A verticality detection system and method for reverse construction process

Abstract

This application provides a verticality detection system and method for reverse construction. The system involves: determining a reference grid for visual inspection during reverse construction; acquiring a sequence of visual inspection images of the target foundation pit support structure based on the reference grid; extracting candidate straight lines from the visual inspection image sequence to identify the edges of the target foundation pit support structure; extracting edge straight lines that conform to the vertical characteristics of the target foundation pit support structure from all candidate straight lines; extracting the structural centerline of the target foundation pit support structure based on the distribution pattern of the edge straight lines; converting the centerline into actual spatial verticality deviation based on the structural centerline and the construction floor height calibration data of the target foundation pit support structure; and extracting the confidence verticality of the target foundation pit support structure by combining the centerline change characteristics of adjacent construction stages with the actual spatial verticality deviation. The technical solution provided by this application can identify the cumulative deviation of the verticality of the foundation pit support structure during reverse construction.

Description

Technical Field

[0001] This application relates to the field of verticality detection technology, and more specifically, to a verticality detection system and method for reverse construction processes. Background Technology

[0002] Perpendicularity is a key item in geometric tolerances, used to measure the degree of deviation of the measured element (such as surface, axis) from the ideal vertical direction (usually the direction of gravity or a specified datum). In precision manufacturing (such as machining, assembly), construction engineering (such as column and wall construction), aerospace and other fields, the vertical accuracy requirements of parts or structures are extremely high, which directly affects the performance, life and safety of products. With the advancement of technology, high-precision electronic tilt sensors, laser trackers and coordinate measuring machines (CMMs) have significantly improved the automation and accuracy of inspection to meet increasingly stringent industrial standards and quality control requirements.

[0003] In existing verticality inspection methods, verticality is determined by measuring the maximum deviation between the target element and the reference element to determine whether it meets the tolerance requirements. During inspection, a reference (such as a plane or straight line) must first be determined, and then an ideal straight line or plane perpendicular to the reference is established using inspection tools. For example, with the help of instruments such as a total station or laser line projector, a laser line perpendicular to the reference is projected, and the deviation value between the measured point and the laser line is measured to assess whether the verticality meets the standard. However, in the visual inspection of verticality during the reverse construction process, the foundation pit support structure needs to be constructed layer by layer downwards. During each layer of excavation and support, factors such as soil stress release, changes in construction load, and equipment vibration will cause slight deviations in the verticality of the support structure under visual inspection. These deviations will accumulate during the layer-by-layer construction. If they are not corrected in time, they will lead to the overall instability of the support structure. Therefore, how to identify the cumulative deviation of the verticality of the foundation pit support structure during the reverse construction process has become a difficult problem for the industry. Summary of the Invention

[0004] This application provides a verticality detection system and method for reverse construction, which can identify the cumulative deviation of the verticality of the foundation pit support structure during reverse construction.

[0005] In a first aspect, this application provides a method for detecting the verticality of a reverse construction process, comprising the following steps:

[0006] Multiple target points are symmetrically arranged at the top and bottom of the target foundation pit support structure during reverse construction to form a reference grid for visual inspection.

[0007] By using an optical camera outside the construction area to acquire images of each target point in the reference grid, a visual inspection image sequence of the target foundation pit support structure is obtained.

[0008] Based on the directional consistency constraint, candidate straight lines of the target foundation pit support structure edge are extracted from the visual detection image sequence, and edge straight lines that conform to the vertical characteristics of the target foundation pit support structure are extracted from all candidate straight lines through preset tilt angle threshold and continuous length threshold.

[0009] The distribution pattern of the edge straight lines is determined, and the structural centerline of the target foundation pit support structure is extracted based on the distribution pattern. Then, the actual spatial verticality deviation of the structural centerline in the pixel coordinate system is determined based on the structural centerline and the construction layer height calibration data of the target foundation pit support structure.

[0010] The confidence verticality of the target foundation pit support structure under visual inspection is obtained by combining the changes in the centerline of the target foundation pit support structure during adjacent construction stages with the actual spatial verticality deviation.

[0011] In some embodiments, the visual inspection image sequence of the target foundation pit support structure is obtained by acquiring images of each target point in the reference grid using an optical camera outside the construction area, specifically including:

[0012] Multiple optical cameras are deployed around the construction area based on the reference grid to form a multi-view image acquisition network;

[0013] The parameters of the optical cameras in the multi-view image acquisition network are calibrated.

[0014] Based on the parameter calibration results of the optical camera, continuous shooting of the target points in the reference grid is triggered, thereby obtaining a visual inspection image sequence of the target foundation pit support structure.

[0015] In some embodiments, extracting candidate straight lines from the edge of the target foundation pit support structure from the visual detection image sequence based on directional consistency constraints specifically includes:

[0016] Smooth each frame of the visual detection image sequence to generate a smoothed image sequence;

[0017] An edge detection operator is used to perform edge detection on the smoothed images in the smoothed image sequence to obtain an edge pixel set;

[0018] All potential structural line segments are extracted from the set of edge pixels based on the Hough transform;

[0019] Establish a histogram of orientation angle distribution based on the orientation angles of each potential structural line segment;

[0020] Based on the histogram of the direction angle distribution, the main direction interval of potential structural line segments under the direction consistency constraint is determined, and potential structural line segments with direction angles located within the main direction interval are selected according to the main direction interval to form a set of direction consistency line segments;

[0021] The potential structural segments in the set of directional consistent line segments are optimized to obtain candidate straight lines for the edge of the target foundation pit support structure.

