A method and system for constructing a three-dimensional model of a shield cutter product

Abstract

The present application relates to the technical field of three-dimensional modeling, and particularly relates to a shield cutter product production three-dimensional model construction method and system, which solves the technical problem of poor modeling quality in the prior art. The method comprises the following steps: obtaining standard three-dimensional model data of a shield cutter and historical three-dimensional model data of a plurality of defective products of the shield cutter; performing region identification based on the standard three-dimensional model data and the plurality of historical three-dimensional model data, and dividing the shield cutter into a plurality of regions; performing modeling analysis on point cloud data in the plurality of regions based on the standard three-dimensional model data and the plurality of historical three-dimensional model data, and obtaining scanning modeling parameters of the shield cutter; the scanning modeling parameters comprise marking point layout parameters and scanning point distance parameters; and performing three-dimensional model construction on a to-be-detected product of the shield cutter based on the scanning modeling parameters of the shield cutter, and obtaining three-dimensional model data of the to-be-detected product.

Description

Technical Field

[0001] This invention relates to the field of 3D modeling technology, specifically to a method and system for constructing 3D models for the production of tunnel boring machine tools. Background Technology

[0002] To achieve quality control and fault analysis of tunnel boring machine (TBM) cutters, it is necessary to construct accurate 3D models using reverse engineering techniques to provide data support for defect detection and performance evaluation. Currently, the industry generally adopts a partitioning strategy for reverse modeling of TBM cutters. By processing different areas of the cutter separately, the accuracy of local scanning and the quality of surface fitting can be improved in a targeted manner, effectively controlling the overall splicing error and making the management of complex modeling tasks more efficient.

[0003] However, in existing technologies, 3D modeling typically determines the scanning point spacing parameters based on the overall structural complexity of the tunnel boring machine (TBM) cutter, without fully considering the differences in structural complexity and defect susceptibility in different cutter regions. This standardized parameter setting method leads to inaccurate acquisition of 3D point cloud data in some key areas, resulting in poor modeling quality. This modeling fails to accurately reflect the actual structure and defect status of the cutter, thus failing to provide reliable reference for fault analysis and impacting the effectiveness of quality control and the safety of engineering construction. Summary of the Invention

[0004] To address the technical problem of poor modeling quality in existing technologies, the present invention aims to provide a method and system for constructing three-dimensional models for the production of tunnel boring machine (TBM) cutting tools. The specific technical solution adopted is as follows: This application provides a method for constructing a three-dimensional model for the production of tunnel boring machine cutting tools, including: Acquire standard 3D model data of tunnel boring machine (TBM) cutters and historical 3D model data of multiple defective TBM cutter products; standard 3D model data is used to characterize the standard geometry of TBM cutters; historical 3D model data is used to characterize the geometry of defective TBM cutters actually produced. Based on standard 3D model data and multiple historical 3D model data, the shield cutterhead is divided into multiple regions; these regions include areas with high incidence of defects and areas without high incidence of defects. Based on standard 3D model data and multiple historical 3D model data, point cloud data in multiple regions are modeled and analyzed to obtain the scanning modeling parameters of the tunnel boring machine cutter. The scanning modeling parameters include marker point layout parameters and scanning point distance parameters. Based on the scanning and modeling parameters of the tunnel boring machine cutter, a three-dimensional model of the product to be inspected is constructed, and the three-dimensional model data of the product to be inspected is obtained.

[0005] In one possible implementation, the method includes: For each historical 3D model data, the historical 3D model data is aligned with the standard 3D model data to identify the individual defect areas on the shield cutter corresponding to the historical 3D model data; Based on the individual defect areas on the shield cutter corresponding to each historical 3D model data, the high-incidence areas of defects and the non-defect-high-incidence areas of the shield cutter are determined.

[0006] In one possible implementation, the method includes: For each historical 3D model data, the historical 3D model data is aligned with the standard 3D model data, and the distance from each 3D point in the historical 3D model data to the model surface represented by the standard 3D model data is determined. Based on the three-dimensional points in the historical three-dimensional model data whose distance is greater than a preset distance threshold, the individual defect areas on the shield cutting tool corresponding to the historical three-dimensional model data are determined.

[0007] In one possible implementation, the method includes: The individual defect areas on the shield cutter corresponding to each historical 3D model data are mapped to the model surface represented by the standard 3D model data to obtain the defect coverage density distribution information of the shield cutter. Based on the defect coverage density distribution information of the shield cutter, boundary extraction and surface fitting are performed to obtain the high-incidence area of ​​defects in the shield cutter. Spatial clustering is performed on the areas of the model surface represented by the standard three-dimensional model data, excluding the high-defect areas, to obtain one or more non-defect-high-incidence areas.

[0008] In one possible implementation, the method includes: For each non-defect-prone high-incidence area, the marker point layout parameters within the non-defect-prone high-incidence area are determined based on the non-defect-prone high-incidence area and the defect-prone areas adjacent to the non-defect-prone high-incidence area. For each region, the corresponding scan point distance parameter is determined based on the point cloud data of the region in the standard 3D model data and multiple historical 3D model data.

[0009] In one possible implementation, the method includes: For each non-defect-prone high-incidence area, the marking bias coefficient of the non-defect-prone high-incidence area is determined based on the area of ​​the non-defect-prone high-incidence area, the number of defect-prone high-incidence areas adjacent to the non-defect-prone high-incidence area, and the distance between each three-dimensional point in the non-defect-prone high-incidence area and the corresponding fitting plane of the non-defect-prone high-incidence area; the marking bias coefficient is used to characterize the intensity of the demand for marking points in the non-defect-prone high-incidence area during three-dimensional scanning. The marking point layout parameters in the non-defect-prone area are determined based on the marking bias coefficient.

