A Ship Point Cloud Dorsal Completion Method and System

Abstract

The present invention discloses a method and system for dorsal complementation of ship point clouds. Among them, the method includes: determining the mid-plane of the original ship point cloud, where the mid-plane is determined according to the normal line of the mid-plane and the reference point of the mid-plane; first, determining the normal line of the mid-plane according to the plane fitted by the point cloud on the outer contour surface of the ship, and the reference point of the mid-plane is determined through two situations: one is when the point cloud scanning of the outer contour surface of the ship is complete, and the other is when the original ship point cloud contains symmetric feature information; after determining the mid-plane, removing the point cloud data on the dorsal side of the mid-plane, and finally obtaining the mirrored ship point cloud by performing a mirror symmetry operation on the front-side point cloud; since there are two methods for determining the mid-plane, two mirrored ship point clouds will be generated. Finally, similarity evaluation is performed on the two generated mirrored ship point clouds, and the optimal mirrored ship point cloud is selected as the target point cloud. The method of the present invention improves the complementation accuracy while ensuring the complementation efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of ships, and more particularly, to a method and system for completing the back side of ship point clouds. Background Art

[0002] With the continuous progress of technology, three-dimensional point cloud technology has become an important tool for capturing and representing the geometric shape of objects. In practical applications, the obtained ship point cloud data often has missing parts, especially in the back side area of the ship. Due to physical occlusion and equipment limitations, data loss has become an urgent problem to be solved.

[0003] To solve the problem of missing point cloud data, various completion methods have been proposed one after another. Currently, most point cloud completion solutions are based on deep neural networks, and these networks are usually trained on large datasets with limited categories, which limits the object categories that these solutions can handle. These completion methods perform well on academic datasets with predefined object categories and very specific defect types; however, their performance significantly degrades in real-world environments and further deteriorates on previously unseen object categories.

[0004] In addition to deep learning-based methods, there are also some attempts to adopt geometry-based methods. These methods utilize geometric features for point cloud completion without the need for any external data. Geometry-based methods play an important role in point cloud completion. Instead of relying on external data, they achieve data completion by analyzing and utilizing the intrinsic geometric features of the point cloud. The surface interpolation method is a common strategy in geometric methods. This method analyzes the spatial distribution of the known point cloud data and uses interpolation techniques to speculate in the data missing area. Interpolation methods usually require the assumption that the distribution of the known data has a certain continuity and smoothness, so they are applicable to scenarios with relatively regular geometric structures. These methods include, but are not limited to, linear interpolation, spline interpolation, and Kriging method, etc. They construct a mathematical model to fit the point cloud surface and then generate new data points in the missing area. The advantage of this method is that the calculation is simple and no additional training data is required, but its effect may be limited in highly complex or irregular geometric shapes.

[0005] Aiming at the problem that the existing methods for completing the back side data of ship point clouds cannot improve the completion accuracy while ensuring the completion efficiency, no effective solution has been proposed yet. Summary of the Invention

[0006] An embodiment of the present invention provides a method for completing the back side of ship point clouds to solve the problem that the existing methods for completing the back side data of ship point clouds cannot improve the completion accuracy while ensuring the completion efficiency.

[0007] To achieve the above object, on the one hand, the present invention provides a method for completing the back side of ship point cloud, which includes: S1. Perform radius filtering preprocessing on the original ship point cloud to obtain a noise-reduced ship point cloud; S2. Screen the noise-reduced ship point cloud according to the Z value, and cluster the screened noise-reduced ship point cloud using the DBSCAN clustering algorithm to extract the class with the largest number of point clouds as the outer contour surface point cloud of the ship; S3. Fit the extracted outer contour surface point cloud of the ship using the RANSAC algorithm to obtain a first fitting plane and the inliers of the first fitting plane; Fit the inliers of the first fitting plane by the least squares method to obtain the normal of the central plane of the original ship point cloud; S41. When the outer contour surface point cloud of the ship is scanned completely, perform a convex hull processing on the extracted outer contour surface point cloud of the ship to obtain a convex hull; Find the two end points of the longest line segment in the convex hull, and find the end point farthest from the outer contour surface point cloud of the ship from the two end points, and use it as the reference point of the central plane; S42. When the original ship point cloud contains symmetric feature information, find the midpoint of the connection line of any two points on the symmetric plane by using the symmetric feature information, and use it as the reference point of the central plane; S5. For each case, generate a central plane according to the normal of the central plane and the reference point of the central plane; Remove the point cloud data on the back side of the central plane; S6. For each case, perform a symmetric operation on the point cloud data in front of the central plane according to the central plane to obtain a mirror ship point cloud; S7. Compare the similarity of the two mirror ship point clouds with the original ship point cloud to select the mirror ship point cloud with the highest similarity as the target point cloud.