[0022] In some embodiments, extracting edge lines that conform to the vertical characteristics of the target foundation pit support structure from all candidate lines by using preset inclination angle thresholds and continuous length thresholds specifically includes:

[0023] Calculate the angle between each candidate line and the vertical direction to obtain the inclination angle value of each candidate line;

[0024] The inclination angle value of each candidate line is compared with the preset inclination angle threshold, and candidate lines with inclination angle values ​​less than or equal to the inclination angle threshold are selected to form a subset of low inclination angle lines.

[0025] Calculate the continuous pixel length of each low-angle straight line in the low-angle straight line subset to obtain the continuous length value of each low-angle straight line;

[0026] The continuous length value of each low-angle straight line is compared with the preset continuous length threshold, and the low-angle straight lines with continuous length values ​​greater than or equal to the continuous length threshold are selected to obtain the edge straight lines that conform to the vertical characteristics of the target foundation pit support structure.

[0027] In some embodiments, determining the distribution pattern of the edge lines specifically includes:

[0028] Extract the endpoint coordinates of the edge lines and construct a spatial coordinate set for the edge lines;

[0029] Clustering algorithms are used to perform cluster analysis on the endpoints of the spatial coordinate set to obtain multiple clusters;

[0030] Calculate the direction and position parameters of the edge lines in each cluster to generate a set of cluster feature parameters;

[0031] The distribution pattern of edge lines is identified based on the cluster feature parameter set.

[0032] In some embodiments, extracting the structural centerline of the target foundation pit support structure based on the distribution pattern specifically includes:

[0033] Obtain the distribution pattern of edge lines, wherein the distribution pattern is divided into parallel distribution and intersecting distribution;

[0034] If the distribution pattern is a parallel distribution, then the symmetrical centerline of the edge line is calculated as the structural centerline of the target foundation pit support structure;

[0035] If the distribution pattern is an intersecting distribution, then the angle bisector corresponding to the edge line is calculated as the structural centerline of the target foundation pit support structure.

[0036] In some embodiments, the reference grid is a spatial reference coordinate system formed by target points symmetrically arranged at the top and bottom of the foundation pit support structure.

[0037] Secondly, this application provides a verticality detection system for reverse construction processes, used to perform a verticality detection method during reverse construction processes. The system includes:

[0038] The layout module is used to symmetrically arrange multiple target points on the top and bottom of the target foundation pit support structure during reverse construction, forming a reference grid for visual inspection.

[0039] The processing module is used to acquire images of each target point in the reference grid through an optical camera outside the construction area, and obtain a visual inspection image sequence of the target foundation pit support structure.

[0040] The processing module is also used to extract candidate straight lines of the edge of the target foundation pit support structure from the visual detection image sequence based on the direction consistency constraint, and to extract edge straight lines that conform to the vertical characteristics of the target foundation pit support structure from all candidate straight lines through preset tilt angle threshold and continuous length threshold.

[0041] The processing module is also used to determine the distribution pattern of the edge straight lines, extract the structural centerline of the target foundation pit support structure based on the distribution pattern, and then determine the actual spatial verticality deviation of the structural centerline in the pixel coordinate system based on the structural centerline and the construction layer height calibration data of the target foundation pit support structure.

[0042] The execution module is used to extract the confidence verticality of the target foundation pit support structure under visual inspection by combining the changes in the centerline of the target foundation pit support structure during adjacent construction stages with the actual spatial verticality deviation.

[0043] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the verticality detection method of the above-described reverse construction process.

[0044] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for detecting the verticality of the reverse construction process.

[0045] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0046] The verticality detection system and method for reverse construction provided in this application firstly involves symmetrically arranging multiple target points at the top and bottom of the target foundation pit support structure during reverse construction to form a reference grid for visual inspection. Secondly, an optical camera located outside the construction area acquires images of each target point in the reference grid to obtain a visual inspection image sequence of the target foundation pit support structure. Further, candidate straight lines representing the edges of the target foundation pit support structure are extracted from the visual inspection image sequence based on directional consistency constraints. Then, edge straight lines conforming to the vertical characteristics of the target foundation pit support structure are extracted from all candidate straight lines using preset tilt angle and continuous length thresholds. Next, the distribution pattern of the edge straight lines is determined, and the structural centerline of the target foundation pit support structure is extracted based on the distribution pattern. Then, the actual spatial verticality deviation corresponding to the structural centerline in the pixel coordinate system is determined based on the structural centerline and the construction floor height calibration data of the target foundation pit support structure. Finally, the confidence verticality of the target foundation pit support structure under visual inspection is extracted by combining the changes in the centerline between adjacent construction stages in the target foundation pit support structure with the actual spatial verticality deviation.

[0047] Therefore, this application can identify the cumulative deviation of the verticality of the foundation pit support structure during reverse construction. First, multiple target points are symmetrically arranged at the top and bottom of the target foundation pit support structure to form a reference grid, providing a stable spatial reference for visual inspection and reducing errors in subsequent image acquisition and feature extraction. The reference grid is a reference system for positioning composed of symmetrically distributed target points. Second, a visual inspection image sequence of the reference grid is acquired by an optical camera, realizing non-contact dynamic monitoring of the target structure and completely recording the morphological changes of the structure during construction. Furthermore, candidate straight lines are extracted and edge straight lines are screened based on directional consistency constraints, effectively filtering out interference information from non-target structures and accurately locking edges that conform to vertical features. The method improves the targeting of feature extraction by identifying the edge straight line distribution pattern and extracting the structural centerline. This is then combined with construction floor height calibration data to convert the deviation into actual spatial verticality, achieving the transformation from image features to actual spatial parameters and providing a core basis for verticality measurement. Finally, the actual spatial verticality deviation is corrected by the centerline change characteristics of adjacent stages to obtain the confidence verticality. This integrates information on dynamic structural changes, reduces the systematic error of single-stage detection, and makes the results closer to the real state. It effectively avoids the slight deviation in the verticality of the foundation pit support structure under visual inspection, providing a reliable basis for construction quality assessment. In summary, the technical solution provided in this application can identify the cumulative deviation of the verticality of the foundation pit support structure during reverse construction. Attached Figure Description