[0010] In one possible implementation, the method includes: For each region, the scanning attention level of the region is determined based on the region's defect status in multiple historical 3D model data and the region's point cloud distribution in the standard 3D model data; the scanning attention level is used to characterize the intensity of attention to the region during 3D scanning. The scan point spacing parameter corresponding to the region is determined based on the scan attention level.

[0011] In one possible implementation, the method includes: Based on the marker layout parameters, determine the position information of each marker on the product to be inspected; Based on the scanning point distance parameters, perform multi-view 3D scanning on the product to be inspected to obtain point cloud data corresponding to multiple views; Based on the position information of each marker point on the product to be inspected, the point cloud data corresponding to the multiple viewpoints are stitched together to obtain the initial three-dimensional model data of the product to be inspected. For each 3D point in the initial 3D model data, the interference coefficient of the 3D point is determined based on the point cloud data corresponding to the 3D point in each viewpoint; Interference points are filtered based on the interference coefficient of each three-dimensional point in the initial three-dimensional model data to obtain the three-dimensional model data of the product to be tested.

[0012] In one possible implementation, the method further includes: Cluster analysis is performed on the three-dimensional model data of the product to be tested to determine multiple cluster ranges of the product to be tested. For each cluster range, the articulation arm supplementation necessity coefficient of the cluster range is determined based on the point cloud distribution characteristics and grayscale characteristics of the cluster range in the 3D model data; the articulation arm supplementation necessity coefficient is used to characterize the degree of necessity for articulation arm supplementation of the cluster range. Articulated arms are supplemented based on the necessary coefficient for articulated arm supplementation for each cluster range.

[0013] This application provides a three-dimensional model construction system for the production of tunnel boring machine cutting tools, including: The data acquisition unit is used to acquire standard 3D model data of shield tunneling cutters and historical 3D model data of multiple defective shield tunneling cutter products; the standard 3D model data is used to characterize the standard geometric shape of shield tunneling cutters; the historical 3D model data is used to characterize the geometric shape of defective shield tunneling cutters actually produced. The region identification unit is used to identify regions based on standard 3D model data and multiple historical 3D model data, dividing the shield cutter into multiple regions; these regions include defect-prone areas and non-defect-prone areas of the shield cutter. The modeling and analysis unit is used to perform modeling and analysis on point cloud data in multiple regions based on standard 3D model data and multiple historical 3D model data to obtain the scanning modeling parameters of the tunnel boring machine cutter. The scanning modeling parameters include marker point layout parameters and scanning point distance parameters. The model building unit is used to build a 3D model of the shield cutter product to be inspected based on the scanning modeling parameters of the shield cutter, and obtain the 3D model data of the product to be inspected.

[0014] The present invention has the following beneficial effects: Based on the above technical solution, this application can obtain standard 3D model data of the tunnel boring machine (TBM) cutter and historical 3D model data of multiple defective TBM products. Based on the standard 3D model data and multiple historical 3D model data, region identification is performed to divide the TBM cutter into multiple regions, achieving region division based on historical defect patterns. This makes the determination of scanning modeling parameters more targeted. Subsequently, based on the standard 3D model data and multiple historical 3D model data, point cloud data within the multiple regions is modeled and analyzed to obtain the scanning modeling parameters of the TBM cutter. This achieves differentiated settings for marker point layout parameters and scanning point distance parameters, avoiding the problem of inaccurate local point clouds caused by uniform parameters in existing technologies. This ensures that areas with high defect incidence and complex structures can obtain denser and more accurate point cloud data. The final constructed 3D model has higher modeling quality and better reflects the actual state of the cutter, providing a reliable reference for fault analysis by personnel and improving the effectiveness of TBM cutter quality control. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a method for constructing a three-dimensional model for the production of tunnel boring machine cutter products, provided in one embodiment of the present invention. Figure 2 This is a system architecture diagram of a three-dimensional model building system for the production of tunnel boring machine cutter products, provided as an embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method and system for constructing a three-dimensional model for the production of tunnel boring machine cutters according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] In all division and logarithmic operations covered in this application, a smoothing mechanism is employed to prevent computer program crashes or invalid values ​​from being generated due to a zero denominator or a zero input. Specifically, a positive correction factor is superimposed on the denominator term of the division operation or the argument term of the logarithmic function. For example, the value is This ensures the robustness and feasibility of the algorithm under extreme conditions.

[0020] The normalization function mentioned in this application Unless otherwise specified, all values ​​are normalized using maximum and minimum values. The maximum and minimum values ​​are preset empirical extreme values ​​derived from a large amount of historical experimental data. If the calculated result exceeds the [0,1] interval, it is restricted to the [0,1] range by a truncation function (i.e., if the result is less than 0, it is taken as 0, and if it is greater than 1, it is taken as 1) to eliminate the influence of outliers on the evaluation index.

[0021] The following description, in conjunction with the accompanying drawings, details the specific scheme of the method and system for constructing a three-dimensional model for the production of tunnel boring machine tools provided by this invention.

[0022] Please see Figure 1 The diagram illustrates a method flowchart for constructing a three-dimensional model of a tunnel boring machine tool product according to an embodiment of the present invention. The method includes the following steps: Step 101: Obtain standard 3D model data of the tunnel boring machine cutter and historical 3D model data of multiple defective tunnel boring machine cutter products.

[0023] Among them, the standard 3D model data is used to characterize the standard geometry of the shield cutter, which is the ideal state model in the cutter design stage. The historical 3D model data is used to characterize the geometry of the defective shield cutters in actual production, reflecting the defects that occur in the actual use or production process of the cutters.

