A construction surveying method, system, electronic device and storage medium

Abstract

The application provides a kind of construction surveying method, system, electronic equipment and storage medium, it is related to the technical field of construction surveying, the application is through image acquisition equipment multi-angle shooting steel structure node, fusion original image obtains target image containing weld contour and bolt spacing;Call twin neural network to compare its with the geometric characteristics of the corresponding node of the preset BIM model, obtain the difference information of weld and bolt spacing;Phased array microwave radar scans component surface, measures distance to form multiple groups of data, generates three-dimensional point cloud grid through three-dimensional reconstruction, compares with the theoretical coordinates of the corresponding node of BIM model to obtain spatial coordinate deviation;Correlate three types of data to obtain comprehensive deviation, generate installation quality evaluation report by fusing point cloud, can collect data through image and radar, combined with twin neural network and three-dimensional reconstruction technology, compare the actual characteristics of steel structure with BIM model, integrate deviation information to generate quality evaluation report, realize construction quality comprehensive measurement and evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building construction measurement, and in particular to a building construction measurement method, system, electronic device and storage medium. BACKGROUND

[0002] In the building construction scene, the installation accuracy of steel structural members directly affects the safety and stability of the project, especially the weld forming quality of the node position, the bolt connection spacing and the overall spatial position deviation, which needs to be controlled in real time through accurate measurement. During the construction process, the node micro features (such as weld contour, bolt spacing) and the component macro spatial coordinates need to be obtained synchronously, and compared with the preset BIM model to quantify the deviation and provide data support for installation adjustment, which requires the measurement method to have the technical capabilities of multi-dimensional, high-precision, and efficient integrated analysis.

[0003] At present, for the above-mentioned needs, the industry commonly uses a measurement scheme based on three-dimensional laser scanning, which scans the steel structural member through a laser scanner, extracts the component surface features after generating point cloud data, combines the point cloud matching algorithm with the BIM model for coordinate comparison, outputs the spatial position deviation data, and supplements the micro feature information such as weld and bolt through manual auxiliary recognition of point cloud details.

[0004] However, the existing scheme has obvious limitations. Although three-dimensional laser scanning can obtain macro spatial coordinates, it lacks precision in capturing micro features such as weld contour edge lines and bolt spacing, and is prone to lose details due to the limitation of point cloud resolution. Moreover, its data processing relies on manual intervention for micro feature recognition, and the comparison with the BIM model lacks intelligent correlation analysis capability, making it difficult to efficiently integrate micro features and macro coordinates, resulting in low efficiency and insufficient completeness of deviation quantification, which cannot meet the needs of multi-dimensional synchronous accurate measurement in the construction scene. SUMMARY

[0005] The present application aims to provide a building construction measurement method, system, electronic device and storage medium to solve the problem of low accuracy in building construction measurement in the prior art.

[0006] To solve the above technical problems, in a first aspect, the present application provides a building construction measurement method, comprising:

[0007] In the building construction process, the node position of the steel structural member is photographed from multiple angles by an image acquisition device to obtain original images from multiple angles, and all original images are fused to obtain a target image containing the weld contour and bolt spacing at the node position.

[0008] The twin neural network is called to compare the weld contour and bolt spacing in the target image with the geometric features of the corresponding node positions in the preset BIM model respectively, to obtain weld contour difference information and bolt spacing difference information;

[0009] The overall surface of the steel structure member is scanned at multiple angles by using a phased array microwave radar array, and the distance between the radar in the phased array microwave radar array and the overall surface is measured during the scanning process to form multiple sets of distance data;

[0010] The multiple sets of distance data are combined by using a three-dimensional reconstruction technology to generate a three-dimensional point cloud grid, and the actual spatial coordinates in the three-dimensional point cloud grid are compared with the theoretical spatial coordinates of the corresponding node positions in the preset BIM model to obtain spatial coordinate deviation data;

[0011] The spatial coordinate deviation data, the weld contour difference information and the bolt spacing difference information are associated to obtain comprehensive deviation data, and the comprehensive deviation data is fused with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structure member.

[0012] Optionally, the twin neural network is called to compare the weld contour and bolt spacing in the target image with the geometric features of the corresponding node positions in the preset BIM model respectively, to obtain weld contour difference information and bolt spacing difference information, including the following steps:

[0013] The edge line information of the weld contour and the standard edge line information in the preset BIM model are respectively processed layer by layer by a first subnetwork of the twin neural network to obtain first local features and second local features, and the coordinate class intermediate features and angle class intermediate features of each point in the first local features and the second local features are respectively calculated to obtain a first feature vector;

[0014] The numerical information of the bolt spacing and the standard bolt spacing numerical information in the preset BIM model are respectively processed layer by layer by a second subnetwork of the twin neural network to obtain first spacing length features and second spacing length features, and the first spacing length features and the second spacing length features are calculated to output a second feature vector;

[0015] Based on the output layer of the twin neural network, the weld contour difference information is determined based on the first feature vector, and the bolt spacing difference information is determined based on the second feature vector.

[0016] Optionally, the first sub-network of the twin neural network processes the edge line information of the weld contour and the standard edge line information in the preset BIM model respectively by layer-by-layer feature processing to obtain first local features and second local features, including the following steps:

[0017] Through the identification layer of the first sub-network, feature points with a bending degree value exceeding a preset bending degree threshold are identified from the edge lines in the edge line information and the edge lines in the standard edge line information, and the edge lines between adjacent feature points are regarded as continuous line segments;

[0018] Through the feature extraction layer of the first sub-network, point positions are selected on the continuous line segments according to a preset interval, and the horizontal coordinate values and the vertical coordinate values of all the point positions in a two-dimensional plane are recorded to form coordinate-type intermediate features;

[0019] Through the calculation layer of the first sub-network, the angle value between the line connecting the two end points on the continuous line segment and the horizontal reference direction and the angle change value of the line connecting adjacent point positions on the continuous line segment are calculated to form angle-type intermediate features;

[0020] Through the feature association layer of the first sub-network, the coordinate-type intermediate features and the angle-type intermediate features of the edge line information of the weld contour are associated to obtain first local features, and the coordinate-type intermediate features and the angle-type intermediate features of the standard edge line information are associated to obtain second local features.

[0021] Optionally, the second sub-network of the twin neural network processes the numerical information of the bolt spacing and the standard bolt spacing numerical information in the preset BIM model respectively by layer-by-layer feature processing to obtain first spacing length features and second spacing length features, including the following steps:

[0022] Through the identification layer of the second sub-network, the position information in the numerical information of the bolt spacing and the position information in the standard bolt spacing numerical information are identified respectively, and two adjacent bolts are paired to form a spacing unit based on the position information, wherein the spacing unit contains the position identification of the two adjacent bolts and the corresponding spacing length value;

[0023] Through the first association layer of the second sub-network, the position identification and the corresponding spacing length value of each adjacent bolt pair in the spacing unit are associated to form numerical-type intermediate features;

[0024] arranging the spacing units in the bolt arrangement sequence according to the bolt arrangement sequence on the steel structure member, calculating the proportional relationship between the continuous spacing length values in the bolt arrangement sequence, and forming the relationship class intermediate feature;

[0025] associating the numerical value information of the bolt spacing and the numerical value information of the standard bolt spacing with the relationship class intermediate feature, to obtain a first spacing length feature and a second spacing length feature.

[0026] Optionally, the first feature vector is used to determine the weld contour difference information, and the second feature vector is used to determine the bolt spacing difference information, including the following steps:

[0027] calculating the difference between the horizontal coordinate value and the vertical coordinate value of each actual point in the first local feature and the corresponding standard point in the second local feature, to obtain the coordinate value difference;

[0028] calculating the difference between the angle value of the first local feature and the angle value of the second local feature, and the difference between the angle change value of the corresponding adjacent points in the first local feature and the second local feature, to obtain the angle value difference;

[0029] associating the coordinate value difference and the angle value difference in the order of continuous line segments, and integrating into the weld contour difference information containing the position deviation and the trend deviation of each continuous line segment;

[0030] calculating the difference between the first spacing length feature and the second spacing length feature in the second feature vector, to obtain the basic spacing difference, and calculating the difference between the proportional relationship in the first spacing length feature and the proportional relationship in the second spacing length feature based on the continuous spacing units in the bolt arrangement sequence, to obtain the proportional difference;

[0031] associating the basic spacing difference and the proportional difference in the order of the bolt arrangement sequence, and integrating into the bolt spacing difference information containing the length deviation of each spacing unit and the sequence proportional deviation.

