A 3D box girder point cloud data topology structure identification method based on persistent homology
By combining continuous coherence technology and mechanics, the problems of difficulty in identifying multi-scale topology and sensitivity to noise in existing technologies are solved, enabling multi-scale, robust and interpretable topological analysis of box girder point cloud data, supporting health assessment and maintenance of box girders.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 中国水利水电第七工程局有限公司
- Filing Date
- 2025-01-15
- Publication Date
- 2026-05-12
AI Technical Summary
Existing technologies struggle to effectively capture multi-scale topological information in 3D box girder point cloud data analysis, especially in complex structures where key topological features are difficult to identify. Furthermore, they are sensitive to noise and have poor interpretability.
A topology identification method based on continuous homology 3D box girder point cloud data is adopted, including preprocessing, continuous homology analysis and topology feature interpretation. By denoising, registering and simplifying the data, a persistent bar chart is generated, and the topology features are identified by combining mechanical analysis.
It achieves multi-scale, robust topology analysis, can distinguish between noise and true features, provides comprehensive and interpretable health assessments, adapts to unlabeled data scenarios, and supports real-time monitoring and maintenance.
Smart Images
Figure CN119942344B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of data structure analysis, specifically relating to a method for identifying the topological structure of 3D box girder point cloud data based on continuous homology. Background Technology
[0002] In bridge engineering, box girders, as key load-bearing components in bridges, buildings, and infrastructure, directly affect the safety of the entire project through their structural stability and deformation. The geometry of the box girder has a decisive influence on its load-bearing capacity and mechanical response under load. Therefore, a comprehensive understanding and accurate analysis of the structural state, deformation, and potential damage of box girders is a crucial prerequisite for ensuring project safety.
[0003] With the continuous development of 3D scanning, remote sensing, and computer vision technologies, 3D point cloud data has been widely applied in various fields such as bridge inspection, architectural modeling, and medical imaging. 3D point cloud data can accurately capture the geometric shape of objects and provide rich spatial information, enabling engineers to conduct detailed structural analyses based on it. These analyses can not only help assess the structural health of box girders but also identify issues such as deformation, damage, and wear, thus providing a basis for subsequent quality assessment, maintenance inspection, and design optimization.
[0004] Currently, existing point cloud data analysis techniques are mainly divided into geometric feature-based analysis and machine learning-based methods. Geometric feature-based methods, such as KD-trees and quadtrees, typically analyze the spatial distribution of point cloud data through local features and distance metrics. However, these methods rely heavily on local geometric information, making it difficult to capture changes in the overall structure of the box girder under complex stress environments. Especially when multi-scale topological features are involved, existing geometric analysis techniques often fall short. This means that when faced with complex structural deformations or potential damage, the results of geometric feature analysis may be limited and unable to fully reflect the actual situation of the box girder under various loads.
[0005] On the other hand, machine learning methods, especially deep learning techniques such as PointNet and PointCNN, have demonstrated excellent performance in feature extraction and classification of point cloud data in recent years. These methods, through end-to-end learning frameworks, can automatically extract useful features from point cloud data, reducing the burden of manual feature engineering. However, machine learning methods also face some challenges. First, deep learning models typically require large amounts of high-quality training data, while obtaining comprehensive and accurately labeled point cloud datasets can be relatively difficult in bridge engineering. Second, although these models perform well in feature extraction, they lack the ability to directly interpret key topological features in box girder structures, making it difficult to reveal topological relationships in point cloud data, such as connectivity and voids.
[0006] In summary, existing technologies for point cloud data analysis share a common limitation: they primarily rely on geometric metrics, making it difficult to effectively capture topological information from box girder point cloud data. This is particularly true in multi-scale analysis and complex structural scenarios, where geometric analysis methods and machine learning models struggle to distinguish between true structural features and noise. Furthermore, they lack the ability to identify crucial topological structures within box girders, which is especially important when assessing the overall structural stability of box girders. Therefore, more comprehensive analytical methods, such as topological data analysis methods based on continuous homology, are needed to overcome these limitations, particularly in the multi-scale topological feature detection of complex structures. Summary of the Invention
[0007] To address the aforementioned shortcomings in existing technologies, this invention provides a method for topological structure recognition of 3D box girder point cloud data based on continuous homology, which solves the problems of difficulty in identifying multi-scale topology, sensitivity to noise, and poor interpretability in existing 3D box girder point cloud data analysis.