[0048] Figure 1This is an exemplary flowchart of a verticality detection method for a reverse construction process according to some embodiments of this application;

[0049] Figure 2 This is an exemplary flowchart illustrating the determination of a visually inspected image sequence according to some embodiments of this application;

[0050] Figure 3 This is an exemplary flowchart illustrating the determination of edge lines according to some embodiments of this application;

[0051] Figure 4 This is a structural schematic diagram of a verticality detection system for a reverse construction process according to some embodiments of this application;

[0052] Figure 5 This is a schematic diagram of the structure of a computer device for implementing a verticality detection method for a reverse construction process, according to some embodiments of this application. Detailed Implementation

[0053] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0054] refer to Figure 1 This figure is an exemplary flowchart of a verticality detection method for a reverse construction process according to some embodiments of this application. The figure mainly includes the following steps:

[0055] In step S101, multiple target points are symmetrically arranged at the top and bottom of the target foundation pit support structure during reverse construction to form a reference grid for visual inspection.

[0056] In practice, during the reverse construction, multiple target points are symmetrically arranged on the design axis of the top and bottom of the target pit support structure (where the target points are set up to satisfy ≥4 groups per layer, with a spacing of 5 to 8 meters) to form a reference grid for visual inspection.

[0057] It should be noted that, in this application, the foundation pit support structure refers to the retaining system constructed during the foundation pit excavation process to maintain the stability of the pit wall and prevent soil collapse and groundwater infiltration, such as concrete columns; the reference grid used for visual inspection in this application is a spatial reference coordinate system composed of target points symmetrically arranged at the top and bottom of the foundation pit support structure. By determining the reference grid, spatial scale benchmarks and structural deformation references can be effectively provided for the visual inspection algorithm.

[0058] In step S102, an optical camera outside the construction area is used to acquire images of each target point in the reference grid to obtain a visual inspection image sequence of the target foundation pit support structure.

[0059] In some embodiments, reference Figure 2 As shown in the figure, this is an exemplary flowchart of determining a visual inspection image sequence according to some embodiments of this application. In this embodiment, the visual inspection image sequence of the target foundation pit support structure is obtained by acquiring images of each target point in the reference grid through an optical camera on the periphery of the construction area using the following steps:

[0060] In step S1021, multiple optical cameras are deployed around the construction area according to the reference grid to form a multi-view image acquisition network;

[0061] In step S1022, the parameters of the optical cameras in the multi-view image acquisition network are calibrated.

[0062] In step S1023, based on the parameter calibration results of the optical camera, continuous shooting of the target points in the reference grid is triggered, thereby obtaining a visual inspection image sequence of the target foundation pit support structure.

[0063] In specific implementation, firstly, multiple optical cameras are deployed around the construction area to ensure that all target points in the reference grid can be captured by at least one optical camera, forming a multi-view image acquisition network. This multi-view image acquisition network refers to a collaborative image acquisition system constructed by deploying multiple optical cameras around the construction area according to the reference grid, used to ensure that target points can be captured from multiple perspectives. Secondly, the optical cameras in the multi-view image acquisition network are calibrated, specifically: the intrinsic parameter matrix (including focal length, principal point coordinates, and distortion coefficients) and extrinsic parameter matrix (including rotation matrix and translation vector) of the optical cameras in the multi-view image acquisition network are calculated using the Zhang Zhengyou calibration method, thereby completing the calibration of the target points. The parameters of the optical cameras in the multi-view image acquisition network are calibrated. The parameters include an intrinsic parameter matrix and an extrinsic parameter matrix. The intrinsic parameter matrix describes the internal optical characteristics of the camera, and the extrinsic parameter matrix describes the position and orientation of the camera in the world coordinate system. Finally, based on the parameter calibration results of the optical cameras, multiple optical cameras are controlled by a synchronous triggering device to continuously capture images of the target points in the reference grid, resulting in an image sequence containing the target points. The distortion of the image sequence is then corrected by combining the intrinsic and extrinsic parameter matrices from the parameter calibration results (this process can be performed using existing camera calibration methods, which will not be elaborated here), ultimately forming a visual inspection image sequence of the target foundation pit support structure.

[0064] It should be noted that the visual inspection image sequence in this application refers to a series of images arranged in chronological order for subsequent visual inspection processing. By determining the visual inspection image sequence, continuous and complete visual data support can be provided for extracting the edge features of the target foundation pit support structure from the images and analyzing its spatial morphology. It also ensures that the collected images are coherent in the time dimension and fully covered in the spatial dimension, which can accurately reflect the state of the target structure at different construction times. This provides reliable raw materials for subsequent processing methods such as line segment detection and line extraction. At the same time, through the collaboration of multi-view images, it can make up for the blind spots that may exist under a single viewpoint and improve the accuracy of capturing the structural edges and target points.

[0065] In step S103, candidate straight lines of the target foundation pit support structure edge are extracted from the visual detection image sequence based on the direction consistency constraint, and edge straight lines that conform to the vertical characteristics of the target foundation pit support structure are extracted from all candidate straight lines through preset tilt angle threshold and continuous length threshold.