[0024] In some embodiments, standard 3D model data can be obtained by requesting the original parametric CAD model of the tunnel boring machine (TBM) cutter from the internal design department. This model contains complete design information such as the cutter's design dimensions, structural features, and surface parameters. Historical 3D model data is obtained by collecting TBM cutters that have defects encountered during actual production or use, comprehensively scanning the defective cutters using 3D scanning equipment, and reconstructing the actual 3D model using reverse engineering techniques—this is the historical 3D model data. For example, the standard CAD model and the historical defect reconstruction model can be imported into the same 3D processing software to provide a unified operating environment for subsequent data processing.

[0025] Step 102: Based on standard 3D model data and multiple historical 3D model data, perform region identification and divide the tunnel boring machine cutter into multiple regions.

[0026] The study includes multiple regions encompassing both high-defect areas and non-defect-prone areas for tunnel boring machine (TBM) cutters. This application utilizes statistical analysis of historical defect data to identify areas with concentrated defects and areas with fewer defects on the cutters, providing a basis for subsequent differentiated modeling.

[0027] In some embodiments, each historical 3D model data is first precisely aligned with standard 3D model data. The alignment reference can be selected from unworn, structurally stable features on the tool (such as mounting holes, locating slots, etc.) to ensure accurate positional correspondence between the two in 3D space. Subsequently, by comparing the surface differences between the historical 3D model data and the standard 3D model data, defect areas on each historical 3D model data are identified. Finally, the defect areas of all historical 3D model data are statistically fused to identify high-defect and non-defect-prone areas.

[0028] Step 103: Based on standard 3D model data and multiple historical 3D model data, perform modeling and analysis on point cloud data in multiple regions to obtain the scanning modeling parameters of the tunnel boring machine cutter.

[0029] The scanning modeling parameters include marker placement parameters and scan point spacing parameters. These parameters are key factors affecting the accuracy of 3D scanning. Marker placement parameters determine the location and number of markers during the scanning process. Proper placement of markers ensures accurate stitching of multi-view scanning data. Scan point spacing parameters determine the sampling interval of the 3D scanner, and the spacing directly affects the density and scanning accuracy of the point cloud data.

[0030] In some embodiments, for the marker placement parameters, this application can determine the intensity of demand for markers in non-defect-prone areas by combining factors such as their positional relationship with defect-prone areas and their own structural characteristics, and then determine the number and distribution of markers.

[0031] Regarding the scan point distance parameter, this application can determine the structural complexity and defect risk of each region (including defect-prone and non-defect-prone regions) by combining the region's historical defect situation with the point cloud distribution characteristics in the standard model, and then determine the differentiated scan point distance.

[0032] Step 104: Based on the scanning and modeling parameters of the tunnel boring machine cutter, construct a three-dimensional model of the product to be inspected, and obtain the three-dimensional model data of the product to be inspected.

[0033] In one possible implementation, this application can determine the position information of each marker on the product to be tested based on the marker layout parameters.

[0034] For example, this application can first perform a preliminary scan of the product to be inspected to obtain its approximate surface contour data. Based on the number and distribution requirements of the marker points, candidate positions of the marker points are planned on the preliminary contour data to ensure that the candidate positions do not cover obvious defect areas or are located at abrupt changes in surface curvature. Subsequently, the marker points are pasted on the candidate positions, and the three-dimensional coordinate information of each marker point is recorded as a reference for subsequent point cloud stitching.

[0035] Subsequently, based on the scanning point distance parameters, a multi-view 3D scan is performed on the product to be inspected to obtain point cloud data corresponding to multiple views.

[0036] For example, this application can plan multiple scanning perspectives based on the size and structural features of the product to be inspected, ensuring that each area is covered by at least two perspectives. During the scanning process of each perspective, the 3D scanner is controlled to sample according to the scanning point distance parameters corresponding to that area, while simultaneously capturing the geometric data and marker point information of the surface of the product to be inspected. For example, the scanning perspectives can be set to six basic perspectives: top, bottom, left, right, front, and rear. For areas with complex structures, a 45° tilting perspective can be added to ensure no scanning blind spots. The scanning equipment can be a laser 3D scanner to meet the modeling accuracy requirements of the tunnel boring machine cutter.

[0037] Thus, this application can stitch together point cloud data from multiple perspectives based on the position information of each marker point on the product to be inspected, thereby obtaining the initial three-dimensional model data of the product to be inspected.

[0038] For example, this application can utilize 3D processing software to identify marker points in point cloud data from different viewpoints, calculate the spatial transformation matrix (including translation and rotation parameters) between different viewpoints using the 3D coordinate information of the marker points, and perform coordinate transformation on the point cloud data of all viewpoints according to the transformation matrix, so that the point cloud data of all viewpoints are unified into the same 3D coordinate system. Subsequently, a point cloud fusion algorithm is used to deduplicate and optimize the point cloud data in the overlapping areas to obtain complete initial 3D model data. During the stitching process, the stitching accuracy can be verified by calculating the distance error of the point cloud in the overlapping area. If the error is greater than a preset threshold (e.g., 0.05mm), the transformation matrix is ​​readjusted to ensure that the stitching accuracy meets the requirements.

[0039] Furthermore, this application can also determine the interference coefficient of each three-dimensional point in the initial three-dimensional model data based on the point cloud data corresponding to the three-dimensional point in each viewpoint, and filter interference points based on the interference coefficient of each three-dimensional point in the initial three-dimensional model data to obtain the three-dimensional model data of the product to be tested.