[0008] To achieve the aforementioned objectives, the present invention employs the following technical solution: a method for identifying the topological structure of 3D box girder point cloud data based on continuous homology, comprising the following steps:
[0009] S1. Obtain the point cloud of the box girder, preprocess the point cloud, and generate a 3D dataset;
[0010] S2. Perform continuous homology analysis based on the three-dimensional dataset to obtain the topological characteristics of the box girder;
[0011] S3. Analyze the topological characteristics of the box girder and mark the areas where structural problems exist;
[0012] S4. Perform a physical interpretation of the topological features based on the marked regions and generate a health assessment report.
[0013] Further: In step S1, the method for preprocessing the point cloud data includes the following steps:
[0014] S11. Calculate the neighborhood density of each point in the point cloud and remove outlier points whose neighborhood density is lower than the neighborhood density threshold.
[0015] S12. Register the point cloud after removing outliers;
[0016] S13. Use a method of coarsening the data granularity to reduce the size of the registered point cloud and generate a 3D dataset.
[0017] The beneficial effects of the above-mentioned further scheme are as follows: To improve the accuracy and efficiency of subsequent continuous coherence analysis, data processing including denoising, registration, and simplification is performed sequentially to ensure its integrity and accuracy. This invention ensures that the simplification process does not result in the loss of key geometric information by measuring the distance error between the simplified point cloud and the original point cloud.
[0018] Further: In S11, the first i Neighborhood density of each point The specific expression is:
[0019]
[0020] In the formula, For the first i The point and the first j Euclidean distance between points The selected cutoff distance.
[0021] Furthermore, S13 specifically refers to:
[0022] S131. Treat the point cloud as a set of points to be partitioned;
[0023] S132. Obtain the two points with the greatest distance between them in the point set, establish a first subset and a second subset based on the obtained points, and assign the remaining points in the point set to the nearest subset.
[0024] S133, Calculate the compactness of the point set, the first subset, and the second subset; compactness The specific expression is:
[0025]
[0026] In the formula, m For point set D The number of midpoints p j For point set D The point in the middle, C For point set D The center of mass;
[0027] Determine whether the splitting condition is met. The specific splitting condition is as follows:
[0028] and
[0029] In the formula, For the compactness of the first subset, For the compactness of the second subset, For the compactness of the point set, For point set D The total number of points, nThe total number of points in the point cloud;
[0030] If yes, then both the first and second subsets are treated as point sets to be partitioned, and return to S132; otherwise, the current point set is treated as the simplified point cloud, and proceed to S134.
[0031] S134. Create a 3D dataset based on all the simplified point clouds.
[0032] The beneficial effects of the above-mentioned further scheme are as follows: In order to improve the computational efficiency of subsequent continuous coherence analysis, this invention introduces a method of coarsening data granularity to reduce the volume of point cloud data, divides the overall dataset into multiple point sets, and the centroid of each point set represents these point sets. In this process, key geometric information is preserved and the integrity and accuracy of the data are ensured.
[0033] Further: S2 includes the following sub-steps:
[0034] S21. Based on each point in the point cloud of the 3D dataset, calculate the Euclidean distance between the points and generate a distance matrix;
[0035] S22. Set the neighborhood radius and create a persistent bar chart to analyze the persistence of the topological features generated from each point cloud, and obtain the topological features of the box girder.
[0036] Furthermore: S22 specifically includes:
[0037] Set the neighborhood radius, gradually increase the neighborhood radius within the framework of continuous homology, calculate the topological features at different scales of neighborhood radius, create a persistent bar chart based on the life cycle of the topological features, and retain the topological features whose life cycle exceeds 10% of the overall time length as the topological features of the box girder.