[0008] Optionally, S2 includes: S21. Sort all the points in the noise-reduced ship point cloud from small to large according to the Z value, and extract the points of the preset percentage at the front to obtain the screened noise-reduced ship point cloud; S22. Cluster the screened noise-reduced ship point cloud using the DBSCAN clustering algorithm to obtain multiple classes; Use the class with the largest number of point clouds among all classes as the outer contour surface point cloud of the ship.

[0009] Optionally, S3 includes: S31. Determine the minimum number of points, the maximum number of iterations, and the distance threshold required for each fitting; S32. Randomly select the minimum number of points required for the current fitting in the extracted outer contour surface point cloud of the ship, and record it as the fitting points; Calculate a candidate plane according to all the fitting points; S33. Calculate the distance from each point in the extracted outer contour surface point cloud of the ship to the candidate plane. If the distance from the current point to the candidate plane is less than the distance threshold, record the current point as an inlier and count the number of inliers; S34. Repeat S32~S33 until the maximum number of iterations is reached; Select the candidate plane with the largest number of inliers as the first fitting plane; S35. Fit the inliers of the first fitting plane by the least squares method to obtain the normal of the central plane of the original ship point cloud.

[0010] Optionally, S42 includes: S421. Removing the point cloud on the outer contour surface of the ship and the points near the point cloud on the outer contour surface of the ship from the original ship point cloud to obtain a target symmetric point cloud; S422. Using the RANSAC algorithm to respectively fit two second fitting planes for the target symmetric point cloud; calculating the center points of the two second fitting planes; S423. Connecting the center points of the two second fitting planes, finding the midpoint of the connection line, and taking the midpoint of the connection line as the reference point of the central plane.

[0011] Optionally, removing the point cloud data on the back side of the central plane includes: calculating the point cloud data on both sides of the central plane, determining the side with fewer points as the back side of the central plane; removing the point cloud data on the back side of the central plane.

[0012] On the other hand, the present invention provides a ship point cloud back side completion system, which includes: a filtering unit for performing radius filtering preprocessing on the original ship point cloud to obtain a noise-reduced ship point cloud; a clustering unit for screening the noise-reduced ship point cloud according to the Z value, and clustering the screened noise-reduced ship point cloud using the DBSCAN clustering algorithm to extract the class with the largest number of points as the point cloud on the outer contour surface of the ship; a fitting unit for fitting the extracted point cloud on the outer contour surface of the ship using the RANSAC algorithm to obtain a first fitting plane and the inliers of the first fitting plane; fitting the inliers of the first fitting plane by the least squares method to obtain the normal of the central plane of the original ship point cloud; a first reference point obtaining unit for performing a convex hull process on the extracted point cloud on the outer contour surface of the ship when the scanning of the point cloud on the outer contour surface of the ship is complete to obtain a convex hull; finding the two end points of the longest line segment in the convex hull, and finding the end point farthest from the point cloud on the outer contour surface of the ship from the two end points, and taking it as the reference point of the central plane; a second reference point obtaining unit for, when the original ship point cloud contains symmetric feature information, finding the midpoint of the connection line between any two points on the symmetric plane by using the symmetric feature information, and taking it as the reference point of the central plane; a removing unit for generating a central plane according to the normal of the central plane and the reference point of the central plane for each case; removing the point cloud data on the back side of the central plane; a symmetric unit for, for each case, performing a symmetric operation on the point cloud data on the front side of the central plane according to the central plane to obtain a mirrored ship point cloud; a similarity comparison unit for comparing the similarity of the two mirrored ship point clouds with the original ship point cloud to select the mirrored ship point cloud with the highest similarity as the target point cloud.

[0013] Optionally, the clustering unit includes: a screening subunit, configured to sort all points in the noise-reduced ship point cloud from smallest to largest according to the Z value, and extract the points in the top preset percentage to obtain the screened noise-reduced ship point cloud; a clustering subunit, configured to cluster the screened noise-reduced ship point cloud by using the DBSCAN clustering algorithm to obtain multiple classes; and use the class with the largest number of points in all classes as the point cloud of the outer contour surface of the ship.