[0040] Interference points mainly originate from light reflection, dust in the scanning environment, and equipment measurement noise, and need to be identified and filtered using the interference coefficient. For example, the interference coefficient satisfies the following formula: in, For three-dimensional points The interference coefficient, To cover this three-dimensional point Number of viewpoints For this three-dimensional point to exist The number of viewpoints corresponding to the point cloud data. For three-dimensional points The distance to the fitted surface corresponding to the initial 3D model data. This represents the maximum distance from all 3D points to the fitted surface corresponding to the initial 3D model data. This is a safety parameter used to correct for fractions with a denominator of 0. Its dimensions are the same as the distance, and its specific value can be determined based on... The value of the value determines the outcome, such as . This is a normalization function (e.g., maximum / minimum normalization) used to map the calculation results to the range of 0 to 1.

[0041] Characterizing three-dimensional points The stability of the existence of a point depends on the number of times that 3D point is detected in the scanning view. The larger the value, the worse the stability. Characterizing three-dimensional points The more abrupt a point is, the smaller the difference in distance between it and the surrounding points (i.e., the more abrupt it is). The larger.

[0042] For example, this application can set an interference coefficient threshold, which can be adjusted by statistically analyzing the scanning data of multiple sets of standard tools and combining the distribution characteristics of interference points. For example, it can be 0.7, when the three-dimensional points... interference coefficient When the interference coefficient threshold is greater than a certain value, the 3D point is determined to be an interference point and removed from the initial 3D model data. 3D points with interference coefficients less than or equal to the interference coefficient threshold are retained to obtain filtered point cloud data. Finally, the filtered point cloud data is reconstructed and optimized to obtain the final 3D model data of the product to be tested.

[0043] Based on the above technical solution, this application can obtain standard 3D model data of the tunnel boring machine (TBM) cutter and historical 3D model data of multiple defective TBM products. Based on the standard 3D model data and multiple historical 3D model data, region identification is performed to divide the TBM cutter into multiple regions, achieving region division based on historical defect patterns. This makes the determination of scanning modeling parameters more targeted. Subsequently, based on the standard 3D model data and multiple historical 3D model data, point cloud data within the multiple regions is modeled and analyzed to obtain the scanning modeling parameters of the TBM cutter. This achieves differentiated settings for marker point layout parameters and scanning point distance parameters, avoiding the problem of inaccurate local point clouds caused by uniform parameters in existing technologies. This ensures that areas with high defect incidence and complex structures can obtain denser and more accurate point cloud data. The final constructed 3D model has higher modeling quality and better reflects the actual state of the cutter, providing a reliable reference for fault analysis by personnel and improving the effectiveness of TBM cutter quality control.

[0044] As a possible embodiment of this application, step 102 above can be implemented through the following steps: Step 201: For each historical 3D model data, align the historical 3D model data with the standard 3D model data, and identify the individual defect areas on the shield cutter corresponding to the historical 3D model data.

[0045] In one possible implementation, this application can align the historical 3D model data with the standard 3D model data for each historical 3D model data, and determine the distance from each 3D point in the historical 3D model data to the model surface represented by the standard 3D model data.

[0046] Alignment between historical 3D model data and standard 3D model data is a prerequisite for ensuring accurate identification of defect areas. For example, this application can adopt a feature-based alignment method. First, common reference features (such as cylindrical surfaces, planes, symmetrical structures, etc.) are extracted from the standard 3D model and the historical 3D model. The spatial transformation relationship between the reference features (including translation, rotation, and scaling parameters) is calculated using the least squares method. This transformation relationship is then applied to the historical 3D model data to achieve accurate alignment between the two.

[0047] After alignment, this application can use a point-to-surface distance calculation method to obtain the distance from each 3D point to the model surface represented by the standard 3D model data. For example, for any 3D point P in the historical 3D model data, find the point closest to the 3D point on the model surface represented by the standard 3D model data, calculate the magnitude of the vector formed by the two points, and this magnitude is the distance from the 3D point to the model surface represented by the standard 3D model data.

[0048] Subsequently, based on the three-dimensional points in the historical three-dimensional model data that are at a distance greater than a preset distance threshold, the individual defect areas on the shield cutting tool corresponding to the historical three-dimensional model data are determined.

[0049] The preset distance threshold can be determined based on the design tolerances, machining errors, and measurement errors of the scanning equipment for the tunnel boring machine cutters. For example, by statistically analyzing the scanning data of multiple sets of standard cutters, the measurement error of the scanning equipment is determined to be ±0.05mm. Combined with the design tolerance of the cutters ±0.1mm, the preset distance threshold is set to 0.2mm.

[0050] This application can filter out 3D points in historical 3D model data whose distance is greater than a preset distance threshold, and use a region growing algorithm to cluster these points, aggregating spatially adjacent points with consistent distance characteristics into a continuous region. This region is the individual defect region corresponding to the historical 3D model data, and records the 3D spatial coordinates, area, shape and other information of the region.

[0051] Step 202: Based on the individual defect areas on the shield cutter corresponding to each historical 3D model data, determine the high-incidence areas of defects and non-defect high-incidence areas of the shield cutter.

[0052] In one possible implementation, this application can map the individual defect area on the shield cutter corresponding to each historical 3D model data to the model surface represented by the standard 3D model data to obtain the defect coverage density distribution information of the shield cutter.

[0053] Since each historical 3D model data is precisely aligned with the standard 3D model data, the 3D coordinates of individual defect areas can be directly mapped onto the surface of the standard model. For example, this application can divide the surface of the standard 3D model into multiple tiny triangular patches (e.g., triangular patches with a side length of 0.1 mm), count the number of times each triangular patch is covered by historical individual defect areas, and the ratio of the number of coverage times to the number of historical 3D model data points is the defect coverage density of that triangular patch. The defect coverage densities of all triangular patches constitute the defect coverage density distribution information of the tunnel boring machine cutter, which can be visualized in the form of a heatmap, intuitively reflecting the distribution pattern of defects.