[0038] The neighborhood radius is used to determine whether points are considered connected. If the Euclidean distance between two points is less than the neighborhood radius, then the two points are connected in the topological space.
[0039] The lifecycle is specifically the time when topological features appear and disappear.
[0040] The beneficial effects of the above-mentioned further solutions are as follows: To ensure the accuracy of the analysis results, this invention filters out noise by setting a threshold. Typically, only topological features with a lifespan exceeding 10% of the overall time are retained, effectively removing short-lived features caused by noise or localized microstructures. The retained persistent topological features need to be physically interpreted in conjunction with the specific structure of the box girder to better guide the health status assessment and structural analysis of the box girder.
[0041] Furthermore: the dimensions of the topological features include 0-dimensional, 1-dimensional, and 2-dimensional;
[0042] Among them, 0-dimensional topological features represent connected branches in point clouds, 1-dimensional topological features represent closed loops or holes, and 2-dimensional topological features represent cavities or enclosed three-dimensional spaces.
[0043] Furthermore: In S3, the method for analyzing the topological characteristics of the box girder is specifically as follows:
[0044] (1) The stress of the box girder is analyzed by finite element analysis, and the stress distribution in different regions under external force is calculated. By comparing the stress distribution with the topological features, the degree of overlap between the topological features and the stress concentration area is identified.
[0045] (2) For the geometric structural region where the topological features are located, fatigue life assessment was carried out in combination with mechanical model, stress state and load history of the geometric structural region were analyzed, and the possibility of fatigue failure under long-term cyclic loading was assessed.
[0046] Furthermore: In S4, the method for physically interpreting the topological features includes crack and damage detection and cavity and material defect identification;
[0047] The specific methods for detecting cracks and damage are as follows:
[0048] Detecting connectivity interruptions or new connecting branches indicates that there is a break or damage in a local area of the box girder. By analyzing the geometric structure and mechanics, the impact of the connecting branch on the overall performance of the box girder can be guided to improve maintenance work.
[0049] The specific methods for material defect identification are as follows:
[0050] The cavity structure is extracted from the 2D topological features. The specific location and shape of the cavity structure are determined by continuous homology analysis, and the potential threat of the cavity structure to structural safety is determined by mechanical analysis.
[0051] Furthermore, in S4, the health assessment report includes the overall connectivity of the box girder, local damage, and areas requiring focused monitoring or repair.
[0052] The beneficial effects of this invention are as follows: This invention provides a method for topological structure identification of 3D box girder point cloud data based on continuous homology, aiming to achieve multi-scale, robust, and highly interpretable topological structure analysis of box girder point cloud data. Through this technology, this invention can comprehensively capture the complex topological features of box girders at different scales, overcoming the shortcomings of traditional geometric analysis methods. Especially when dealing with problems such as noise and uneven point cloud data, it exhibits higher robustness and reliability. Compared with existing technologies, it has the following advantages:
[0053] (1) Multi-scale topological feature capture: This invention utilizes continuous cohomology technology to analyze 3D point cloud data at different scales and capture topological features such as connected branches, holes, and cavities in box girder structures. This method can reveal the geometric and topological information of the structure from multiple scales, both global and local, providing a more comprehensive topological analysis.
[0054] (2) High noise robustness: Through the persistence analysis of topological features, the present invention can effectively distinguish noise features from persistent features that reflect the real structure.
[0055] (3) Enhanced interpretability: This invention visually displays the topological features in box girder point cloud data by generating persistent bar charts, and assesses the importance of the topological features by their lifecycle length. Combined with actual geometric morphology and mechanical analysis, engineers can clearly determine the physical meaning of the topological features, such as holes and cracks.
[0056] (4) Strong adaptability to unlabeled data: Continuous homology analysis does not rely on large-scale labeled data, but is purely based on topological analysis to mine the intrinsic structure of the data. Therefore, this invention can adapt to unlabeled or limited labeled box girder 3D point cloud data scenarios.
[0057] (5) Analysis of the combination of topology and mechanical properties: This invention can not only identify the topological features in the point cloud data of box girders, but also combine the geometric structure and mechanical properties to conduct in-depth physical interpretation and structural health assessment of these topological features.
