Deformation visual monitoring method and system applied to tunnels

Abstract

The present application relates to the field of deformation visual monitoring technology, and specifically to a deformation visual monitoring method and system for tunnels, the method comprising: obtaining three-dimensional point cloud data of each tunnel section at each acquisition time; obtaining a three-dimensional coordinate system of each tunnel section at each acquisition time and dividing each voxel; obtaining a point feature histogram of each voxel based on the three-dimensional point cloud data; then obtaining a difference discrimination value of each voxel in each tunnel section at each acquisition time; obtaining abnormal voxels based on the average of the difference discrimination values; clustering the voxels based on the number of abnormal voxels to obtain suspected abnormal clusters; calculating the average curvature change rate of each suspected abnormal cluster of each tunnel section; then obtaining a deformation discrimination value of each suspected abnormal cluster of each tunnel section; and monitoring the deformation of the tunnel based on the deformation discrimination value. The present application improves the accuracy of deformation monitoring results.

Description

Technical Field

[0001] The present application relates to the technical field of deformation visual monitoring, and in particular to a deformation visual monitoring method and system applied to tunnels. Background Art

[0002] Tunnels, due to their unique underground spatial structure, are susceptible to deformation due to a variety of factors, including geological conditions, groundwater, and construction techniques. If these deformations are not addressed promptly, they can lead to tunnel instability and even serious safety accidents.

[0003] In order to ensure the normal use of the tunnel, real-time monitoring and regular maintenance of tunnel deformation are required. The data obtained by traditional leveling instruments, convergence meters, and total stations are relatively discrete and sparse, making it difficult to make timely and effective reflections on the overall deformation trend of the tunnel. With the development of laser scanning technology, the method of using point clouds to construct the inner wall surface of the tunnel can monitor the overall deformation of the tunnel. At present, when using point cloud data to analyze tunnel deformation, the difference in data before and after is observed to determine whether the tunnel has deformed or not. However, in actual calculations, due to the presence of obstacles, the calculation results cannot accurately reflect the deformation of the tunnel, resulting in inaccurate deformation monitoring results for the tunnel. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a deformation visual monitoring method and system for tunnels. The technical solutions adopted are as follows:

[0005] In a first aspect, an embodiment of the present application provides a deformation visual monitoring method applied to a tunnel, the method comprising the following steps:

[0006] Acquire three-dimensional point cloud data of each tunnel section at each acquisition time; acquire a three-dimensional coordinate system of each tunnel section at each acquisition time, and divide the three-dimensional coordinate system of each tunnel section into voxels of a preset size; and acquire a point feature histogram of each voxel based on the three-dimensional point cloud data;

[0007] Based on the correlation between the point feature histograms of each voxel and the difference between the number of three-dimensional point cloud data in each voxel, the difference discrimination value of each voxel in each tunnel section at each acquisition time is obtained;

[0008] Abnormal voxels are obtained based on the average of the difference discriminant values; voxels are clustered based on the number of abnormal voxels to obtain suspected abnormal clusters;

[0009] Based on the change of curvature of the three-dimensional point cloud data points in the suspected abnormal clusters on the fitting surface of the three-dimensional point cloud data, the average curvature change rate of each suspected abnormal cluster in each tunnel section is obtained;

[0010] Based on the average curvature change rate and the average curvature of each 3D point cloud data point in the suspected abnormal cluster on the 3D point cloud data fitting surface, the deformation discrimination value of each suspected abnormal cluster in each tunnel section is obtained;

[0011] The deformation of the tunnel is monitored based on the deformation discrimination value.

[0012] Furthermore, the obtaining of the three-dimensional coordinate system of each tunnel section at each acquisition time includes:

[0013] For each tunnel segment, the three-dimensional point cloud data of the tunnel segment at each acquisition moment and the three-dimensional point cloud data of the tunnel segment at each historical acquisition moment are projected into the same three-dimensional coordinate system to obtain the three-dimensional coordinate system of each tunnel segment at each acquisition moment.