[0054] Subsequently, based on the defect coverage density distribution information of the tunnel boring machine cutter, boundary extraction and surface fitting were performed to obtain the high-incidence area of ​​defects in the tunnel boring machine cutter.

[0055] For example, this application can set a defect coverage density threshold, which can be adjusted according to actual needs, such as 70%. Then, triangular facets with a coverage density greater than 70% are selected, and these facets are clustered using a density-based clustering algorithm (such as DBSCAN). Spatially continuous clustering results are identified as initial high-defect regions. Subsequently, the boundaries of the initial high-defect regions are extracted, and B-spline surface fitting is used to smooth the boundaries, resulting in high-defect regions with clear contours and continuous surfaces, ensuring that the region boundaries match the actual structural characteristics of the tool.

[0056] Thus, this application can further perform spatial clustering on the areas of the model surface represented by standard 3D model data, excluding areas with high defect incidence, to obtain one or more non-defect-prone areas.

[0057] It should be noted that the remaining area on the surface of the standard model, excluding the high-defect-occurrence region, may contain multiple structurally independent parts. Therefore, this application also requires spatial clustering for partitioning. For example, the DBSCAN density clustering algorithm is used, with three-dimensional spatial distance as the clustering basis, to cluster the three-dimensional point cloud of the remaining region. Each cluster obtained after clustering represents a geometrically continuous and independent region, i.e., a non-defect-occurrence region. Each non-defect-occurrence region can serve as the basic unit for subsequent marker placement and scanning parameter setting.

[0058] Based on the above technical solution, this application can align historical 3D model data with standard 3D model data for each historical 3D model data, identify individual defect areas on the shield cutter corresponding to the historical 3D model data, ensure that the location and range of each historical defect can be accurately mapped onto the standard model, and determine the high-incidence defect areas and non-high-incidence defect areas of the shield cutter based on the individual defect areas on the shield cutter corresponding to each historical 3D model data, making the division of high-incidence defect areas more objective and reliable, providing a precise basis for regional division for subsequent differentiated setting of scanning modeling parameters, and further improving the targeting and accuracy of 3D model construction.

[0059] As a possible embodiment of this application, step 103 above can be implemented through the following steps: Step 301: For each non-defect-prone high-incidence area, determine the marker layout parameters within the non-defect-prone high-incidence area based on the non-defect-prone high-incidence area and the defect-prone areas adjacent to it.

[0060] The marker placement parameters include the number and distribution of markers. These parameters determine the spatial distribution of markers during 3D scanning. A reasonable placement ensures accurate stitching of multi-view scans while avoiding interference with defect detection. Markers should typically be placed in non-defect-prone areas to avoid obscuring potential defects, but the placement density needs to consider the distribution of surrounding defect areas.

[0061] In addition, the distribution of marker points should avoid linear, circular, or symmetrical arrangements. This application can adopt a random staggered arrangement to ensure that the marker points can be clearly identified from different scanning angles, and the normal directions of the marker points are different, which improves the accuracy of splicing.

[0062] In one possible implementation, this application can determine the labeling bias coefficient of each non-defect-prone high-incidence area based on the area of ​​the non-defect-prone high-incidence area, the number of defect-prone areas adjacent to the non-defect-prone high-incidence area, and the distance between each three-dimensional point in the non-defect-prone high-incidence area and the corresponding fitting plane of the non-defect-prone high-incidence area.

[0063] Among them, the marking bias coefficient is used to characterize the intensity of the demand for marking points in 3D scanning in non-defect-prone areas.

[0064] For example, the labeling bias coefficient satisfies the following formula: in, Non-defect high-incidence areas The labeling bias coefficient, Non-defect high-incidence areas The area of ​​the region can be calculated by examining the non-defect-prone high-incidence area. The area of ​​the curved surface enclosed by the boundary of the region is obtained. This represents the number of regions with a high incidence of defects. To be in areas with high incidence of non-defects The number of adjacent defect-prone areas can be determined by comparing the boundaries of the defect-prone areas with those of non-defect-prone areas. The adjacency relationship is determined by whether the boundaries of the two sides intersect. Non-defect high-incidence areas The Middle A three-dimensional point, Non-defect high-incidence areas Inner Three-dimensional points and areas with high incidence of non-defects Based on the distance to the corresponding fitted plane, this application can use the least squares method for non-defect-prone areas. The plane is obtained by fitting a plane to all three-dimensional points within the plane. This is a safety parameter used to correct fractions where the denominator is 0; its specific value can be determined based on... The value of the value determines the outcome, such as . This is a normalization function (e.g., maximum / minimum normalization) used to map the calculation results to the range of -0.5 to 0.5.

[0065] The size of the non-defect-prone high-incidence area is characterized by the combined effect of the area size and the scarcity of adjacent defect-prone high-incidence areas; the larger the area, the more adjacent defect-prone high-incidence areas there are (i.e., (smaller) The larger, The degree of surface fluctuation characterizes areas with high non-defect incidence; the greater the fluctuation, the better. The larger the area, the more severe the surface fluctuations, and the more adjacent areas with high incidence of defects, the higher the demand for marker placement.

[0066] Then, the marking point layout parameters in non-defect-prone areas are determined based on the marking bias coefficient.

[0067] For example, the marker placement parameters can be represented by the number of markers, which satisfies the following formula: in, Non-defect high-incidence areas The number of marker points, The number of baseline markers can be adjusted based on the average area of ​​all non-defect-prone areas; for example, it could be 5. Non-defect high-incidence areas The labeling bias coefficient. The rounding operator is used to round up. If the result is a decimal, this application can obtain the final non-defect-prone area by rounding up. The number of marker points.