[0058] (6) Real-time monitoring and maintenance support: The systematic analysis process of this invention provides structural health assessment and maintenance support. Through long-term monitoring, the system can dynamically track changes in the box girder structure, promptly detect potential damage or deformation, and generate health assessment reports, providing engineers with strong decision support.
[0059] In summary, this invention, through continuous coherence technology, combined with topology analysis, noise robustness, geometric and mechanical interpretation, and multi-scale feature capture of point cloud data, solves the problems of difficult multi-scale topology identification, noise sensitivity, and poor interpretability in existing technologies, and provides a more robust, accurate, and interpretable method for box girder structure analysis. Attached Figure Description
[0060] Figure 1 This is a flowchart of a method for topological structure recognition of 3D box girder point cloud data based on continuous homology. Detailed Implementation
[0061] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.
[0062] like Figure 1 As shown, in one embodiment of the present invention, a method for identifying the topological structure of 3D box girder point cloud data based on continuous homology includes the following steps:
[0063] S1. Obtain the point cloud of the box girder, preprocess the point cloud, and generate a 3D dataset;
[0064] S2. Perform continuous homology analysis based on the three-dimensional dataset to obtain the topological characteristics of the box girder;
[0065] S3. Analyze the topological characteristics of the box girder and mark the areas where structural problems exist;
[0066] S4. Perform a physical interpretation of the topological features based on the marked regions and generate a health assessment report.
[0067] In this embodiment, the present invention uses a suitable lidar device to scan the box girder structure, acquiring its high-precision 3D point cloud data, and preprocessing this data. The preprocessing steps include denoising, data alignment, and simplification to ensure the quality and accuracy of the analyzed data. Based on this, continuous coherence analysis is applied to the processed point cloud data to generate persistent bar charts, revealing key topological features in the point cloud, such as connected branches, voids, and cavities. By analyzing the duration of these topological features, the importance of each feature can be effectively assessed, distinguishing between key elements in the global structure and transient features that may be caused by noise. Next, the identified topological features are combined with the box girder's geometry and mechanical properties for in-depth physical interpretation. This process helps engineers assess the health status of the box girder and detect potential structural problems, such as cracks, voids, or other damage. This systematic topological analysis method provides a scientific basis for the maintenance, repair, and monitoring of box girders, and supports long-term tracking and risk management of the structural condition.
[0068] In step S1, the method for preprocessing point cloud data includes the following steps:
[0069] S11. Calculate the neighborhood density of each point in the point cloud and remove outlier points whose neighborhood density is lower than the neighborhood density threshold.
[0070] S12. Register the point cloud after removing outliers;
[0071] S13. Use a method of coarsening the data granularity to reduce the size of the registered point cloud and generate a 3D dataset.
[0072] In this embodiment, after acquiring the 3D point cloud data of the box girder, in order to improve the accuracy and efficiency of subsequent continuous coherence analysis, data processing including denoising, registration, and simplification is performed sequentially to ensure its integrity and accuracy. This invention ensures that the simplification process does not result in the loss of key geometric information by measuring the distance error between the simplified point cloud and the original point cloud.
[0073] In S11, the first i Neighborhood density of each point The specific expression is:
[0074]
[0075] In the formula, For the first i The point and the first j Euclidean distance between points The selected cutoff distance should ensure that the average number of samples contained in the neighborhood of each data point accounts for 2% of the total number of samples.
[0076] In this embodiment, the present invention sets a neighborhood density threshold. If the number of other points in the local neighborhood of a certain point is too small, it is usually considered to be a noise point or a false detection point. The present invention will determine the threshold with the goal of eliminating 10% of low density points. Through this density filtering method, noise points caused by surface reflection, occlusion or sensor error during the scanning process can be effectively removed, thereby improving the overall quality of point cloud data.
[0077] In step S12, the present invention uses the GPS and IMU data built into the lidar to perform preliminary registration of point cloud data. The purpose of registration is to align point cloud data collected from multiple angles into a unified spatial coordinate system to construct a complete box girder model. In this step, by matching the position information of each scan segment, approximate alignment can be achieved, laying the foundation for subsequent fine registration and analysis.
