset, and then perform nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor;

[0104] A registration optimization module 300 is configured to determine a registration error of the reflective marker points based on the local deformation descriptor and the spatial image information, and then iteratively eliminate and reconstruct the laser point cloud data based on the registration error to obtain a local deformation matching model;

[0105] A feature reference determination module 400 is configured to perform local laser scanning on the reflective marking points on the inner wall of the tunnel to obtain local reflection data, and determine the deformation feature reference of the inner wall of the tunnel based on the local reflection data;

[0106] The deformation determination module 500 is used for the local deformation matching model to determine the deviation abnormal value of the tunnel inner wall based on the deformation feature benchmark, and then obtain the convergent deformation data of the tunnel inner wall by screening the deviation abnormal value.

[0107] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0108] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program. The program can be stored in a computer-readable storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disc storage, magnetic disk storage, or magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0109] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

Claims

1. A method for collecting tunnel convergence deformation data, characterized in that: The data collection method comprises the following steps: Obtain the reflection marking points on the inner wall of the tunnel and determine the spatial image information of the inner wall of the tunnel; Collecting laser point cloud data of the reflective marker point, performing local modeling on the laser point cloud data to obtain a local point cloud subset, and then performing nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor; Determining a registration error of the reflective marker point based on the local deformation descriptor and the spatial image information, and then iteratively eliminating and reconstructing the laser point cloud data based on the registration error to obtain a local deformation matching model; Performing local laser scanning on the reflective marking points on the inner wall of the tunnel to obtain local reflection data, and determining a deformation characteristic benchmark of the inner wall of the tunnel based on the local reflection data; The local deformation matching model determines the deviation anomaly value of the tunnel inner wall based on the deformation feature benchmark, and then obtains the convergent deformation data of the tunnel inner wall through screening the deviation anomaly value.

2. A tunnel convergence deformation data acquisition method according to claim 1, characterized in that: The reflective marking points are reflective signs set on the surface of the inner wall of the tunnel, and the spatial image information of the inner wall of the tunnel is determined by a multi-eye vision camera installed on the tunnel detection equipment.

3. The method for collecting tunnel convergence deformation data according to claim 1, characterized in that: The laser point cloud data of the reflective marking points are collected by a laser scanner.

4. The method for collecting tunnel convergence deformation data according to claim 1, characterized in that: Performing local modeling on the laser point cloud data to obtain a local point cloud subset specifically includes: Performing neighborhood indexing on the laser point cloud data to obtain a neighborhood center point; The neighborhood center point is subjected to neighborhood modeling using a preset neighborhood radius to obtain a local point cloud subset.

5. The method for collecting tunnel convergence deformation data according to claim 1, characterized in that: Performing nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor specifically includes: Determine a fitting coefficient vector based on the point cloud data in the local point cloud subset; Performing nonlinear fitting on the point cloud data in the local point cloud subset to obtain a fitting surface; Determine the fitting residual of the point cloud data in the local point cloud subset by using the fitting surface and the fitting coefficient vector; The point cloud data in the local point cloud subset is fitted based on the fitting residual to obtain a local deformation descriptor.

6. The method for collecting tunnel convergence deformation data according to claim 1, characterized in that: Determining the registration error of the reflective marker point according to the local deformation descriptor and the spatial image information specifically includes: Performing feature description on the local deformation descriptor to obtain local deformation features of the tunnel inner wall; Determine the spatial position of the reflective marker point using the spatial image information, and then determine the surface features and normal vector of the reflective marker point using the spatial position; An error evaluation is performed on the curved surface feature and the local deformation feature according to the normal vector to obtain a registration error of the reflective marker point.

7. The method for collecting tunnel convergence deformation data according to claim 1, characterized in that: Perform local laser scanning on the reflective marking points on the inner wall of the tunnel to obtain local reflection data, including: The inner wall of the tunnel is locally scanned by laser scanning equipment to obtain a local scanning point set; Light sensor is used to collect light data reflected by local scanning points in the local scanning point set, and the light data is used as local reflection data.

8. The method for collecting tunnel convergence deformation data according to claim 1, characterized in that: Determining the deformation characteristic benchmark of the tunnel inner wall according to the local reflection data specifically includes: Performing contour detection on the local reflection data to obtain local contour features of the tunnel inner wall; A light fluctuation curve of the inner wall of the tunnel is determined, and then a deformation feature reference of the inner wall of the tunnel is determined based on the light fluctuation curve and the local contour feature.

9. The method for collecting tunnel convergence deformation data according to claim 1, characterized in that: Obtaining the convergent deformation data of the tunnel inner wall by screening the deviation abnormal values ​​means performing vector determination on the deviation abnormal values, and then screening to obtain the convergent deformation data of the tunnel inner wall.

10. A tunnel convergence deformation data acquisition system, used to execute a tunnel convergence deformation data acquisition method according to any one of claims 1 to 9, characterized in that: The data acquisition system includes: The data acquisition module is used to obtain the reflection marking points on the inner wall of the tunnel and determine the spatial image information of the inner wall of the tunnel; a point cloud fitting module, configured to collect laser point cloud data of the reflective marker point, perform local modeling on the laser point cloud data to obtain a local point cloud subset, and then perform nonlinear fitting on the local point cloud subset to obtain a local deformation descriptor; a registration optimization module, configured to determine a registration error of the reflective marker points based on the local deformation descriptor and the spatial image information, and then iteratively eliminate and reconstruct the laser point cloud data based on the registration error to obtain a local deformation matching model; A feature reference determination module is used to perform local laser scanning on the reflective marking points on the inner wall of the tunnel to obtain local reflection data, and determine the deformation feature reference of the inner wall of the tunnel based on the local reflection data; The deformation determination module is used for the local deformation matching model to determine the deviation abnormal value of the tunnel inner wall based on the deformation feature benchmark, and then obtain the convergent deformation data of the tunnel inner wall by screening the deviation abnormal value.

Images