[0032] Optionally, the three-dimensional reconstruction technology is used to combine the multiple groups of distance data to generate a three-dimensional point cloud grid, and the actual space coordinates in the three-dimensional point cloud grid are compared with the theoretical space coordinates of the corresponding node positions in the preset BIM model to obtain space coordinate deviation data, including the following steps:

[0033] The distance information of any node position in the multiple sets of distance data collected at different angles is associated through a three-dimensional reconstruction technology to obtain actual spatial coordinates of the node position in a three-dimensional space, and the actual spatial coordinates of all the node positions are integrated to form a three-dimensional point set, and adjacent node positions in the three-dimensional point set are connected to form a three-dimensional point cloud grid.

[0034] The actual spatial coordinates of each node position in the three-dimensional point cloud grid in the three-dimensional space are compared with the theoretical spatial coordinates of the corresponding node position in the preset BIM model, and the numerical difference between the actual spatial coordinates and the theoretical spatial coordinates of the node position in three axial directions is calculated, and the numerical difference in the three axial directions is integrated to generate spatial coordinate deviation data of the node position.

[0035] Optionally, the spatial coordinate deviation data, the weld contour difference information and the bolt spacing difference information are associated to obtain comprehensive deviation data, and the comprehensive deviation data is fused with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structure component, including the following steps:

[0036] The spatial coordinate deviation, the weld contour difference and the bolt spacing difference of the same node position of the steel structure component are integrated to form comprehensive deviation data representing the overall deviation value and distribution of the steel structure component;

[0037] The comprehensive deviation data is associated with the actual spatial coordinates of the corresponding node position in the three-dimensional point cloud grid to form a three-dimensional model containing deviation information;

[0038] Based on the three-dimensional model, the comprehensive deviation data in the same installation area of the steel structure component is classified and arranged according to the installation area of the steel structure component, and the overall deviation value and distribution of all the comprehensive deviation data in each installation area are recorded to form an installation quality comprehensive evaluation report of the steel structure component.

[0039] In a second aspect, the present application provides a building construction measurement system, comprising:

[0040] The acquisition module is configured to capture multiple-angle original images of the node part of the steel structure component through an image acquisition device during the building construction process, and fuse all the original images to obtain a target image containing the weld contour and bolt spacing of the node part;

[0041] The first comparison module is configured to call a twin neural network to compare the weld contour and bolt spacing in the target image with the geometric features of the corresponding node part in the preset BIM model, respectively, to obtain weld contour difference information and bolt spacing difference information.

[0042] a measurement module configured to perform multi-angle scanning on an overall surface of the steel structural member by using a phased array microwave radar array, and measure distances between radars in the phased array microwave radar array and the overall surface during the scanning to form a plurality of sets of distance data;

[0043] a second comparison module configured to combine the plurality of sets of distance data by using a three-dimensional reconstruction technique to generate a three-dimensional point cloud grid, and compare actual spatial coordinates in the three-dimensional point cloud grid with theoretical spatial coordinates of corresponding node positions in a preset BIM model to obtain spatial coordinate deviation data;

[0044] a fusion module configured to associate the spatial coordinate deviation data, the weld contour difference information and the bolt spacing difference information to obtain comprehensive deviation data, and fuse the comprehensive deviation data with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structural member.

[0045] In a third aspect, the present application provides an electronic device, comprising:

[0046] a memory configured to store a computer program;

[0047] a processor configured to execute the computer program to implement the steps of the building construction measurement method according to the first aspect.

[0048] In a fourth aspect, the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executable by a processor to implement the steps of the building construction measurement method according to the first aspect.

[0049] The method for building construction measurement provided in the application comprises the following steps: in the process of building construction, a node part of a steel structure component is photographed at multiple angles by an image acquisition device to obtain original images at multiple angles, all the original images are fused to obtain a target image containing a weld contour and a bolt spacing at the node part; the weld contour and the bolt spacing in the target image are compared with geometric features of a corresponding node part in a preset BIM model respectively by calling a twin neural network to obtain weld contour difference information and bolt spacing difference information; the overall surface of the steel structure component is scanned at multiple angles by a phased array microwave radar array, the distance between a radar in the phased array microwave radar array and the overall surface is measured during the scanning process to form multiple sets of distance data; the multiple sets of distance data are combined by using a three-dimensional reconstruction technology to generate a three-dimensional point cloud grid, actual space coordinates in the three-dimensional point cloud grid are compared with theoretical space coordinates of a corresponding node position in the preset BIM model to obtain space coordinate deviation data; the space coordinate deviation data, the weld contour difference information and the bolt spacing difference information are associated to obtain comprehensive deviation data, and the comprehensive deviation data are fused with the three-dimensional point cloud grid for processing to generate an installation quality comprehensive evaluation report of the steel structure component.

[0050] The technical scheme of the application has the following beneficial effects:

[0051] The complete process from image acquisition and fusion, twin neural network comparison, radar scanning, three-dimensional reconstruction to data association and fusion is used to realize multi-dimensional coordination of steel structure construction measurement: multi-angle image fusion ensures that microscopic features such as weld contours and bolt spacings of node parts are fully captured; the twin neural network intelligently compares the microscopic features with the BIM model to accurately output difference information of the welds and the bolt spacings; the phased array microwave radar scanning combined with three-dimensional reconstruction obtains overall macroscopic space coordinates of the component and quantifies the deviation from the BIM model; finally, the microscopic and macroscopic deviation data are associated and the three-dimensional point cloud is fused to generate a comprehensive evaluation report, which not only realizes synchronous measurement of node details and overall space, but also improves the integrity and accuracy of deviation quantification through multi-source data integration, thereby providing efficient and accurate technical support for comprehensive control of steel structure installation quality.

[0052] Further, the application processes the weld contour edge line information and the standard information layer by layer through the first sub-network, identifies the feature points first and determines the continuous line segments, extracts the intermediate features in the coordinate and angle categories and associates to obtain the first and second local features, and then calculates the corresponding coordinate and angle differences and integrates them into the weld contour difference information; processes the bolt spacing numerical information and the standard information layer by layer through the second sub-network, identifies the position information first to form the spacing units, associates the numerical values and the position identifiers and calculates the proportional relationship of the arrangement sequence to obtain the first and second spacing length features, and then calculates the basic spacing and the proportional difference and integrates them into the bolt spacing difference information; and finally determines the two types of difference information based on the feature vectors output by the two sub-networks.

[0053] The application realizes the accurate capture of the coordinates and angle features of the weld contour edge lines and the length and proportional relationship of the bolt spacing through the hierarchical feature extraction and association of the twin neural network, and through the targeted difference calculation, compares the micro features with the corresponding features of the preset BIM model in fine granularity, and finally obtains the weld contour difference information containing the position and direction deviations and the bolt spacing difference information containing the length and proportional deviations, thereby providing comprehensive and accurate micro feature deviation data support for the steel structure node quality evaluation.

[0054] These and other aspects of the application will become more apparent from the following description of the embodiments taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0056] Figure 1 A flowchart of a building construction measurement method provided by an embodiment of the application;

[0057] Figure 2 A scene diagram of a building construction measurement method provided by an embodiment of the application;

[0058] Figure 3 A structural schematic diagram of a building construction measurement system provided by an embodiment of the application;

[0059] Figure 4 A structural schematic diagram of an electronic device provided by an embodiment of the application. DETAILED DESCRIPTION

[0060] The existing measurement scheme based on three-dimensional laser scanning can obtain the macro spatial coordinates of the components in the steel structure construction measurement, but the accuracy of capturing the micro features such as the edge details of the weld contour and the subtle values of the bolt spacing is limited, and the details are often lost due to insufficient scanning resolution. At the same time, the data processing needs manual assistance to identify the micro features, and the comparison with the BIM model lacks intelligent integration capability, making it difficult to efficiently correlate the micro feature deviations with the macro spatial coordinate deviations, resulting in insufficient completeness and efficiency of the measurement results, and unable to meet the demand for simultaneous and accurate control of multi-dimensional data in construction.

[0061] To solve the above problems, the present application provides a building construction measurement method, which can accurately capture micro features such as weld contours and bolt spacings by multi-angle shooting of node parts by image acquisition equipment and fusion of images. With the help of a twin neural network (an intelligent comparison model), these features are automatically compared with the corresponding parts of the BIM model, and the micro feature differences are output. At the same time, the component surface is scanned by a phased array microwave radar, and a three-dimensional model is generated by three-dimensional reconstruction, obtaining macro spatial coordinates and comparing them with the BIM model to obtain coordinate deviations. Finally, the micro and macro deviation data are correlated, and a three-dimensional model containing comprehensive information is fused to generate a quality evaluation report. This scheme focuses on micro details through images, captures macro coordinates through radar, automatically compares intelligent models, and correlates multi-source data, solving the problems of existing schemes such as inaccurate micro feature capture and dependence on manual work, and achieving efficient integration of micro and macro deviations, improving the accuracy, completeness and efficiency of the measurement.

