Ship 3D scanning point cloud data processing system and method thereof

Abstract

The application discloses a kind of ship 3D scanning point cloud data processing system and method thereof, it is related to ship safety analysis technical field, including point cloud data management center, the point cloud data management center communication connection has ship point cloud data acquisition module, hull surface reconstruction module, ship deformation analysis module and visual interactive module, wherein, each module electric signal connection between each other.This application can comprehensively cover each part of ship through multi-view point cloud data acquisition and registration, effectively avoids scanning blind area, at the same time, accurate registration is carried out to multi-view point cloud data using iterative closest point algorithm, unified to the same coordinate system, ensure the accurate alignment of data collected at different angles in space, not only greatly improve the precision of ship deformation detection, but also can significantly reduce the misjudgment caused by data misalignment, so as to improve the efficiency of the whole detection process.

Description

Technical Field

[0001] This invention relates to the field of ship safety analysis technology, specifically to a ship 3D scanning point cloud data processing system and method. Background Technology

[0002] As a vital pillar of the national economy, the shipbuilding industry shoulders multiple missions, including maritime transportation, resource development, and national defense. With the rapid development of information technology, the shipbuilding industry is undergoing a profound transformation from traditional to digital and intelligent models. During long-term use, ships are affected by various factors such as seawater erosion and collisions, leading to component damage or hull deformation. In order to detect and address these problems in a timely manner, ships need to undergo regular maintenance and modifications. Ship 3D scanning can quickly acquire three-dimensional data of the ship before and after use. By comparing and analyzing the data, the degree of deformation and damaged parts can be identified, providing accurate data for the formulation of maintenance plans.

[0003] For example, a method for processing point cloud data of ship 3D scanning in Chinese Patent Publication No. CN111968224A includes: a three-dimensional laser scanning step: using a scanner at multiple stations to obtain point cloud data of their respective scanning segments through 3D laser scanning.

[0004] In existing technologies, although 3D laser scanning technology can be used to quickly acquire 3D models of various parts inside and outside the ship, the ship will be affected by various factors during long-term use, resulting in local or overall deformation of the hull. Point cloud data needs to be collected from different perspectives to improve the accuracy of deformation detection. Therefore, how to accurately identify the deformed area from the point cloud data, especially in the case of small deformation, is a problem we need to solve. To this end, a ship 3D scanning point cloud data processing system and method are proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a ship 3D scanning point cloud data processing system and method to solve the problems mentioned in the background art.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: In a first aspect, a ship 3D scanning point cloud data processing system includes a point cloud data management center, wherein the point cloud data management center is communicatively connected to a ship point cloud data acquisition module, a hull surface reconstruction module, a ship deformation analysis module, and a visualization interaction module, wherein the modules are electrically connected to each other. The point cloud data management center is responsible for the storage, backup and management of point cloud data, ensuring data security and accessibility, providing efficient data storage and management solutions, and supporting long-term data preservation and fast retrieval. The ship point cloud data acquisition module is used to scan the ship from different perspectives using a 3D laser scanner to obtain multi-view point cloud data, and to perform registration processing on the multi-view point cloud data. The hull surface reconstruction module is used to reconstruct a three-dimensional model of the ship based on the processed point cloud data, generate a complete hull surface, and provide a detailed three-dimensional representation of the ship's shape. The ship deformation analysis module is used to detect deformations on the surface of the hull, and to perform quantitative analysis on the detected deformation areas to calculate the degree and direction of the deformation. The visualization interaction module is used to summarize the processed point cloud data and deformation analysis results to generate a visualization report and display it visually. At the same time, it displays the deformation of the hull in real time. Through intuitive 3D model display and interactive operation, it is convenient for users to understand and analyze the deformation, thus improving the usability of the system.

[0007] A further improvement of the technical solution of the present invention is that: the ship point cloud data acquisition module includes a data acquisition unit and a point cloud data registration unit; The data acquisition unit is used to collect multi-view point cloud data of the ship from different perspectives of each station on the ship, ensuring that the point cloud data covers all parts of the ship, and preprocessing the collected multi-view point cloud data to improve data quality, achieve accurate alignment and integration of multi-view point cloud data, and then divide the preprocessed multi-view point cloud data into multiple point cloud datasets according to the station. The point cloud data registration unit is used to register multiple point cloud datasets using the Iterative Closest Point (ICP) algorithm, unifying them into the same coordinate system, achieving accurate registration of point cloud data from multiple perspectives, and ensuring that point cloud data collected from different perspectives are accurately aligned in space.

[0008] A further improvement to the technical solution of the present invention is that the acquisition process of multi-view point cloud data in the data acquisition unit includes: Based on the size and shape of the ship, plan and arrange the measuring stations around the ship, determine the location of multiple measuring stations, ensure that the measuring stations can cover all parts of the ship, avoid scanning blind spots, and formulate a scanning plan, including the scanning angle, scanning distance and scanning sequence of each measuring station, to ensure the efficiency of the scanning process and the integrity of the data. After the survey stations are set up, a 3D laser scanner is installed, and the ship is scanned from different perspectives at each survey station according to the predetermined scanning plan to collect multi-view point cloud data. The collected multi-view point cloud data is preprocessed, including denoising, filtering, and cropping. Denoising eliminates random noise points generated during the scanning process, filtering smooths the point cloud data, reduces data fluctuations and irregularities, and makes the point cloud surface smoother. Cropping removes unwanted parts such as the scanned background or debris to improve the quality of the point cloud data. The preprocessed multi-view point cloud data is integrated to form a complete dataset, and then divided into multiple point cloud datasets according to the station. Each dataset corresponds to the scanning results of one station.