[0014] Furthermore, the method for obtaining the point feature histogram is:

[0015] For each acquisition moment, the three-dimensional point cloud data within each voxel of each tunnel section is used as the input of the PFH point feature histogram algorithm, a spherical area with a preset length as the radius is used as the nearest neighbor neighborhood of each three-dimensional point cloud data point, and the point feature histogram is output as the point feature histogram of the three-dimensional point cloud data within each voxel of each tunnel section.

[0016] Furthermore, the calculation formula of the difference discrimination value is: Where, represents the difference discriminant value between the j-th voxel of the i-th tunnel segment at each acquisition moment and the j-th voxel of the i-th tunnel segment at each historical acquisition moment; represents the point feature histogram of the jth voxel in the i-th tunnel segment at each acquisition moment, Represents the point feature histogram of the jth voxel in the i-th tunnel segment at each historical acquisition time; is the distance function; represents the number of 3D point cloud data within the jth voxel of the i-th tunnel segment at each acquisition time; represents the number of 3D point cloud data within the jth voxel of the i-th tunnel segment at each historical acquisition moment; is the preset parameter factor.

[0017] Furthermore, the method for obtaining the abnormal voxels is:

[0018] Calculate the average difference discrimination value between the j-th voxel of the i-th tunnel segment at each acquisition time and the j-th voxel of the i-th tunnel segment at all previous acquisition times as the average difference discrimination value of the j-th voxel of the i-th tunnel segment at each acquisition time;

[0019] For the average difference discrimination value of each voxel of each tunnel section at each acquisition moment, when the average difference discrimination value is greater than a first preset threshold, the voxel is an abnormal voxel.

[0020] Furthermore, the method for obtaining the suspected abnormal cluster is:

[0021] The center coordinates of all abnormal voxels in each tunnel section are used as input to the density clustering algorithm, and each cluster is output;

[0022] A cluster in which the number of abnormal voxels is greater than a second preset threshold is regarded as a suspected abnormal cluster.

[0023] Furthermore, the method for obtaining the average curvature change rate is:

[0024] The three-dimensional coordinates of the three-dimensional point cloud data of each tunnel section at each acquisition time are used as the input of the least squares method to obtain the three-dimensional point cloud data fitting surface, and the curvature of each three-dimensional point cloud data point mapped from the xoy plane in the three-dimensional coordinate system to the three-dimensional point cloud data fitting surface is calculated as the curvature of each three-dimensional point cloud data point;

[0025] For each suspected abnormal cluster, the curvatures of all three-dimensional point cloud data points with the same z coordinate in the suspected abnormal cluster are arranged from small to large according to the x coordinate, and each curvature sequence is obtained, and the first-order difference sequence of each curvature sequence is calculated;

[0026] The calculation formula of the average curvature change rate is: Where, represents the average curvature change rate of the mth suspected abnormal cluster in the i-th tunnel section at each acquisition time, Represents the first-order difference sequence of the sth curvature sequence in the mth suspected anomaly cluster in the ith tunnel segment at each acquisition moment; Represents the first-order difference sequence The average value of the absolute values ​​of all elements in , p is the number of first-order difference sequences of the mth suspected abnormal cluster in the i-th tunnel segment.

[0027] Furthermore, the calculation formula of the deformation discrimination value is: Where, represents the deformation discrimination value of the mth suspected abnormal cluster in the i-th tunnel section; Represents the average value of the curvature of the 3D point cloud data points in all abnormal voxels in the mth suspected abnormal cluster in the i-th tunnel segment at each acquisition time; Represents the average value of the y-coordinates of the three-dimensional point cloud data points in all abnormal voxels in the m-th suspected abnormal cluster in the i-th tunnel segment at each acquisition time.