[0062] In order for those skilled in the art to better understand the present application, the present application will be further described in detail below in conjunction with the drawings and specific embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0063] The core of the present application is to provide a building construction measurement method, and a flowchart of one specific embodiment thereof is shown in Figure 1 The method comprises:

[0064] S101, in the building construction process, the node parts of the steel structure components are multi-angle shot by the image acquisition equipment, and the original three-dimensional images of multiple angles are obtained. All original three-dimensional images are fused to obtain a target image containing the weld contour and bolt spacing at the node parts.

[0065] In the above scheme, the image acquisition device refers to a device for shooting images; the node part of the steel structural member refers to the connecting part of different members in the steel structure, which usually has welds and bolt connections; multi-angle shooting refers to shooting the same node part from different directions and angles; multi-angle original images refer to the initial images obtained by multi-angle shooting without processing; fusion refers to the integration and processing of multiple original images to eliminate overlapping information and retain key details; the target image refers to the image obtained after fusion, which fully presents the details of the node part; the weld contour refers to the shape and lines of the weld edge; the bolt spacing refers to the distance between two adjacent bolts in the node part.

[0066] In the examples of the present application, the shooting object and device are first determined by step S101. In building construction, the node part of the steel structural member to be measured (such as the connecting node of the column and beam) is clearly defined, and an appropriate image acquisition device (such as an industrial camera with anti-backlight function) is selected to ensure that the device resolution meets the requirements of capturing the details of the weld contour and bolt spacing. For example, for the welds (about 5-10 mm wide) and bolts (about 12-20 mm in diameter) of the node part, a camera with a resolution of not less than 5 million pixels is selected to ensure clear details.

[0067] Then, multi-angle shooting is performed. The node part is shot from different angles (such as front, left 45°, right 45°, top view, and bottom view, usually 4-8 angles are selected) to ensure that each angle covers a different area of the node, and the images of adjacent angles have some overlap (about 30%-50% overlap rate) to provide a matching basis for subsequent fusion. For example, for node A (the connecting part of the column and beam), it is shot from the front, left 30°, right 30°, top view, and bottom view. Each image contains part of the weld and bolt, and about 40% of the area of adjacent images overlaps.

[0068] Finally, the original images are fused to obtain the target image. Image stitching technology (such as an algorithm based on SIFT feature point matching) is used to process the multi-angle original images. First, feature points (such as the inflection points of the weld edge and the edge corner points of the bolt) are extracted from each image. The position relationship between the images is determined through feature point matching, and then the overlapping areas are fused at the pixel level (to eliminate the stitching marks). Finally, a target image that fully covers the node part is generated. This image clearly presents the complete contour of the weld and the relative positions of all bolts (i.e., the bolt spacing). For example, for the 5 original images of node A, 10-15 feature points of the weld and 8-12 feature points of the bolt are extracted from each image. After matching, the image stitching order is determined, and the target image obtained after fusion has continuous and unbroken edge lines of the weld, and the positions of all bolts are clear and distinguishable, so that the distance between adjacent bolts can be directly observed.

[0069] In practical application, during the steel structure construction stage of a certain building project, the construction personnel need to collect images of the connection joint (node M) of the steel column and the steel beam before measurement. An E brand industrial camera (resolution 6 million pixels) is selected as the image collection equipment, and 5 angles are selected around the node M for shooting, including the front, left 45°, right 45°, top view, and bottom view. One image is shot at each angle, and 5 original images are obtained (the imaging area of node M in each image accounts for about 60% of the image, and the overlapping area of adjacent angle images accounts for about 45%). Subsequently, the 5 original images are fused through image processing software: first, the edge feature points of the weld (a total of 62) and the corner feature points of the bolt (a total of 45) in each image are identified, the matching feature points are determined to splice the coordinates between the images, and then the pixels in the overlapping area are smoothed, and finally the target image is generated. The target image fully presents the overall appearance of node M, the contour line of the weld is continuous and clear, the positions of the 8 bolts are clear, and the distance between adjacent bolts can be intuitively distinguished.

[0070] The above S101 overall scheme ensures that all key details of the node part are captured through multi-angle shooting, avoiding information omission caused by single-angle shooting; the multi-image fusion technology integrates multiple original images into a complete target image, eliminating the perspective limitations of multi-angle shooting, making the details of the weld and the bolt clearly presented in the same image, providing a comprehensive and accurate image basis for subsequent feature comparison using a twin neural network, and ensuring that the subsequent comparison process can be based on complete visual information.

[0071] S102, calling a twin neural network to compare the weld contour and bolt spacing in the target image with the geometric features of the corresponding node part in the preset BIM model respectively, to obtain weld contour difference information and bolt spacing difference information, wherein the corresponding node part refers to the model node part in the preset BIM model that corresponds to the actual node part of the steel structure member photographed by the target image, that is, the target image photographs a specific node of a certain steel structure member, and the complete model of the member is designed in advance in the preset BIM model, wherein the model node completely matches the actual photographed node in position and function.

[0072] Optionally, the step S102 of calling a twin neural network to compare the weld contour and bolt spacing in the target image with the geometric features of the corresponding node part in the preset BIM model respectively to obtain weld contour difference information and bolt spacing difference information comprises the following steps:

[0073] In step 1021, the edge line information of the weld contour and the standard edge line information in the preset BIM model are respectively subjected to layer-by-layer feature processing through a first subnetwork of the twin neural network to obtain first local features and second local features. The coordinate type intermediate features and the angle type intermediate features of each point in the first local features and the second local features are respectively subjected to difference calculation to obtain a first feature vector.

[0074] In step 1021, the edge line information of the weld contour and the standard edge line information in the preset BIM model are respectively subjected to layer-by-layer feature processing through a first subnetwork of the twin neural network to obtain first local features and second local features. The coordinate type intermediate features and the angle type intermediate features of each point in the first local features and the second local features are respectively subjected to difference calculation to obtain a first feature vector.

[0075] The coordinate type intermediate features are specifically intermediate data formed by recording the horizontal coordinate values and the vertical coordinate values of a plurality of point positions on a continuous line segment at a preset interval. The intermediate data directly reflect the spatial position and morphological details of the continuous line segment and are the basis for subsequent feature comparison. For example, three points are selected on a continuous line segment, and the coordinates of the three points are (x1, y1), (x2, y2) and (x3, y3). The set of the coordinates constitutes the coordinate type intermediate features of the line segment. The horizontal reference direction is a unified horizontal reference direction used when processing the target image and the preset BIM model. The horizontal reference direction is usually a fixed and globally applicable horizontal direction. For example, the bottom horizontal line of the target image (such as the edge of a component kept horizontal during shooting) or the global horizontal coordinate axis in the preset BIM model (such as the positive direction of the X-axis in the model) is used as the reference, and all angle calculations are performed with reference to the direction.

[0076] Step 1022, respectively, performing layer-by-layer feature processing on the numerical information of the bolt spacing and the standard bolt spacing numerical information in the preset BIM model through a second subnetwork of the twin neural network, to obtain a first spacing length feature and a second spacing length feature, and performing difference calculation on the first spacing length feature and the second spacing length feature to output a second feature vector. The spacing length value is the core component of the numerical information and is the key data for describing the bolt spacing, but the numerical information is a more general concept, which may include not only the spacing length value (such as 100 mm) of adjacent bolts, but also auxiliary numerical information related to the spacing (such as the bolt number corresponding to the spacing and the position sequence number of the spacing in the arrangement sequence, etc.). The spacing length value only refers to the actual distance value between two adjacent bolts and is a direct quantification of the physical length of the spacing.

[0077] In step 1022, the process may specifically include the following steps: through the identification layer of the second subnetwork, identifying the position information in the numerical information of the bolt spacing and the position information in the standard bolt spacing numerical information, respectively, based on the position information, pairing two adjacent bolts to form a spacing unit, wherein the spacing unit contains the position identification of the two adjacent bolts and the corresponding spacing length value; through the first association layer of the second subnetwork, associating the position identification and the corresponding spacing length value of the two adjacent bolts in each spacing unit to form a numerical intermediate feature; through the calculation layer of the second subnetwork, arranging the spacing units according to the bolt arrangement sequence on the steel structure member to form a bolt arrangement sequence, and calculating the proportional relationship between the continuous spacing length values in the bolt arrangement sequence to form a relationship intermediate feature; through the second association layer of the second subnetwork, associating the numerical intermediate features and the relationship intermediate features of the numerical information of the bolt spacing and the standard bolt spacing numerical information, respectively, to obtain a first spacing length feature and a second spacing length feature.