[0066] In some embodiments, the extraction of candidate straight lines from the edge of the target foundation pit support structure from the visual detection image sequence based on directional consistency constraints is specifically achieved through the following steps:

[0067] Smooth each frame of the visual detection image sequence to generate a smoothed image sequence;

[0068] An edge detection operator is used to perform edge detection on the smoothed images in the smoothed image sequence to obtain an edge pixel set;

[0069] All potential structural line segments are extracted from the set of edge pixels based on the Hough transform;

[0070] Establish a histogram of orientation angle distribution based on the orientation angles of each potential structural line segment;

[0071] Based on the histogram of the direction angle distribution, the main direction interval of potential structural line segments under the direction consistency constraint is determined, and potential structural line segments with direction angles located within the main direction interval are selected according to the main direction interval to form a set of direction consistency line segments;

[0072] The potential structural segments in the set of directional consistent line segments are optimized to obtain candidate straight lines for the edge of the target foundation pit support structure.

[0073] In specific implementation, firstly, when smoothing each frame of the visual inspection image sequence using existing Gaussian filtering, a smoothed image sequence is generated. This smoothed image sequence refers to a series of visual inspection images that retain the main structural information after filtering. The smoothed image sequence contains multiple smoothed images, each of which is a visual inspection image that retains the main structural information after filtering. Secondly, the Canny edge detection operator is used to perform edge detection on the smoothed images in the smoothed image sequence to obtain an edge pixel set. This edge pixel set is a set composed of the coordinates of all edge points in the smoothed image. Further, based on the Hough transform, all potential structural line segments are extracted from the edge pixel set. That is, each edge pixel in the edge pixel set is mapped to a parameter space through polar coordinate transformation using the Hough transform, and the vote count of the edge pixels in the parameter space is counted using an accumulator. When the vote count exceeds a preset threshold, the line segment corresponding to the edge pixel represents a potential structural line segment. This potential structural line segment refers to a line segment that may constitute a foundation pit support structure. The target edge line segments are further analyzed. A histogram of direction angle distribution is established based on the direction angles of each potential structural line segment. This involves counting the number of potential structural line segments within different direction angle intervals and presenting the results in a histogram. This histogram is a statistical chart describing the direction angle distribution characteristics of potential structural line segments. Then, the direction angle interval corresponding to the peak value in the histogram is selected as the main direction interval for potential structural line segments under direction consistency constraints. Potential structural line segments with direction angles within the main direction interval are then selected as direction-consistent line segments, forming a set of direction-consistent line segments. This set refers to a collection of potential structural line segments with similar direction angles that conform to the main direction characteristics. Finally, by calculating the distance and angle between potential structural line segments in the direction-consistent line segment set, potential structural line segments with distances less than a distance threshold and angles less than an angle threshold are identified as different parts of the same line segment and fused using existing line segment fusion methods. The resulting potential structural line segments are then used as candidate straight lines for the edge of the target foundation pit support structure.

[0074] It should be noted that, in this application, candidate straight lines refer to straight line segments that are candidate for the edge of the target foundation pit support structure. By determining the candidate straight lines, the straight line segments most likely to belong to the edge of the target foundation pit support structure can be screened from the visual inspection image sequence. This provides accurate processing objects for the subsequent extraction of edge straight lines that conform to vertical features, and further lays a reliable foundation for determining the distribution pattern of the structure edge and extracting the structure's central axis. This enables the entire verticality detection process to focus on the feature line segments that truly reflect the morphology of the target foundation pit support structure, thereby improving the accuracy and reliability of the detection results.

[0075] In some embodiments, reference Figure 3As shown in the figure, this is an exemplary flowchart of determining edge lines according to some embodiments of this application. In this embodiment, the extraction of edge lines that conform to the vertical characteristics of the target foundation pit support structure from all candidate lines by using preset inclination angle thresholds and continuous length thresholds can be achieved by the following steps:

[0076] In step S1031, the angle between each candidate line and the vertical direction is calculated to obtain the inclination angle value of each candidate line.

[0077] In step S1032, the inclination angle value of each candidate line is compared with the preset inclination angle threshold, and candidate lines with inclination angle values ​​less than or equal to the inclination angle threshold are selected to form a subset of low inclination angle lines.

[0078] In step S1033, the continuous pixel length of each low-angle straight line in the low-angle straight line subset is calculated to obtain the continuous length value of each low-angle straight line.

[0079] In step S1034, the continuous length value of each low-angle straight line is compared with a preset continuous length threshold, and low-angle straight lines with continuous length values ​​greater than or equal to the continuous length threshold are selected to obtain edge straight lines that conform to the vertical characteristics of the target foundation pit support structure.

[0080] In specific implementation, firstly, the angle between each candidate line and the vertical direction is calculated to obtain the inclination value of each candidate line. This is achieved by taking the dot product of the direction vector of the candidate line and the unit vector of the vertical direction, and then solving for the inclination value using inverse trigonometric functions. The inclination value refers to the angle between the candidate line and the vertical direction, which is the direction of gravity. Secondly, the inclination value of each candidate line is compared with a preset inclination threshold. Candidate lines with inclination values ​​less than or equal to the inclination threshold are retained, forming a low-inclination subset. This process selects approximately vertical line segments based on the vertical characteristics of the target foundation pit support structure and the required degree of inclination. The inclination threshold can be set according to actual needs, which will not be elaborated here. The low-inclination subset refers to the subset of lines selected from the candidate lines. A set of candidate straight lines whose angles meet the preset tilt requirements is formed. The low-angle straight line subset contains multiple low-angle straight lines, which are candidate straight lines whose tilt angles meet the preset tilt requirements. Then, the number of consecutive pixels on each low-angle straight line in the low-angle straight line subset is calculated to obtain the continuous length value of each low-angle straight line. The continuous length value refers to the length of the line segment formed by consecutive pixels of the low-angle straight line in the image. Finally, the continuous length value of each low-angle straight line is compared with a preset continuous length threshold, and low-angle straight lines with continuous length values ​​greater than or equal to the continuous length threshold are selected to obtain the edge straight lines that meet the vertical characteristics of the target foundation pit support structure. This step excludes short and fragmented line segments through length screening and retains continuous line segments that can reflect the edge of the main structure.