[0068] Step 302: For each region, determine the corresponding scan point distance parameters based on the point cloud data of the region in the standard 3D model data and multiple historical 3D model data.

[0069] Among them, the scanning point spacing parameter directly affects the density and scanning accuracy of point cloud data, and needs to be set differently according to the structural complexity and defect risk of the region.

[0070] In some embodiments, for areas with high defect incidence, due to the high frequency of historical defects, complex defect types, and possible minor structural damage, a smaller scan point spacing is required to obtain dense point cloud data and accurately capture defect details. For areas without high defect incidence, if the point cloud distribution in the standard model is dense (i.e., the structure is complex), a smaller scan point spacing is also required. If the structure is relatively simple, a larger scan point spacing can be used to improve scanning efficiency while ensuring accuracy.

[0071] In one possible implementation, this application can determine the scanning attention of each region based on the region's defects in multiple historical 3D model data and the point cloud distribution of the region in standard 3D model data.

[0072] Among them, scan attention intensity is used to characterize the intensity of attention a region receives during 3D scanning.

[0073] For example, the intensity of attention satisfies the following formula: in, For the region The scanning attention For the region The amount of historical 3D model data containing defects in memory. The amount of historical 3D model data, For the region Internal defect area and region The area of ​​intersection For the region The average distance between each 3D point and its nearest neighbor. This is a safety parameter used to correct fractions where the denominator is 0; its specific value can be determined based on... The value of the value determines the outcome, such as . This is a normalization function (e.g., maximum / minimum normalization) used to map the calculation results to the range of 0 to 0.5.

[0074] For example, the region The average distance between each 3D point and its nearest neighbor 3D point satisfies the following formula: in, For the region The average distance between each 3D point and its nearest neighbor. For the region The number of internal 3D points For the region Inner The distance between a 3D point and its nearest neighbor 3D point.

[0075] Characterization region The higher the frequency of defects, the greater the scanning attention. Characterization region The greater the sum of the severity of the defects, the more severe the defects, and the higher the level of attention to the scan. Characterization region The degree of structural complexity, The smaller the size, the more complex the structure. The more frequent, severe, and complex the regional defects, the higher the intensity of attention required during the scanning process.

[0076] Then, the scan point spacing parameters corresponding to the scan focus area are determined.

[0077] For example, the scan point spacing parameter can be represented by the scan point spacing, which satisfies the following formula: in, For the region The scan point distance, The reference scan dot spacing can be adjusted according to the accuracy level of the scanning equipment; for example, it can be 0.5mm. For the region The scanning attention.

[0078] Based on the above technical solution, this application determines the layout parameters of marker points in non-defect-prone areas, taking into full account the characteristics of the area itself and the influence of adjacent defect-prone areas, ensuring that the layout of marker points meets the requirements of splicing accuracy without affecting defect identification. The differentiated setting of scanning point distance parameters achieves a balance between accurate scanning of key areas and efficient scanning of ordinary areas. While ensuring the scanning accuracy of defect areas and complex structure areas, it avoids unnecessary waste of scanning resources and improves the overall efficiency and quality of 3D modeling.

[0079] As one possible embodiment of this application, the method further includes the following steps: Step 401: Perform cluster analysis on the 3D model data of the product to be tested to determine multiple cluster ranges of the product to be tested.

[0080] It should be noted that 3D scanning technology is based on optical principles. For blind spots such as deep holes, internal cavities, and densely packed teeth that optical equipment cannot directly access, data loss or low quality often occurs. Articulated arms, as contact measurement devices, can penetrate these optical measurement blind spots to obtain accurate geometric data. This embodiment analyzes the point cloud distribution and grayscale characteristics of each region of the 3D model to automatically identify areas requiring supplementary measurements by the articulated arm.

[0081] For example, this application can use the DBSCAN density clustering algorithm to cluster the point cloud of the 3D model data, with the clustering parameters adjusted according to the point cloud density. The purpose of cluster analysis is to divide the point cloud data into multiple relatively independent cluster ranges, each corresponding to a local area of ​​the product to be inspected, facilitating subsequent analysis of the point cloud distribution characteristics and grayscale characteristics of each area.

[0082] Step 402: For each cluster range, determine the necessary coefficients for supplementing the articulated arms of the cluster range based on the point cloud distribution characteristics and grayscale characteristics of the cluster range in the 3D model data.

[0083] The articulated arm supplementation necessity coefficient is used to characterize the necessity of articulated arm supplementation within the cluster range. Gray-scale features can be represented by the average gray-scale value, which reflects the optical measurement quality of the region. Lower gray-scale values ​​generally indicate weaker optical reflection signals, potentially indicating measurement difficulties or missing data. Point cloud distribution features can be represented by point cloud distribution density, which reflects the data integrity of the region. Lower density indicates a sparser point cloud, potentially indicating measurement blind spots. The density difference with adjacent regions reflects the relative anomaly of the region; a larger difference is more likely to indicate measurement difficulties caused by special geometric structures.

[0084] For example, the necessary coefficient for joint arm supplementation satisfies the following formula: in, Clustering range The joint arm supplements the necessary coefficient. Clustering range The average grayscale value of all three-dimensional points within the area, which is obtained by converting light intensity or reflectivity information captured by the scanning device. The smaller the value, the weaker the optical reflectivity of the cluster area, potentially indicating the presence of blind spots in the scan. To the clustering range The number of adjacent cluster ranges, and the adjacency relationship is determined by judging whether the boundaries of the cluster ranges intersect. Clustering range The average distance between each 3D point and its nearest neighbor can indirectly characterize the point cloud distribution density within that cluster. The larger the size, the sparser the point cloud. For adjacent cluster range The average distance between each 3D point and its nearest neighbor. This is a safety parameter used to correct fractions where the denominator is 0; its specific value can be determined based on... The value of the value determines the outcome, such as . This is a normalization function (e.g., maximum / minimum normalization) used to map the calculation results to the range of 0 to 1.