[0009] A further improvement to the technical solution of the present invention is that the process of registering multiple point cloud datasets in the point cloud data registration unit includes: A preliminary check is performed on the multiple point cloud datasets to be registered to ensure the integrity and consistency of the datasets. A reference point cloud dataset is selected as the reference point cloud for registration, and the remaining point cloud datasets are used as target point clouds. The target point cloud is aligned with the reference point cloud dataset. After determining the reference point cloud and the target point cloud, the target point cloud is placed near the reference point cloud according to the preliminary alignment result to obtain an approximate transformation matrix. Start the Iterative Closest Point (ICP) algorithm and initialize relevant parameters, including the maximum number of iterations and the convergence threshold. The maximum number of iterations is used to limit the running time of the algorithm to prevent it from getting stuck in an infinite loop. The convergence threshold is used to determine whether the algorithm has converged to a stable state. The approximate transformation matrix obtained from the preliminary alignment results is used as the initial transformation matrix. Fine registration is performed using the iterative nearest point algorithm. In each iteration, the nearest point from each point in the target point cloud to the reference point cloud is calculated to establish the correspondence between point pairs. Based on the correspondence between point pairs, the rotation and translation transformation matrix from the target point cloud to the reference point cloud is calculated. The optimal transformation matrix is ​​obtained by minimizing the sum of squared distances between the target point cloud and the reference point cloud. After obtaining the new transformation matrix, the target point cloud is updated by applying the transformation matrix to make it closer to the reference point cloud. The above process is repeated until the convergence condition is met or the maximum number of iterations is reached. The final transformation matrix is ​​applied to the point cloud dataset to be registered, so that it is accurately aligned with the reference point cloud dataset in space, and then the registered point cloud dataset and transformation matrix are saved.

[0010] A further improvement to the technical solution of this invention lies in that: in the hull surface reconstruction module, the process of reconstructing the three-dimensional model of the ship includes: The multi-view point cloud data after registration is imported into the hull surface reconstruction module. Based on the structural features of the hull, the point cloud data is segmented into different components including the bow, stern, side and deck. The point cloud data is then organized to establish the topological relationship between the point clouds. Based on the characteristics of the hull surface and the reconstruction requirements, a surface reconstruction algorithm for surface fitting is applied to process the point cloud data to generate a preliminary hull surface model. The initially generated hull surface model is optimized and adjusted through smoothing and detail enhancement. In particular, the smoothing algorithm is applied to remove noise from the model surface to make it smoother, and the detail enhancement process is applied to highlight the lines and uneven features of the hull surface to improve the visual effect and realism of the model. This process generates a complete 3D hull surface model to display the complete hull surface. Finally, 3D visualization technology is used to display the complete 3D hull surface model in 3D space.

[0011] A further improvement of the technical solution of the present invention is that the ship deformation analysis module includes a feature extraction unit, a deformation detection unit, and a deformation quantification evaluation unit; The feature extraction unit is used to extract key feature points from the processed point cloud data, obtain point cloud feature sequences, simplify the complexity of point cloud data, facilitate deformation analysis and visualization, and improve processing efficiency. The deformation detection unit is used to compare point cloud differences using point cloud feature sequences, and to identify the deformation area of ​​the hull by comparing the normal hull model with the reconstructed three-dimensional ship model, accurately detecting the local or overall deformation of the hull, and providing reliable data for maintenance and modification. The deformation quantification assessment unit is used to quantify and analyze the detected deformation area, calculate the deformation size, direction and depth parameters, and then comprehensively calculate the deformation assessment index to quantify and analyze the degree of deformation in the ship's deformation area, providing accurate deformation quantification data, providing specific guidance for maintenance and modification, and helping to assess the overall structural safety of the ship.

[0012] A further improvement to the technical solution of the present invention is that the point cloud feature sequence acquisition process in the feature extraction unit includes: Based on the characteristics of ship point cloud data and the needs of deformation analysis, key feature points are selected from the preprocessed point cloud data. Through curvature calculation and normal vector analysis, local geometric features reflecting ship deformation are extracted. Among them, key feature points include the corner points, edge points and areas with large curvature changes of the hull. By calculating the geometric relationship between each point in the preprocessed point cloud data and its neighboring points, the curvature value of the point is estimated, and the normal vector of each point is calculated. The shape changes of the hull surface are analyzed, and each point is assigned a curvature value and a normal vector, thereby obtaining the local geometric features of each point in the point cloud data. Based on the calculated local geometric features, set curvature threshold and normal vector change threshold, traverse each point in the point cloud data, check whether its curvature value and normal vector change meet the set threshold conditions, and filter out the points that meet the conditions as key feature points. Based on the location of key feature points in the point cloud data, the key feature points are sorted, and the sorted key feature points are combined into a point cloud feature sequence to facilitate subsequent processing and analysis.