[0081] It should be noted that the edge lines that conform to the vertical characteristics of the target foundation pit support structure refer to the straight line segments that constitute the edge of the target structure. Specifically, they are straight line segments that both meet the requirement of approximate verticality and have sufficient continuous length to constitute the edge of the target structure. In existing technologies, the extraction of edge lines often adopts single threshold screening or general feature matching, which easily includes non-vertical interference line segments (such as horizontal supports and the edges of temporary components) in the results. However, this implementation scheme uses a dual constraint of "tilt angle threshold + continuous length threshold". First, the tilt angle threshold strictly limits the deviation range between the straight line and the vertical direction to ensure that the screened object conforms to the inherent attribute of the vertical extension of the support structure. Then, the continuous length threshold is used to remove short and fragmented noise line segments and retain the continuous edges that can reflect the main shape of the structure. This layered screening logic specifically solves the problem of edge confusion between the support structure and the surrounding complex environment (such as temporary facilities for cross-operations and scattered construction materials) in reverse construction scenarios, and significantly improves the correlation between edge lines and the target structure.

[0082] In step S104, the distribution pattern of the edge straight lines is determined, the structural centerline of the target foundation pit support structure is extracted based on the distribution pattern, and then the actual spatial verticality deviation of the structural centerline in the pixel coordinate system is determined based on the structural centerline and the construction layer height calibration data of the target foundation pit support structure.

[0083] In some embodiments, determining the distribution pattern of the edge lines is specifically achieved through the following steps:

[0084] Extract the endpoint coordinates of the edge lines and construct a spatial coordinate set for the edge lines;

[0085] Clustering algorithms are used to perform cluster analysis on the endpoints of the spatial coordinate set to obtain multiple clusters;

[0086] Calculate the direction and position parameters of the edge lines in each cluster to generate a set of cluster feature parameters;

[0087] The distribution pattern of edge lines is identified based on the cluster feature parameter set.

[0088] In specific implementation, firstly, by traversing each edge line, the coordinate values ​​of the two endpoints of the edge line in the image pixel coordinate system are obtained. These coordinate values ​​are then grouped and organized according to the line to obtain a spatial coordinate set of the edge lines. This spatial coordinate set refers to a dataset composed of the coordinate values ​​of the endpoints of the edge lines. Secondly, the number of clusters is set according to actual needs, and K-means clustering, a common clustering algorithm, is used to cluster the endpoints in the spatial coordinate set, resulting in multiple clusters. Each cluster is a group formed by aggregating endpoints with similar spatial locations using a clustering algorithm. Then, the direction and position parameters of the edge lines in each cluster are calculated to generate a cluster feature parameter set. The direction parameter can be represented by the direction angle of the edge line, and the position parameter... The distribution pattern of edge lines can be determined by the mean of the endpoint coordinates within the cluster. The cluster feature parameter set refers to a set of parameters containing the orientation and position parameters of each cluster. Finally, the distribution pattern of edge lines is identified based on the cluster feature parameter set. That is, the orientation and position parameters of each cluster are extracted from the cluster feature parameter set. By calculating the difference in orientation parameters of different clusters, if the difference is within a preset parallel threshold range, the corresponding cluster is determined to be a homo-directional cluster. If the spatial positions of homo-directional clusters are equidistant and have no intersection, the distribution pattern of edge lines is determined to be a parallel distribution. If the difference in orientation parameters of different clusters exceeds the parallel threshold range, and the edge lines have intersections calculated by the line equation, the distribution pattern of edge lines is determined to be an intersecting distribution.

[0089] It should be noted that the distribution pattern of edge lines refers to the arrangement of edge lines in space. By determining the distribution pattern, the overall morphological pattern of the target foundation pit support structure can be extracted from the spatial arrangement and clustering characteristics of the edge lines. This provides a clear feature basis for the subsequent extraction of the structural centerline and ensures that the subsequent extraction of the centerline focuses on the core features that can truly reflect the structural morphology. This lays the foundation for accurately calculating the verticality deviation of the structure and improves the adaptability of the entire detection method to complex construction environments.

[0090] In some embodiments, extracting the structural centerline of the target foundation pit support structure based on the distribution pattern is specifically achieved through the following steps:

[0091] Obtain the distribution pattern of edge lines, wherein the distribution pattern is divided into parallel distribution and intersecting distribution;

[0092] If the distribution pattern is a parallel distribution, then the symmetrical centerline of the edge line is calculated as the structural centerline of the target foundation pit support structure;

[0093] If the distribution pattern is an intersecting distribution, then the angle bisector corresponding to the edge line is calculated as the structural centerline of the target foundation pit support structure.

[0094] In specific implementation, firstly, the distribution pattern of the edge lines is obtained. The distribution pattern is divided into parallel distribution and intersecting distribution. The parallel distribution refers to a spatial arrangement in which the edge lines are in the same direction and have no intersection points. The intersecting distribution refers to a spatial arrangement in which the edge lines are in different directions and have intersection points. Then, if the distribution pattern is parallel, the symmetrical centerline of the edge lines is calculated as the structural centerline of the target foundation pit support structure. Specifically, the perpendicular bisector of the parallel edge lines is taken, and then a linear regression algorithm is used to fit the perpendicular bisector. The fitted line obtained is the symmetrical centerline, where the symmetrical centerline is a line that is equidistant from the parallel edge lines. Finally, if the distribution pattern is intersecting, the angle bisector corresponding to the edge lines is calculated as the structural centerline of the target foundation pit support structure. Specifically, the direction angles of the intersecting edge lines are solved, the average of the two direction angles is taken, and the line at the average value is taken as the angle bisector. The angle bisector is a line that divides the included angle formed by the intersecting edge lines into two equal angles.