[0078] Step 1023, based on the output layer of the twin neural network, determining the weld contour difference information based on the first feature vector and determining the bolt spacing difference information based on the second feature vector.

[0079] The step 1023 can specifically include the following processes: calculating the difference between the horizontal coordinate value and the vertical coordinate value between each actual point in the first local feature and the corresponding standard point in the second local feature to obtain the coordinate value difference; calculating the difference between the angle value in the first local feature and the angle value in the second local feature, and the difference between the respective angle change values between the corresponding adjacent points in the first local feature and the second local feature to obtain the angle value difference; associating the coordinate value difference and the angle value difference according to the order of the continuous line segments, and integrating into the welding seam contour difference information containing the position deviation and the direction deviation of each continuous line segment; calculating the difference between the first interval length feature and the second interval length feature in the second feature vector to obtain the basic interval difference, and calculating the difference between the proportional relationship in the first interval length feature and the proportional relationship in the second interval length feature based on the continuous interval units in the bolt arrangement sequence to obtain the proportion difference; associating the basic interval difference and the proportion difference according to the order of the bolt arrangement sequence, and integrating into the bolt interval difference information containing the length deviation of each interval unit and the sequence proportion deviation.

[0080] The coordinate value difference refers to the horizontal coordinate value difference (such as Δx=xactual-xstandard) and the vertical coordinate value difference (such as Δy=yactual-ystandard) between the actual point on the continuous line segment in the target image and the corresponding standard point in the preset BIM model, which can be understood as a specific and local numerical difference. The position deviation is the overall situation of the entire continuous line segment deviating from the standard line segment in the spatial position after integrating the coordinate value differences in the order of the continuous line segment (such as the line segment deviating to the left or upward as a whole). The position deviation is calculated based on the coordinate value difference, but the coordinate value difference is a numerical result of a local point, and the position deviation is a comprehensive description of the overall position deviation of the line segment. The angle value difference includes the angle difference (such as Δθ=θactual-θstandard) between the line connecting the two end points of the continuous line segment and the horizontal reference direction, and the angle change difference (such as Δα=αactual-αstandard) between the positions of adjacent points, which is a specific angle value difference. The trend deviation is the overall situation of the entire continuous line segment deviating from the standard line segment in the extension direction (trend) after integrating the angle value differences (such as the line segment deviating upward with an excessively large inclination angle as a whole). The trend deviation is calculated based on the angle value difference, but the angle value difference is a numerical result of a local angle, and the trend deviation is a comprehensive description of the overall trend deviation of the line segment. The length deviation of the spacing unit is the difference between the actual measured length value of the bolt spacing in each spacing unit (formed by a pair of adjacent two bolts, including the position identification of the two bolts and the corresponding length value of the spacing) and the standard length value of the corresponding spacing unit in the preset BIM model, which directly reflects the deviation degree of the actual length of a single adjacent bolt pair from the standard length. The sequence proportion deviation refers to the difference between the proportional relationship between the actual length values of the continuous multiple spacing units in the bolt arrangement sequence (a plurality of spacing units arranged according to the arrangement order of the bolts on the steel structure member) and the proportional relationship between the standard length values of the corresponding continuous spacing units in the preset BIM model, which reflects the deviation degree of the overall arrangement proportion of the multiple adjacent bolt spacings from the standard proportion and embodies the overall coordination deviation of the bolt arrangement.

[0081] In the above scheme, the twin neural network refers to a model containing two subnetworks with the same structure, which is used to compare similar features; the target image refers to an image containing the weld contour and bolt spacing of the node position; the weld contour refers to the edge line shape of the weld; the bolt spacing refers to the distance between two adjacent bolts; the preset BIM model refers to a pre-constructed building information model containing the standard geometric features of the node position; the geometric feature refers to the feature data describing the shape and position in the model; the first subnetwork refers to a feature related to the weld contour; the second subnetwork refers to a feature related to the bolt spacing; the first local feature and the second local feature refer to the feature data of the weld contour in the target image and the BIM model, respectively; the first spacing length feature and the second spacing length feature refer to the feature data of the bolt spacing in the target image and the BIM model, respectively; the feature vector is a vector form of integrated feature data; the weld contour difference information refers to the deviation data of the actual weld from the standard feature; and the bolt spacing difference information refers to the deviation data of the actual bolt spacing from the standard feature.

[0082] In the examples of the present application, the weld contour features are processed by step 1021. In the first step, the recognition layer of the first sub-network uses an edge detection algorithm to extract the edge line information of the weld in the target image and the standard edge line information in the BIM model, and identifies feature points with a bending degree exceeding a preset threshold (for example, the preset bending degree threshold is 30°, and the formula for calculating the curvature is the bending degree of the curve at a certain point, which is simplified here as the angle between the adjacent 3-point connecting line being less than 30°, which is determined as bending), and determines the straight line segment between adjacent feature points as a continuous line segment. For example, the weld edge of node M has 4 feature points (P1, P2, P3, P4), among which the bending degrees of the line segments between P1 and P2, P2 and P3, and P3 and P4 are all less than 30°, thus forming 3 continuous line segments (S1: P1-P2, S2: P2-P3, S3: P3-P4). In the second step, the feature extraction layer selects point positions on each continuous line segment at a preset interval, records the X-axis and Y-axis coordinates (units: millimeters) of each point, and forms the coordinate class intermediate features. For example, the 10 point coordinates on S1 are (x1, y1), (x2, y2)…(x10, y10), among which x1=100, y1=200; x2=105, y2=200, etc. These coordinates form the coordinate class intermediate features of S1. In the third step, the calculation layer calculates the angle between the connecting line of the two end points of each continuous line segment and the horizontal reference direction (where (x1, y1) and (x2, y2) are the coordinates of the two end points of the line segment) by the formula θ=arctan[(y2-y1) / (x2-x1)], and calculates the angle change Δθ=|θafter-θbefore| of the connecting line of adjacent points (where θbefore is the angle between the connecting line of the previous two points and the horizontal line, and θafter is the angle between the connecting line of the latter two points and the horizontal line), and forms the angle class intermediate features. For example, the two end points of S1 are P1 (100, 200) and P2 (150, 250), then y2-y1=50, x2-x1=50, θ=arctan(50 / 50)=45°; the θbefore of the connecting line of adjacent points (x1, y1) and (x2, y2) on S1 is 45°, the θafter of the connecting line of (x2, y2) and (x3, y3) is 47°, then Δθ=|47°-45°|=2°; in the fourth step, the feature association layer and difference calculation, the feature association layer associates the coordinate class intermediate features and the angle class intermediate features of each continuous line segment in point order, obtains the first local feature (target image weld feature) and the second local feature (BIM model standard weld feature), and calculates the coordinate difference (Δx=xactual-xstandard, Δy=yactual-ystandard) and the angle difference (Δθangle=θactual-θstandard, Δθchange=Δθactual-Δθstandard) of the corresponding points, and integrates these differences in line segment order into the first feature vector.For example, the actual coordinates of a point in S1 are (102, 201), and the standard coordinates are (100, 200), so Δx = 2 and Δy = 1; the actual angle of S1 is 45°, and the standard angle is 43°, so Δθ = 2°. These data form part of the first feature vector in sequence.

[0083] Next, the bolt spacing feature is processed by step 1022. In the first step, a recognition layer is processed, the recognition layer of the second sub-network uses a target detection algorithm to identify the position information of the bolts in the target image and the standard position information of the bolts in the BIM model, and based on the principle of proximity (for example, two bolts with a distance less than 100 mm are determined to be adjacent), adjacent bolts are paired to form spacing units (format: "position identifier: spacing length"), wherein the spacing length is calculated by the formula , wherein, is the spacing length, and (x1, y1) and (x2, y2) are the coordinates of the centers of adjacent bolts. For example, node M has 8 bolts (B1-B8), and after recognition, 7 spacing units are formed, in which the spacing L between B1 (200, 300) and B2 (250, 300) is 50 mm, and this unit is recorded as "B1-B2: 50 mm"; in the second step, a first association layer is processed, the first association layer corresponds the position identifier and the spacing length of each spacing unit one by one (for example, "B1-B2" corresponds to "50 mm"), forming a numerical class intermediate feature. For example, the numerical class intermediate feature of the 7 units is "B1-B2: 50 mm; B2-B3: 52 mm; …; B7-B8: 49 mm"; in the third step, a calculation layer is processed, the calculation layer arranges the spacing units according to the bolt arrangement order (for example, according to the X-axis coordinates from small to large, B1 to B8 are arranged in turn), calculates the length ratio R of the continuous spacing units in the sequence (wherein Lfront is the spacing length of the previous unit, and Lback is the spacing length of the next unit), and forms a relationship class intermediate feature. For example, the ratio R of B1-B2 (50 mm) and B2-B3 (52 mm) in the sequence is 50 / 52 ≈ 0.96, the ratio R of B2-B3 and B3-B4 (51 mm) is 52 / 51 ≈ 1.02, etc.; in the fourth step, a second association layer and a difference calculation are performed, the second association layer associates the numerical class intermediate feature and the relationship class intermediate feature of each spacing unit (for example, "B1-B2: 50 mm, and the ratio of B2-B3 is 0.96"), to obtain a first spacing length feature (target image bolt feature) and a second spacing length feature (BIM model standard bolt feature); and then the spacing difference (ΔL = Lactual - Lstandard) and the ratio difference (ΔR = Ractual - Rstandard) of the two are calculated, and integrated into a second feature vector. For example, the standard spacing of B1-B2 is 52 mm, so ΔL = 50-52 = -2 mm; and the standard ratio is 0.98, so ΔR = 0.96-0.98 = -0.02. These data form part of the second feature vector in sequence.