[0028] Furthermore, the deformation monitoring of the tunnel based on the deformation discrimination value includes:

[0029] For each suspected abnormal cluster in each tunnel section at each acquisition time, if the deformation discrimination value of the suspected abnormal cluster is greater than the third preset threshold, the abnormal voxel in the suspected abnormal cluster is a large obstacle; if the deformation discrimination value is less than or equal to the third preset threshold, the abnormal voxel in the suspected abnormal cluster is tunnel deformation.

[0030] In a second aspect, an embodiment of the present application also provides a deformation visual monitoring system for tunnels, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of any one of the above methods when executing the computer program.

[0031] This application has at least the following beneficial effects:

[0032] The present application constructs a difference discrimination value by analyzing the difference characteristics of the three-dimensional point cloud images of each historical acquisition moment and each single tunnel segment at each acquisition moment, which is used to judge whether the voxels in the tunnel segment at each acquisition moment have changed compared with the historical acquisition moments of the acquisition moment, and the voxels with larger difference discrimination values ​​are regarded as abnormal voxels; according to the density characteristics of the abnormal voxels, the abnormal voxels belonging to small obstacles are screened out; according to the curvature and spatial position of the midpoint of the abnormal voxels, the abnormal voxels belonging to large obstacles are distinguished from the abnormal voxels belonging to tunnel deformation; the present application not only takes into account the difference characteristics of the tunnel at each acquisition moment and each historical acquisition moment of the acquisition moment, but also takes into account the interference of obstacles in the tunnel, so that the deformation monitoring results of the tunnel can more accurately reflect the actual deformation of the tunnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or 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 some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0034] Figure 1 A flowchart of a method for visually monitoring deformation of a tunnel according to an embodiment of the present application;

[0035] Figure 2 A schematic diagram of histogram changes provided for one embodiment of the present application;

[0036] Figure 3 A schematic diagram of a fitting surface for three-dimensional point cloud data provided in one embodiment of the present application. DETAILED DESCRIPTION

[0037] To further illustrate the technical means and effectiveness of this application to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effectiveness of the method and system for visual deformation monitoring in tunnels proposed in this application. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0038] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0039] The specific scheme of the deformation visual monitoring method and system for tunnels provided by the present application is described in detail below with reference to the accompanying drawings.

[0040] See also Figure 1 , which shows a flowchart of a deformation visual monitoring method for a tunnel provided by an embodiment of the present application, the method comprising the following steps:

[0041] Step S1, obtaining three-dimensional point cloud data of each tunnel section at each acquisition time.

[0042] Tunnel projects are long and narrow structures, and the scanning range of 3D laser scanners is limited. Typically, a segmented approach is used to acquire point cloud data within the tunnel, requiring a certain degree of overlap between segments to ensure continuous monitoring data. To detect deformation in a mountain tunnel, this embodiment uses a 3D laser scanner with a maximum scanning distance of 36 meters. A set of targets is affixed every 30 meters within the tunnel. The 3D laser scanner is positioned 3 meters in front of each target position and begins scanning. This ensures that the 3D point cloud data obtained from each scan includes the target at both ends. Implementers can choose other distances based on actual conditions. The point cloud data obtained from each scan is registered and spliced ​​based on the target positions, ultimately generating point cloud data for the entire tunnel.

[0043] According to the above method, the three-dimensional point cloud data of the entire tunnel at each acquisition time is obtained. For each acquisition time, every 10 cm of the entire tunnel is regarded as a tunnel segment, and the three-dimensional point cloud data of each tunnel segment at each acquisition time is obtained.

[0044] Step S2, obtaining the three-dimensional coordinate system of each tunnel section at each acquisition moment, dividing the three-dimensional coordinate system of each tunnel section into voxels of preset size; obtaining the point feature histogram of each voxel based on the three-dimensional point cloud data; obtaining the difference discrimination value of each voxel in each tunnel section at each acquisition moment based on the correlation between the point feature histograms of each voxel and the difference between the amount of three-dimensional point cloud data in each voxel; obtaining abnormal voxels based on the average of the difference discrimination values; clustering the voxels based on the number of abnormal voxels to obtain suspected abnormal clusters; obtaining the average curvature change rate of each suspected abnormal cluster of each tunnel section based on the change of the curvature of the three-dimensional point cloud data points in the suspected abnormal cluster on the fitting surface of the three-dimensional point cloud data; obtaining the deformation discrimination value of each suspected abnormal cluster of each tunnel section based on the average curvature change rate and the average of the curvature of each three-dimensional point cloud data point in the suspected abnormal cluster on the fitting surface of the three-dimensional point cloud data.