[0085] The influence of gray-scale features on the clustering range, average gray-scale value The smaller, The larger the size, the greater the need for supplementation. The difference in point cloud density between the cluster range and adjacent regions is characterized by the greater the difference. The larger the size, the greater the need for supplementation.

[0086] Step 403: Perform articulation arm supplementation based on the necessary coefficients for articulation arm supplementation for each cluster range.

[0087] For example, this application can set an articulated arm supplementation threshold, which can be adjusted according to the measurement accuracy and supplementation efficiency of the articulated arm, for example, it can be 0.7.

[0088] When the necessary coefficient for articulated arm supplementation in a cluster range is greater than the articulated arm supplementation threshold, it is determined that there are scanning blind spots or missing features in the cluster range, and articulated arm supplementation measurement is required.

[0089] For example, this application can fix the product to be inspected on a measurement platform, accurately measure the cluster range using an articulated arm measurement device, obtain the three-dimensional coordinate data of the missing area, fuse the point cloud data obtained from the supplementary measurement with the original three-dimensional model data, fill in the scanning blind spots and missing features, and obtain complete three-dimensional model data of the product to be inspected.

[0090] Based on the above technical solution, this embodiment performs cluster analysis on the 3D scanning model. Based on the grayscale characteristics and point cloud distribution density differences of each cluster range, it automatically identifies areas difficult to measure optically and quantifies the necessity of articulated arm supplementation. Then, it performs articulated arm supplementation measurements on the necessary areas. This technical solution effectively solves the problem of measurement blind spots in optical 3D scanning under complex geometric structures (such as deep holes and densely packed teeth). By complementing the advantages of optical measurement and contact measurement, it ensures the integrity and accuracy of the 3D model of the tunnel boring machine cutter. Especially for tunnel boring machines with complex internal structures, it significantly improves the model's data coverage and geometric accuracy.

[0091] Please see Figure 2 The diagram illustrates a system architecture diagram of a 3D model building system for shield tunneling cutter product manufacturing, provided by an embodiment of the present invention. The 3D model building system 20 for shield tunneling cutter product manufacturing includes: The data acquisition unit 21 is used to acquire standard three-dimensional model data of shield cutters and historical three-dimensional model data of multiple defective shield cutter products; the standard three-dimensional model data is used to characterize the standard geometric shape of shield cutters; the historical three-dimensional model data is used to characterize the geometric shape of defective shield cutters actually produced. The region identification unit 22 is used to identify regions based on standard 3D model data and multiple historical 3D model data, dividing the shield cutter into multiple regions; the multiple regions include defect-prone areas and non-defect-prone areas of the shield cutter. The modeling and analysis unit 23 is used to perform modeling and analysis on point cloud data in multiple regions based on standard 3D model data and multiple historical 3D model data to obtain the scanning modeling parameters of the shield cutter. The scanning modeling parameters include marker point layout parameters and scanning point distance parameters. Model building unit 24 is used to build a three-dimensional model of the shield cutter product to be inspected based on the scanning modeling parameters of the shield cutter, and obtain the three-dimensional model data of the product to be inspected.

[0092] It should be noted that the various embodiments of this application can be referenced or learned from each other. For example, the same or similar steps, method embodiments, system embodiments and device embodiments can be referenced from each other without limitation.

[0093] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0094] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for constructing a three-dimensional model for the production of tunnel boring machine (TBM) cutterheads, characterized in that, include: Obtain standard 3D model data of the tunnel boring machine cutter and historical 3D model data of multiple defective products of the tunnel boring machine cutter; The standard 3D model data is used to characterize the standard geometric shape of the tunnel boring machine cutter; the historical 3D model data is used to characterize the geometric shape of defective tunnel boring machine cutters actually produced. Based on standard 3D model data and multiple historical 3D model data, the shield cutter is divided into multiple regions for region identification; the multiple regions include high-defect regions and non-defect-high-incidence regions of the shield cutter. Based on standard 3D model data and multiple historical 3D model data, point cloud data in the multiple regions are modeled and analyzed to obtain the scanning modeling parameters of the shield cutter; the scanning modeling parameters include marker point layout parameters and scanning point distance parameters. Based on the scanning and modeling parameters of the tunnel boring machine cutter, a three-dimensional model of the product to be inspected is constructed, and the three-dimensional model data of the product to be inspected is obtained.

2. The method for constructing a three-dimensional model for the production of tunnel boring machine cutter products according to claim 1, characterized in that, Based on standard 3D model data and multiple historical 3D model data, region identification is performed, and the tunnel boring machine cutter is divided into multiple regions, including: For each historical 3D model data, the historical 3D model data is aligned with the standard 3D model data to identify the individual defect areas on the shield cutter corresponding to the historical 3D model data; Based on the individual defect areas on the shield cutter corresponding to each historical 3D model data, the high-incidence areas of defects and the non-defect-high-incidence areas of the shield cutter are determined.

3. The method for constructing a three-dimensional model for the production of tunnel boring machine cutter products according to claim 2, characterized in that, For each historical 3D model data, the historical 3D model data is aligned with the standard 3D model data, and individual defect areas on the shield cutting tool corresponding to the historical 3D model data are identified, including: For each historical 3D model data, the historical 3D model data is aligned with the standard 3D model data, and the distance from each 3D point in the historical 3D model data to the model surface represented by the standard 3D model data is determined. Based on the three-dimensional points in the historical three-dimensional model data whose distance is greater than a preset distance threshold, the individual defect areas on the shield cutting tool corresponding to the historical three-dimensional model data are determined.