[0078] Specifically, S13 is:
[0079] S131. Treat the point cloud as a set of points to be partitioned;
[0080] S132. Obtain the two points with the greatest distance between them in the point set, establish a first subset and a second subset based on the obtained points, and assign the remaining points in the point set to the nearest subset.
[0081] S133, Calculate the compactness of the point set, the first subset, and the second subset; compactness The specific expression is:
[0082]
[0083] In the formula, m For point set D The number of midpoints p j For point set D The point in the middle, C For point set D The center of mass;
[0084] Determine whether the splitting condition is met. The specific splitting condition is as follows:
[0085] and
[0086] In the formula, For the compactness of the first subset, For the compactness of the second subset, For the compactness of the point set, For point set D The total number of points, n The total number of points in the point cloud;
[0087] If yes, then both the first and second subsets are treated as point sets to be partitioned, and return to S132; otherwise, the current point set is treated as the simplified point cloud, and proceed to S134.
[0088] S134. Create a 3D dataset based on all the simplified point clouds.
[0089] In this embodiment, since the point cloud data generated by high-precision scanning is large in volume and complex to process, in order to improve the computational efficiency of subsequent continuous coherence analysis, the present invention introduces a method of coarsening the data granularity to reduce the volume of the point cloud data. The overall dataset is divided into multiple point sets, and the centroid of each point set represents these point sets. In this process, key geometric information is preserved and the integrity and accuracy of the data are ensured.
[0090] S2 includes the following steps:
[0091] S21. Based on each point in the point cloud of the 3D dataset, calculate the Euclidean distance between the points and generate a distance matrix;
[0092] S22. Set the neighborhood radius and create a persistent bar chart to analyze the persistence of the topological features generated from each point cloud, and obtain the topological features of the box girder.
[0093] In this embodiment, the distance matrix provides the basis for subsequent topology calculations. Based on this, the present invention sets a neighborhood radius. This radius determines whether points are considered connected. If the distance between two points is less than this radius, they are considered connected in the topological space, forming an edge between them. As this neighborhood radius gradually increases, a complex body composed of points and edges begins to form. In this process, changes in the neighborhood radius correspond to changes in connectivity and structure in the topological space. Different radii will produce different topological structures, including connected branches, loop structures, and closed cavities.
[0094] Specifically, S22 is:
[0095] Set the neighborhood radius, gradually increase the neighborhood radius within the framework of continuous homology, calculate the topological features at different scales of neighborhood radius, create a persistent bar chart based on the life cycle of the topological features, and retain the topological features whose life cycle exceeds 10% of the overall time length as the topological features of the box girder.
[0096] The neighborhood radius is used to determine whether points are considered connected. If the Euclidean distance between two points is less than the neighborhood radius, then the two points are connected in the topological space.
[0097] The lifecycle is specifically the time when topological features appear and disappear.
[0098] In this embodiment, topological features gradually appear and disappear as the neighborhood radius increases. This means that the lifecycle of different features can be recorded to determine their importance. Longer-lasting topological structures typically represent more global or stable geometric structures in the data, while short-lived features may be noise or localized minor structures. Based on this, a persistent bar chart can be created according to the appearance and disappearance times of topological features, with bar length representing the lifecycle of the topological feature. The horizontal axis is the neighborhood radius, and each bar corresponds to the lifecycle of a topological feature—the starting point of the bar represents the moment the feature is "born," and the ending point represents the moment it "dies." The longer the bar, the more stable the topological feature is over a larger scale, meaning the feature is more important. For example, a persistent connected branch or hole may correspond to some critical structure in a box girder, while a short-lived bar may simply be localized noise or an unimportant detail.
[0099] After generating the persistent bar chart, the importance of each topological feature needs to be assessed by analyzing its persistence (i.e., the length of its lifetime). To ensure the accuracy of the analysis results, this invention filters out noise by setting a threshold. Typically, only topological features with a lifetime exceeding 10% of the overall time are retained, effectively removing short-lived features caused by noise or localized minor structures. The retained persistent topological features need to be physically interpreted in conjunction with the specific structure of the box girder to better guide the health status assessment and structural analysis of the box girder.
