Shield tunnel segment segmentation method and system

Abstract

The invention discloses a segment segmentation method and system for a shield tunnel, and is applied to the field of shield tunnels, and the method comprises the steps: obtaining original point cloud data of the shield tunnel, and carrying out the preprocessing, and obtaining tunnel point cloud data; for each circular seam of the shield tunnel, identifying the circular seam according to the tunnel point cloud data to obtain a fitting plane and target point cloud data of the circular seam; dividing the tunnel point cloud data into independent point cloud data of a plurality of segment rings of the shield tunnel according to the fitting plane of each circular seam; for the longitudinal joints of the segment rings, the longitudinal joints of the segment rings are identified according to the independent point cloud data of the segment rings, and longitudinal joint point cloud data of the segment rings are obtained; and performing segment segmentation on the shield tunnel according to the target point cloud data of each circular seam and the longitudinal seam point cloud data of each segment ring to obtain a plurality of segment areas of the shield tunnel. According to the method, the segment segmentation precision of the shield tunnel can be effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of shield tunnels, and in particular to a method and system for segmenting shield tunnel segments. Background Art

[0002] A shield tunnel is a tunnel obtained by using a shield machine for tunnel excavation and lining operations. It is usually composed of multiple segment rings connected in sequence. Among them, the segment ring is the main lining body of the shield tunnel. It is assembled from prefabricated reinforced concrete segments. These segments include standard blocks, adjacent blocks and capping blocks; the annular seam refers to the connection seam between two adjacent segment rings; the longitudinal seam refers to the connection seam between each segment in the segment ring. In the related technology, laser radar scanning technology is used to capture the point cloud data of the shield tunnel, and the geometric information of each segment ring is extracted from the point cloud data of the shield tunnel. Then, based on the geometric information of each segment ring, the shield tunnel is segmented, thereby obtaining the segment distribution information of the shield tunnel. The segment distribution information can be used to analyze the structure of the shield tunnel. However, the implementation of related technologies relies on data fitting methods or spatial distance measurement methods, such as Hough transform, region growing, random sampling consistency and fast point feature histogram, which results in low processing efficiency and high spatial complexity of related technologies when facing large-scale tunnel point cloud data, resulting in phenomena such as loss of important point cloud features. The segment segmentation accuracy of related technologies needs to be improved. Summary of the invention

[0003] The present application provides a shield tunnel segment segmentation method and system for improving the segment segmentation accuracy of the shield tunnel.

[0004] On the one hand, an embodiment of the present application provides a method for segmenting a shield tunnel, comprising the following steps: Obtain the original point cloud data of the shield tunnel and perform preprocessing to obtain the tunnel point cloud data; For each annular seam of the shield tunnel, the annular seam is identified according to the tunnel point cloud data to obtain a fitting plane and target point cloud data of the annular seam; Dividing the tunnel point cloud data into independent point cloud data of a plurality of segment rings of the shield tunnel according to the fitting planes of the annular joints; For each of the longitudinal seams of the segment rings, the longitudinal seams of the segment rings are identified according to the independent point cloud data of the segment rings to obtain the point cloud data of the longitudinal seams of the segment rings; The shield tunnel is segmented according to the target point cloud data of each annular seam and the point cloud data of the longitudinal seam of each segment ring to obtain a plurality of segment areas of the shield tunnel.

[0005] On the other hand, an embodiment of the present application provides a segment segmentation system for a shield tunnel, comprising: Point cloud scanning module, used to obtain the original point cloud data of the shield tunnel and perform preprocessing to obtain tunnel point cloud data; A ring seam identification module, for identifying each ring seam of the shield tunnel according to the tunnel point cloud data, and obtaining a fitting plane and target point cloud data of the ring seam; A division processing module, used for dividing the tunnel point cloud data into independent point cloud data of a plurality of segment rings of the shield tunnel according to the fitting planes of the annular joints; A longitudinal seam identification module is used to identify the longitudinal seam of each segment ring according to the independent point cloud data of the segment ring, so as to obtain the longitudinal seam point cloud data of the segment ring; The segment segmentation module is used to segment the shield tunnel according to the target point cloud data of each annular seam and the longitudinal seam point cloud data of each segment ring to obtain multiple segment areas of the shield tunnel.

[0006] According to the shield tunnel segment segmentation method and system of the embodiment of the present application, first, the original point cloud data of the shield tunnel is obtained and preprocessed to obtain tunnel point cloud data; secondly, for each annular joint of the shield tunnel, the annular joint is identified according to the tunnel point cloud data to obtain the fitting plane and target point cloud data of the annular joint; then, the tunnel point cloud data is divided into independent point cloud data of multiple segment rings of the shield tunnel according to the fitting plane of each annular joint; thereafter, for the longitudinal joint of each segment ring, the longitudinal joint of the segment ring is identified according to the independent point cloud data of the segment ring to obtain the longitudinal joint point cloud data of the segment ring; finally, the shield tunnel is segmented according to the target point cloud data of each annular joint and the longitudinal joint point cloud data of each segment ring to obtain multiple segment areas of the shield tunnel, which can not only improve the processing efficiency and reduce the spatial complexity when facing large-scale tunnel point cloud data, but also reduce the influence of feature loss on segment segmentation, thereby improving the segment segmentation accuracy of the shield tunnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 It is a flow chart of the segment segmentation method of the shield tunnel provided by the present application; Figure 2 It is a structural diagram of the three-dimensional cylindrical coordinate system provided by this application; Figure 3 is a graph of the relationship between radial distance and cross-sectional width at different azimuth angles provided by the present application; Figure 4 It is a schematic diagram of the annular seam identification provided by the present application; Figure 5 It is a schematic diagram of the annular seam plane fitting provided by this application; Figure 6 is a structural expansion diagram of the segment ring provided by the present application; Figure 7 is a graph of the relationship between azimuth and radial distance provided by the present application; Figure 8 This is an example diagram of the effect of segment segmentation provided in this application. DETAILED DESCRIPTION

[0008] The present application is further described below in conjunction with the accompanying drawings and specific embodiments. The described embodiments should not be regarded as limiting the present application, and all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0009] In response to the defects existing in the relevant technology, the present application provides a method and system for segmenting the shield tunnel segment, which aims to accurately extract the relevant point clouds of the annular joints and the relevant point clouds of the longitudinal joints from the point cloud data of the shield tunnel, and then use the relevant point clouds of the annular joints and the relevant point clouds of the longitudinal joints to realize the segmentation of the shield tunnel, thereby improving the segment segmentation accuracy of the shield tunnel.

[0010] It should be noted that the shield tunnel in the embodiment of the present application adopts a single-layer flat-plate segment for lining, and the outer diameter, inner diameter and ring width of the segment can be flexibly set according to actual conditions; the line fitting and correction adopt a universal double-sided wedge control, and the wedge amount can be flexibly set according to actual conditions; for each segment ring, the segment is staggered assembled in a block mode, that is, each segment ring includes multiple standard blocks, 2 adjacent blocks and 1 capping block, the number of standard blocks is flexibly set according to actual conditions, and multiple bent bolts are used to connect the segments; a number of bent bolts are used to connect the segment rings; the number of bent bolts is flexibly set according to actual conditions.

[0011] First, the implementation steps of the shield tunnel segment segment segmentation method provided by the present application will be described in detail below with reference to the accompanying drawings.

[0012] The segment segmentation method of the shield tunnel provided in the present application can be applied to a terminal, a server, or software running in a terminal or a server. The terminal can be a tablet computer, a laptop computer, a desktop computer, etc., but is not limited thereto. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms. In addition, the server can also be a node server in a blockchain network, but is not limited thereto. Among them, blockchain is a new application model of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.

[0013] Reference Figure 1 , Figure 1 It is a flow chart of the segment segmentation method of the shield tunnel provided in the present application. The segment segment segmentation method may include the following steps S101-S105.

[0014] S101, obtaining the original point cloud data of the shield tunnel and preprocessing it to obtain the tunnel point cloud data.

[0015] In this step, first, a three-dimensional laser scan is performed on the interior of the shield tunnel to obtain the original point cloud data of the shield tunnel; then, the original point cloud data of the shield tunnel is preprocessed to improve the accuracy of the point cloud data, thereby obtaining tunnel point cloud data, so as to facilitate the recognition of the annular seam and longitudinal seam of the shield tunnel through the tunnel point cloud data in subsequent steps.

[0016] S102, for each annular joint of the shield tunnel, the annular joint is identified according to the tunnel point cloud data, and a fitting plane and target point cloud data of the annular joint are obtained.

[0017] In this step, the shield tunnel is identified for annular seams, which refer to the connecting seams between different segment rings. The specific manifestation of annular seam identification is to use the tunnel point cloud data to identify each annular seam, and obtain the fitting plane of each annular seam and the target point cloud data of each annular seam. Among them, the fitting plane of the annular seam refers to the plane where the annular seam is located, and the fitting plane of each annular seam is used for longitudinal seam identification and tunnel point cloud data division. The target point cloud data of the annular seam refers to the point cloud data related to the height of the annular seam. The target point cloud data of each annular seam indicates the distribution of each annular seam in the shield tunnel, and the target point cloud data of each annular seam is used for segment segmentation.

[0018] S103, dividing the tunnel point cloud data into independent point cloud data of multiple segment rings of the shield tunnel according to the fitting planes of each annular joint.

[0019] In this step, since the longitudinal seam of the shield tunnel refers to the connection seam between the segments in a single segment ring, before realizing the longitudinal seam identification, it is necessary to first divide the tunnel point cloud data into independent point cloud data of multiple segment rings, so as to realize the longitudinal seam identification for a single segment ring in the subsequent steps and improve the accuracy of longitudinal seam identification. Specifically, the tunnel point cloud data is divided using the fitting planes of each annular seam obtained in the previous steps. The fitting plane of the annular seam can indicate the plane where the annular seam is located, so that the independent point cloud data of multiple segment rings can be accurately divided. Among them, the independent point cloud data of the segment ring refers to the point cloud data that is highly correlated with the segment ring.