[0013] A further improvement to the technical solution of the present invention is that the process of identifying the deformation area of ​​the hull in the deformation detection unit includes: Point cloud feature sequences are extracted from the reconstructed 3D ship model and the normal ship model respectively. The key feature points in the point cloud feature sequences are used to reflect the shape changes and geometric features of the ship surface. The point cloud feature sequences of the reconstructed 3D ship model and the normal ship model are aligned so that the two point cloud feature sequences completely overlap in space. The geometric properties of the position, curvature, and normal vector of corresponding key feature points in the reconstructed 3D ship model and the normal hull model are compared. Feature difference values ​​are calculated, including Euclidean distance, curvature difference, and normal vector angle between feature points, to reflect the degree of change of the hull surface in local areas. A difference threshold is set, and key feature points exceeding the difference threshold are marked as potential deformation areas. Here, Euclidean distance is the spatial distance between corresponding key feature points in the two models, reflecting positional changes; curvature difference is the curvature difference between corresponding key feature points in the two models, reflecting local shape changes; and normal vector angle is the angle between the normal vectors of corresponding key feature points in the two models, reflecting surface orientation changes. By analyzing the spatial distribution of feature difference values, the range and location of deformed regions are determined. The clustering algorithm (DBSCAN) is used to cluster key feature points with high feature difference values ​​to identify independent deformed regions. The baseline values ​​of the curvature and normal vector angle of key feature points are determined by combining the normal hull model. The average local area of ​​the normal hull model corresponding to the key feature points is used as the baseline value. Then, the deformation region discrimination index is calculated by combining the structural features and geometric information of the hull, and the identified deformation regions are classified to distinguish between local deformation and overall deformation.

[0014] A further improvement to the technical solution of this invention lies in the fact that the process of quantifying and analyzing the degree of deformation in the deformation area of ​​the ship in the deformation quantification evaluation unit includes: Parameters are extracted from the detected deformed areas, including deformation size, direction, and depth. Deformation size reflects the extent and degree of the deformed area, deformation direction indicates the direction in which the deformation occurs, and deformation depth reflects the change in the depth or height of the deformation. Calculate the change in Euclidean distance between corresponding key feature points before and after deformation to obtain the displacement in each direction. Perform statistical analysis on the displacement in all directions to calculate the average expansion or contraction of the deformed area and obtain the deformation size. For each key feature point, calculate its displacement vector before and after deformation. Perform principal component analysis on all displacement vectors to obtain the principal direction vector of the deformed area. This principal direction vector is the deformation direction. Use a polynomial fitting surface fitting algorithm to fit the point cloud data of the deformed area to obtain the surface model of the deformed area. Calculate the maximum curvature of the surface model and its corresponding depth change. The depth change is the corresponding deformation depth. The safety margins of deformation magnitude, deformation direction, and deformation depth parameters are analyzed to obtain the safety margins for deformation magnitude, deformation direction, and deformation depth. These safety margins are then weighted and summed to calculate the deformation assessment index, quantifying the degree of deformation in the ship's deformation area. Here, the safety margin refers to the difference between the actual degree of deformation in the deformation area and the maximum allowable degree of deformation. The safety margin for deformation magnitude is the difference between the calculated deformation magnitude and the maximum allowable deformation magnitude. The safety margin for deformation direction is the deviation between the analyzed deformation direction and the allowable deformation direction of the hull structure. The safety margin for deformation depth is the difference between the calculated deformation depth and the maximum allowable deformation depth. Based on the magnitude of the deformation assessment index, the deformation area is divided into different deformation levels: slight deformation, moderate deformation, and severe deformation, with each deformation level corresponding to a range of deformation severity.

[0015] Secondly, a method for processing ship 3D scanning point cloud data, based on the aforementioned ship 3D scanning point cloud data processing system, includes the following steps: S1. Use a 3D laser scanner to scan the ship from different perspectives to obtain multi-view point cloud data, and perform preprocessing operations such as noise reduction, filtering, and cropping to ensure coverage of all parts of the ship. S2. The iterative nearest point algorithm is used to register the preprocessed multi-view point cloud data and unify them to the same coordinate system to achieve precise alignment. S3. Based on the registered point cloud data, apply the surface fitting algorithm to reconstruct the three-dimensional model of the hull and generate the complete hull surface. S4. Extract key feature points from the reconstructed model and the normal model, compare the differences between the normal hull model and the reconstructed model, and identify and quantify the deformation area. S5. Summarize the processed point cloud data and deformation analysis results to generate a visualization report. The report displays the hull deformation in real time through a 3D model and interactive operation, making it easier for users to understand and analyze.

[0016] Due to the adoption of the above technical solution, the technical progress achieved by this invention compared to the prior art is as follows: 1. This invention provides a ship 3D scanning point cloud data processing system and method. Through multi-view point cloud data acquisition and registration, it can comprehensively cover all parts of the ship, effectively avoiding scanning blind spots. At the same time, it uses an iterative nearest point algorithm to accurately register multi-view point cloud data and unify them to the same coordinate system, ensuring accurate spatial alignment of data acquired from different perspectives. This not only greatly improves the accuracy of ship deformation detection but also significantly reduces misjudgments caused by data misalignment, thereby improving the efficiency of the entire detection process.