[0045] Due to the massive amount of collected 3D point cloud data, directly analyzing the entire tunnel's point cloud data can lead to low computational efficiency and slow processing. Deformation in mountain tunnels is often caused by geological conditions or environmental factors, and the deformation typically covers a large area. To reduce the computational complexity of the 3D point cloud data, only deformation within each small segment needs to be detected to determine the exact location in the tunnel. The 3D point cloud data for each tunnel segment collected at different times is analyzed to determine whether any tunnel segment has deformed.

[0046] For each collection time, other collection times before the collection time are taken as historical collection times of the collection time.

[0047] The 3D point cloud data for each tunnel segment is projected into 3D space according to the 3D coordinates of each point, generating a 3D point cloud image. The x-axis, y-axis, and z-axis represent the width, length, and height of the tunnel, respectively. By comparing the 3D point cloud image features of the same tunnel segment at each acquisition time with those at previous acquisition times, it is possible to determine whether the tunnel segment has undergone any deformation.

[0048] If the tunnel at each acquisition moment has deformed compared to previous historical acquisition moments, the location of the deformation might manifest as vault settlement, sidewall deformation, or bottom bulging. In the 3D point cloud image corresponding to the tunnel segment at that location, the deformed location will appear as a bulge or depression. Furthermore, because obstacles such as fallen objects may exist within the tunnel, some bulges that appear in the point cloud plane of the 3D point cloud image may not be tunnel deformation, but rather obstacles.

[0049] For each tunnel segment, the three-dimensional point cloud data of the tunnel segment at each acquisition moment and the three-dimensional point cloud data of the tunnel segment at each historical acquisition moment are projected into the same three-dimensional coordinate system. Then, the points in the undeformed area of ​​the tunnel will overlap, while the points in the deformed area will not overlap.

[0050] The 3D coordinate system of each tunnel segment is divided into voxels of size a×a×a. In this embodiment, a is set to 0.5 cm; implementers may select other values ​​based on actual conditions. The point cloud data for the same voxel at each acquisition time is compared with the point cloud data for each historical acquisition time to determine whether the track area corresponding to the voxel has changed.

[0051] At each acquisition moment, the three-dimensional point cloud data within each voxel of each tunnel section is used as the input of the PFH point feature histogram algorithm. With each three-dimensional point cloud data point at each acquisition moment as the center of the circle and a spherical area with a radius of r as the nearest neighbor of each three-dimensional point cloud data point, the geometric features within the nearest neighbor of each point are calculated. The calculation process is a well-known process and will not be described in detail here. Based on the geometric features, the three-dimensional point cloud data point feature histogram within each voxel of each tunnel section is output as the point feature histogram of each voxel. In this embodiment, the value of r is 0.2 cm. The implementer can select other values ​​according to the actual situation. The PFH point feature histogram algorithm is a well-known technology and will not be described in detail here. The histogram change diagram is shown as follows: Figure 2 As shown in , for each voxel of each tunnel section at each acquisition time, if the position of the voxel is deformed, the number of 3D point cloud data points in the voxel will be different from the number of 3D point cloud data points in other voxels, such as Figure 2 The height variation of the black rectangle is shown in the figure.

[0052] If the position of each voxel in each tunnel section is deformed, then the number of 3D point cloud data points at each acquisition moment in the voxel will be significantly different from the number of 3D point cloud data points at each historical acquisition moment, and the feature histogram of the 3D point cloud data points in the voxel will also be significantly different.