[0014] Optionally, the fitting unit includes: a determination subunit, configured to determine the minimum number of points, the maximum number of iterations, and the distance threshold required for each fitting; a first fitting subunit, configured to randomly select the minimum number of points required for the current fitting from the point cloud of the outer contour surface of the extracted ship, and denote them as fitting points; calculate a candidate plane based on all the fitting points; a marking subunit, configured to calculate the distance from each point in the point cloud of the outer contour surface of the extracted ship to the candidate plane, and if the distance from the current point to the candidate plane is less than the distance threshold, denote the current point as an inlier and count the number of inliers; a repetition subunit, configured to repeat the first fitting subunit and the marking subunit until the maximum number of iterations is reached; select the candidate plane with the largest number of inliers as the first fitting plane; a second fitting subunit, configured to fit the inliers of the first fitting plane by using the least squares method to obtain the normal line of the central axis plane of the original ship point cloud.

[0015] Optionally, the second reference point obtaining unit includes: a removal subunit, configured to remove the point cloud of the outer contour surface of the ship and the points near the point cloud of the outer contour surface of the ship from the original ship point cloud to obtain a target symmetric point cloud; a third fitting subunit, configured to respectively fit two second fitting planes from the target symmetric point cloud by using the RANSAC algorithm; calculate the center points of the two second fitting planes; a connection subunit, configured to connect the center points of the two second fitting planes, find the midpoint of the connection line, and use the midpoint of the connection line as the reference point of the central axis plane.

[0016] Optionally, removing the point cloud data on the back side of the central axis plane includes: calculating the point cloud data on both sides of the central axis plane, determining the side with the smaller quantity as the back side of the central axis plane; and removing the point cloud data on the back side of the central axis plane.

[0017] Advantages of the present invention:

[0018] The present invention provides a method and system for backside completion of ship point clouds. Among them, through the mirror driving technology and the full utilization of the inherent symmetry of the ship, the method can complete the backside data of the ship point cloud under limited sample conditions, reducing the impact of data loss on analysis and application. Through two strategies for determining the medial plane, it processes the cases of complete scanning and cases with symmetric feature information respectively, ensuring accurate completion effects in different situations and improving the overall completion accuracy. This method does not rely on a large amount of training data, but realizes the completion of point clouds through geometric features and symmetry, which is suitable for application under the condition of limited data acquisition, reducing the cost and complexity of collecting and maintaining large-scale data sets. Through the methods of symmetry driving and geometric analysis, compared with methods such as deep learning, this method is more efficient in terms of computational resource requirements and is suitable for use in resource-constrained environments. Brief Description of the Drawings

[0019] Figure 1 is a flowchart of a method for backside completion of ship point clouds provided by an embodiment of the present invention;

[0020] Figure 2 is a schematic structural diagram of a system for backside completion of ship point clouds provided by an embodiment of the present invention;

[0021] Figure 3 is a schematic structural diagram of denoised ship point clouds provided by an embodiment of the present invention;

[0022] Figure 4 is a schematic structural diagram of the screened denoised ship point clouds provided by an embodiment of the present invention;

[0023] Figure 5 is a schematic structural diagram of the outer contour surface point clouds of a ship provided by an embodiment of the present invention;

[0024] Figure 6 is a schematic structural diagram of the first fitting plane provided by an embodiment of the present invention;

[0025] Figure 7 is a schematic structural diagram of the convex hull provided by an embodiment of the present invention;

[0026] Figure 8 is a schematic structural diagram of the original ship point cloud containing symmetric feature information provided by an embodiment of the present invention;

[0027] Figure 9 is a schematic structural diagram of the target symmetric point cloud provided by an embodiment of the present invention. Detailed Embodiments

[0028] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0029] Figure 1 is a flowchart of a method for dorsal completion of ship point cloud provided by an embodiment of the present invention. As Figure 1 shown, the method includes:

[0030] S1. Perform radius filtering preprocessing on the original ship point cloud to obtain a denoised ship point cloud;

[0031] Figure 3 is a schematic structural diagram of the denoised ship point cloud provided by an embodiment of the present invention. As Figure 3 shown, radius filtering is a preprocessing technique for removing noise and outliers, and its main principle is based on neighborhood analysis of spatial distances. The following are the basic principle and operation steps of radius filtering:

[0032] First, for each point in the original ship point cloud, it is processed as follows:

[0033] Taking the current point in the original ship point cloud as the center, a radius r is defined. This radius r determines a spherical region, called a neighborhood. Then, search for all points in the original ship point cloud that are located within this spherical neighborhood. This process usually uses a spatial data structure (such as a KD tree) to accelerate the search process in order to effectively find the points within the neighborhood. Then, calculate the number of points contained in this spherical neighborhood. If the number is lower than a preset threshold, the current point in the original ship point cloud is considered an isolated point or a noise point. Finally, mark the current points with a number of points within the neighborhood lower than the threshold as noise points and remove them from the original ship point cloud dataset.