[0100] The dimensions of the topological features include 0-dimensional, 1-dimensional, and 2-dimensional.
[0101] Among them, 0-dimensional topological features represent connected branches in point clouds, 1-dimensional topological features represent closed loops or holes, and 2-dimensional topological features represent cavities or enclosed three-dimensional spaces.
[0102] In this embodiment, the dimensionality of the topological features can be used to classify and identify these features:
[0103] Zero-dimensional topological features primarily represent the connectivity of point cloud data. For box girder structures, zero-dimensional topological features can reflect the overall connectivity of the box girder. If there are breaks or discontinuities in certain areas, changes in zero-dimensional features can accurately detect these break points or structural defects.
[0104] One-dimensional topological features represent ring structures or voids in point cloud data. In box girder applications, one-dimensional features can reflect voids or cracks that may exist on or inside the girder surface. These structures may be formed due to manufacturing defects or long-term stress. By analyzing the duration of one-dimensional features, the extent of these voids or cracks and their importance in the box girder structure can be determined.
[0105] Two-dimensional topological features describe closed three-dimensional spaces or cavities. For example, there may be incompletely filled areas inside a box girder, or the structure itself may form some kind of enclosed space. By analyzing the durability of these cavities, it is possible to determine whether they are key features in the box girder structure, and thus to effectively assess the health of the box girder.
[0106] In S3, the method for analyzing the topological characteristics of the box girder is as follows:
[0107] (1) The stress of the box girder is analyzed by finite element analysis, and the stress distribution in different regions under external force is calculated. By comparing the stress distribution with the topological features, the degree of overlap between the topological features and the stress concentration area is identified.
[0108] (2) For the geometric structural region where the topological features are located, fatigue life assessment was carried out in combination with mechanical model, stress state and load history of the geometric structural region were analyzed, and the possibility of fatigue failure under long-term cyclic loading was assessed.
[0109] In this embodiment, stress analysis of the box girder is performed using finite element analysis to calculate the stress distribution in different regions under external forces. By comparing the stress analysis results with topological features, the degree of overlap between the topological features and stress concentration areas is identified. For example, if severe stress concentration occurs around holes revealed by 1D topological features, this may indicate a risk of material fatigue or crack propagation in that area. This type of information is of significant value for structural health monitoring and predictive maintenance.
[0110] Furthermore, for the geometric structural regions where topological features are located, this invention combines mechanical models to conduct fatigue life assessments. By analyzing the stress state and load history of these regions, the likelihood of fatigue failure under long-term cyclic loading is evaluated. For example, the cavity structure identified by 2D topological features may lead to localized stress concentration under prolonged dynamic loading; these areas of stress concentration are often the initiation points for fatigue cracks. Analysis of these regions can provide important basis for predicting the fatigue life of box girders.
[0111] In S4, the methods for physically interpreting topological features include crack and damage detection and cavity and material defect identification;
[0112] The specific methods for detecting cracks and damage are as follows:
[0113] Detecting connectivity interruptions or new connecting branches indicates that there is a break or damage in a local area of the box girder. By analyzing the geometric structure and mechanics, the impact of the connecting branch on the overall performance of the box girder can be guided to improve maintenance work.
[0114] The specific methods for material defect identification are as follows:
[0115] The cavity structure is extracted from the 2D topological features. The specific location and shape of the cavity structure are determined by continuous homology analysis, and the potential threat of the cavity structure to structural safety is determined by mechanical analysis.
[0116] In this embodiment, the present invention provides a physical interpretation of topological features and provides a basis for the health assessment of box girders, for example:
[0117] Crack and Damage Detection: Disruptions in connectivity or the appearance of new connecting branches may indicate localized fractures or damage in the box girder. In most cases, these fractures initially manifest as small cracks, which may gradually widen over time with stress accumulation. Continuous harmonic analysis allows for timely detection of cracks before they extend to a level that threatens structural safety. Combined with geometric and mechanical analysis, this helps determine the impact on the overall performance of the box girder, thus guiding maintenance efforts.