[0053] Based on the above characteristics, a single voxel difference discrimination value is constructed to determine whether the j-th voxel of the i-th tunnel segment at each acquisition time has changed compared with the historical acquisition time. The calculation formula is: Where, represents the difference discriminant value between the j-th voxel of the i-th tunnel segment at each acquisition moment and the j-th voxel of the i-th tunnel segment at each historical acquisition moment; represents the point feature histogram of the j-th voxel at each acquisition moment, Represents the point feature histogram of the j-th voxel at each historical acquisition time; is the Bhattacharyya distance function, which is used to calculate the Bhattacharyya distance of the input histogram; represents the number of points contained in the point cloud of the j-th voxel at each acquisition moment; represents the number of points contained in the point cloud of the j-th voxel at each historical acquisition time; To preset the parameter adjustment factor and prevent the absolute value from being zero, the value is taken as 1 in this embodiment. The implementer can select other values ​​according to actual conditions.

[0054] It should be noted that when The larger the value of , the greater the difference between the feature of the point cloud at each acquisition moment and the feature of the point cloud at each historical acquisition moment in the j-th voxel of the i-th tunnel segment. This further indicates that the difference between the feature of the point cloud at each acquisition moment and the feature of the historical acquisition moment at the geographical location of the j-th voxel of the i-th tunnel segment is greater. The larger the value of , the greater the difference between the number of points at each acquisition moment in the j-th voxel in the i-th tunnel segment and the number of points at each historical acquisition moment. This further indicates that the difference between each acquisition moment and each historical acquisition moment at the geographical location of the j-th voxel in the i-th tunnel segment is greater.

[0055] Furthermore, the average difference discrimination value between the j-th voxel of the i-th tunnel segment at each acquisition moment and the j-th voxel of the i-th tunnel segment at all previous acquisition moments is calculated as the average difference discrimination value of the j-th voxel of the i-th tunnel segment at each acquisition moment.

[0056] For the average difference discrimination value of each voxel in each tunnel section at each acquisition moment, when the average difference discrimination value is greater than a first preset threshold, the voxel is an abnormal voxel; when the average difference discrimination value is less than or equal to the first preset threshold, the voxel is a normal voxel. In this embodiment, the value of the first preset threshold is 10, and the implementer may select other values ​​according to actual conditions.

[0057] At this point, all abnormal voxels in each tunnel segment are obtained. These abnormal voxels may contain tunnel deformation pixels or obstacle pixels. Further judgment is needed on the abnormal voxels to distinguish tunnel deformation pixels from obstacle pixels.

[0058] Because tunnel vault settlement, sidewall deformation, or bottom bulging typically affects a large area rather than just a single point, the deformation occurs across more than just a single voxel. Obstacles within tunnels are categorized into large and small obstacles. Large obstacles include cargo accidentally dropped by passing trucks, while small obstacles include small spots of mud splashed onto the tunnel walls by passing cars.

[0059] When the abnormal voxel is a small obstacle like a small mud dot, there are fewer abnormal voxels adjacent to the abnormal voxel; when the abnormal voxel is a large obstacle or tunnel deformation, there are more abnormal voxels adjacent to the abnormal voxel.

[0060] The center point coordinates of all abnormal voxels in each tunnel segment are used as the input of the DCP density clustering algorithm, and the cutoff distance is 1% of the total number of abnormal voxels to obtain each cluster. The DCP density clustering algorithm is a well-known technology and the specific process will not be repeated here.

[0061] If the number of outlier voxels in a cluster is less than or equal to the second preset threshold e, the locations of all outlier voxels in the cluster are considered to be small obstacles. If the number of outlier voxels in a cluster is greater than the second preset threshold e, the outliers in the cluster may be large obstacles or tunnel deformation voxels. In this embodiment, the second preset threshold e is set to 12; implementers may select other values ​​based on actual circumstances.