[0095] It should be noted that, in this application, the structural centerline refers to the reference straight line that reflects the extension trend of the target foundation pit support structure. By determining the structural centerline, a core benchmark can be provided for the spatial morphology analysis of the target foundation pit support structure. On the one hand, it can serve as a reference for measuring the spatial attitude of the structure, providing a direct feature carrier for converting the centerline in the pixel coordinate system into the actual spatial verticality deviation, thus providing a clear positioning benchmark for the conversion of two-dimensional image information into three-dimensional spatial parameters. On the other hand, the central axis can integrate scattered edge line features to form a unified description of the overall morphology of the structure, effectively reducing the impact of interference information in complex construction environments on the structural status assessment, and laying the foundation for accurately detecting the verticality of the structure and judging whether the construction quality meets the specifications.

[0096] In some embodiments, determining the actual spatial verticality deviation of the structural centerline in the pixel coordinate system based on the structural centerline and the construction floor height calibration data of the target foundation pit support structure is achieved through the following steps:

[0097] Extract the coordinate parameters of the central axis of the structure in the pixel coordinate system to obtain the pixel coordinate sequence;

[0098] Based on the intrinsic and extrinsic parameter matrices of the optical camera, the pixel coordinate sequence is converted into a three-dimensional coordinate sequence in the camera coordinate system;

[0099] Based on the construction floor height calibration data of the target foundation pit support structure, establish the transformation relationship between the camera coordinate system and the actual construction coordinate system;

[0100] Based on the transformation relationship, the three-dimensional coordinate sequence in the camera coordinate system is converted into the three-dimensional coordinates of the centerline in the actual construction coordinate system, thereby obtaining the structural centerline in the actual construction coordinate system;

[0101] Calculate the spatial angle between the structural centerline and the theoretical vertical axis in the actual construction coordinate system to obtain the actual spatial perpendicularity deviation of the structural centerline in the pixel coordinate system.

[0102] In specific implementation, firstly, the pixels of the central axis of the structure in the image are traversed, and their horizontal and vertical coordinate values ​​are recorded and arranged in order to obtain a pixel coordinate sequence. The pixel coordinate sequence refers to the set of coordinates of all points of the central axis of the structure in the two-dimensional image pixel coordinate system in order. Secondly, using an existing pinhole camera model, the pixel coordinates in the pixel coordinate sequence are first converted into normalized image coordinates in the camera coordinate system through the intrinsic parameter matrix. Then, the rotation matrix and translation vector in the extrinsic parameter matrix are combined to perform coordinate transformation to obtain the three-dimensional coordinates corresponding to each pixel coordinate, thus obtaining a three-dimensional coordinate sequence. The three-dimensional coordinate sequence in the camera coordinate system refers to the set of coordinates of all points of the central axis of the structure in the three-dimensional coordinate system in order. Further, based on the construction floor height calibration data of the target foundation pit support structure, a conversion relationship between the camera coordinate system and the actual construction coordinate system is established. That is, the construction floor height calibration data of the target foundation pit support structure is obtained. The construction floor height calibration data is the known value of each floor height of the target foundation pit support structure in the actual construction coordinate system. The existing matching algorithm is used to convert the construction floor height... The calibration data is matched with the corresponding height coordinates in the camera coordinate system to solve for the transformation matrix between the two coordinate systems (including rotation and translation parameters). This transformation matrix is ​​used as the transformation relationship between the camera coordinate system and the actual construction coordinate system. The transformation relationship describes the mathematical relationship between the coordinate mapping between the camera coordinate system and the actual construction coordinate system. Then, based on the transformation relationship, the three-dimensional coordinate sequence in the camera coordinate system is converted into the three-dimensional coordinates of the central axis in the actual construction coordinate system. That is, the three-dimensional coordinates in the camera coordinate system are rotated and translated using the transformation matrix corresponding to the transformation relationship to obtain the spatial coordinates of the central axis of the structure in the actual construction coordinate system (i.e., the three-dimensional coordinates of the central axis). The obtained three-dimensional coordinates of the central axis are then connected in sequence to form the three-dimensional coordinates of the central axis of the structure in the actual construction coordinate system. The three-dimensional coordinates of the central axis of the structure in the actual construction coordinate system refer to the coordinates of the central axis of the structure in a three-dimensional coordinate system based on the construction area. Finally, the spatial angle between the central axis of the structure in the actual construction coordinate system and the theoretical vertical axis is calculated to obtain the actual spatial verticality deviation of the central axis of the structure in the pixel coordinate system.

[0103] It should be noted that the actual spatial verticality deviation in this application refers to the spatial angle between the central axis of the structure and the theoretical vertical direction. This actual spatial verticality deviation reflects the verticality deviation of the structure's tilt. In the prior art, the actual spatial verticality deviation mostly relies on single-point measurement or local coordinate transformation, which is easily affected by insufficient representativeness of the measurement points or the accumulation of coordinate system transformation errors. However, this scheme establishes the transformation relationship between the camera coordinate system and the actual construction coordinate system through the construction floor height calibration data, and accurately maps the central axis in the pixel coordinate system to the actual construction coordinate system through the camera coordinate system. This achieves a coherent transformation from two-dimensional image features to three-dimensional actual spatial parameters. Moreover, the transformation process uses the overall central axis of the structure as the carrier, avoiding the one-sidedness of single-point measurement and improving the engineering practicality and accuracy reliability of verticality detection.

[0104] In step S105, the confidence verticality of the target foundation pit support structure under visual inspection is obtained by combining the changes in the centerline of the target foundation pit support structure during adjacent construction stages with the actual spatial verticality deviation.