[0020] S104, for the longitudinal seam of each segment ring, the longitudinal seam of the segment ring is identified according to the independent point cloud data of the segment ring to obtain the point cloud data of the longitudinal seam of the segment ring.

[0021] In this step, the longitudinal seam of the shield tunnel is identified. The longitudinal seam of the segment ring refers to the connection seam between the segments in the segment ring. Since the segment ring is composed of multiple segments, the segment ring has multiple longitudinal seams. The specific manifestation of longitudinal seam identification is to identify the longitudinal seam of each segment ring using the independent point cloud data of each segment ring to obtain the longitudinal seam point cloud data of each segment ring. Among them, the longitudinal seam point cloud data of the segment ring is the point cloud data highly related to the longitudinal seam of the segment ring. The longitudinal seam point cloud data of each segment ring indicates the distribution of the longitudinal seams of each segment ring in the shield tunnel. The longitudinal seam point cloud data of each segment ring is used for segment segmentation.

[0022] S105 , segmenting the shield tunnel according to the target point cloud data of each annular seam and the point cloud data of the longitudinal seam of each segment ring to obtain a plurality of segment areas of the shield tunnel.

[0023] In this step, the target point cloud data of each annular joint obtained in the previous step is used to indicate the distribution of each annular joint in the shield tunnel, and the longitudinal joint point cloud data of each segment ring obtained in the previous step is used to indicate the distribution of each longitudinal joint in the shield tunnel. According to the distribution of each annular joint in the shield tunnel and the distribution of each longitudinal joint in the shield tunnel, the multiple segments of the shield tunnel are divided to obtain multiple segment areas of the shield tunnel. The multiple segment areas of the shield tunnel can be understood as the segment distribution information of the shield tunnel, which is used to indicate the distribution of each segment in the shield tunnel.

[0024] It can be seen that the embodiment of the present application first performs annular joint identification and longitudinal joint identification on the shield tunnel based on the point cloud data of the shield tunnel, so as to fully explore the distribution and morphological characteristics of the annular joints and longitudinal joints in the shield tunnel, thereby obtaining point cloud data related to the annular joint height and point cloud data related to the longitudinal joint height, and then uses the point cloud data related to the annular joint height and the point cloud data related to the longitudinal joint height to realize the segment segmentation of the shield tunnel. In this way, it can not only effectively improve the processing efficiency when facing large-scale tunnel point cloud data and reduce the spatial complexity, but also effectively reduce the impact of feature loss on segment segmentation, thereby improving the segment segmentation accuracy of the shield tunnel.

[0025] In some embodiments, in the above step S101, the original point cloud data of the shield tunnel is obtained and preprocessed to obtain the tunnel point cloud data, which may include the following steps: Scan the interior of the shield tunnel through several preset measuring stations to obtain the measuring station cloud data of several measuring stations; The point cloud data of several measuring stations are spliced ​​and processed to obtain the original point cloud data of the shield tunnel; Perform denoising on the original point cloud data to obtain denoised original point cloud data; The denoised original point cloud data is converted from the Cartesian coordinate system to the three-dimensional cylindrical coordinate system to obtain the tunnel point cloud data.

[0026] In this embodiment, since the shield tunnel has the characteristics of narrow and long structure and unclear feature surface, and the three-dimensional laser scanner can only ensure high-quality point cloud data within a short distance, before scanning the shield tunnel, it is necessary to reasonably select a reasonable station spacing inside the shield tunnel and set a corresponding scanning mode. Specifically, multiple stations are pre-set along the axial direction of the shield tunnel, and the inside of the shield tunnel is scanned and processed by continuously setting up three-dimensional laser scanners at the stations to obtain the station cloud data of several stations. Optionally, the distance between two adjacent stations, the model of the three-dimensional laser scanner of each station, the scanner resolution of the three-dimensional laser scanner of each station, and the coordinate control points of each station can be flexibly set according to actual conditions. After obtaining the station cloud data of each station, the station cloud data of all stations are spliced ​​to accurately obtain the original point cloud data used to describe the internal structure of the shield tunnel. For example, the Leica Cyclone software supporting the three-dimensional laser scanner is used to splice the point cloud stations, so as to obtain the original point cloud data of the shield tunnel.

[0027] Since there are a large number of noisy point cloud data in the original point cloud data, these noisy point cloud data usually include point cloud data associated with pedestrians, point cloud data associated with tunnel construction equipment, and point cloud data associated with the backfill layer of the shield tunnel, etc. Therefore, in order to ensure the accuracy of the point cloud data, it is necessary to denoise the original point cloud data to remove the noisy point cloud data in the original point cloud data, so as to obtain the denoised original point cloud data, which can improve the accuracy of the original point cloud data and ensure the accuracy of subsequent annular seam identification and longitudinal seam identification. It can be understood that each point in the denoised original point cloud data is in the Cartesian coordinate system, and the coordinates of each point in the denoised original point cloud data are expressed as .

[0028] After the noise point cloud is removed, in order to facilitate the recognition of annular seams, each point in the denoised original point cloud data needs to be converted from the Cartesian coordinate system to the three-dimensional cylindrical coordinate system to obtain the tunnel point cloud data. This can improve the accuracy of the original point cloud data, thereby ensuring the accuracy of subsequent annular seam recognition and longitudinal seam recognition. Among them, each point in the tunnel point cloud data is located in the three-dimensional cylindrical coordinate system, and the coordinates of each point in the tunnel point cloud data are expressed as , ,like Figure 2 As shown, Indicates that any point in the tunnel point cloud data is XOY The projection point on the plane is relative to The rotation angle in the positive direction of the axis is referred to as the azimuth in the embodiment of the present application; Indicates that any point in the tunnel point cloud data is XOY Projection point on the plane and coordinate origin O The distance between any point in the tunnel point cloud data and The measured radius of the axis is referred to as the radial distance in the embodiment of the present application; The meaning of is the same as in the Cartesian coordinate system The meaning of is the same, and it is referred to as the axis distance in the embodiment of the present application. In addition, the Cartesian coordinate system and the three-dimensional cylindrical coordinate system satisfy the conversion relationship shown in the following formula (1): (1).

[0029] In some embodiments, in the above step S102, the process of identifying the annular seam according to the tunnel point cloud data and obtaining the fitting plane of the annular seam and the target point cloud data mainly includes the following steps S1021-S1024.

[0030] S1021, based on the axial direction, number of rings and segment width of the shield tunnel and a preset cross-sectional width, point cloud data associated with the annular joint is intercepted from the tunnel point cloud data as annular joint cross-sectional point cloud data.

[0031] In this step, first, along the three-dimensional cylindrical coordinate system The axial direction selects the approximate area where the annular joint is located, that is, the area with a preset cross-sectional width near the annular joint along the axial direction of the shield tunnel is intercepted as the annular joint identification area. It can be understood that the axial direction of the shield tunnel refers to the shield tunnel in the three-dimensional cylindrical coordinate system. axial direction; then, from the tunnel point cloud data, the point cloud data of the identification area of ​​the annular seam is selected as the point cloud data associated with the annular seam, that is, the annular seam cross-section point cloud data, so as to improve the accuracy and efficiency of subsequent annular seam identification. It should be noted that the point set of the annular seam cross-section point cloud data refers to the set of points in the annular seam cross-section point cloud data. Since the annular seam is the connecting seam between the segment rings, the point set mainly includes the annular seam points and their adjacent segment points, that is, the annular seam cross-section point cloud data includes multiple initial annular seam point cloud data and multiple segment point cloud data. The initial annular seam point cloud data refers to the point cloud data moderately related to the annular seam, and the segment point cloud data refers to the point cloud data highly related to the segment. Among them, the point set of the annular seam cross-section point cloud data satisfies the following formula (2): , (2); In formula (2), Indicates the segment width of the shield tunnel, which refers to the width of a single segment ring, which can be pre-calibrated according to actual conditions; Indicates the preset section width, which can be flexibly set according to actual conditions; Indicates the axis distance of each annular seam section point in the annular seam section point cloud data; Indicates the number of rings of a shield tunnel, which refers to the total number of segment rings.

[0032] S1022, removing the segment point cloud data in the annular seam cross-section point cloud data to obtain initial annular seam point cloud data.

[0033] In this step, since this embodiment aims to extract point cloud data related to the annular seam height from the annular seam cross-section point cloud data, the segment point cloud data in the annular seam cross-section point cloud data is useless point cloud data, and the useless point cloud data should be eliminated before the annular seam identification is performed, so as to improve the accuracy and efficiency of the subsequent annular seam identification. Specifically, by eliminating the segment point cloud data in the annular seam cross-section point cloud data, the initial annular seam point cloud data can be obtained.

[0034] S1023, performing feature point recognition processing on the annular seam according to the initial annular seam point cloud data to obtain an average width value of the annular seam and a plurality of feature points.

[0035] S1024, performing plane fitting processing on the annular seam according to the average width value and a plurality of feature points to obtain a fitting plane of the annular seam and target point cloud data.

[0036] In the above steps, according to the prior shield tunnel point cloud image, the initial annular point cloud data distributed in an annular shape along the cross section of the shield tunnel can be regarded as the intersection of a plane of a certain thickness and the tunnel point cloud data. Based on this idea, annular seam identification can be transformed into solving the plane parameters where the initial annular seam point cloud data is located. Among them, solving the plane parameters where the initial annular seam point cloud data is located is mainly divided into two parts: feature point extraction and plane fitting. Specifically, in the process of feature point extraction, the initial annular seam point cloud data obtained in the above steps is used to identify the feature points associated with the annular seam, thereby obtaining several feature points of the annular seam, and then the average width value of the annular seam is obtained using several feature points of the annular seam, wherein several feature points of the annular seam together constitute the morphological characteristics of the annular seam; and the average width value of the annular seam is used to indicate the width characteristics of the annular seam, which is mainly obtained by processing several feature points of the annular seam. In the process of plane fitting, the average width value and several characteristic points obtained in the above steps are used to perform plane fitting on the annular seam, aiming to determine the plane where the annular seam is located, that is, the fitting plane of the annular seam, and determine the point cloud data highly related to the annular seam, that is, the target point cloud data of the annular seam, so as to realize the annular seam recognition. In this way, by converting the annular seam recognition into solving the plane parameters where the initial annular seam point cloud data is located, the accuracy of the annular seam recognition can be effectively improved, and a better annular seam recognition effect can be achieved.