[0017] 2. This invention provides a ship 3D scanning point cloud data processing system and method. By identifying the deformed areas of the hull and performing detailed quantitative analysis of the deformed areas, including parameters such as deformation size, direction, and depth, it provides specific data basis for maintenance and modification, and helps to formulate more scientific and reasonable maintenance plans. At the same time, by calculating deformation evaluation indicators, it comprehensively evaluates the degree of deformation of the ship's deformed areas, helping users to understand the deformation of the hull more intuitively, so as to ensure the overall structural safety of the ship. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a schematic diagram of the system functional modules of the present invention; Figure 2 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1, as Figure 1As shown, the present invention provides a ship 3D scanning point cloud data processing system, including a point cloud data management center. The point cloud data management center is communicatively connected to a ship point cloud data acquisition module, a hull surface reconstruction module, a ship deformation analysis module, and a visualization interaction module, wherein the modules are electrically connected. The point cloud data management center is responsible for the storage, backup and management of point cloud data, ensuring data security and accessibility, providing efficient data storage and management solutions, and supporting long-term data preservation and fast retrieval. The ship point cloud data acquisition module is used to scan the ship from different perspectives using a 3D laser scanner to obtain multi-view point cloud data, and to perform registration processing on the multi-view point cloud data. The ship point cloud data acquisition module includes a data acquisition unit and a point cloud data registration unit. The data acquisition unit is used to collect multi-view point cloud data of the ship from different perspectives at various monitoring stations, ensuring that the point cloud data covers all parts of the ship. It preprocesses the collected multi-view point cloud data to improve data quality and achieve precise alignment and integration. The preprocessed multi-view point cloud data is then divided into multiple point cloud datasets according to the monitoring stations. Based on the ship's size and shape, monitoring stations are planned and arranged around the ship, determining their positions to ensure coverage of all parts of the ship while avoiding blind spots. A scanning plan is developed, including the scanning angle, scanning distance, and scanning sequence for each station, to ensure efficiency and data integrity. After the monitoring stations are arranged, a 3D laser scanner is installed, and the ship is scanned from different perspectives at each station according to the predetermined scanning plan to collect multi-view point cloud data. Operators sequentially activate the 3D laser scanner according to the predetermined monitoring station positions. During data acquisition and scanning, the 3D laser scanner emits a laser beam. After the laser beam hits the surface of the ship, it reflects back. The 3D laser scanner calculates the spatial coordinates of the scanned points by measuring the flight time and angle information of the laser, thereby generating point cloud data. To ensure the integrity and accuracy of the data, the scanning time of each station must be long enough to obtain high-density point cloud data. The acquired multi-view point cloud data is preprocessed, including noise reduction, filtering, and cropping. Noise reduction eliminates random noise points generated during the scanning process. Filtering smooths the point cloud data, reducing data fluctuations and irregularities, making the point cloud surface smoother. Cropping removes unwanted parts such as background or debris from the scanned area to improve the quality of the point cloud data. The preprocessed multi-view point cloud data is integrated to form a complete dataset. The integrated multi-view point cloud data is then divided into multiple point cloud datasets according to the station, with each dataset corresponding to the scanning results of one station. The point cloud data registration unit is used to register multiple point cloud datasets using the Iterative Closest Point (ICP) algorithm, unifying them to the same coordinate system to achieve accurate registration of multi-view point cloud data. This ensures that point cloud data acquired from different perspectives are spatially aligned accurately. The unit performs a preliminary check on the multiple point cloud datasets to be registered, ensuring their integrity and consistency. A reference point cloud dataset is selected as the reference point cloud for registration, and the remaining point cloud datasets are used as target point clouds. Alignment is performed with the reference point cloud dataset. After determining the reference and target point clouds, the target point cloud is placed near the reference point cloud according to the preliminary alignment result, resulting in an approximate transformation matrix. The Iterative Closest Point (ICP) algorithm is then started, and relevant parameters are initialized, including the maximum number of iterations and the convergence threshold. The maximum number of iterations limits the algorithm's running time to prevent it from getting stuck in an infinite loop, and the convergence threshold is used to determine... The algorithm is checked to see if it has converged to a stable state. The approximate transformation matrix obtained from the initial alignment results is used as the initial transformation matrix. Fine registration is performed using the iterative nearest point algorithm. In each iteration, the nearest point from each point in the target point cloud to the reference point cloud is calculated to establish the correspondence between point pairs. Based on the correspondence between point pairs, the rotation and translation transformation matrix from the target point cloud to the reference point cloud is calculated. The optimal transformation matrix is ​​solved by minimizing the sum of squared distances between the target point cloud and the reference point cloud. After obtaining the new transformation matrix, the target point cloud is updated by applying the transformation matrix to make it closer to the reference point cloud. The above process is repeated until the convergence condition is met or the maximum number of iterations is reached. The final transformation matrix is ​​applied to the point cloud dataset to be registered to make it accurately aligned with the reference point cloud dataset in space. The registered point cloud dataset and transformation matrix are then saved. The hull surface reconstruction module is used to reconstruct a three-dimensional model of the ship based on the processed point cloud data, generating a complete hull surface and providing a detailed three-dimensional representation of the ship's shape. The module imports the registered multi-view point cloud data into the hull surface reconstruction module and segments the point cloud data according to the structural features of the hull, dividing it into different components including the bow, stern, side, and deck. The point cloud data is then organized, establishing topological relationships between the point clouds. Based on the characteristics of the hull surface and reconstruction requirements, a surface reconstruction algorithm using surface fitting is applied to process the point cloud data, generating a preliminary hull surface model. This preliminary model is then optimized and adjusted through smoothing and detail enhancement. Specifically, a smoothing algorithm removes noise from the model surface, making it smoother, while detail enhancement highlights the lines and uneven features of the hull surface, improving the model's visual effect and realism. This generates a complete three-dimensional hull surface model to display the complete hull surface. Finally, using three-dimensional visualization technology, the complete three-dimensional hull surface model is displayed in three-dimensional space. The ship deformation analysis module is used to detect deformations on the surface of the ship and to perform quantitative analysis on the detected deformation areas, calculating the degree and direction of the deformation. The visualization interaction module is used to summarize the processed point cloud data and deformation analysis results to generate a visualization report and display it visually. At the same time, it displays the deformation of the hull in real time. Through intuitive 3D model display and interactive operation, it is convenient for users to understand and analyze the deformation, thus improving the usability of the system.