[0062] If the number of clusters with abnormal voxels greater than the second preset threshold e is 0, it means that the i-th tunnel segment does not contain tunnel deformation positions or large obstacles; as long as there is a cluster with the number of abnormal voxels greater than the second preset threshold e, it means that the i-th tunnel segment contains tunnel deformation positions or large obstacles, and the cluster with the number of abnormal voxels greater than the second preset threshold e is regarded as a suspected abnormal cluster.

[0063] Furthermore, the suspected abnormal clusters are distinguished to determine whether the location of the abnormal voxels in the cluster is a large obstacle or a tunnel deformation.

[0064] Tunnel deformation is usually a slight deformation of a large area, while a large obstacle is a severe deformation of a large area. Therefore, among the abnormal voxels at each acquisition time, the curvature of the points in the abnormal voxels belonging to the tunnel deformation is smaller than the curvature of the points in the abnormal voxels belonging to the large obstacle. In addition, since the shape of a large obstacle is usually irregular, while tunnel deformation is a slight arch of the tunnel arm, the surface of the tunnel deformation is relatively flat relative to the surface of the large obstacle. Therefore, if the point cloud in the cluster belongs to the large obstacle, the curvature of the adjacent point cloud data will change significantly. If the point cloud in the cluster belongs to the tunnel deformation, the curvature of the adjacent point cloud data will change slightly. Large obstacles are usually located at the bottom of the tunnel, so the y coordinates of the points in the abnormal voxels belonging to the large obstacle will be smaller.

[0065] The 3D coordinates of the 3D point cloud data at each acquisition time of each tunnel section are used as the input of the least squares method to obtain the fitting surface of the 3D point cloud data. The schematic diagram of the fitting surface of the 3D point cloud data is shown in the figure. Figure 3As shown, the curvature of each 3D point cloud data point on the 3D point cloud data fitting surface can be obtained according to the fitting surface and curvature calculation formula. The calculation method is a well-known technology and will not be described in detail here.

[0066] For each suspected abnormal cluster, the curvatures of all three-dimensional point cloud data points with the same z coordinate in the suspected abnormal cluster are arranged from small to large according to the x coordinate, and each curvature sequence is obtained, and the first-order difference sequence of each curvature sequence is calculated.

[0067] In order to reflect the average change of curvature, the average curvature change rate is constructed, and the formula is: Where, represents the average curvature change rate of the mth cluster in the i-th tunnel segment at each acquisition moment, Represents the first-order difference sequence of the sth curvature sequence in the mth cluster in the ith tunnel segment at each acquisition moment; Represents the first-order difference sequence The average value of the absolute value of each element in , p is the number of first-order difference sequences of the mth suspected abnormal cluster in the i-th tunnel section at each acquisition time.

[0068] It should be noted that when When the value of is large, it means It means that the adjacent data in the sth curvature sequence in the mth cluster in the i-th tunnel segment have a larger change. The larger the value of the average curvature change rate is, the greater the curvature difference of the sequence in the three-dimensional point cloud data points is, which indicates that the sequence in the three-dimensional point cloud data points may be a large obstacle.

[0069] Furthermore, in order to determine whether the abnormal voxels in the mth cluster in the i-th tunnel segment at each acquisition time belong to tunnel deformation or large obstacles, a deformation judgment value is constructed, and the formula is: Where, represents the deformation discriminant value of the mth cluster in the i-th tunnel section at each acquisition moment; represents the average value of the curvature of the 3D point cloud data points in all abnormal voxels in the mth cluster in the i-th tunnel segment at each acquisition time; Represents the average value of the y-coordinates of the three-dimensional point cloud data points in all abnormal voxels in the m-th cluster in the i-th tunnel segment at each acquisition time.

[0070] It should be noted that when When the value of is large, it means that the average curvature change rate of all points in the mth cluster in the i-th tunnel segment is large, which further indicates that the points in the mth cluster in the i-th tunnel segment may be large obstacles.