[0037] In some embodiments, in the annular seam cross-section point cloud data, at position Select Point cloud data within the range, and section width As the horizontal axis and with radial distance As the vertical axis, the first two-dimensional coordinate system is constructed, and then the point cloud data selected in the first two-dimensional coordinate system is plotted, such as Figure 3 As shown. Figure 3 It can be seen that the radial distance of the point cloud data at different positions increases significantly near the midpoint of the cross-section width and presents an irregular sharp angle of a certain width, while the points densely distributed outside the sharp angle have a small change in the ordinate, which is close to a straight line distribution. It can be seen that in the point cloud data at the same position, the point with a significant change in the ordinate is the initial annular seam point, while the point that is relatively smooth and close to a straight line distribution along the cross-section width is the segment point. Figure 3 The geometric relationship between the initial annular seam points and the segment points is shown. This embodiment provides a segment point cloud extraction method based on the Hough transform straight line detection method, in which the mapping law based on the Hough transform parameter space is applied to the straight line detection of the annular seam section point cloud data in the form of a discrete point cloud.

[0038] Specifically, in the above step S1022, the segment point cloud data in the annular seam cross-section point cloud data is eliminated, and the implementation process of obtaining the initial annular seam point cloud data mainly includes the following steps S01-S04.

[0039] S01, performing straight line detection processing on the annular seam section point cloud data using a straight line detection method of Hough transform, and obtaining a straight line associated with the segment point cloud data in the annular seam section point cloud data as a first straight line.

[0040] In this step, for ease of understanding, Figure 3 The horizontal axis of the first two-dimensional coordinate system (i.e., the cross-sectional width) is shown ) named axis, and its longitudinal axis (i.e. radial distance ) named Axis, then Figure 3 The first two-dimensional coordinate system shown can be named Coordinate system. For The point cloud data of the annular seam section in the coordinate system The coordinates of the annular seam section points are The countless straight lines can be expressed as cosine curves in Hough space , the cosine curve set mapped to the annular seam cross-section point cloud data in the Hough space is expressed as , is the Hough space parameter, is the total number of annular seam section points in the annular seam section point cloud data. It can be understood that the annular seam section point refers to the point of the annular seam section point cloud data. If the intersection point of the curve in the cosine curve set is Then the intersection coordinate set of the annular seam section point cloud data in the Hough space satisfies the following formula (3). The intersection coordinate set can be understood as the Hough space parameter set of the annular seam section point cloud data: , (3); In formula (3), Indicates The Hough space parameters of the intersection points include the axial angle parameters Distance from origin parameter Formula (3) shows that the number of straight lines detected based on Hough transform in the point cloud data of the annular seam section is However, the purpose of this step is to eliminate the segment point cloud data in the annular seam section point cloud data by straight line detection. Which of these straight lines is most relevant to the segment point cloud data still needs to be further determined, so certain constraints need to be added to determine the best straight line.

[0041] Based on this, this step proposes an improved Hough transform straight line detection method to determine the best straight line associated with the segment point cloud data, thereby improving the efficiency and accuracy of straight line detection. The implementation process may include the following steps S011-S013.

[0042] S011, determine the possible parameter values ​​of the segment point cloud data in Hough space , and select the value range of the segment point cloud data from the possible parameter values.

[0043] Specifically, firstly, the origin distance parameters of multiple annular seam section points in the annular seam section point cloud data are obtained as the distance set It can be understood that the annular seam section point refers to the point in the annular seam section point cloud data, and the origin distance parameter of the annular seam section point refers to the distance between the annular seam section point and the origin of the annular seam section point. The distance between the annular seam section point and the origin of the coordinate system. At the same time, the axial angle parameters of multiple annular seam section points in the annular seam section point cloud data are obtained as the angle set The axial angle parameter of the annular seam section point refers to The line segment connecting the annular seam section point and the origin of the coordinate system is relative to The angle between the positive direction of the axis. and angle set The initial Hough space parameter set and distance set can be obtained and angle set It can be expressed as the following formula (4): , (4); In formula (4), Indicated in Coordinate system The coordinates of the annular seam section points, express The origin of the coordinate system, For The minimum value of the ordinates of all annular seam section points in the coordinate system, For The minimum value of the horizontal coordinates of all annular seam section points in the coordinate system.

[0044] Then, according to the distribution characteristics of the segment point cloud data, the distance set The value range is reduced to , and obtain the distance set after the range is reduced as the distance set , and set the angle The value range is reduced to , and the angle set with a reduced range is obtained as the angle set , and then according to the distance set The value range and angle set of The value range of is selected from the point cloud data of the annular seam section as a two-dimensional point set, which can be expressed as , ,in Represents the first two-dimensional point set A two-dimensional point in Coordinates in the coordinate system; The total number of points representing the two-dimensional point set.

[0045] S012, according to the distance set and angle set , and obtain multiple Hough space parameter sets of uniform intervals. Specifically, the distance set and angle set Divided from small to large into , uniform intervals, that is, Hough space is divided into grid, thus obtaining The Hough space parameter set of uniform intervals, and All of them can be flexibly set according to actual conditions. The Hough space parameter set of each uniform interval includes multiple origin distance parameters and axial angle parameters within the uniform interval. Among them, the Hough space parameter set of each uniform interval can be expressed as the following formula (5): (5); In formula (5), , , , ; To fall into the angle set at the same time No. Interval and distance sets No. The parameter set of intervals, namely, the angle set No. Interval and distance sets No. The Hough space parameter set of uniform intervals composed of intervals; Angle set The minimum value in Angle set The maximum value in ; The distance set The minimum value in The distance set The maximum value in ; Angle set The normalized parameter of The distance set The normalized parameter of .

[0046] S013, due to the huge number of points in the discrete point cloud, it is difficult to select an optimal straight line parameter from the many parameters mapped to the Hough space. In this regard, according to the principle of Hough transform, we consider using voting to determine the optimal straight line parameter, that is, according to The Hough space parameter set of the uniform interval, the Hough space parameter set of the annular seam section point cloud data and a preset accumulator are used to obtain a straight line associated with the segment point cloud data in the annular seam section point cloud data as the first straight line.

[0047] Specifically, first, obtain a The two-dimensional matrix is ​​used as an accumulator ,accumulator The initial value of each element in is 0, and each element corresponds to the Hough space parameter set of each uniform interval. Then, the Hough space parameters of each section point in the Hough space parameter set of the annular seam section point cloud data shown in the above formula (3) are traversed, that is, If there is a Hough space parameter set of the cross-section point cloud data whose Hough space parameters belong to any uniform interval of the Hough space parameter set, then let the accumulator The value of the element corresponding to the Hough space parameter set of the uniform interval to which the Hough space parameter of the section point belongs is increased by 1. For example, if , then let the accumulator Elements in Increase by 1. The resulting accumulator Reflects The parameter distribution of the straight line. Obviously, the accumulator The Hough space parameter set corresponding to the position of the largest element in is the optimal straight line parameter. Afterwards, the accumulator The element with the largest value in is determined as the target element, and the Hough space parameter set corresponding to the target element is taken as the average value of the upper and lower bounds of the uniform interval in which it is located to obtain the target Hough space parameter set of the target element. Finally, the target Hough space parameter set of the target element is converted into the corresponding straight line parameters, and a straight line associated with the segment point cloud data in the annular seam section point cloud data is constructed as the first straight line according to the straight line parameters. For example, when the accumulator Medium Element When the value of is the largest, the element The corresponding Hough space parameter set takes the element The average value of the upper and lower bounds of the uniform interval, and the element The target Hough space parameter set , the target Hough space parameter set Include elements Target axial angle parameter and elements The target origin distance parameter , as shown in the following formula (6): (6).

[0048] Further simplifying the above formula (6), we can get the following formula (7): (7).

[0049] The element The target Hough space parameter set Convert to Line Parameters , as follows formula (8): , (8); In formula (8), Indicated in The slope of the straight line detected in the coordinate system, Indicated in According to the slope and intercept of the straight line shown in formula (8), we can construct The first straight line in the coordinate system. For ease of understanding, The first straight line in the coordinate system can be expressed as .

[0050] The specific implementation process of step S01 can effectively improve the efficiency and accuracy of line detection, and enhance the accuracy of the best line associated with the segment point cloud data.

[0051] S02, eliminating the segment point cloud data in the annular seam cross-section point cloud data according to the first straight line to obtain remaining cross-section point cloud data.

[0052] In this step, each 2D point in the 2D point set is traversed. If the distance from the 2D point to the first straight line is less than a preset distance threshold, where the distance threshold can be flexibly set according to the actual situation, it means that the 2D point is a pipe segment point, and the 2D point is removed at this time. Otherwise, it means that the 2D point is not a pipe segment point, and the 2D point is retained at this time. After traversing all the 2D points in the 2D point set, the remaining 2D points in the 2D point set are output as the remaining cross-section point cloud data, so that the shallow pipe segment point cloud can be removed. Among them, the remaining cross-section point cloud data can be expressed as , , represents the total number of remaining two-dimensional points in the two-dimensional point set. In addition, the distance from the two-dimensional point to the first straight line satisfies the following formula (9): (9); In formula (9), ; Indicates 2D points to the first line distance; Represents the first two-dimensional point set A two-dimensional point in Coordinates in the coordinate system.

[0053] S03, performing straight line detection processing on the remaining cross-section point cloud data using a straight line detection method of Hough transform, and obtaining a straight line associated with the segment point cloud data in the remaining cross-section point cloud data as a second straight line.