[0022] Example 2, as Figure 1 As shown, based on Embodiment 1, the present invention provides a technical solution: preferably, the ship deformation analysis module includes a feature extraction unit, a deformation detection unit, and a deformation quantification evaluation unit; The feature extraction unit is used to extract key feature points from the processed point cloud data, obtain point cloud feature sequences, simplify the complexity of point cloud data, facilitate deformation analysis and visualization, and improve processing efficiency. Based on the characteristics of ship point cloud data and the needs of deformation analysis, key feature points are selected from the preprocessed point cloud data. Through curvature calculation and normal vector analysis, local geometric features reflecting ship deformation are extracted. Key feature points include corner points, edge points, and areas with significant curvature changes in the hull. The curvature value of each point is estimated by calculating the geometric relationship between each point and its neighboring points in the preprocessed point cloud data, and the normal vector of each point is calculated. The shape changes on the hull surface are analyzed, and each point is assigned a curvature value and a normal vector, thereby obtaining the key feature points from the preprocessed point cloud data. The local geometric features of each point are calculated, and curvature thresholds and normal vector change thresholds are set. Each point in the point cloud data is traversed, and its curvature value and normal vector change are checked to see if they meet the set threshold conditions. Points that meet the conditions are selected as key feature points. The curvature threshold is set according to the statistical characteristics of ship point cloud data. Points with curvature higher than the curvature threshold are considered potential key feature points. The normal vector change threshold is determined by calculating the change in the angle between normal vectors. When the normal vector change exceeds the normal vector change threshold, the point is considered a key feature point. The key feature points are sorted according to their position in the point cloud data, and the sorted key feature points are combined into a point cloud feature sequence for subsequent processing and analysis. The deformation detection unit is used to compare point cloud differences using point cloud feature sequences, and to identify deformed areas of the hull by comparing a normal hull model with a reconstructed 3D ship model. This accurately detects local or overall hull deformation, providing reliable data for maintenance and modification. Point cloud feature sequences are extracted from both the reconstructed 3D ship model and the normal hull model. Key feature points in these sequences reflect the shape changes and geometric features of the hull surface. The point cloud feature sequences of the reconstructed 3D ship model and the normal hull model are aligned so that they completely overlap in space. The geometric properties of the corresponding key feature points, such as position, curvature, and normal vectors, are compared between the reconstructed 3D ship model and the normal hull model. Feature difference values ​​are calculated, including the Euclidean distance, curvature difference, and normal vector angle between feature points, to reflect the degree of change on the hull surface in local areas. A difference threshold is set, and key feature points exceeding the threshold are marked as potential deformation areas. The Euclidean distance is the spatial distance between corresponding key feature points in the two models. The curvature difference is calculated by taking the curvature difference of corresponding key feature points in two models to reflect local shape changes, and the normal vector angle is calculated by taking the angle between the normal vectors of corresponding key feature points in two models to reflect surface direction changes. By analyzing the spatial distribution of feature difference values, the range and location of the deformation area are determined. The clustering algorithm (DBSCAN) is used to cluster key feature points with high feature difference values ​​to identify independent deformation areas. The baseline values ​​of the curvature and normal vector angle of key feature points are determined by combining the normal hull model. The average local area of ​​the normal hull model corresponding to the key feature points is used as the baseline value. Then, combined with the structural features and geometric information of the hull, the deformation area discrimination index is calculated to classify the identified deformation areas and distinguish between local deformation and overall deformation. The clustering algorithm groups adjacent high feature difference value points together to form deformation areas based on the spatial location and feature difference value of key feature points. Local deformation is manifested as a cluster of high feature difference values ​​in a small area, while overall deformation involves shape changes in a larger area. The formula for calculating the deformation region differentiation index is as follows: ; ; In the formula, An index for distinguishing deformation regions, used to differentiate between localized and overall deformation. This represents the number of key feature points within the deformation region. For the first Euclidean distance of key feature points The standard deviation of the Euclidean distance is used for normalization. , For the first The curvature of key feature points The average curvature of a normal ship hull model is used as a reference value. For the first The angle between the normal vectors of the key feature points The average angle between the normal vectors of a normal hull model is used as the reference value. and These are the normal vectors of the corresponding key feature points in the two models. This is the distance reference value, used for the exponential decay term. For the first The area of ​​the local region where each key feature point is located. The average local area of ​​a normal ship hull model is used as a baseline value. The larger the value, the more severe the deformation. The larger, the better The greater the contribution, the more it is affected by the exponentially decaying term. The impact of points that are too far apart The contribution will decrease. and The larger, the better The greater the contribution, the more significant the local shape change. The larger, the better The greater the contribution, the wider the range of the deformation region; The deformation quantification assessment unit is used to quantify and analyze the detected deformation area, calculate the deformation magnitude, direction, and depth parameters, and then comprehensively calculate the deformation assessment index to quantify and analyze the degree of deformation in the ship's deformation area, providing accurate deformation quantification