[0084] Finally, the difference information is generated by step 1023, the output layer receives the first feature vector and the second feature vector, based on the first feature vector, the coordinate value difference of each continuous line segment (such as Δx, Δy of S1, Δx, Δy of S2, etc.) and the angle value difference (such as Δθ angle, Δθ change of S1, Δθ angle, Δθ change of S2, etc.) are associated in sequence according to the line segment, and integrated into the weld contour difference information containing the position deviation (represented by Δx, Δy) and the trend deviation (represented by Δθ angle, Δθ change) of each line segment. For example, the position deviation of S1 is Δx = 2 mm, Δy = 1 mm, and the trend deviation is Δθ angle = 2°, Δθ change = 0.5°, and these data are associated to become part of the weld contour difference information. At the same time, based on the second feature vector, the basic pitch difference ΔL and the proportion difference ΔR of each pitch unit are associated in sequence according to the bolt arrangement sequence, and integrated into the bolt pitch difference information containing the length deviation (represented by ΔL) and the sequence proportion deviation (represented by ΔR) of each unit. For example, the length deviation of B1-B2 is -2 mm, and the proportion deviation is -0.02, and these data are associated to become part of the bolt pitch difference information.

[0085] In practical application, taking the steel structure node M (column and beam connection part) in a certain building project as an example, the first sub-network identifies 4 feature points (P1-P4) on the weld edge, forming 3 continuous line segments (S1-S3). S1 is 50 mm long, 10 points are selected at 5 mm intervals, with coordinates (100, 200), (105, 200)…(150, 200); the angle θ between the two end points of S1 is calculated by the formula θ = arctan[(y2-y1) / (x2-x1)] = 0° (horizontal line segment), and the angle change Δθ between adjacent points is 0°. The standard coordinates of S1 in the BIM model are (100, 200), (105, 199)…(150, 199), the standard angle θ is 0°, and Δθ is 0°. After association, the coordinate difference Δx = 0 mm, Δy = 0-1 mm, the angle difference Δθ angle = 0°, Δθ change = 0° between the first local feature and the second local feature are calculated, and integrated into the first feature vector; the second sub-network identifies 8 bolts (B1-B8) of node M, arranged in sequence along the X axis as B1 (200, 300) to B8 (550, 300), and the formula Seven spacing units are calculated: B1-B2=50mm, B2-B3=52mm, B3-B4=51mm, B4-B5=50mm, B5-B6=53mm, B6-B7=51mm, B7-B8=49mm; the continuous unit ratio is 50 / 52≈0.96, 52 / 51≈1.02, 51 / 50≈1.02, 50 / 53≈0.94, 53 / 51≈1.04, 51 / 49≈1.04. The standard spacing in the BIM model is 52mm, 52mm, 50mm, 50mm, 52mm, 52mm, 50mm, and the standard ratio is 1.00, 1.04, 1.00, 1.00, 1.00, 1.04. After correlation, ΔL is calculated as -2mm, 0mm, 1mm, 0mm, 1mm, -1mm, -1mm, and ΔR is -0.04, -0.02, 0.02, -0.06, 0.04, -0.00, respectively, which are integrated into the second feature vector; the output layer correlates the coordinate difference and angle difference of the first feature vector in the order of S1-S3 to obtain the weld contour difference information (e.g., the position deviation of S1 is Δy=0-1mm, and there is no deviation in the strike direction); and the ΔL and ΔR of the second feature vector are correlated according to the bolt sequence to obtain the bolt spacing difference information (e.g., the length deviation of B1-B2 is -2mm, and the ratio deviation is -0.04).

[0086] The above S102 overall scheme realizes accurate extraction and comparison of weld contour and bolt spacing features through hierarchical processing of the twin neural network. Among them, the subdivision processing of the coordinate and angle features of the weld contour and the length and ratio features of the bolt spacing ensures the comprehensiveness of the difference information; the automatic correlation and calculation of the sub-network reduces manual intervention and improves the comparison efficiency; and the finally obtained weld contour difference information and bolt spacing difference information clearly reflect the deviation between the actual construction and the BIM model, providing detailed and reliable microscopic feature deviation data for subsequent quality evaluation.

[0087] S103, using a phased array microwave radar array to perform multi-angle scanning on the overall surface of the steel structure member, measuring the distance between the radar in the phased array microwave radar array and the overall surface during the scanning process to form a plurality of sets of distance data.

[0088] In the above scheme, the phased array microwave radar array refers to a device composed of multiple microwave radar units arranged in a certain manner, which can realize beam direction regulation by controlling the phase of microwave emission / reception of each unit; the overall surface of the steel structure member refers to the entire external surface of the member exposed to the outside; multi-angle scanning refers to scanning the overall surface from different directions and angles; and the plurality of sets of distance data refers to a set of distance values measured by different scanning angles and different radar units, each set of data containing a measurement position identifier and a corresponding distance value.

[0089] In the examples of the present application, the scanning range and parameters are first determined by step S103. According to the size of the steel structural member, the scanning coverage range of the phased array microwave radar array is set to ensure that the whole surface of the member can be covered. At the same time, the scanning angle interval, the microwave signal frequency and the sampling frequency are set. For example, for a steel column member with a length of 3 m, the scanning angle is set to 0° (front), 30°, 60°…330°, a total of 12 angles, and 16 units of the radar array work synchronously at each angle.

[0090] Then, multi-angle scanning is performed. The phased array microwave radar array controls the phase difference of each radar unit by beamforming technology to make the microwave beam point to the preset angle, and then emits the microwave signal from each set angle to the whole surface of the member in turn. After the signal contacts the surface, it is reflected, and the radar array receives the reflected signal. By calculating the signal propagation time t, the formula is used, where d is the distance between the radar and the surface, and c is the speed of light. For example, at an angle of 0°, 16 radar units emit signals to the front of the steel column, and after receiving the reflected signals, the distance values of 1.2 m, 1.21 m…1.23 m are calculated. The measurement results of all radar units at each scanning angle are recorded in the format of “angle-radar unit number-distance value”, and the data of all angles are integrated to form multiple sets of distance data. For example, 12 angles x 16 units get a total of 192 sets of data, and each set of data is like “30°-unit 5-1.18 m”, which means that the distance measured by the 5th radar unit at an angle of 30° is 1.18 m. These data will be used for subsequent three-dimensional reconstruction.

[0091] In actual application, in a steel structure plant construction, a steel beam member C (length 6 m, height 0.8 m, width 0.3 m) is taken as an example, a phased array microwave radar array composed of 20 radar units is selected, the scanning angle is set to 0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°, a total of 8 angles, the microwave frequency is 24 GHz, and the sampling frequency is 50 times per second. During scanning, the radar array emits microwave signals to the whole surface of the steel beam C (including the upper and lower flanges, the web and the connecting nodes) from 8 angles in turn, and 20 units measure synchronously at each angle: the distance value range measured at an angle of 0° (facing the web) is 1.5-1.55 m, the distance value range measured at an angle of 45° (side to the flange) is 1.3-1.33 m… All data are recorded in the format of “angle-unit number-distance”, and finally 8 x 20 = 160 sets of distance data are formed, such as “0°-unit 3-1.52 m” “45°-unit 10-1.31 m” and the like. These data will be used for subsequent generation of three-dimensional point cloud grid.

[0092] The overall scheme S103 achieves comprehensive coverage measurement of the overall surface of the steel structural member through multi-angle scanning of the phased array microwave radar array, avoids information blind area of single-angle scanning, uses microwave signal characteristics and beamforming technology to ensure distance measurement accuracy at different positions, forms multiple sets of distance data containing spatial position information of each point on the member surface, provides complete and reliable original data support for generating accurate three-dimensional point cloud grid by subsequent three-dimensional reconstruction technology, and guarantees comprehensiveness and accuracy of macro spatial coordinate measurement.