[0054] In this step, if the segment point clouds on both sides of the annular seam are detected as the same straight line, that is, the segment point clouds on both sides are basically not misaligned, then this process can completely separate the segment point cloud data from the initial annular seam point cloud data. However, in the actual annular seam cross-section point cloud data, in most cases, the point cloud is The distribution shape in the plane is as follows Figure 4 As shown, the segments on both sides of the annular seam are not on the same straight line. In view of this situation, this step uses the straight line detection method of Hough transform to perform a second straight line detection on the remaining cross-section point cloud data, so as to remove the segment point cloud again, thereby ensuring that all segment point cloud data are removed. Specifically, according to the above formula (3), the Hough space parameter set of the remaining cross-section point cloud data is obtained as the Hough space parameter set of the new annular seam cross-section point cloud data, and then the accumulator All elements in are set to 0, and step S013 is executed once to generate a new first straight line as the second straight line.

[0055] S04, eliminating the segment point cloud data in the remaining cross-section point cloud data according to the second straight line to obtain initial annular seam point cloud data.

[0056] In this step, after completing the second straight line detection, the remaining two-dimensional points in the remaining cross-sectional point cloud data are traversed again. If the distance from the remaining two-dimensional point to the second straight line is less than the preset distance threshold, it means that the remaining two-dimensional point is a pipe segment point, and the remaining two-dimensional point is removed at this time. Otherwise, it means that the remaining two-dimensional point is not a pipe segment point, and the remaining two-dimensional point is retained at this time. After traversing all the two-dimensional points in the remaining cross-sectional point cloud data, the remaining two-dimensional points in the remaining cross-sectional point cloud data are output as the initial annular seam point cloud data, so that the deep pipe segment point cloud can be removed, thereby improving the removal efficiency and removal accuracy of the pipe segment point cloud data. Among them, the initial annular seam point cloud data is represented as , , It is understood that the distance from the two-dimensional point to the second straight line can be calculated by the above formula (9), which will not be described in detail.

[0057] In some embodiments, in the above step S1023, the process of performing feature point recognition processing on the annular seam according to the initial annular seam point cloud data to obtain the average width value of the annular seam and a number of feature points mainly includes the following steps S11-S14: S11, obtaining a plurality of first characteristic points of the annular seam in a three-dimensional cylindrical coordinate system according to the maximum value of the radial distances of all initial annular seam points in the initial annular seam point cloud data; S12, obtaining a plurality of second characteristic points of the annular seam in a three-dimensional cylindrical coordinate system according to the median of the radial distances of all initial annular seam points in the initial annular seam point cloud data; S13, converting the plurality of first feature points and the plurality of second feature points from the three-dimensional cylindrical coordinate system into a three-dimensional rectangular coordinate system, and obtaining a plurality of converted first feature points and a plurality of converted second feature points as a plurality of feature points of the annular seam; S14, obtaining an average width value of the annular seam according to a plurality of characteristic points of the annular seam.

[0058] In this embodiment, refer to Figure 4 , the initial annular seam point cloud data and segment point cloud data are There are generally two forms in the coordinate system. The first form shows that in the area close to the 3D laser scanner, the laser beam can fully enter the annular gap width, and the radial distance of the initial annular gap point cloud data is significantly greater than the radial distance of the segment point cloud data. The second form shows that at a position far away from the 3D laser scanner, due to the influence of the scanning angle, the laser beam cannot enter the annular gap width, resulting in the radial distance of the initial annular gap point cloud data being between the segment point cloud data on both sides.

[0059] For the above two different forms of initial annular seam point cloud data, this embodiment adopts two different methods to determine the characteristic points and coordinates of the annular seam. The first form of initial annular seam point cloud data is often a sharp-angle point cloud with a certain width. Therefore, for the first form of initial annular seam point cloud data, each initial annular seam point in the first form of initial annular seam point cloud data is first used as the first characteristic point of the annular seam, and then the radial distance of all initial annular seam points in the initial annular seam point cloud data is calculated. The maximum value of the first feature point Axis coordinates are obtained to obtain multiple first characteristic points of the annular seam and their coordinates in the three-dimensional cylindrical coordinate system. The second form of initial annular seam point cloud data presents a gentle slope transition between the segment point cloud data on both sides. Therefore, for the second form of initial annular seam point cloud data, each initial annular seam point in the second form of initial annular seam point cloud data is first used as the second characteristic point of the annular seam, and then the radial distance of all initial annular seam points in the initial annular seam point cloud data is calculated. The median of the second feature point Axis coordinates, thereby obtaining multiple second feature points of the annular seam and their coordinates in the three-dimensional cylindrical coordinate system. In this way, while fully considering the morphological features of the initial annular seam point cloud data, multiple first feature points and multiple second feature points are accurately extracted from the initial annular seam point cloud data and the coordinates of the first feature points and the second feature points in the three-dimensional cylindrical coordinate system are determined, which can effectively improve the extraction accuracy of the feature points. Among them, the coordinate set of the first feature point and the coordinate set of the second feature point can be expressed as the following formula (10): , (10); In formula (10), A coordinate set representing a first feature point, which includes coordinates of a plurality of first feature points in a three-dimensional cylindrical coordinate system; Represents the radial distance of all initial annular gap points in the initial annular gap point cloud data The maximum value of Indicates The coordinates of the first feature point in the three-dimensional cylindrical coordinate system; The total number of points representing the initial annular seam point cloud data of the first form; A coordinate set representing a second feature point, which includes coordinates of a plurality of second feature points in a three-dimensional cylindrical coordinate system; Represents the radial distance of all initial annular gap points in the initial annular gap point cloud data the median; Indicates The coordinates of the second feature point in the three-dimensional cylindrical coordinate system; , The total number of points representing the initial annular seam point cloud data of the second form.

[0060] Then, in order to facilitate direct plane fitting in the point cloud data and identify the target point cloud data of the annular seam, the multiple first feature points and the multiple second feature points are converted from the three-dimensional cylindrical coordinate system to the three-dimensional rectangular coordinate system, so as to obtain multiple converted first feature points and multiple converted second feature points as several feature points of the annular seam, as shown in the following formula (11): (11); In formula (11), Represents the coordinates of the characteristic points of the annular seam in the three-dimensional rectangular coordinate system; Represents the coordinates of the first and second feature points of the annular seam in the three-dimensional cylindrical coordinate system.

[0061] Finally, according to several feature points and their coordinates, the average width of the annular seam can be obtained for subsequent plane fitting processing to ensure the accuracy of plane fitting. The average width of the annular seam is shown in the following formula (12): (12); In formula (12), Indicates the average width of the annular gap; Indicates the three-dimensional rectangular coordinate system feature points Axis coordinates, .

[0062] In some embodiments, in the above step S1024, the process of performing plane fitting processing on the annular seam according to the average width value and a number of feature points to obtain the fitting plane of the annular seam and the target point cloud data may include the following steps S21-S23.

[0063] S21, selecting a plurality of feature points at preset azimuth angles from a plurality of feature points as feature points to be fitted.

[0064] It should be noted that the feature points to be fitted refer to feature points at preset azimuths.

[0065] In this step, characteristic points at a preset azimuth angle are obtained as characteristic points to be fitted, and the characteristic points at the preset azimuth angle can fully reflect the distribution of the annular seams in the shield tunnel. It can be understood that the characteristic points to be fitted are the objects of plane fitting processing.

[0066] Optionally, the preset azimuth angle can be set according to actual conditions, and this embodiment does not specifically limit this. , but not limited to this.

[0067] S22, performing plane fitting processing on the feature points to be fitted using the M-estimation sample consistency method to obtain a fitting plane of the annular seam.

[0068] In this step, refer to Figure 5, a number of feature points are plane-fitted, and the intersection of the fitted plane and the tunnel point cloud data within a certain thickness is the complete annular seam point cloud, that is, the target point cloud data of the annular seam. In terms of the implementation of plane fitting, this step adopts the M-estimate Sample Consensus (MSAC) method for plane fitting, which can effectively improve the fitting accuracy and efficiency of the annular seam plane. MSAC is an improved method of the random sampling consistency method, which is specifically implemented in the following steps S221-S225.

[0069] S221, randomly select three feature points from multiple feature points to be fitted as the first The target feature points for wheel fitting.

[0070] Among them, the coordinate values ​​of the three target feature points in the three-dimensional rectangular coordinate system can be expressed as , and .

[0071] S222, assuming that the fitted target plane equation is , according to The coordinate values ​​of the target feature points of the round fitting in the three-dimensional rectangular coordinate system are obtained. The initial fitting plane and initial fitting plane parameters of the wheel fitting. The initial fitting plane parameters may include the first plane parameters , Second plane parameters , third plane parameters and the fourth plane parameters It can be understood that in a three-dimensional rectangular coordinate system, the first plane parameter Determines the plane Changes in the axis direction, second plane parameters Determines the plane Changes in the axis direction, third plane parameters Determines the plane Changes in the axis direction, fourth plane parameters Determine the position of the plane at the origin, then the initial fitting plane can be expressed as For ease of understanding, the initial fitting plane parameters are expressed as follows (13): (13).

[0072] S223, traverse each feature point to be fitted, if the feature point to be fitted reaches If the distance of the initial fitting plane of the first round of fitting is less than the preset fitting distance threshold, it means that the feature point to be fitted is an inner point. In this case, the feature point to be fitted is added to the first round of fitting. The preset internal point set of the first round of fitting is the default internal point set. Otherwise, it means that the feature point to be fitted is an outlier. In this case, the feature point to be fitted is added to the first round of fitting. The preset outlier point set for round fitting. Optionally, the fitting distance threshold can be flexibly set according to the actual situation. It can be understood that the inliers refer to the data points that can be correctly described by the model; the outliers refer to the data points that deviate significantly from the model.

[0073] S224, if No. If the number of elements in the preset internal point set of the round fitting is greater than three, The preset internal point set of the round of fitting is determined as the target internal point set, and the first The initial fitting plane parameters of the wheel fitting are determined as the target plane parameters, and then the process goes to step S225, otherwise the process goes directly to step S225.