[0071] when When the value of is large, it means that the average curvature of all abnormal voxels in the mth cluster in the i-th tunnel segment is large, which further indicates that the possibility that the abnormal voxels in the mth cluster in the i-th tunnel segment belong to large obstacles is greater; when When the value of is small, it means that all abnormal voxels in the mth cluster in the i-th tunnel segment are located at a lower position, which further indicates that the abnormal voxels in the mth cluster in the i-th tunnel segment are more likely to belong to large obstacles.

[0072] At this point, the deformation discrimination value of each suspected abnormal cluster in each tunnel section at each acquisition time is obtained.

[0073] Step S3: monitoring the deformation of the tunnel based on the deformation discrimination value.

[0074] For each suspected abnormal cluster in each tunnel section at each acquisition moment, if the deformation discrimination value of the suspected abnormal cluster is greater than the third preset threshold, the abnormal voxel in the suspected abnormal cluster is a large obstacle; if the deformation discrimination value is less than or equal to the third preset threshold, the abnormal voxel in the suspected abnormal cluster is tunnel deformation; in this embodiment, the value of the third preset threshold is 0.4, and the implementer can select other values ​​according to actual conditions.

[0075] Furthermore, each tunnel section of the entire tunnel is judged according to the above steps to determine whether deformation occurs, and the deformation position of the entire tunnel can be monitored based on the positions of all tunnel sections that have been deformed.

[0076] Based on the same inventive concept as the above-mentioned method, an embodiment of the present application also provides a deformation visual monitoring system applied to tunnels, comprising a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned deformation visual monitoring methods applied to tunnels.

[0077] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0078] The various embodiments in this application are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0079] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A deformation visual monitoring method applied to tunnels, characterized in that: The method comprises the following steps: Acquire three-dimensional point cloud data of each tunnel section at each acquisition time; acquire a three-dimensional coordinate system of each tunnel section at each acquisition time, and divide the three-dimensional coordinate system of each tunnel section into voxels of a preset size; and acquire a point feature histogram of each voxel based on the three-dimensional point cloud data; Based on the correlation between the point feature histograms of each voxel and the difference between the number of three-dimensional point cloud data in each voxel, the difference discrimination value of each voxel in each tunnel section at each acquisition time is obtained; Abnormal voxels are obtained based on the average of the difference discriminant values; voxels are clustered based on the number of abnormal voxels to obtain suspected abnormal clusters; Based on the change of curvature of the three-dimensional point cloud data points in the suspected abnormal clusters on the fitting surface of the three-dimensional point cloud data, the average curvature change rate of each suspected abnormal cluster in each tunnel section is obtained; Based on the average curvature change rate and the average curvature of each 3D point cloud data point in the suspected abnormal cluster on the 3D point cloud data fitting surface, the deformation discrimination value of each suspected abnormal cluster in each tunnel section is obtained; The deformation of the tunnel is monitored based on the deformation discrimination value.

2. The deformation visual monitoring method for tunnels according to claim 1, characterized in that: The obtaining of the three-dimensional coordinate system of each tunnel section at each acquisition moment includes: For each tunnel segment, the three-dimensional point cloud data of the tunnel segment at each acquisition moment and the three-dimensional point cloud data of the tunnel segment at each historical acquisition moment are projected into the same three-dimensional coordinate system to obtain the three-dimensional coordinate system of each tunnel segment at each acquisition moment.

3. The deformation visual monitoring method for tunnels according to claim 1, characterized in that: The method for obtaining the point feature histogram is: For each acquisition moment, the three-dimensional point cloud data within each voxel of each tunnel section is used as the input of the PFH point feature histogram algorithm, a spherical area with a preset length as the radius is used as the nearest neighbor neighborhood of each three-dimensional point cloud data point, and the point feature histogram is output as the point feature histogram of the three-dimensional point cloud data within each voxel of each tunnel section.