[0105] In some embodiments, the confidence verticality of the target foundation pit support structure under visual inspection is obtained by combining the characteristics of the change in the centerline of the target foundation pit support structure in adjacent construction stages with the actual spatial verticality deviation. This is specifically achieved through the following steps:

[0106] Extract the structural centerline of the target foundation pit support structure in adjacent construction stages to obtain the centerline sequence of adjacent stages;

[0107] The characteristics of the change of the centerline in adjacent construction stages of the target foundation pit support structure are determined based on the sequence of the centerlines in adjacent stages.

[0108] A correction model for actual spatial verticality deviation is constructed based on the characteristics of central axis variation;

[0109] The actual spatial verticality deviation is input into the correction model for error compensation to obtain the confidence verticality of the target foundation pit support structure under visual inspection.

[0110] In specific implementation, firstly, the structural centerlines of adjacent construction stages of the target foundation pit support structure are extracted to obtain a sequence of adjacent stage centerlines. Specifically, using the aforementioned method for extracting structural centerlines based on a distribution pattern, the structural centerlines of two adjacent construction stages (e.g., the completion of the nth floor and the completion of the n+1th floor) are obtained respectively. These corresponding structural centerlines are then arranged according to the construction sequence to form a sequence of adjacent stage centerlines. This sequence refers to a set of structural centerlines of adjacent construction stages recorded in chronological order. Secondly, the change characteristics of the centerlines in adjacent construction stages of the target foundation pit support structure are determined based on the adjacent stage centerline sequence. Specifically, the spatial position deviation (or angular deviation) of adjacent structural centerlines in the adjacent stage centerline sequence is calculated, and this spatial position deviation (or angular deviation) is used as the change characteristic of the centerlines in adjacent construction stages of the target foundation pit support structure. This change characteristic refers to… The characteristics of the change in the spatial position of the central axis of the structure during adjacent construction stages are analyzed. Then, based on the changes in the central axis, a correction model for the actual spatial verticality deviation is constructed. That is, a linear regression method can be used, with the central axis change characteristic parameters (i.e., spatial position deviation and orientation angle deviation) as independent variables and the difference between the verticality deviation and the verticality deviation under actual visual inspection as the dependent variable. An error compensation function is obtained by fitting the data using the least squares method, and the error compensation function is used as the correction model for the actual spatial verticality deviation. The correction model is a mathematical model used to adjust the error of the actual spatial verticality deviation. Finally, the actual spatial verticality deviation is input into the correction model for error compensation to obtain the confidence verticality of the target foundation pit support structure under visual inspection. That is, the calculated actual spatial verticality deviation is substituted into the error compensation function of the correction model to obtain the confidence verticality after error correction.

[0111] It should be noted that in this application, the confidence verticality refers to the visual inspection verticality result that is closest to the true state. By determining the confidence verticality, a more reliable quantitative basis can be provided for the construction quality assessment of the target foundation pit support structure. The confidence verticality is the visual inspection verticality result that is closer to the true state after being corrected by the change characteristics of the centerline in adjacent stages.

[0112] In another aspect, in some embodiments, this application provides a verticality detection system for reverse construction processes, referring to... Figure 4 The figure is a structural schematic diagram of a verticality detection system for a reverse construction process according to some embodiments of this application. The verticality detection system for the reverse construction process includes: a layout module 201, a processing module 202, and an execution module 203, which are described below:

[0113] Arrangement module 201, in this application, is mainly used to symmetrically arrange multiple target points on the top and bottom of the target foundation pit support structure in reverse construction to form a reference grid for visual inspection.

[0114] Processing module 202, in this application, is mainly used to acquire images of each target point in the reference grid through an optical camera on the periphery of the construction area, and obtain a visual inspection image sequence of the target foundation pit support structure;

[0115] The processing module 202 is further configured to extract candidate straight lines of the edge of the target foundation pit support structure from the visual detection image sequence based on the direction consistency constraint, and extract edge straight lines that conform to the vertical characteristics of the target foundation pit support structure from all candidate straight lines through preset tilt angle threshold and continuous length threshold.

[0116] In addition, the processing module 202 is also used to determine the distribution pattern of the edge straight line, extract the structural centerline of the target foundation pit support structure based on the distribution pattern, and then determine the actual spatial verticality deviation of the structural centerline in the pixel coordinate system based on the structural centerline and the construction layer height calibration data of the target foundation pit support structure.

[0117] The execution module 203 in this application is mainly used to extract the confidence verticality of the target foundation pit support structure under visual inspection by combining the changes in the central axis of the target foundation pit support structure in adjacent construction stages with the actual spatial verticality deviation.

[0118] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the verticality detection method of the above-described reverse construction process.

[0119] In some embodiments, reference Figure 5 The figure is a schematic diagram of the structure of a computer device for implementing a verticality detection method for a reverse construction process according to some embodiments of this application. The verticality detection method for the reverse construction process in the above embodiments can be achieved through... Figure 5 The computer device shown is used to implement this, and the computer device includes at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

[0120] The processor 301 may be a general-purpose central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more methods for controlling the execution of the verticality detection method in the reverse construction process of this application.

[0121] The communication bus 302 can be used to transmit information between the aforementioned components.

[0122] The memory 303 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via the communication bus 302. The memory 303 may also be integrated with the processor 301.

[0123] The memory 303 stores program code for executing the scheme of this application, and its execution is controlled by the processor 301. The processor 301 executes the program code stored in the memory 303. The program code may include one or more software modules. In the above embodiments, the determination of the verticality detection method in the reverse construction process can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0124] Communication interface 304 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0125] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0126] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0127] In addition, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the verticality detection method for the above-described reverse construction process.

[0128] Although preferred embodiments of this application have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0129] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application.