[0074] S225, Judgment Is it equal to , is the total number of fittings; if not, let , and return to step S221 to loop fitting; if so, construct the fitting plane of the annular seam according to the target plane parameters Among them, the total number of fittings satisfies , Indicates passing The probability that the three feature points to be fitted are selected as internal points after the second selection is generally set to a value close to 1, such as 0.99; It represents the probability of an inlier point being selected in the entire data set, and its value can be flexibly set according to actual conditions.

[0075] Optionally, the cost function of the MSAC method is Satisfies the following formula (14): , , (14); In formula (14), is the loss function of the MSAC method, Indicates the distance threshold; represents the error function of the MSAC method.

[0076] S23, obtaining target point cloud data of the annular seam according to the average width value and the distance from each feature point to the fitting plane of the annular seam.

[0077] In this step, after determining the fitting plane of the annular seam and its target plane parameters, if the plane thickness is set to the average width value of the annular seam, the points within the plane thickness range are the points of the annular seam, where the plane thickness range refers to less than or equal to half of the average width value of the annular seam. Based on this, the target point cloud data of the annular seam can be extracted according to the conditions shown in the following formula (15), that is, retaining the feature points that satisfy the following formula (15) and filtering out the feature points that do not satisfy the following formula (15), thereby obtaining the target point cloud data of the annular seam, which can effectively improve the extraction accuracy and efficiency of the annular seam point cloud: (15); In formula (15), The fitting plane representing the annular gap; are the target plane parameters of the fitting plane of the annular seam respectively; Represents the coordinates of the characteristic points of the annular seam in the three-dimensional rectangular coordinate system.

[0078] In some embodiments, in the above step S103, the process of dividing the tunnel point cloud data into independent point cloud data of multiple segment rings of the shield tunnel according to the fitting planes of each annular joint may include the following steps: Traversing multiple annular seams of the shield tunnel, according to the fitting planes of two adjacent annular seams in the axial direction of the shield tunnel, the tunnel point cloud data located between the fitting planes of the two annular seams are obtained as the independent point cloud data of the segment ring, thereby obtaining the independent point cloud data of multiple segment rings.

[0079] In this embodiment, the longitudinal seam refers to the connection seam between each segment on the segment ring, which means that the point cloud data of the longitudinal seam is highly correlated with the point cloud data of the segment ring. Based on this, before the longitudinal seam identification, this embodiment divides the tunnel point cloud data into independent point cloud data based on the fitting planes of multiple annular seams, which can effectively improve the efficiency of extracting the longitudinal seam point cloud of each segment ring. Specifically, traverse multiple annular seams of the shield tunnel, for the shield tunnel in the three-dimensional cylindrical coordinate system, The two adjacent ring seams in the axial direction are determined, and the fitting planes of the two ring seams are determined, and the tunnel point cloud data located between the fitting planes of the two ring seams are determined as the independent point cloud data of a single segment ring, so that the independent point cloud data of multiple segment rings can be obtained. For ease of understanding, it is assumed that the shield tunnel in the three-dimensional cylindrical coordinate system The two adjacent annular gaps in the axial direction are the first annular gap and the second annular gap, respectively. The tunnel point cloud data between the fitting plane of the first annular gap and the fitting plane of the second annular gap satisfies the following formula (16): (16); In formula (16), The target plane parameters representing the fitting plane of the first annular seam; The target plane parameters representing the fitting plane of the second annular seam; Indicates the average width of the first annular gap; Indicates the average width of the second annular gap.

[0080] In some embodiments, in the above step S104, the longitudinal seam of the segment ring is identified according to the independent point cloud data of the segment ring, and the implementation process of obtaining the longitudinal seam point cloud data of the segment ring may include the following steps S1041-S1043.

[0081] S1041, based on the front end axis direction and the rear end axis direction of the segment ring and in combination with a preset cross-sectional width, point cloud data associated with the longitudinal seam of the segment ring is intercepted from the independent point cloud data as the longitudinal seam cross-sectional point cloud data of the segment ring.

[0082] In this step, similar to the distribution characteristics of the target point cloud data of the annular seam, the longitudinal seam point cloud data obtained by scanning is morphologically manifested as a protrusion that is obviously protruding from the segment, so the longitudinal seam point cloud data can be extracted based on the difference between the position of the longitudinal seam in the independent point cloud data and the radial distance between the surrounding segments. Based on this, after completing the division of the independent point cloud data, this step first uses the point cloud data of the preset interception width along the axis direction of the shield tunnel at the front and rear ends of the independent point cloud data (i.e., the front axis direction and the rear axis direction of the segment ring) as the point cloud data associated with the longitudinal seam of the segment ring, i.e., the longitudinal seam section point cloud data of the segment ring, so as to facilitate the subsequent longitudinal seam identification. It can be understood that the interception of the longitudinal seam section point cloud data is the same as the interception of the annular seam section point cloud data, and this step will not be repeated, but it should be noted that the interception width used for the two interceptions is the same.

[0083] S1042, performing longitudinal seam point cloud extraction processing on the longitudinal seam cross-section point cloud data according to a preset central angle feature set of the segment ring and a fitting plane of at least one annular seam, to obtain initial longitudinal seam point cloud data of the segment ring.

[0084] It should be noted that, refer to Figure 6 , the design radius of the observation tunnel in the embodiment of the present application is R Each segment ring is made of It is assembled from pipe segments. The pipe segment includes 1 standard block, 2 adjacent blocks and 1 capping block, Figure 6 The number of standard blocks in is 3. Block standard blocks are named , and name the two adjacent blocks as and , name the capping block Among them, the standard block The inner arc length is , the corresponding center angle is the standard center angle ; For adjacent blocks and , the inner arc length of the end with the larger arc length is the first adjacent arc length , the center angle corresponding to the first adjacent arc length is the first adjacent center angle , the inner arc length of the end with smaller arc length is the second adjacent arc length , the central angle corresponding to the second adjacent arc length is the second adjacent central angle ; For the capping block , the inner arc length of the end with smaller arc length is the first capping arc length , the central angle corresponding to the first capping arc length is the first capping central angle , the inner arc length of the end with the larger arc length is the second capping arc length , the central angle corresponding to the second capping arc length is the second capping central angle The preset center angle feature set can include standard center angles , the first adjacent central angle , the second adjacent central angle , the first capping center angle and the second capping center angle , these central angles satisfy: .

[0085] It can be understood that the preset center angle feature set is used to indicate the positional relationship between each segment and the center of the segment ring.

[0086] In this step, after intercepting the longitudinal seam cross-section point cloud data, the longitudinal seam cross-section point cloud data is preliminarily extracted using the preset central angle feature set of the segment ring and the fitting plane of at least one annular seam obtained in the previous step to obtain the initial longitudinal seam point cloud data of the segment ring.

[0087] S1043, screening and additional processing are performed on the initial longitudinal seam point cloud data according to a preset central angle feature set to obtain longitudinal seam point cloud data of the segment ring.

[0088] In this step, although most of the longitudinal seam point cloud data can be extracted through the previous step, some of the longitudinal seam point cloud data are still blocked by objects in the tunnel and affected by the scanning distance, so that they cannot be displayed in the tunnel point cloud data. In addition, there is a phenomenon that some pipe segment point cloud data are misjudged as longitudinal seam point cloud data. Therefore, after the initial longitudinal seam point cloud data is initially extracted, it is necessary to use the preset center angle feature set to process the initial longitudinal seam point cloud data, aiming to filter out the pipe segment point cloud data in the initial longitudinal seam point cloud data, and add some longitudinal seam point cloud data that have not been extracted in the previous steps, so as to effectively improve the extraction accuracy and extraction efficiency of the longitudinal seam point cloud data, thereby improving the accuracy and efficiency of longitudinal seam identification.

[0089] In some embodiments, in the above step S1042, the longitudinal seam point cloud extraction processing is performed on the longitudinal seam section point cloud data according to the preset central angle feature set of the segment ring and the fitting plane of at least one annular seam to obtain the initial longitudinal seam point cloud data of the segment ring, and the implementation process may include the following steps S31-S34: S31, projecting the longitudinal seam section point cloud data onto a fitting plane that is closest to the longitudinal seam section point cloud data to obtain projected point cloud data; S32, converting the projected point cloud data from the three-dimensional rectangular coordinate system into a three-dimensional cylindrical coordinate system to obtain converted projected point cloud data; S33, denoising the converted projected point cloud data to obtain denoised projected point cloud data; S34, extracting initial longitudinal seam point cloud data of the segment ring from the denoised projection point cloud data according to the preset central angle feature set and the denoised projection point cloud data.

[0090] In this embodiment, first, the longitudinal seam cross-section point cloud data is projected onto the fitting plane closest to the longitudinal seam cross-section point cloud data to obtain the projected point cloud data, so that all longitudinal seam cross-section points can be unified to the same data space position to improve the extraction accuracy of the initial longitudinal seam point cloud data. Assume that the fitting plane closest to the longitudinal seam cross-section point cloud data is , then the projected point cloud data obtained by projecting the longitudinal seam section point cloud data onto the fitting plane satisfies the following formula (17): (17); In formula (17), is the coordinate of the point projected into the three-dimensional plane, that is, the three-dimensional rectangular coordinate value of the projection point in the three-dimensional plane; is the coordinate of the longitudinal seam section point in the three-dimensional rectangular coordinate system, and the coordinate of the longitudinal seam section point can be obtained by converting the longitudinal seam section point cloud data from the three-dimensional cylindrical coordinate system to the three-dimensional rectangular coordinate system. It can be understood that the projection point refers to the point of the projection point cloud data, and the longitudinal seam section point refers to the point of the longitudinal seam section point cloud data.

[0091] Then, in order to conveniently express the relative position of the longitudinal seam and the segment ring, the projection point cloud data is converted from the 3D rectangular coordinate system to the 3D cylindrical coordinate system according to formula (1) to obtain the converted projection point cloud data, which is conducive to expressing the relative position of the longitudinal seam and the segment ring. Among them, the 3D cylindrical coordinate value of each point in the converted projection point cloud data is expressed as , azimuth and radial distance The relationship is as Figure 7 As shown. Figure 7 It can be seen that the radial distance As the azimuth changes, it changes slowly in most areas, but at certain specific angles, the radial distance A large sudden increase or decrease has occurred, so it can be inferred that the converted projected point cloud data includes the initial longitudinal joint point cloud data and the internal noise point cloud data, among which the points with sudden increase are the longitudinal joint points, and the points with sudden decrease are the internal noise points of the shield tunnel, such as wires, etc. It can be seen that as long as these sudden increase points are extracted from the converted projected point cloud data and their coordinate positions are obtained, the initial longitudinal joint point cloud data can be obtained.