data to provide specific guidance for maintenance and modification. It also helps to assess the overall structural safety of the ship. Parameters are extracted from the detected deformation area, including deformation magnitude, direction, and depth. Deformation magnitude reflects the range and extent of the deformation area, deformation direction indicates the direction of deformation, and deformation depth reflects the change in depth or height of the deformation. The Euclidean distance change between corresponding key feature points before and after deformation is calculated to obtain the displacement in each direction. Statistical analysis is performed on the displacements in all directions to calculate the average expansion or contraction of the deformation area, thus obtaining the deformation magnitude. For each key feature point, its displacement vector before and after deformation is calculated. Principal component analysis is performed on all displacement vectors to obtain the principal direction vector of the deformation area, which is the deformation direction. A polynomial-fitted surface model is then used for further analysis. The algorithm fits the point cloud data of the deformed region to obtain the surface model of the deformed region, calculates the maximum curvature of the surface model and its corresponding depth change, and the depth change corresponds to the deformation depth. It analyzes the safety margins of the deformation magnitude, deformation direction and deformation depth parameters, obtains the deformation magnitude safety margin, deformation direction safety margin and deformation depth safety margin, and then calculates the deformation assessment index by weighted summation to quantify the deformation degree of the ship's deformation region. Among them, the safety margin refers to the difference between the actual deformation degree of the deformation region and the maximum allowable deformation degree. The deformation magnitude safety margin is the difference between the calculated deformation magnitude and the maximum allowable deformation magnitude. The deformation direction safety margin is the deviation between the analyzed deformation direction and the allowable deformation direction of the hull structure. The deformation depth safety margin is the difference between the calculated deformation depth and the maximum allowable deformation depth. According to the magnitude of the deformation assessment index, the deformation region is divided into different deformation levels, namely slight deformation level, moderate deformation level and severe deformation level, and each deformation level corresponds to a deformation degree range. It should be noted that the process of performing principal component analysis on all displacement vectors to obtain the principal direction vectors of the deformed region is as follows: The displacement vectors of all key feature points are collected and arranged into a matrix of size n*3. The data is centered by calculating the mean of each displacement vector and obtaining the mean vector. The mean vector is then subtracted from each displacement vector to obtain the centered data matrix. The covariance matrix of the centered data matrix is ​​calculated, and the eigenvalues ​​and eigenvectors of the covariance matrix are solved to obtain the principal direction of the deformation region. Here, the eigenvalues ​​represent the variance of each principal component, and the eigenvectors represent the direction of the principal components. The eigenvalues ​​are sorted from largest to smallest. The eigenvector corresponding to the largest eigenvalue is the first principal component, representing the principal direction of the deformation region. The eigenvector corresponding to the second largest eigenvalue is the second principal component, representing the secondary direction. The eigenvector corresponding to the smallest eigenvalue is the third principal component, representing the least important direction. From the sorted eigenvalues ​​and eigenvectors, the eigenvector corresponding to the largest eigenvalue is extracted as the principal direction vector of the deformation region. The principal direction vector represents the main direction of the deformation region and reflects the main trend of deformation. The formula for calculating the deformation evaluation index is as follows: ; In the formula, As a deformation evaluation index, To provide a safety margin for the size of the deformation, For safety margin in the deformation direction, For the safety margin of deformation depth, , and These are weighting coefficients, representing the importance of deformation magnitude, deformation direction, and deformation depth in the comprehensive evaluation. The maximum allowable deformation direction deviation, The maximum allowable deformation size, The maximum allowable deformation depth; The formula for calculating the safety margin of the deformation size is: ; In the formula, The size of the deformation; The formula for calculating the safety margin in the deformation direction is: ; In the formula, Let be the principal direction vector of the deformable region. This is the allowable deformation direction vector of the hull structure; The formula for calculating the safety margin of the deformation depth is: ; In the formula, The deformation depth; The formula for calculating the deformation magnitude is: ; In the formula, and These are the first and second parts after deformation and before deformation, respectively. Spatial distance of key feature points It is the number of key feature points; The formula for calculating the displacement vector before and after deformation is: ; In the formula, Let be the displacement vector before and after deformation. and These are the first and second parts after deformation and before deformation, respectively. The coordinates of the key feature points; The formula for calculating the deformation depth is: ; ; In the formula, For maximum curvature, For the first Normal vectors of key feature points For the first Displacement vectors of key feature points; The multiple deformation levels correspond one-to-one with the multiple deformation degree ranges, specifically: The deformation range of the slight deformation level is as follows: This indicates that the deformation is relatively minor and has little impact on the structural safety of the ship. The deformation range of the medium deformation level is as follows: This indicates a moderate degree of deformation, requiring attention and appropriate repair measures. The deformation range of the severe deformation level is as follows: ; indicates severe deformation, which may have a significant impact on the structural safety of the ship and requires immediate repair; in, As a deformation evaluation index, These are the upper threshold for slight deformation and the lower threshold for moderate deformation. These represent the upper threshold for moderate deformation and the lower threshold for severe deformation. , .