[0093] S104, combining the multiple sets of distance data by using three-dimensional reconstruction technology to generate a three-dimensional point cloud grid, comparing actual spatial coordinates in the three-dimensional point cloud grid with theoretical spatial coordinates of corresponding node positions in the preset BIM model to obtain spatial coordinate deviation data.

[0094] Optionally, the step S104 of combining the multiple sets of distance data by using three-dimensional reconstruction technology to generate a three-dimensional point cloud grid, comparing actual spatial coordinates in the three-dimensional point cloud grid with theoretical spatial coordinates of corresponding node positions in the preset BIM model to obtain spatial coordinate deviation data, comprises the following steps:

[0095] Step 1041, associating distance information of any node position in the multiple sets of distance data collected at different angles by using three-dimensional reconstruction technology to obtain actual spatial coordinates of the node position in three-dimensional space, integrating actual spatial coordinates of all node positions to form a three-dimensional point set, connecting adjacent node positions in the three-dimensional point set to form a three-dimensional point cloud grid.

[0096] Step 1042, numerically comparing actual spatial coordinates of each node position in the three-dimensional point cloud grid in three-dimensional space with theoretical spatial coordinates of corresponding node positions in the preset BIM model, respectively calculating numerical differences between the actual spatial coordinates and the theoretical spatial coordinates of the node positions in three axial directions, and integrating the numerical differences in the three axial directions to generate spatial coordinate deviation data of the node position.

[0097] In the above scheme, the three-dimensional reconstruction technology refers to a technology of constructing a three-dimensional model of an object by integrating multi-source spatial data; the three-dimensional point cloud grid is a grid-shaped model formed by connecting a large number of three-dimensional space points according to position relationship, wherein each point contains three-dimensional coordinate information; the node position refers to a key connection point on the steel structural member; the actual spatial coordinates are coordinates of the node position in the three-dimensional point cloud grid in three-dimensional space; the theoretical spatial coordinates of the corresponding node position in the preset BIM model are standard three-dimensional coordinates of the node pre-designed in the model; and the spatial coordinate deviation data is a set of differences between the actual spatial coordinates and the theoretical spatial coordinates in three axial directions, reflecting a spatial deviation degree of the node position.

[0098] In the examples of the present application, first, a three-dimensional point cloud grid is generated by step 1041, a plurality of sets of distance data are associated by using the principle of triangulation, distance values of each node position are extracted from distance data at different scanning angles, and three-dimensional coordinates are calculated by combining scanning angle parameters through a formula 、 、 , wherein d is the distance value, is the horizontal rotation angle, is the vertical inclination angle, actual spatial coordinates of all node positions are integrated into a three-dimensional point set, and adjacent node positions are connected by a point cloud stitching algorithm (such as the ICP algorithm) to form a three-dimensional point cloud grid.

[0099] Secondly, spatial coordinate deviation data is calculated by step 1042, actual spatial coordinates of each node position are extracted from the three-dimensional point cloud grid, theoretical spatial coordinates of the corresponding node are obtained from the preset BIM model, and numerical differences of three axial directions are calculated respectively 、 、 , and the three differences are integrated into the spatial coordinate deviation data of the node.

[0100] In actual application, in the A project, the node group (including nodes M, N, O, etc.) of the steel structure member D is taken as an example, 160 sets of distance data are processed by using the three-dimensional reconstruction technology, and two sets of key data of node N in 8 scanning angles are taken as an example, which are angle 1, wherein the horizontal rotation angle the vertical inclination angle the measured distance is calculated by the formula 、 、 , the three-dimensional coordinates 、 、 are calculated, angle 2, wherein the horizontal rotation angle the vertical inclination angle the measured distance is calculated 、 、 , after the comprehensive 8 angle data are optimized by the fusion algorithm, the actual spatial coordinates of node N are (2.30, 1.10, 0.50), and the coordinates of all nodes are obtained by the same method and integrated into a three-dimensional point set, and the three-dimensional point cloud grid is formed by connecting adjacent nodes (the distance less than 0.5 m is determined as adjacent) through the ICP algorithm; the actual coordinates (2.30, 1.10, 0.50) of node N are extracted, compared with the theoretical coordinates (2.32, 1.08, 0.50) of node N in the preset BIM model, and , , , the spatial coordinate deviation data of node N is integrated, and the same operation is performed on nodes M, O, etc., and finally the spatial coordinate deviation data of all nodes is obtained.

[0101] The above S104 overall scheme realizes the visual presentation of the spatial form of the steel structure member by converting multiple sets of distance data into a three-dimensional point cloud grid through three-dimensional reconstruction technology; the deviation of each node position in three axial directions is accurately quantified through the comparison of actual spatial coordinates and theoretical coordinates, and the spatial coordinate deviation data formed reflects the overall spatial installation accuracy of the member; the entire process converts scattered distance data into systematic three-dimensional models and deviation information, providing key data support in the macro spatial dimension for subsequent integration of microscopic feature differences and generation of a comprehensive quality evaluation report.

[0102] S105, correlate the spatial coordinate deviation data, the weld contour difference information and the bolt spacing difference information to obtain comprehensive deviation data, and fuse the comprehensive deviation data with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structure member.

[0103] Optionally, the step S105 of correlating the spatial coordinate deviation data, the weld contour difference information and the bolt spacing difference information to obtain comprehensive deviation data, and fusing the comprehensive deviation data with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structure member comprises the following steps:

[0104] Step 1051, integrating the spatial coordinate deviation, the weld contour difference and the bolt spacing difference of the same node position of the steel structure member to form comprehensive deviation data representing the overall deviation value and distribution of the steel structure member.

[0105] Step 1052, correlating the comprehensive deviation data with the actual spatial coordinates of the corresponding node position in the three-dimensional point cloud grid to form a three-dimensional model containing deviation information, wherein the deviation information specifically refers to the three types of specific deviation content contained in the comprehensive deviation data, i.e. the difference details of the same node position in the spatial coordinate, the weld contour and the bolt spacing from the preset standard.

[0106] Step 1053, based on the three-dimensional model, classifying and arranging the comprehensive deviation data in the same installation area of the steel structure member according to the installation area of the steel structure member as a whole, and recording the overall deviation value and distribution of all the comprehensive deviation data in each installation area, to form an installation quality comprehensive evaluation report of the steel structure member.

[0107] In the above scheme, the spatial coordinate deviation data refers to the difference between the actual coordinates and the theoretical coordinates of the node position in three-dimensional space; the weld contour difference information refers to the position and orientation deviation of the actual weld edge and the standard edge; the bolt spacing difference information refers to the length and proportion deviation of the actual bolt spacing and the standard spacing; the comprehensive deviation data is the overall deviation record formed by integrating the spatial coordinate deviation, the weld contour difference, and the bolt spacing difference of the same node; the three-dimensional point cloud grid is a grid model formed by connecting the three-dimensional coordinates of the nodes; the corresponding node position refers to the specific node associated with the deviation data in the three-dimensional point cloud grid; the installation area refers to the construction area divided according to the component structure, such as the top, middle, and bottom; the installation quality comprehensive evaluation report is a summary document recording the deviation distribution and overall quality of each area.

[0108] In the example of the present application, the comprehensive deviation data is integrated by step 1051, and a data correlation algorithm (such as node ID matching) is used to bind the three types of deviation data of the same node position (such as the spatial coordinate deviation, the weld contour difference, and the bolt spacing difference of node N are all associated with the "node N" identifier), and the comprehensive deviation data is integrated in the format of "node ID-spatial coordinate deviation-weld contour difference-bolt spacing difference" to form the comprehensive deviation data. For example, the comprehensive deviation data of node M is "node M: spatial coordinate deviation (-0.02m, 0.02m, 0m); weld contour difference (position deviation 1-2mm, orientation deviation 1-3°); bolt spacing difference (length deviation -2-+1mm, proportion deviation -0.03-+0.01)".

[0109] The three-dimensional model containing deviation information is generated by step 1052, and the node ID in the comprehensive deviation data is associated with the actual spatial coordinates of the corresponding node in the three-dimensional point cloud grid using coordinate matching technology (such as the ID of node M corresponds to the node with coordinates (1.18m, 0.32m, 0m) in the three-dimensional point cloud grid), so that the three-dimensional coordinates of each node are accompanied by its comprehensive deviation data, forming a visual three-dimensional model (the deviation size can be marked by color, such as red representing larger deviation). For example, the comprehensive deviation data of node M is marked beside its coordinates in the three-dimensional model, and clicking on the node can view the detailed spatial, weld, and bolt deviation.