[0092] Before extracting the initial longitudinal seam point cloud data, it is necessary to remove the internal noise points with similar shapes to the longitudinal seam points, which is conducive to removing the internal noise points with similar shapes to the longitudinal seam points and improving the accuracy and quality of the projected point cloud data. The principle is to transform the projected point cloud data into a three-dimensional cylindrical coordinate system. The axis is divided into M intervals, M =360, then the three-dimensional cylindrical coordinate value of each point in each interval in the transformed projected point cloud data satisfies the following formula (18): (18); In formula (18), Indicates that each point in the transformed projected point cloud data is The set of three-dimensional cylindrical coordinate values ​​within the interval. It can be understood that Axis coordinates do not satisfy The points that are internal noise points are considered as abnormal points and removed, while those that meet point.

[0093] After the denoising process is implemented, based on 3 The criterion further separates the initial longitudinal seam point cloud data. When the longitudinal seam section point cloud data is projected onto the fitting plane of the annular seam, the radial distance There is a certain gap. If the M value is set too large, the average interval is too small. If the M value is set too small, the average interval is too large. In either case, it may cause multiple longitudinal seam point clouds to be included in the same interval, which is not conducive to accurate point cloud extraction and processing. Therefore, different from the situation during denoising, when separating the initial longitudinal seam point cloud data, The M value in the criterion must depend on the preset central angle feature set, that is, the M value must satisfy: By setting the M value to 3 The criterion can extract the initial longitudinal seam point cloud data of the segment ring from the denoised projection point cloud data, thereby effectively improving the extraction accuracy of the initial longitudinal seam point cloud data.

[0094] In some embodiments, in the above step S1043, the initial longitudinal seam point cloud data is screened and additionally processed according to a preset central angle feature set to obtain the longitudinal seam point cloud data of the segment ring, which may include the following steps S41-S44.

[0095] S41, obtaining an endpoint set and an azimuth set of a first capping end of the segment ring, and obtaining an endpoint set and an azimuth set of a second capping end of the segment ring.

[0096] It should be noted that, refer to Figure 6 , the first capping end refers to the first capping arc length The first end of the seal is named t end, t The end points of the longitudinal seams are ; The second capping end refers to the second capping arc length The second end is named h end, h The end points of the longitudinal seams are . Longitudinal seams usually include straight seams and oblique seams, such as Figure 6 As shown, standard block Between or standard blocks With two adjacent blocks , The longitudinal seam between them is perpendicular to the normal section of the tunnel, that is, h、t Azimuth of the end longitudinal seam point in the three-dimensional cylindrical coordinate system equal, for the convenience of expression, it is called a straight seam; similarly, the two adjacent blocks , With capping block The longitudinal seam between them is called an oblique seam. In the three-dimensional cylindrical coordinate system, the difference in azimuth between the longitudinal seam points is essentially the azimuth Distance in direction.

[0097] It can be understood that the endpoint set includes a plurality of longitudinal suture point clusters. Further,h The endpoint set includes h End A cluster of longitudinal suture points, t The endpoint set of the end point includes t End A cluster of longitudinal suture-like points, . Set the azimuth As the horizontal axis, the radial distance As the vertical axis, a third two-dimensional coordinate system is constructed. The third two-dimensional coordinate system can be named Coordinate system. Each longitudinal seam linear point cluster includes multiple longitudinal seam endpoints. In the coordinate system, h The coordinates of the longitudinal seam endpoints in each longitudinal seam linear point cluster are expressed as , ,Will t The coordinates of the longitudinal seam endpoints in each longitudinal seam linear point cluster are expressed as , ,for h Duanhe t The median of the azimuth values ​​of all the longitudinal seam endpoints in each longitudinal seam linear point cluster is used as the azimuth of the longitudinal seam endpoint representing the longitudinal seam linear point cluster, and each longitudinal seam linear point cluster is sorted from small to large according to the azimuth of the longitudinal seam endpoint of each longitudinal seam linear point cluster, so as to obtain h The azimuth set of the end t It is understandable that h The azimuth set of the end includes h End The azimuth of the longitudinal seam endpoints of the longitudinal seam linear point clusters, t The azimuth set of the end includes t End The azimuth of the longitudinal seam endpoints of a longitudinal seam linear point cluster is shown in the following formula (19): (19); In formula (19), represents the median operator; express h The azimuth set of the end; express t The azimuth set of the end.

[0098] In this step, although most of the longitudinal seam point cloud data can be extracted through the previous step, some of the longitudinal seam point cloud data cannot be displayed in the tunnel point cloud data due to the obstruction of objects in the tunnel and the influence of the scanning distance. In addition, there is a phenomenon that some segment point cloud data is misjudged as longitudinal seam point cloud data. Therefore, after the initial longitudinal seam point cloud data is initially extracted, the initial longitudinal seam point cloud data needs to be processed secondary. In the secondary processing, first obtain the endpoint set and azimuth set of the first capping end of the segment ring, and obtain the endpoint set and azimuth set of the second capping end of the segment ring, so as to facilitate subsequent screening and additional processing.

[0099] S42, based on the endpoint set of the first sealing top end, the azimuth set of the first sealing top end, the endpoint set of the second sealing top end and the azimuth set of the second sealing top end, combined with the preset central angle feature set, obtain the azimuth angle of the longitudinal seam endpoint of the first sealing top end and the azimuth angle of the longitudinal seam endpoint of the second sealing top end as the longitudinal seam endpoint feature of the segment ring.

[0100] In this step, the azimuth set of the first sealing end point and the azimuth set of the second sealing end point may not include the azimuths of the longitudinal joint endpoints of the entire segment ring, and there may be misjudged longitudinal joint endpoints in these coordinates. Therefore, it is necessary to further verify and screen the endpoint set and azimuth set of the first sealing end point and the endpoint set and azimuth set of the second sealing end point, and calculate the azimuths of the longitudinal joint endpoints that are not displayed in the tunnel point cloud data.

[0101] Specifically, according to the distribution characteristics of the longitudinal seams in the ring, the capping block The azimuth of the oblique seam endpoint determines the position of the longitudinal seam endpoint in the entire ring. Therefore, when locating the longitudinal seam of the entire ring, the capping block should be determined first. Specifically, if the azimuth set of the first sealing top (i.e., the set shown in formula (19) t ) and the azimuth set of the second capping vertex (i.e., the set shown in formula (19) h ) respectively contain elements and elements All satisfy the orientation conditions shown in the following formula (20), where is the preset tolerance threshold, then the element and elements Determined as capping block The azimuth of the end points of the oblique seams on both sides, Store them in the collection in ascending order , and the elements Store them in the collection in ascending order : (20); In formula (20), means h The azimuth of the end is concentrated The azimuth of the longitudinal seam endpoints of the longitudinal seam linear point clusters, means t The azimuth of the end is concentrated The azimuth of the longitudinal seam end points of a longitudinal seam linear point cluster.

[0102] Then, for h End, if the collection The length is 2, that is, the capping block If both longitudinal seams on both sides are successfully identified, then it is assumed that the set The elements contained in and , then the initial longitudinal seam endpoint azimuth of the entire ring is identified, which satisfies the following formula (21): (twenty one); In formula (21), express h Minimum end azimuth, Indicates rounding down; express h Maximum end azimuth angle.

[0103] At the same time, for t End, if the collection The length is 2, that is, the capping block If both longitudinal seams on both sides are successfully identified, then it is assumed that the set The elements contained in and , then the initial longitudinal seam endpoint azimuth of the entire ring is identified, which satisfies the following formula (22): (twenty two); In formula (22), express t Minimum end azimuth, express t Maximum end azimuth angle.

[0104] Afterwards, the above identified h The initial longitudinal seam end azimuth of the end is stored in h The azimuth set of the end, and the above identified t The initial longitudinal seam end azimuth of the end is stored in t The azimuth set of the end, update h The azimuth set of the end t The elements in the azimuth set of the end are arranged from small to large. A certain tolerance is set. In this implementation, the tolerance is set to , by comparisonh The size of the adjacent elements in the azimuth set of the end points is used to delete the redundant and repeated azimuths of the longitudinal seam end points, and the h The intermediate azimuth set of the end At the same time, by comparing t The size of the adjacent elements in the azimuth set of the end points is used to delete the redundant and repeated azimuths of the longitudinal seam end points, and the t The intermediate azimuth set of the end . It should be noted that the principle of redundant coordinate deletion is that when two elements are approximately equal, the endpoint azimuth obtained by longitudinal seam identification is preferred to the result obtained by geometric reasoning, so as to retain the original information of the measurement as much as possible. It can be understood that if the difference between two elements is less than the tolerance, the two elements are judged to be approximately equal; in addition, the result obtained by geometric reasoning refers to the element in the original azimuth set; the endpoint azimuth obtained by longitudinal seam identification refers to the initial longitudinal seam endpoint azimuth.

[0105] Finally, after deleting redundant data, it is necessary to ensure that each longitudinal seam t End coordinate point and h The end coordinate points correspond one to one, so the above formula (20) is used again to h The intermediate azimuth set of the end t The intermediate azimuth set of the end is checked to verify t End and h The mapping relationship between the two ends. h The middle azimuth of the end is concentrated on each element. If the element cannot be obtained by the above formula (20) t If the corresponding element is found in the middle azimuth set of the end, it will be deleted, otherwise it will be retained, so we get h The target azimuth set at the end includes h The azimuth of the longitudinal seam end point at the end; t The middle azimuth of the end is concentrated on each element. If the element cannot be obtained by the above formula (20) h If the corresponding element is found in the middle azimuth set of the end, it will be deleted, otherwise it will be retained, so we get t The target azimuth set at the end includes t The azimuth of the longitudinal seam end point. t The target azimuth set and h The target azimuth sets at the ends of the target set establish a bijective relationship between the sets. , it can be calculated t The azimuth angle of the longitudinal seam end point and h The azimuth of the longitudinal seam endpoint at the end is used as the longitudinal seam endpoint feature of the segment ring, and the coordinates of the longitudinal seam that is not displayed in the point cloud can be calculated.