[0118] Cavity and Material Defect Identification: Cavity structures extracted from 2D topological features in actual box girders may be caused by material defects, uneven casting, or omissions during construction. These problems can lead to localized failures during the long-term service of the box girder, especially when stress concentration occurs around these areas. Through continuous coherence analysis, engineers can determine the specific location and shape of these cavities and, combined with mechanical analysis, assess their potential threat to structural safety.
[0119] In S4, the health assessment report includes the overall connectivity of the box girder, local damage, and areas requiring key monitoring or repair.
[0120] The beneficial effects of this invention are as follows: This invention provides a method for topological structure identification of 3D box girder point cloud data based on continuous homology, aiming to achieve multi-scale, robust, and highly interpretable topological structure analysis of box girder point cloud data. Through this technology, this invention can comprehensively capture the complex topological features of box girders at different scales, overcoming the shortcomings of traditional geometric analysis methods. Especially when dealing with problems such as noise and uneven point cloud data, it exhibits higher robustness and reliability. Compared with existing technologies, it has the following advantages:
[0121] (1) Multi-scale topological feature capture: This invention utilizes continuous cohomology technology to analyze 3D point cloud data at different scales and capture topological features such as connected branches, holes, and cavities in box girder structures. This method can reveal the geometric and topological information of the structure from multiple scales, both global and local, providing a more comprehensive topological analysis.
[0122] (2) High noise robustness: Through the persistence analysis of topological features, the present invention can effectively distinguish noise features from persistent features that reflect the real structure.
[0123] (3) Enhanced interpretability: This invention visually displays the topological features in box girder point cloud data by generating persistent bar charts, and assesses the importance of the topological features by their lifecycle length. Combined with actual geometric morphology and mechanical analysis, engineers can clearly determine the physical meaning of the topological features, such as holes and cracks.
[0124] (4) Strong adaptability to unlabeled data: Continuous homology analysis does not rely on large-scale labeled data, but is purely based on topological analysis to mine the intrinsic structure of the data. Therefore, this invention can adapt to unlabeled or limited labeled box girder 3D point cloud data scenarios.
[0125] (5) Analysis of the combination of topology and mechanical properties: This invention can not only identify the topological features in the point cloud data of box girders, but also combine the geometric structure and mechanical properties to conduct in-depth physical interpretation and structural health assessment of these topological features.
[0126] (6) Real-time monitoring and maintenance support: The systematic analysis process of this invention provides structural health assessment and maintenance support. Through long-term monitoring, the system can dynamically track changes in the box girder structure, promptly detect potential damage or deformation, and generate health assessment reports, providing engineers with strong decision support.
[0127] In summary, this invention, through continuous coherence technology, combined with topology analysis, noise robustness, geometric and mechanical interpretation, and multi-scale feature capture of point cloud data, solves the problems of difficult multi-scale topology identification, noise sensitivity, and poor interpretability in existing technologies, and provides a more robust, accurate, and interpretable method for box girder structure analysis.
[0128] In the description of this invention, it should be understood that the terms "center," "thickness," "upper," "lower," "horizontal," "top," "bottom," "inner," "outer," and "radial," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying the relative importance or the number of technical features implicitly specified. Therefore, a feature defined by "first," "second," and "third" may explicitly or implicitly include one or more of that feature.