4. The deformation visual monitoring method for tunnels according to claim 1, characterized in that: The calculation formula of the difference discrimination value is: Where, represents the difference discriminant value between the j-th voxel of the i-th tunnel segment at each acquisition moment and the j-th voxel of the i-th tunnel segment at each historical acquisition moment; represents the point feature histogram of the jth voxel in the i-th tunnel segment at each acquisition moment, Represents the point feature histogram of the jth voxel in the i-th tunnel segment at each historical acquisition time; is the distance function; represents the number of 3D point cloud data within the jth voxel of the i-th tunnel segment at each acquisition time; represents the number of 3D point cloud data within the jth voxel of the i-th tunnel segment at each historical acquisition moment; is the preset parameter factor.

5. The deformation visual monitoring method for tunnels according to claim 1, characterized in that: The method for obtaining the abnormal voxels is: Calculate the average difference discrimination value between the j-th voxel of the i-th tunnel segment at each acquisition time and the j-th voxel of the i-th tunnel segment at all previous acquisition times as the average difference discrimination value of the j-th voxel of the i-th tunnel segment at each acquisition time; For the average difference discrimination value of each voxel of each tunnel section at each acquisition moment, when the average difference discrimination value is greater than a first preset threshold, the voxel is an abnormal voxel.

6. The deformation visual monitoring method for tunnels according to claim 1, characterized in that: The method for obtaining the suspected abnormal cluster is: The center coordinates of all abnormal voxels in each tunnel section are used as input to the density clustering algorithm, and each cluster is output; A cluster in which the number of abnormal voxels is greater than a second preset threshold is regarded as a suspected abnormal cluster.

7. The deformation visual monitoring method for tunnels according to claim 1, characterized in that: The method for obtaining the average curvature change rate is: The three-dimensional coordinates of the three-dimensional point cloud data of each tunnel section at each acquisition time are used as the input of the least squares method to obtain the three-dimensional point cloud data fitting surface, and the curvature of each three-dimensional point cloud data point mapped from the xoy plane in the three-dimensional coordinate system to the three-dimensional point cloud data fitting surface is calculated as the curvature of each three-dimensional point cloud data point; For each suspected abnormal cluster, the curvatures of all three-dimensional point cloud data points with the same z coordinate in the suspected abnormal cluster are arranged from small to large according to the x coordinate, and each curvature sequence is obtained, and the first-order difference sequence of each curvature sequence is calculated; The calculation formula of the average curvature change rate is: Where, represents the average curvature change rate of the mth suspected abnormal cluster in the i-th tunnel section at each acquisition time, Represents the first-order difference sequence of the sth curvature sequence in the mth suspected anomaly cluster in the ith tunnel segment at each acquisition moment; Represents the first-order difference sequence The average value of the absolute values ​​of all elements in , p is the number of first-order difference sequences of the mth suspected abnormal cluster in the i-th tunnel segment.

8. The deformation visual monitoring method for tunnels according to claim 7, characterized in that: The calculation formula of the deformation discrimination value is: Where, represents the deformation discrimination value of the mth suspected abnormal cluster in the i-th tunnel section; Represents the average value of the curvature of the 3D point cloud data points in all abnormal voxels in the mth suspected abnormal cluster in the i-th tunnel segment at each acquisition time; Represents the average value of the y-coordinates of the three-dimensional point cloud data points in all abnormal voxels in the m-th suspected abnormal cluster in the i-th tunnel segment at each acquisition time.

9. The deformation visual monitoring method for tunnels according to claim 1, characterized in that: The monitoring of the deformation of the tunnel based on the deformation discrimination value includes: For each suspected abnormal cluster in each tunnel section at each acquisition time, if the deformation discrimination value of the suspected abnormal cluster is greater than the third preset threshold, the abnormal voxel in the suspected abnormal cluster is a large obstacle; if the deformation discrimination value is less than or equal to the third preset threshold, the abnormal voxel in the suspected abnormal cluster is tunnel deformation.

10. A deformation visual monitoring system for a tunnel, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the deformation visual monitoring method applied to a tunnel as described in any one of claims 1 to 9 are implemented.

Images