[0106] S43, according to the endpoint characteristics of the longitudinal seam, a straight line where the longitudinal seam of the segment ring is located in the second two-dimensional coordinate system is obtained as a third straight line.

[0107] In this step, the axis distance As the horizontal axis, with the azimuth As the vertical axis, construct a second two-dimensional coordinate system and name it Since the longitudinal seams are distributed linearly between the annular seams on both sides of the segment ring, the longitudinal seams are first determined The straight line equation in the coordinate system is used to extract the final longitudinal seam point cloud data. h The azimuth angle of the longitudinal seam end point is , the cylindrical coordinate value of the annular seam point cloud in the three-dimensional cylindrical coordinate system is , segment ring t The azimuth angle of the longitudinal seam end point is , the cylindrical coordinate value of the annular seam point cloud in the three-dimensional cylindrical coordinate system is ,but h The longitudinal seam ends at The coordinate values ​​and t The longitudinal seam ends at The coordinate values ​​in the coordinate system are shown in the following formula (23): (twenty three); In formula (23), express h The longitudinal seam ends at Coordinate values ​​in the coordinate system; express t The longitudinal seam ends at Coordinate values ​​in the coordinate system.

[0108] Then, according to h The longitudinal seam endpoints and t The longitudinal seam ends at The coordinate value in the coordinate system is obtained The straight line where the longitudinal seam is located in the coordinate system is taken as the third straight line .

[0109] S44, extracting the longitudinal seam point cloud data of the segment ring from the initial longitudinal seam point cloud data according to the average width value of the longitudinal seam of the segment ring and the distance from each initial longitudinal seam point in the initial longitudinal seam point cloud data to the third straight line.

[0110] In this step, it is assumed h The number of longitudinal suture point clusters at the end is ,and t The number of longitudinal suture point clusters at the end is ,exist In the coordinate system,h The coordinates of the longitudinal seam endpoints in each longitudinal seam linear point cluster are expressed as , ,Will t The coordinates of the longitudinal seam endpoints in each longitudinal seam linear point cluster are expressed as , , then according to h The number of longitudinal suture point clusters at the end, t The number of longitudinal suture point clusters at the end, Coordinate system h The coordinates of the longitudinal seam endpoints in each longitudinal seam linear point cluster in the end and Coordinate system t The coordinates of the longitudinal seam endpoints in each longitudinal seam linear point cluster are calculated to obtain the average width of the longitudinal seam, as shown in the following formula (24): (twenty four); In formula (24), Indicates the average width of the longitudinal seam; express h The maximum value of the horizontal coordinates of all longitudinal seam endpoints in each longitudinal seam linear point cluster; express h The minimum value of the horizontal coordinates of all longitudinal seam endpoints in each longitudinal seam linear point cluster; express t The maximum value of the horizontal coordinates of all longitudinal seam endpoints in each longitudinal seam linear point cluster; express t The minimum value of the horizontal coordinates of all longitudinal seam endpoints in each longitudinal seam linear point cluster.

[0111] After the average width of the longitudinal seam is determined, the point cloud whose linear distance from the third straight line in the initial longitudinal seam point cloud data is less than half of the average width of the longitudinal seam is regarded as the final longitudinal seam point cloud, that is, the longitudinal seam point cloud data of the segment ring. Specifically, first, for each initial longitudinal seam point in the initial longitudinal seam point cloud data, if the distance from the initial longitudinal seam point to the third straight line is less than half of the average width of the longitudinal seam, that is , If is the distance from the initial longitudinal seam point to the third straight line, it means that the initial longitudinal seam point is highly correlated with the longitudinal seam, and the initial longitudinal seam point is retained. Otherwise, it means that the initial longitudinal seam point is poorly correlated or uncorrelated with the longitudinal seam, and the initial longitudinal seam point is removed to obtain the longitudinal seam point cloud data of the segment ring. Among them, the distance from the initial longitudinal seam point to the third straight line satisfies the following formula (25): (25); In formula (25), represents the coordinates of the initial longitudinal seam point in the three-dimensional cylindrical coordinate system, Indicates the initial longitudinal seam point to the third straight line distance.

[0112] Finally, the longitudinal seam point cloud data of the segment ring is converted from the three-dimensional cylindrical coordinate system to the three-dimensional rectangular coordinate system, and the converted longitudinal seam point cloud data is output as the final longitudinal seam point cloud data. The final longitudinal seam point cloud data satisfies the following formula (26): (26); In formula (26), Represents the horizontal coordinate of the point of the final longitudinal seam point cloud data in the three-dimensional rectangular coordinate system; Represents the ordinate of the point of the final longitudinal seam point cloud data in a three-dimensional rectangular coordinate system; Represents the depth coordinate of the point of the final longitudinal seam point cloud data in a three-dimensional rectangular coordinate system.

[0113] Finally, the embodiment of the present application also provides a segment segmentation system for a shield tunnel, which includes: a point cloud scanning module, a ring seam identification module, a division processing module, a longitudinal seam identification module and a segment segmentation module. Among them, the point cloud scanning module is used to obtain the original point cloud data of the shield tunnel and perform preprocessing to obtain tunnel point cloud data; the ring seam identification module is used to identify the ring seams of each ring seam of the shield tunnel according to the tunnel point cloud data, and obtain the fitting plane and target point cloud data of the ring seam; the division processing module is used to divide the tunnel point cloud data into independent point cloud data of multiple segment rings of the shield tunnel according to the fitting plane of each ring seam; the longitudinal seam identification module is used to identify the longitudinal seams of each segment ring according to the independent point cloud data of the segment ring, and obtain the longitudinal seam point cloud data of the segment ring; the segment segmentation module is used to segment the shield tunnel according to the target point cloud data of each ring seam and the longitudinal seam point cloud data of each segment ring, and obtain multiple segment areas of the shield tunnel.

[0114] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0115] The following is an example to illustrate the principle of the shield tunnel segment segmentation method and system according to the embodiment of the present application.

[0116] In this example, the tunnel is lined with a single-layer flat segment with an outer diameter of 6.2 m, an inner diameter of 5.5 m, and a ring width of 1.5 m. Line fitting and deviation correction are controlled by a universal double-sided wedge with a wedge amount of 40 mm. The segments are assembled in a 1+5 block mode, that is, each ring segment contains 3 standard blocks, 2 adjacent blocks, and 1 capping block. The segments within the ring are connected with 12 bent bolts, and the rings are connected with 16 bent bolts. The ring width of the segment ring isD s =1.5 m, the distance between the two ends of the segment and the nearest bolt hole is 0.2 m. To prevent the bolt hole and the annular seam point cloud from being confused in the automatic recognition, the recognition area width is set to d r The 3D laser scanners at each station are all Leica P40. When the 3D laser scanning is performed, the tunnel shield excavation has been completed, and there are no dust, vibration and other interference factors in the tunnel. Considering the quality of the point cloud and the scanning efficiency, the scanner resolution of the 3D laser scanners at each station is set to 3.1 mm, the station spacing is 10-15 m, and three Leica 4.5'' black and white targets are set on the left and right sides and the bottom of the tunnel between the stations as coordinate control points.

[0117] The process of segment segmentation in this example is as follows: steps S1-S5.

[0118] S1, point cloud scanning and preprocessing: Scan the inside of the shield tunnel through several measuring stations to obtain the point cloud data of several measuring stations, and splice the point cloud data of all measuring stations into the original point cloud data of the shield tunnel. After the original point cloud data is exported, it is subjected to denoising and coordinate system conversion processing. The coordinate system conversion follows the above formula (1) to obtain the tunnel point cloud data.

[0119] S2, annular seam identification: For each annular seam of the shield tunnel, firstly, the point cloud data associated with the annular seam is intercepted in the tunnel point cloud data as the annular seam section point cloud data. , Then, the target point cloud data of the annular seam and the fitting plane are obtained through the above steps S1022-S1023. The target plane parameters of the fitting plane are shown in Table 1 below.

[0120] S3, segment point cloud segmentation: divide the tunnel point cloud data into independent point cloud data based on the segment ring according to the fitting planes of multiple annular joints, and obtain independent point cloud data of each segment ring.

[0121] S4, longitudinal seam identification: For each segment ring of the shield tunnel, if the designed arc length and radius of the segment are known, the central angle of each segment at both ends can be calculated. The designed arc length and the corresponding center angle of the segment are shown in Table 2. The possible distance between the end points of the straight seams at both ends of the segment is calculated to be ; h The possible distance between the end straight seam and the inclined seam is , the possible distance between the oblique slits is ; t The possible distance between the end straight seam and the inclined seam is , the possible distance between the oblique slits is . Substituting M The value is set to 11. Combined with the fitting plane of the annular seam and its target plane parameters, the theoretical distance between the endpoints of the longitudinal seam of the segment and the calculation parameters are substituted into steps S1041-S1043 to obtain the longitudinal seam point cloud data of the segment annular seam and the azimuth of the longitudinal seam endpoint. The azimuth of the longitudinal seam endpoint is shown in Table 3. NaN in Table 3 is the coordinate of the point that is not identified.