Claims

1. A method for detecting the verticality of a reverse construction process, characterized in that, Includes the following steps: Multiple target points are symmetrically arranged at the top and bottom of the target foundation pit support structure during reverse construction to form a reference grid for visual inspection. By using an optical camera outside the construction area to acquire images of each target point in the reference grid, a visual inspection image sequence of the target foundation pit support structure is obtained. Each frame of the visual detection image sequence is smoothed to generate a smoothed image sequence; an edge detection operator is used to perform edge detection on the smoothed images in the smoothed image sequence to obtain an edge pixel set; all potential structural line segments are extracted from the edge pixel set based on the Hough transform; and a histogram of orientation angle distribution is established according to the orientation angle of each potential structural line segment. The direction angle interval corresponding to the peak value in the direction angle distribution histogram is selected as the main direction interval of the potential structural line segments under the direction consistency constraint. Then, potential structural line segments whose direction angles are located within the main direction interval are selected as direction consistency line segments, thus forming a direction consistency line segment set. The direction consistency line segment set refers to the set of potential structural line segments whose direction angles conform to the main direction characteristics. The potential structural line segments in the direction consistency line segment set are optimized to obtain candidate straight lines of the edge of the target foundation pit support structure. The edge straight lines that conform to the vertical characteristics of the target foundation pit support structure are extracted from all candidate straight lines through preset inclination angle thresholds and continuous length thresholds. The endpoint coordinates of the edge lines are extracted to construct a spatial coordinate set of the edge lines. A clustering algorithm is used to perform cluster analysis on the endpoints in the spatial coordinate set to obtain multiple clusters. The direction parameters and position parameters of the edge lines in each cluster are calculated to generate a cluster feature parameter set. The distribution pattern of the edge lines is identified based on the cluster feature parameter set. The structural centerline of the target foundation pit support structure is extracted based on the distribution pattern. Then, the actual spatial verticality deviation of the structural centerline in the pixel coordinate system is determined based on the structural centerline and the construction layer height calibration data of the target foundation pit support structure. The confidence verticality of the target foundation pit support structure under visual inspection is obtained by combining the changes in the centerline of the target foundation pit support structure during adjacent construction stages with the actual spatial verticality deviation.

2. The method as described in claim 1, characterized in that, The visual inspection image sequence of the target foundation pit support structure is obtained by acquiring images of each target point in the reference grid using an optical camera outside the construction area. Specifically, the sequence includes: Multiple optical cameras are deployed around the construction area based on the reference grid to form a multi-view image acquisition network; The parameters of the optical cameras in the multi-view image acquisition network are calibrated. Based on the parameter calibration results of the optical camera, continuous shooting of the target points in the reference grid is triggered, thereby obtaining a visual inspection image sequence of the target foundation pit support structure.

3. The method as described in claim 1, characterized in that, By using preset inclination angle thresholds and continuous length thresholds, edge lines that conform to the vertical characteristics of the target foundation pit support structure are extracted from all candidate lines. Specifically, these include: Calculate the angle between each candidate line and the vertical direction to obtain the inclination angle value of each candidate line; The inclination angle value of each candidate line is compared with the preset inclination angle threshold, and candidate lines with inclination angle values ​​less than or equal to the inclination angle threshold are selected to form a subset of low inclination angle lines. Calculate the continuous pixel length of each low-angle straight line in the low-angle straight line subset to obtain the continuous length value of each low-angle straight line; The continuous length value of each low-angle straight line is compared with the preset continuous length threshold, and the low-angle straight lines with continuous length values ​​greater than or equal to the continuous length threshold are selected to obtain the edge straight lines that conform to the vertical characteristics of the target foundation pit support structure.

4. The method as described in claim 1, characterized in that, Extracting the structural centerline of the target foundation pit support structure based on the aforementioned distribution pattern specifically includes: Obtain the distribution pattern of edge lines, wherein the distribution pattern is divided into parallel distribution and intersecting distribution; If the distribution pattern is a parallel distribution, then the symmetrical centerline of the edge line is calculated as the structural centerline of the target foundation pit support structure; If the distribution pattern is an intersecting distribution, then the angle bisector corresponding to the edge line is calculated as the structural centerline of the target foundation pit support structure.

5. The method as described in claim 1, characterized in that, The reference grid is a spatial reference coordinate system formed by target points symmetrically arranged at the top and bottom of the foundation pit support structure.

6. A verticality detection system for a reverse construction process, used to perform the verticality detection method for a reverse construction process as described in any one of claims 1 to 5, the system comprising: The layout module is used to symmetrically arrange multiple target points on the top and bottom of the target foundation pit support structure during reverse construction, forming a reference grid for visual inspection. The processing module is used to acquire images of each target point in the reference grid through an optical camera outside the construction area, and obtain a visual inspection image sequence of the target foundation pit support structure. The processing module is also used to extract candidate straight lines of the edge of the target foundation pit support structure from the visual detection image sequence based on the direction consistency constraint, and to extract edge straight lines that conform to the vertical characteristics of the target foundation pit support structure from all candidate straight lines through preset tilt angle threshold and continuous length threshold. The processing module is also used to determine the distribution pattern of the edge straight lines, extract the structural centerline of the target foundation pit support structure based on the distribution pattern, and then determine the actual spatial verticality deviation of the structural centerline in the pixel coordinate system based on the structural centerline and the construction layer height calibration data of the target foundation pit support structure. The execution module is used to extract the confidence verticality of the target foundation pit support structure under visual inspection by combining the changes in the centerline of the target foundation pit support structure during adjacent construction stages with the actual spatial verticality deviation.

7. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the verticality detection method for the reverse construction process as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the verticality detection method for the reverse construction process as described in any one of claims 1 to 5.

Images