[0023] Example 3, as Figure 2 As shown, based on Examples 1-2, the present invention also provides a method for processing ship 3D scanning point cloud data, which is implemented based on a ship 3D scanning point cloud data processing system, and includes the following steps: S1. Use a 3D laser scanner to scan the ship from different perspectives to obtain multi-view point cloud data, and perform preprocessing operations such as noise reduction, filtering, and cropping to ensure coverage of all parts of the ship. S2. The iterative nearest point algorithm is used to register the preprocessed multi-view point cloud data and unify them to the same coordinate system to achieve precise alignment. S3. Based on the registered point cloud data, apply the surface fitting algorithm to reconstruct the three-dimensional model of the hull and generate the complete hull surface. S4. Extract key feature points from the reconstructed model and the normal model, compare the differences between the normal hull model and the reconstructed model, and identify and quantify the deformation area. S5. Summarize the processed point cloud data and deformation analysis results to generate a visualization report. The report displays the hull deformation in real time through a 3D model and interactive operation, making it easier for users to understand and analyze.

[0024] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A ship 3D scanning point cloud data processing system, comprising a point cloud data management center, characterized in that: The point cloud data management center is connected to a ship point cloud data acquisition module, a hull surface reconstruction module, a ship deformation analysis module, and a visualization interaction module, wherein the modules are connected by electrical signals. The ship point cloud data acquisition module is used to scan the ship from different perspectives using a 3D laser scanner to obtain multi-view point cloud data, and to perform registration processing on the multi-view point cloud data. The hull surface reconstruction module is used to reconstruct a three-dimensional model of the ship based on the processed point cloud data, and generate a complete hull surface. The ship deformation analysis module is used to detect deformations on the surface of the ship and to perform quantitative analysis on the detected deformation areas, calculating the degree and direction of the deformation. The ship deformation analysis module includes a feature extraction unit, a deformation detection unit, and a deformation quantification evaluation unit. The feature extraction unit is used to extract key feature points from the processed point cloud data to obtain a point cloud feature sequence. The deformation detection unit is used to compare point cloud differences using point cloud feature sequences, and to identify the deformation area of ​​the hull by comparing the normal hull model with the reconstructed three-dimensional ship model, thereby detecting local or overall deformation of the hull. The deformation quantification assessment unit is used to quantify and analyze the detected deformation area, calculate the deformation magnitude, direction, and depth parameters, and then comprehensively calculate the deformation assessment index to quantify and analyze the degree of deformation in the ship's deformation area, specifically including: Parameters are extracted from the detected deformed areas, including deformation size, direction, and depth. Deformation size reflects the extent and degree of the deformed area, deformation direction indicates the direction in which the deformation occurs, and deformation depth reflects the change in the depth or height of the deformation. Calculate the change in Euclidean distance between corresponding key feature points before and after deformation to obtain the displacement in each direction. Perform statistical analysis on the displacement in all directions to calculate the average expansion or contraction of the deformed area and obtain the deformation size. For each key feature point, calculate its displacement vector before and after deformation. Perform principal component analysis on all displacement vectors to obtain the principal direction vector of the deformed area. This principal direction vector is the deformation direction. Use a polynomial fitting surface fitting algorithm to fit the point cloud data of the deformed area to obtain the surface model of the deformed area. Calculate the maximum curvature of the surface model and its corresponding depth change. The depth change is the corresponding deformation depth. The safety margins of deformation magnitude, deformation direction, and deformation depth parameters are analyzed to obtain the safety margins of deformation magnitude, deformation direction, and deformation depth. Then, the deformation assessment index is calculated by weighted summation to quantify the degree of deformation in the deformation area of ​​the ship. Based on the magnitude of the deformation assessment index, the deformation area is divided into different deformation levels, namely slight deformation level, moderate deformation level and severe deformation level, and each deformation level corresponds to a range of deformation degree. The visualization interaction module is used to summarize the processed point cloud data and deformation analysis results to generate a visualization report and display it visually.

2. The ship 3D scanning point cloud data processing system according to claim 1, characterized in that: The ship point cloud data acquisition module includes a data acquisition unit and a point cloud data registration unit; The data acquisition unit is used to collect multi-view point cloud data of the ship from different perspectives of each station on the ship, and to preprocess the collected multi-view point cloud data, and then divide the preprocessed multi-view point cloud data into multiple point cloud datasets according to the station. The point cloud data registration unit is used to register multiple point cloud datasets using an iterative nearest point algorithm, unifying them into the same coordinate system.

3. The ship 3D scanning point cloud data processing system according to claim 2, characterized in that: The data acquisition unit includes the following process for acquiring multi-view point cloud data: Based on the size and shape of the ship, plan and arrange the measuring stations around the ship, determine the positions of multiple measuring stations, and formulate a scanning plan, including the scanning angle, scanning distance and scanning sequence of each measuring station; After the survey stations are set up, a 3D laser scanner is installed, and the ship is scanned from different perspectives at each survey station according to the predetermined scanning plan to collect multi-view point cloud data. Preprocessing of the acquired multi-view point cloud data includes noise reduction, filtering, and cropping operations; The preprocessed multi-view point cloud data is integrated to form a complete dataset, and then divided into multiple point cloud datasets according to the station. Each dataset corresponds to the scanning results of one station.