[0122] S5, segment segmentation: extract the annular joints and longitudinal joints in the tunnel point cloud data according to the target point cloud data of each annular joint, the target plane parameters of the fitting plane shown in Table 1, the longitudinal joint point cloud data of each segment ring and the azimuth of the longitudinal joint endpoint shown in Table 3. Combined with the preset software, the segmentation of a single segment in the tunnel point cloud data can be realized. The segmentation result is shown in FIG. Figure 8 As shown. Figure 8 It can be seen that the embodiment of the present application can correctly identify most of the annular seams and longitudinal seams in the point cloud, and only a few areas with a long scanning distance fail to identify or the identification results are inaccurate, such as only one longitudinal seam in the longitudinal seam of the first ring is successfully identified, and the longitudinal seam between the 20th and 21st rings is obviously distorted. This is mainly due to the long scanning distance, and fewer laser beams enter the annular seam, which reduces the number of points in the annular seam point cloud accordingly, affecting the accuracy of its feature point coordinate extraction. If the distance between the measuring stations can be shortened during the scanning process, the accuracy of segment segmentation can be effectively improved.

[0123] Table 1: Example data for annular gap identification

[0124] Table 2: Example of the designed arc length and central angle of the segment

[0125] Table 3: Example table of longitudinal seam end point azimuth angles

[0126] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present application, and that the scope of the present application is defined by the claims and their equivalents.

[0127] The above is a specific description of the preferred implementation of the present application, but the present application is not limited to the described embodiments. Technical personnel familiar with the field may make various equivalent modifications or substitutions without violating the spirit of the present application. These equivalent modifications or substitutions are all included in the scope defined by the claims of the present application.

Claims

1. A method for segmenting a shield tunnel, characterized in that: The following steps are involved: Obtain the original point cloud data of the shield tunnel and perform preprocessing to obtain the tunnel point cloud data; For each annular seam of the shield tunnel, the annular seam is identified according to the tunnel point cloud data to obtain a fitting plane and target point cloud data of the annular seam; Dividing the tunnel point cloud data into independent point cloud data of a plurality of segment rings of the shield tunnel according to the fitting planes of the annular joints; For each of the longitudinal seams of the segment rings, the longitudinal seams of the segment rings are identified according to the independent point cloud data of the segment rings to obtain the point cloud data of the longitudinal seams of the segment rings; The shield tunnel is segmented according to the target point cloud data of each annular seam and the point cloud data of the longitudinal seam of each segment ring to obtain a plurality of segment areas of the shield tunnel.

2. The shield tunnel segment segment segmentation method according to claim 1, characterized in that: The method of obtaining the original point cloud data of the shield tunnel and preprocessing the data to obtain the tunnel point cloud data includes: Scanning the interior of the shield tunnel through a plurality of preset measuring stations to obtain measuring station cloud data of the plurality of measuring stations; Splicing the point cloud data of several measuring stations to obtain the original point cloud data of the shield tunnel; Performing denoising processing on the original point cloud data to obtain the denoised original point cloud data; The denoised original point cloud data is converted from a Cartesian coordinate system into a three-dimensional cylindrical coordinate system to obtain the tunnel point cloud data.

3. The shield tunnel segment segment segmentation method according to claim 1, characterized in that: The step of identifying the annular seam according to the tunnel point cloud data to obtain the fitting plane and target point cloud data of the annular seam includes: According to the axial direction, the number of rings and the segment width of the shield tunnel, combined with the preset cross-sectional width, the point cloud data associated with the annular seam in the tunnel point cloud data is intercepted as the annular seam cross-sectional point cloud data; Eliminating the segment point cloud data in the annular seam cross-section point cloud data to obtain initial annular seam point cloud data; Performing feature point recognition processing on the annular seam according to the initial annular seam point cloud data to obtain an average width value and a plurality of feature points of the annular seam; A plane fitting process is performed on the annular seam according to the average width value and a plurality of the characteristic points to obtain a fitting plane and target point cloud data of the annular seam.

4. The shield tunnel segment segment segmentation method according to claim 3, characterized in that: The step of removing the segment point cloud data in the annular seam cross-section point cloud data to obtain initial annular seam point cloud data includes: Performing straight line detection processing on the annular seam cross-section point cloud data using a straight line detection method of Hough transform, and obtaining a straight line associated with the segment point cloud data in the annular seam cross-section point cloud data as a first straight line; Eliminating the segment point cloud data in the annular seam cross-section point cloud data according to the first straight line to obtain remaining cross-section point cloud data; Performing straight line detection processing on the remaining cross-section point cloud data using a straight line detection method of Hough transform, and obtaining a straight line associated with the segment point cloud data in the remaining cross-section point cloud data as a second straight line; The segment point cloud data in the remaining cross-section point cloud data are eliminated according to the second straight line to obtain the initial annular seam point cloud data.

5. The shield tunnel segment segment segmentation method according to claim 3, characterized in that: The feature point recognition processing is performed on the annular seam according to the initial annular seam point cloud data to obtain the average width value of the annular seam and a plurality of feature points, including: Obtaining a plurality of first characteristic points of the annular seam in a three-dimensional cylindrical coordinate system according to the maximum value of the radial distances of all initial annular seam points in the initial annular seam point cloud data; Obtaining a plurality of second characteristic points of the annular seam in a three-dimensional cylindrical coordinate system according to the median of radial distances of all initial annular seam points in the initial annular seam point cloud data; Converting the first characteristic points and the second characteristic points from a three-dimensional cylindrical coordinate system to a three-dimensional rectangular coordinate system to obtain a plurality of converted first characteristic points and a plurality of converted second characteristic points as characteristic points of the annular seam; According to a plurality of characteristic points of the annular gap, an average width value of the annular gap is obtained.

6. The shield tunnel segment ... The process of performing plane fitting processing on the annular seam according to the average width value and the plurality of feature points to obtain the fitting plane and target point cloud data of the annular seam includes: Selecting a plurality of the feature points at preset azimuth angles from the plurality of the feature points as feature points to be fitted; Performing plane fitting processing on the feature points to be fitted using the M-estimation sample consistency method to obtain a fitting plane of the annular seam; The target point cloud data of the annular seam is obtained according to the average width value and the distance from each of the feature points to the fitting plane of the annular seam.

7. The shield tunnel segment ... The step of identifying the longitudinal seam of the segment ring according to the independent point cloud data of the segment ring to obtain the longitudinal seam point cloud data of the segment ring includes: According to the front end axis direction and the rear end axis direction of the segment ring, combined with the preset cross-sectional width, the point cloud data associated with the longitudinal seam of the segment ring is intercepted from the independent point cloud data as the longitudinal seam cross-sectional point cloud data of the segment ring; Performing longitudinal seam point cloud extraction processing on the longitudinal seam cross-section point cloud data according to the preset central angle feature set of the segment ring and at least one fitting plane of the annular seam to obtain initial longitudinal seam point cloud data of the segment ring; The initial longitudinal seam point cloud data is screened and additionally processed according to the preset central angle feature set to obtain the longitudinal seam point cloud data of the segment ring.

8. The shield tunnel segment segment segmentation method according to claim 7, characterized in that: The step of performing longitudinal seam point cloud extraction processing on the longitudinal seam cross-section point cloud data according to the preset central angle feature set of the segment ring and at least one fitting plane of the annular seam to obtain initial longitudinal seam point cloud data of the segment ring includes: Projecting the longitudinal seam cross-section point cloud data to the fitting plane that is closest to the longitudinal seam cross-section point cloud data to obtain projected point cloud data; Converting the projected point cloud data from a three-dimensional rectangular coordinate system to a three-dimensional cylindrical coordinate system to obtain the converted projected point cloud data; Obtaining denoised projection point cloud data from the converted projection point cloud data; According to the preset central angle feature set and the denoised projected point cloud data, initial longitudinal seam point cloud data of the segment ring is extracted from the denoised projected point cloud data.

9. The shield tunnel segment ... The step of performing point cloud screening and additional processing on the initial longitudinal seam point cloud data according to the preset central angle feature set to obtain the longitudinal seam point cloud data of the segment ring includes: Obtaining an endpoint set and an azimuth set of a first capping end of the segment ring, and obtaining an endpoint set and an azimuth set of a second capping end of the segment ring; wherein the first capping end is an end where a first capping arc length is located, and the first capping arc length is an inner arc of an end with a smaller arc length in a capping block of the segment ring, and the second capping end is an end where a second capping arc length is located, and the second capping arc length is an inner arc of an end with a larger arc length in a capping block, and the endpoint set includes a plurality of longitudinal seam linear point clusters, each of the longitudinal seam linear point clusters includes a plurality of longitudinal seam endpoints, and the azimuth set includes longitudinal seam endpoint azimuths of a plurality of longitudinal seam linear point clusters; According to the endpoint set of the first sealing top end, the azimuth set of the first sealing top end, the endpoint set of the second sealing top end and the azimuth set of the second sealing top end, in combination with the preset central angle feature set, the azimuth angle of the longitudinal seam endpoint of the first sealing top end and the azimuth angle of the longitudinal seam endpoint of the second sealing top end are obtained as the longitudinal seam endpoint features of the segment ring; According to the endpoint features of the longitudinal seam, a straight line where the longitudinal seam of the segment ring is located in a second two-dimensional coordinate system is obtained as a third straight line; wherein the horizontal axis of the second two-dimensional coordinate system is the axis distance and the vertical axis is the azimuth; The longitudinal seam point cloud data of the segment ring are extracted from the initial longitudinal seam point cloud data according to the average width value of the longitudinal seam of the segment ring and the distance from each initial longitudinal seam point in the initial longitudinal seam point cloud data to the third straight line.

10. A segment segmentation system for a shield tunnel, characterized in that: include: Point cloud scanning module, used to obtain the original point cloud data of the shield tunnel and perform preprocessing to obtain tunnel point cloud data; A ring seam identification module, for identifying each ring seam of the shield tunnel according to the tunnel point cloud data, and obtaining a fitting plane and target point cloud data of the ring seam; A division processing module, used for dividing the tunnel point cloud data into independent point cloud data of a plurality of segment rings of the shield tunnel according to the fitting planes of the annular joints; A longitudinal seam identification module is used to identify the longitudinal seam of each segment ring according to the independent point cloud data of the segment ring, so as to obtain the longitudinal seam point cloud data of the segment ring; The segment segmentation module is used to segment the shield tunnel according to the target point cloud data of each annular seam and the longitudinal seam point cloud data of each segment ring to obtain multiple segment areas of the shield tunnel.

Images