[0110] The evaluation report is generated by step 1053, the installation area is divided according to the component structure (such as the top area, middle area, and bottom area of the steel column), and the regional clustering algorithm is used to classify the node comprehensive deviation data in the same area; the deviation distribution of each area is counted (such as the number of nodes with large deviation and the main deviation type), and the regional quality state is described in combination with the engineering quality standard, and finally the report containing the summary of the deviation of each area and the overall quality evaluation is formed. For example, the deviation in the top area is mainly the bolt spacing deviation, and the deviation in the middle area is mainly the spatial coordinate deviation, which are recorded and adjustment suggestions are given in the report.

[0111] In practical applications, in the construction of steel structure project A, when detecting the installation quality of the node N (the connection node of steel column and steel beam) of component D, the image acquisition device obtains the coordinates P1 (100mm, 200mm) and P2 (150mm, 250mm) of the two endpoints of the continuous line segment S1 of the weld edge, and calculates the actual included angle of 43° by using the formula However, the standard included angle of the line segment in the preset BIM model is 43°, so the angle deviation is ; the adjacent point A (105mm, 200mm) and the point B (110mm, 201mm) are selected on S1, and the included angle of the line connecting the two points is about 11.3°, the included angle of the line connecting the point B (110mm, 201mm) and the point C (115mm, 201mm) is about 11.8°, and the actual angle change is 0.5°. Since the standard angle change at this position in the BIM model is 0°, the angle change deviation is 0.5°; the target detection recognizes the bolts B1 (200mm, 300mm) and B2 (250mm, 300mm), and the actual distance of 50mm is calculated by using the formula However, the standard distance in the BIM model is 52mm, and the length deviation is 50-52=-2mm; the bolts of the node N are arranged as B1-B2-B3, the actual distance of B2-B3 is 52mm, the actual proportion is , the standard distance of B1-B2 in the BIM model is 52mm, the standard distance of B2-B3 is 53mm, and the standard proportion is The proportion deviation is 0.96-0.98=-0.02, and these data are integrated into the comprehensive deviation data of the node N, which is associated with the three-dimensional point cloud grid and included in the installation quality comprehensive evaluation report of component D.

[0112] The above S105 overall scheme integrates three types of deviation data to form comprehensive deviation data, realizes the association of micro feature deviation and macro spatial deviation, avoids the one-sidedness of evaluation caused by scattered data; the comprehensive deviation data is fused with the three-dimensional point cloud grid to make the deviation information visualized, which is convenient for intuitive understanding of the spatial distribution of the deviation; the data is arranged according to the installation area and a report is generated, which clearly presents the quality state of each area and the overall installation precision, provides a comprehensive and accurate decision basis for construction adjustment, and ensures the systematization and reliability of the steel structure installation quality evaluation.

[0113] The following is a complete example for steps S101-S105 Figure 2 ​As shown, firstly, in the steel structure construction of Project A, a 6-megapixel industrial camera (resolution 3072×2048) was used to take pictures of the steel column and steel beam connection node M from five angles: front, left 45°, right 45°, top view, and bottom view. Adjacent images overlapped by 40% (e.g., the overlap area between the left 45° image and the front image accounted for 40% of their respective areas). After importing the five original images, the SIFT algorithm was used to extract weld feature points (12 per image, such as P1(100, 200) as the weld inflection point) and bolt feature points (10 per image, such as the center of B1(200, 300)). After matching, the images were stitched together, and the pixels in the overlapping area were weighted and averaged (the weights were distributed according to the distance from the edge) to generate the target image, which clearly showed the three weld contours (S1-S3) and the positions of the eight bolts (B1-B8) of node M.

[0114] Next, the generated target image is input into the Siamese neural network. The first sub-network extracts the coordinates of the two endpoints of weld S1, P1 (100mm, 200mm) and P2 (150mm, 250mm), and uses the formula... Calculate the actual included angle, where , ,have to Compared with the standard angle of 43° for S1 in the BIM model, the angle difference is... Ten points are taken on S1 at 5mm intervals. The difference between the actual coordinates (110mm, 201mm) and the standard coordinates (108mm, 200mm) of the third point is... , This information is integrated into weld contour difference information. The second sub-network identifies bolts B1 (200mm, 300mm) and B2 (250mm, 300mm) using a formula. Calculate the actual spacing, where , ,have to The difference from the standard spacing of 52mm in the BIM model is The ratio of B1-B2 to B2-B3 (actual 52mm) The difference between this and the standard ratio 52 / 53≈0.98 is This information is integrated into bolt spacing difference information.

[0115] Then, a 20-element phased array radar was used to scan the steel column where node M is located, setting 8 angles (0°, 45°…315°) and a microwave frequency of 24GHz (wavelength 12.5mm); at a 30° angle… Next, the 5th unit transmits a signal to the surface of node M, and the reception time is... Using formula Calculate the distance , and 8 angles x 20 units, a total of 160 groups of data (such as "30°-unit 5-1.18m") are measured in the same way, which will be used for three-dimensional reconstruction in step 4.

[0116] Then, the distance data is processed, and the node M is taken as 30° angle distance 1.18m, 60° angle distance 1.22m, combined with , , the formula is used to calculate 30° , , ; after the optimization of 8 angle data, the actual coordinates of node M are (2.30m, 1.10m, 0.50m). The point set is formed by integrating all node coordinates, and the ICP algorithm is used to connect adjacent points (distance <0.5m) to generate a three-dimensional point cloud grid; compared with the theoretical coordinates (2.32m, 1.08m, 0.50m) of node M in the BIM model, it is , , that is, the spatial coordinate deviation data.

[0117] Finally, the bolt spacing difference information and the spatial coordinate deviation data are integrated, and the comprehensive deviation of node M is "space (-0.02m, 0.02m, 0m); weld ; bolt ", which is associated with its actual coordinates (2.30m, 1.10m, 0.50m) and marked in yellow (moderate deviation) in the three-dimensional model. After regional division, the bolt deviation and spatial deviation of the top area (including nodes M and N) are counted, and finally a report is generated, suggesting that the top bolt spacing and the middle node spatial position should be adjusted first. All data are used to evaluate the quality of node installation.

[0118] Figure 3 is a structural schematic diagram of a specific embodiment of a building construction measurement system provided by the embodiment of the application, referring to Figure 3 , the system can include:

[0119] The acquisition module 31 is configured to capture, by an image acquisition device, a node part of a steel structure component at multiple angles during the building construction process, to obtain multiple-angle original images, and to fuse all the original images to obtain a target image containing a weld contour and a bolt spacing at the node part.

[0120] The first comparison module 32 is configured to call a twin neural network to compare the weld contour and the bolt spacing in the target image with the geometric features of the corresponding node part in a preset BIM model, respectively, to obtain weld contour difference information and bolt spacing difference information.

[0121] The measurement module 33 is configured to perform multi-angle scanning on the overall surface of the steel structure member by using a phased array microwave radar array, and measure the distance between the radar in the phased array microwave radar array and the overall surface during the scanning to form a plurality of sets of distance data.

[0122] The second comparison module 34 is configured to combine the plurality of sets of distance data by using a three-dimensional reconstruction technology to generate a three-dimensional point cloud grid, compare the actual space coordinates in the three-dimensional point cloud grid with theoretical space coordinates of corresponding node positions in a preset BIM model to obtain space coordinate deviation data.

[0123] The fusion module 35 is configured to associate the space coordinate deviation data, the weld contour difference information and the bolt spacing difference information to obtain comprehensive deviation data, and fuse the comprehensive deviation data with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structure member.

[0124] The building construction measurement system according to the embodiment of the present application is used to implement the building construction measurement method described above, and the specific embodiments of the building construction measurement system can refer to the embodiment part of the building construction measurement method described above, and the specific embodiments can refer to the description of the corresponding embodiment part, which will not be repeated here.

[0125] As Figure 4 described above, the present application further provides an electronic device, which comprises a memory 41 configured to store a computer program, and a processor 42 configured to execute the computer program to implement the steps of any one of the building construction measurement methods described above.

[0126] The present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of any one of the building construction measurement methods described above.

[0127] In an exemplary embodiment, the computer readable storage medium described above can include but is not limited to a U disk, a read-only memory, a random access memory, a mobile hard disk, a magnetic disk or an optical disk, and various media that can store computer programs.

[0128] The embodiment of the present application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the steps in any one of the building construction measurement method embodiments described above.

[0129] Those skilled in the art will further realize that the mere concepts, teachings, and embodiments described herein are merely meant to provide an enabling description of embodiments of the present application and are not intended to limit the scope of the present application. Accordingly, embodiments as described herein contemplate all modifications that come within the scope of the present application.