4. The ship 3D scanning point cloud data processing system according to claim 3, characterized in that: The point cloud data registration unit performs the registration process for multiple point cloud datasets, including: A preliminary check is performed on the multiple point cloud datasets to be registered. A reference point cloud dataset is selected as the reference point cloud for registration, and the remaining point cloud datasets are used as target point clouds. The target point cloud is aligned with the reference point cloud dataset. After determining the reference point cloud and the target point cloud, the target point cloud is placed near the reference point cloud according to the preliminary alignment result to obtain an approximate transformation matrix. Start the iterative nearest point algorithm, initialize relevant parameters, including the maximum number of iterations and the convergence threshold, and use the approximate transformation matrix obtained from the preliminary alignment results as the initial transformation matrix; Fine registration is performed using the iterative nearest point algorithm. In each iteration, the nearest point from each point in the target point cloud to the reference point cloud is calculated to establish the correspondence between point pairs. Based on the correspondence between point pairs, the rotation and translation transformation matrix from the target point cloud to the reference point cloud is calculated. The optimal transformation matrix is ​​obtained by minimizing the sum of squared distances between the target point cloud and the reference point cloud. After obtaining the new transformation matrix, the target point cloud is updated by applying the transformation matrix to make it closer to the reference point cloud. The above process is repeated until the convergence condition is met or the maximum number of iterations is reached. The final transformation matrix is ​​applied to the point cloud dataset to be registered, aligning it spatially with the reference point cloud dataset, and then the registered point cloud dataset and transformation matrix are saved.

5. A ship 3D scanning point cloud data processing system according to claim 1, characterized in that: The process of reconstructing the three-dimensional model of the ship in the hull surface reconstruction module includes: The multi-view point cloud data after registration is imported into the hull surface reconstruction module. Based on the structural features of the hull, the point cloud data is segmented into different components including the bow, stern, side and deck. The point cloud data is then organized to establish the topological relationship between the point clouds. Based on the characteristics of the hull surface and the reconstruction requirements, a surface reconstruction algorithm for surface fitting is applied to process the point cloud data to generate a preliminary hull surface model. The initial hull surface model is optimized and adjusted through smoothing and detail enhancement to generate a complete 3D hull surface model, which displays the complete hull surface. Then, 3D visualization technology is used to display the complete 3D hull surface model in 3D space.

6. The ship 3D scanning point cloud data processing system according to claim 1, characterized in that: The point cloud feature sequence acquisition process in the feature extraction unit includes: Based on the characteristics of ship point cloud data and the needs of deformation analysis, key feature points are selected from the preprocessed point cloud data. Through curvature calculation and normal vector analysis, local geometric features reflecting ship deformation are extracted. Among them, key feature points include the corner points, edge points and areas with large curvature changes of the hull. By calculating the geometric relationship between each point in the preprocessed point cloud data and its neighboring points, the curvature value of the point is estimated, and the normal vector of each point is calculated. The shape changes of the hull surface are analyzed, and each point is assigned a curvature value and a normal vector, thereby obtaining the local geometric features of each point in the point cloud data. Based on the calculated local geometric features, set curvature threshold and normal vector change threshold, traverse each point in the point cloud data, check whether its curvature value and normal vector change meet the set threshold conditions, and filter out the points that meet the conditions as key feature points. Based on the location of key feature points in the point cloud data, the key feature points are sorted, and the sorted key feature points are combined into a point cloud feature sequence.

7. A ship 3D scanning point cloud data processing system according to claim 6, characterized in that: The deformation detection unit includes the following process for identifying the deformation areas of the hull: Point cloud feature sequences are extracted from the reconstructed 3D ship model and the normal ship model respectively. The key feature points in the point cloud feature sequences are used to reflect the shape changes and geometric features of the ship surface. The point cloud feature sequences of the reconstructed 3D ship model and the normal ship model are aligned so that the two point cloud feature sequences completely overlap in space. The geometric properties of the position, curvature, and normal vector of the corresponding key feature points in the reconstructed 3D ship model and the normal hull model are compared. The feature difference value is calculated, including the Euclidean distance, curvature difference, and normal vector angle between feature points, to reflect the degree of change of the hull surface in local areas. A difference threshold is set, and key feature points that exceed the difference threshold are marked as potential deformation areas. By analyzing the spatial distribution of feature difference values, the range and location of the deformed region are determined. Clustering algorithms are used to cluster key feature points with high feature difference values ​​to identify independent deformed regions. The baseline values ​​of the curvature and normal vector angle of key feature points are determined by combining the normal hull model. The average local area of ​​the normal hull model corresponding to the key feature points is used as the baseline value. Then, the deformation region discrimination index is calculated by combining the structural features and geometric information of the hull, and the identified deformation regions are classified to distinguish between local deformation and overall deformation.

8. A method for processing 3D scanning point cloud data of a ship, implemented based on the 3D scanning point cloud data processing system of a ship as described in any one of claims 1-7, characterized in that, Includes the following steps: S1. Use a 3D laser scanner to scan the ship from different perspectives to obtain multi-view point cloud data and perform preprocessing operations. S2. The iterative nearest point algorithm is used to register the preprocessed multi-view point cloud data and unify them to the same coordinate system. S3. Based on the registered point cloud data, apply the surface fitting algorithm to reconstruct the three-dimensional model of the hull and generate the complete hull surface. S4. Extract key feature points from the reconstructed model and the normal model, compare the differences between the normal hull model and the reconstructed model, and identify and quantify the deformation area. S5. Summarize the processed point cloud data and deformation analysis results, generate a visualization report, and display the hull deformation in real time through 3D model display and interactive operation.

Images