4. The method for constructing a three-dimensional model for the production of tunnel boring machine cutter products according to claim 2, characterized in that, Based on the individual defect areas on the shield cutterhead corresponding to each historical 3D model data, the high-incidence defect areas and non-high-incidence defect areas of the shield cutterhead are determined, including: The individual defect areas on the shield cutter corresponding to each historical 3D model data are mapped to the model surface represented by the standard 3D model data to obtain the defect coverage density distribution information of the shield cutter. Based on the defect coverage density distribution information of the shield cutter, boundary extraction and surface fitting are performed to obtain the high-incidence area of ​​defects in the shield cutter. Spatial clustering is performed on the areas of the model surface represented by the standard three-dimensional model data, excluding the high-defect areas, to obtain one or more non-defect-high-incidence areas.

5. The method for constructing a three-dimensional model for the production of tunnel boring machine cutter products according to claim 1, characterized in that, Based on standard 3D model data and multiple historical 3D model data, point cloud data within the multiple regions are modeled and analyzed to obtain the scanning modeling parameters of the tunnel boring machine cutter, including: For each non-defect-prone high-incidence area, the marker point layout parameters within the non-defect-prone high-incidence area are determined based on the non-defect-prone high-incidence area and the defect-prone areas adjacent to the non-defect-prone high-incidence area. For each region, the corresponding scan point distance parameter is determined based on the point cloud data of the region in the standard 3D model data and multiple historical 3D model data.

6. The method for constructing a three-dimensional model for the production of tunnel boring machine cutter products according to claim 5, characterized in that, For each non-defect-prone high-incidence area, the marker layout parameters within the non-defect-prone high-incidence area are determined based on the non-defect-prone high-incidence area and the defect-prone areas adjacent to it, including: For each non-defect-prone high-incidence area, the marking bias coefficient of the non-defect-prone high-incidence area is determined based on the area of ​​the non-defect-prone high-incidence area, the number of defect-prone high-incidence areas adjacent to the non-defect-prone high-incidence area, and the distance between each three-dimensional point in the non-defect-prone high-incidence area and the corresponding fitting plane of the non-defect-prone high-incidence area; the marking bias coefficient is used to characterize the intensity of the demand for marking points in the non-defect-prone high-incidence area during three-dimensional scanning. The marking point layout parameters in the non-defect-prone area are determined based on the marking bias coefficient.

7. The method for constructing a three-dimensional model for the production of tunnel boring machine cutter products according to claim 5, characterized in that, For each region, the scan point distance parameter corresponding to the region is determined based on the point cloud data of the region in the standard 3D model data and multiple historical 3D model data, including: For each region, the scanning attention level of the region is determined based on the region's defect status in multiple historical 3D model data and the region's point cloud distribution in the standard 3D model data; the scanning attention level is used to characterize the intensity of attention to the region during 3D scanning. The scan point spacing parameter corresponding to the region is determined based on the scan attention level.

8. The method for constructing a three-dimensional model for the production of tunnel boring machine cutter products according to claim 1, characterized in that, Based on the scanning modeling parameters of the tunnel boring machine cutter, a three-dimensional model of the product to be inspected is constructed, resulting in three-dimensional model data of the product to be inspected, including: Based on the marker layout parameters, determine the position information of each marker on the product to be inspected; Based on the scanning point distance parameters, perform multi-view 3D scanning on the product to be inspected to obtain point cloud data corresponding to multiple views; Based on the position information of each marker point on the product to be inspected, the point cloud data corresponding to the multiple viewpoints are stitched together to obtain the initial three-dimensional model data of the product to be inspected. For each 3D point in the initial 3D model data, the interference coefficient of the 3D point is determined based on the point cloud data corresponding to the 3D point in each viewpoint; Interference points are filtered based on the interference coefficient of each three-dimensional point in the initial three-dimensional model data to obtain the three-dimensional model data of the product to be tested.

9. The method for constructing a three-dimensional model for the production of tunnel boring machine cutter products according to claim 1, characterized in that, The method further includes: Cluster analysis is performed on the three-dimensional model data of the product to be tested to determine multiple cluster ranges of the product to be tested. For each cluster range, the articulation arm supplementation necessity coefficient of the cluster range is determined based on the point cloud distribution characteristics and grayscale characteristics of the cluster range in the 3D model data; the articulation arm supplementation necessity coefficient is used to characterize the degree of necessity for articulation arm supplementation of the cluster range. Articulated arms are supplemented based on the necessary coefficient for articulated arm supplementation for each cluster range.

10. A three-dimensional model construction system for the production of tunnel boring machine (TBM) cutterheads, characterized in that, include: The data acquisition unit is used to acquire standard three-dimensional model data of the tunnel boring machine cutter and historical three-dimensional model data of multiple defective products of the tunnel boring machine cutter. The standard 3D model data is used to characterize the standard geometric shape of the tunnel boring machine cutter; the historical 3D model data is used to characterize the geometric shape of defective tunnel boring machine cutters actually produced. The region identification unit is used to identify regions based on standard 3D model data and multiple historical 3D model data, and divide the shield cutter into multiple regions; the multiple regions include defect-prone regions and non-defect-prone regions of the shield cutter. The modeling and analysis unit is used to perform modeling and analysis on point cloud data in the multiple regions based on standard 3D model data and multiple historical 3D model data to obtain the scanning modeling parameters of the shield cutter; the scanning modeling parameters include marker point layout parameters and scanning point distance parameters; The model building unit is used to build a three-dimensional model of the shield cutter to be inspected based on the scanning modeling parameters of the shield cutter, and obtain the three-dimensional model data of the product to be inspected.

Images