[0130] The building construction measurement method, system, electronic device and storage medium provided by the present application are described in detail above. The principles and implementation modes of the present application are described by applying specific examples in this paper. The above description of the embodiments is only used to help understand the method of the present application and its core idea. It should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, some improvements and modifications can be made to the present application, and these improvements and modifications also fall within the protection scope of the present application.

Claims

1. A method of construction surveying, characterised in that, The method comprises the following steps: During the construction process, the node part of the steel structure component is photographed by an image acquisition device from multiple angles to obtain original images from multiple angles, and all original images are fused to obtain a target image containing the weld contour and bolt spacing at the node part; The twin neural network is called to compare the weld contour and bolt spacing in the target image with the geometric features of the corresponding node part in the preset BIM model, respectively, to obtain weld contour difference information and bolt spacing difference information; The overall surface of the steel structure component is scanned by a phased array microwave radar array from multiple angles, and the distance between the radar in the phased array microwave radar array and the overall surface is measured during the scanning process to form multiple sets of distance data; The three-dimensional reconstruction technology is used to combine the multiple sets of distance data to generate a three-dimensional point cloud grid, and the actual space coordinates in the three-dimensional point cloud grid are compared with the theoretical space coordinates of the corresponding node position in the preset BIM model to obtain space coordinate deviation data; The space coordinate deviation data, the weld contour difference information and the bolt spacing difference information are associated to obtain comprehensive deviation data, and the comprehensive deviation data is fused with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structure component; The twin neural network is called to compare the weld contour and bolt spacing in the target image with the geometric features of the corresponding node part in the preset BIM model, respectively, to obtain weld contour difference information and bolt spacing difference information, comprising the following steps: The edge line information of the weld contour and the standard edge line information in the preset BIM model are processed by the first subnetwork of the twin neural network layer by layer, respectively, to obtain first local features and second local features, and the coordinates of each point in the first local features and the second local features are calculated to obtain first feature vectors; The numerical information of the bolt spacing and the standard bolt spacing numerical information in the preset BIM model are processed by the second subnetwork of the twin neural network layer by layer, respectively, to obtain first spacing length features and second spacing length features, and the first spacing length features and the second spacing length features are calculated to output second feature vectors; Based on the output layer of the twin neural network, the weld contour difference information is determined based on the first feature vectors, and the bolt spacing difference information is determined based on the second feature vectors; The space coordinate deviation data, the weld contour difference information and the bolt spacing difference information are associated to obtain comprehensive deviation data, and the comprehensive deviation data is fused with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structure component, comprising the following steps: The space coordinate deviation, the weld contour difference and the bolt spacing difference of the same node position of the steel structure component are integrated to form comprehensive deviation data representing the overall deviation value and distribution of the steel structure component; The comprehensive deviation data is associated with the actual spatial coordinates of the corresponding node positions in the three-dimensional point cloud grid to form a three-dimensional model containing deviation information; Based on the three-dimensional model, the comprehensive deviation data in the same installation area is classified and arranged according to the installation area of the steel structure member as a whole, and the overall deviation value and distribution of all the comprehensive deviation data in each installation area are recorded to form an installation quality comprehensive evaluation report of the steel structure member.

2. The method of claim 1, wherein, The first local feature and the second local feature are obtained by performing layer-by-layer feature processing on the edge line information of the weld contour and the standard edge line information in the preset BIM model respectively through the first subnetwork of the twin neural network, including the following steps: Through the identification layer of the first subnetwork, feature points with a bending degree value exceeding a preset bending degree threshold are identified in the edge lines in the edge line information and the edge lines in the standard edge line information respectively, and the edge lines between adjacent feature points are regarded as continuous line segments; Through the feature extraction layer of the first subnetwork, point positions are selected on the continuous line segments according to a preset interval, and the transverse coordinate values and longitudinal coordinate values of all the point positions in the two-dimensional plane are recorded to form coordinate-type intermediate features; Through the calculation layer of the first subnetwork, the included angle value between the line connecting the two end points on the continuous line segment and the horizontal reference direction and the angle change value of the line connecting adjacent point positions on the continuous line segment are calculated to form angle-type intermediate features; Through the feature association layer of the first subnetwork, the coordinate-type intermediate features and the angle-type intermediate features of the edge line information of the weld contour are associated to obtain the first local feature, and the coordinate-type intermediate features and the angle-type intermediate features of the standard edge line information are associated to obtain the second local feature.

3. The method of claim 1, wherein, The first interval length feature and the second interval length feature are obtained by performing layer-by-layer feature processing on the numerical information of the bolt interval and the standard bolt interval numerical information in the preset BIM model respectively through the second subnetwork of the twin neural network, including the following steps: Through the identification layer of the second subnetwork, the position information in the numerical information of the bolt interval and the position information in the standard bolt interval numerical information are identified respectively, and two adjacent bolts are paired based on the position information to form an interval unit, wherein the interval unit contains the position identification and the corresponding interval length value of the two adjacent bolts; Through the first association layer of the second subnetwork, the position identification and the corresponding interval length value of two adjacent bolts in each interval unit are associated to form numerical-type intermediate features; Through the calculation layer of the second subnetwork, the interval units are arranged according to the bolt arrangement sequence on the steel structure member to form a bolt arrangement sequence, and the proportional relationship between the continuous interval length values in the bolt arrangement sequence is calculated to form relationship-type intermediate features; Correlate the numerical value of the bolt spacing and the numerical value of the standard bolt spacing with the relationship class intermediate feature through the second association layer of the second sub-network, to obtain the first and second interval length features.

4. The method of claim 1, wherein, The method comprises the following steps: Calculate the difference between the horizontal coordinate value and the longitudinal coordinate value of each actual point in the first local feature and the corresponding standard point in the second local feature to obtain the coordinate value difference; Calculate the angle value difference between the angle value of the first local feature and the angle value of the second local feature, and the difference between the angle change value of each corresponding adjacent point position in the first local feature and the second local feature to obtain the angle value difference; Correlate the coordinate value difference and the angle value difference in the order of continuous line segments to integrate the welding seam profile difference information containing the position deviation and the trend deviation of each continuous line segment; Calculate the difference between the first interval length feature and the second interval length feature in the second feature vector to obtain the basic interval difference, and calculate the difference between the proportional relationship in the first interval length feature and the proportional relationship in the second interval length feature based on the continuous interval unit in the bolt arrangement sequence to obtain the proportion difference; Correlate the basic interval difference and the proportion difference in the order of the bolt arrangement sequence to integrate the bolt spacing difference information containing the length deviation of each interval unit and the sequence proportion deviation.

5. The method of claim 1, wherein, The method comprises the following steps: Correlate the distance information of any node position in the plurality of distance data collected at different angles by three-dimensional reconstruction technology to obtain the actual spatial coordinates of the node position in three-dimensional space, integrate the actual spatial coordinates of all node positions to form a three-dimensional point set, and connect adjacent node positions in the three-dimensional point set to form a three-dimensional point cloud grid; Numerically compare the actual spatial coordinates of each node position in the three-dimensional point cloud grid in the three-dimensional space with the theoretical spatial coordinates of the corresponding node position in the preset BIM model, respectively calculate the numerical difference between the actual spatial coordinates and the theoretical spatial coordinates of the node position in the three axial directions, and integrate the numerical differences in the three axial directions to generate the spatial coordinate deviation data of the node position.

6. A construction surveying system, characterised in that, The method comprises the following steps: The collection module is used for multi-angle shooting of a node part of a steel structure component through an image collection device in a building construction process, obtaining multi-angle original images, fusing all the original images, and obtaining a target image containing a weld contour and a bolt spacing at the node part; The first comparison module is used for calling a twin neural network to compare the weld contour and the bolt spacing in the target image with geometric features of a corresponding node part in a preset BIM model respectively, and obtain weld contour difference information and bolt spacing difference information; The measurement module is used for multi-angle scanning of an overall surface of the steel structure component by using a phased array microwave radar array, measuring distances between radars in the phased array microwave radar array and the overall surface in a scanning process, and forming a plurality of sets of distance data; The second comparison module is used for combining the plurality of sets of distance data by using a three-dimensional reconstruction technology, generating a three-dimensional point cloud grid, comparing actual space coordinates in the three-dimensional point cloud grid with theoretical space coordinates of a corresponding node position in a preset BIM model, and obtaining space coordinate deviation data; The fusion module is used for associating the space coordinate deviation data, the weld contour difference information and the bolt spacing difference information, obtaining comprehensive deviation data, and fusing the comprehensive deviation data with the three-dimensional point cloud grid to generate an installation quality comprehensive evaluation report of the steel structure component.

7. An electronic device, comprising: Comprise: A memory for storing a computer program; A processor for executing the computer program to realize the steps of the building construction measurement method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the building construction measurement method according to any one of claims 1 to 5.

Images