Claims
1. A method for topological structure recognition of 3D box girder point cloud data based on continuous homology, characterized in that, Includes the following steps: S1. Obtain the point cloud of the box girder, preprocess the point cloud, and generate a 3D dataset; the method for preprocessing the point cloud data includes the following steps: S11. Calculate the neighborhood density of each point in the point cloud and remove outlier points whose neighborhood density is lower than the neighborhood density threshold. S12. Register the point cloud after removing outliers; S13. A method of coarsening data granularity is used to reduce the size of the registered point cloud. The overall dataset is divided into multiple point sets, with the centroid of each point set representing these point sets, generating a 3D dataset. Specifically, S13 involves: S131. Treat the point cloud as a set of points to be partitioned; S132. Obtain the two points with the greatest distance between them in the point set, establish a first subset and a second subset based on the obtained points, and assign the remaining points in the point set to the nearest subset. S133, Calculate the compactness of the point set, the first subset, and the second subset; compactness The specific expression is: In the formula, m For point set D The number of midpoints p j For point set D The point in the middle, C For point set D The center of mass; Determine whether the splitting condition is met. The specific splitting condition is as follows: and In the formula, For the compactness of the first subset, For the compactness of the second subset, For the compactness of the point set, For point set D The total number of points, n The total number of points in the point cloud; If yes, then both the first and second subsets are treated as point sets to be partitioned, and return to S132; otherwise, the current point set is treated as the simplified point cloud, and proceed to S134. S134. Create a 3D dataset based on all simplified point clouds; S2. Perform continuous homology analysis based on the three-dimensional dataset to obtain the topological characteristics of the box girder; The dimensions of the topological features include 0-dimensional, 1-dimensional, and 2-dimensional. Among them, 0-dimensional topological features represent connected branches in point clouds, 1-dimensional topological features represent closed loops or holes, and 2-dimensional topological features represent cavities or closed three-dimensional spaces. S3. Analyze the topological characteristics of the box girder and mark the areas where structural problems exist; the specific method for analyzing the topological characteristics of the box girder is as follows: The stress of the box girder was analyzed by finite element analysis, and the stress distribution in different regions under external force was calculated. By comparing the stress distribution with the topological features, the degree of overlap between the topological features and the stress concentration areas was identified. For the geometric region containing the topological features, fatigue life assessment was conducted using a mechanical model. The stress state and load history of the geometric region were analyzed to evaluate its potential for fatigue failure under long-term cyclic loading. S4. Perform physical interpretation of topological features based on the marked areas and generate a health assessment report; methods for physical interpretation of topological features include crack and damage detection and cavity and material defect identification; The specific methods for detecting cracks and damage are as follows: Detecting connectivity interruptions or new connecting branches indicates that there is a break or damage in a local area of the box girder. By analyzing the geometric structure and mechanics, the impact of the connecting branch on the overall performance of the box girder can be guided to improve maintenance work. The specific methods for material defect identification are as follows: The cavity structure is extracted from the 2D topological features. The specific location and shape of the cavity structure are determined by continuous homology analysis, and the potential threat of the cavity structure to structural safety is determined by mechanical analysis.
2. The method for topological structure recognition of 3D box girder point cloud data based on continuous homology according to claim 1, characterized in that, In S11, the first i Neighborhood density of each point The specific expression is: In the formula, For the first i The point and the first j Euclidean distance between points The selected cutoff distance.
3. The method for topological structure recognition of 3D box girder point cloud data based on continuous homology according to claim 1, characterized in that, S2 includes the following steps: S21. Based on each point in the point cloud of the 3D dataset, calculate the Euclidean distance between the points and generate a distance matrix; S22. Set the neighborhood radius and create a persistent bar chart to analyze the persistence of the topological features generated from each point cloud, and obtain the topological features of the box girder.
4. The method for topological structure recognition of 3D box girder point cloud data based on continuous homology according to claim 3, characterized in that, Specifically, S22 is: Set the neighborhood radius, gradually increase the neighborhood radius within the framework of continuous homology, calculate the topological features at different scales of neighborhood radius, create a persistent bar chart based on the life cycle of the topological features, and retain the topological features whose life cycle exceeds 10% of the overall time length as the topological features of the box girder. The neighborhood radius is used to determine whether points are considered connected. If the Euclidean distance between two points is less than the neighborhood radius, then the two points are connected in the topological space. The lifecycle is specifically the time when topological features appear and disappear.
5. The method for topological structure recognition of 3D box girder point cloud data based on continuous homology according to claim 1, characterized in that, In S4, the health assessment report includes the overall connectivity of the box girder, local damage, and areas requiring key monitoring or repair.