[0034] By radius filtering, outliers and noise points in the original ship point cloud can be effectively removed to obtain a denoised ship point cloud, improving the quality of the ship point cloud data.

[0035] S2. Screen the denoised ship point cloud according to the Z value, and cluster the screened denoised ship point cloud using the DBSCAN clustering algorithm to extract the class with the largest number of points as the outer contour surface point cloud of the ship;

[0036] In an optional embodiment, Figure 4 is a schematic structural diagram of the screened denoised ship point cloud provided by an embodiment of the present invention; Figure 5 is a schematic structural diagram of the outer contour surface point cloud of the ship provided by an embodiment of the present invention. AsFigure 4 and Figure 5 As shown in and

[0037] , S2 includes:

[0037] S21. Sort all the points in the noise-reduced ship point cloud from smallest to largest according to the Z value, and extract the points in the top preset percentage to obtain the filtered noise-reduced ship point cloud;

[0038] Sorting all the points in the noise-reduced ship point cloud from smallest to largest according to the Z value is to reorganize the data according to the position in the vertical direction. Due to the limitations of lidar acquisition, there is no point cloud data on the back side of the ship. By sorting all the points in the noise-reduced ship point cloud from smallest to largest according to the Z value and extracting the top 60% of the data, the point cloud of the front outer contour surface of the ship can be effectively separated, that is, the filtered noise-reduced ship point cloud is obtained.

[0039] S22. Cluster the filtered noise-reduced ship point cloud using the DBSCAN clustering algorithm to obtain multiple classes; take the class with the largest number of points in all classes as the point cloud of the outer contour surface of the ship.

[0040] In the filtered noise-reduced ship point cloud (the point cloud of the front outer contour surface of the ship), there may be mixed with other irrelevant data points. In order to obtain a pure outer contour surface point cloud, the DBSCAN clustering algorithm is applied to cluster the point cloud (i.e., the point cloud of the front outer contour surface of the ship) to obtain multiple classes (i.e., clusters), and the class with the largest number of points in all classes is returned. This algorithm uses the density characteristics of the point group for clustering analysis. The required class is the pure outer contour surface point cloud (i.e., the point cloud of the outer contour surface of the ship).

[0041] DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is a density-based spatial clustering algorithm, which is widely used in data analysis and is especially suitable for processing point group data with a lot of noise and irregular shapes. Its basic principle and implementation steps are as follows:

[0042] (1) Core points, border points and noise points:

[0043] Core point: If a point contains at least a specified number of other points (called the minimum number of points MinPts) in its neighborhood, then this point is a core point.

[0044] Border point: If a point does not meet the conditions of a core point but falls within the neighborhood of a certain core point, then this point is a border point.

[0045] Noise point: A point that is neither a core point nor a border point is regarded as a noise point.

[0046] (2) Neighborhood definition:

[0047] The neighborhood of each point is defined by a specified radius (ε), called the ε-neighborhood. The number of points contained within the neighborhood is used to determine whether the point is a core point.

[0048] (3) Cluster construction:

[0049] First, start from any unvisited point and check whether the point is a core point. Then, if it is a core point, starting from this core point, gradually add the density-reachable points to the same cluster. Density-reachability means all points that can be connected through a series of core points. Then, if the point is not a core point, mark it as a noise point or a boundary point.

[0050] (4) Expand the cluster:

[0051] For each core point, try to expand its cluster. For each point in the neighborhood, if the point is an unvisited core point, continue to expand. If the point is a boundary point or has been visited, only add it to the current cluster.

[0052] (5) Termination condition:

[0053] When all points have been visited, the algorithm terminates, obtaining several clusters and possible noise points.

[0054] S3. Use the RANSAC algorithm to fit the outer contour surface point cloud of the extracted ship to obtain the first fitting plane and the inliers of the first fitting plane; fit the inliers of the first fitting plane by the least squares method to obtain the normal of the medial surface of the original ship point cloud;

[0055] Figure 6 is the structural schematic diagram of the first fitting plane provided by the embodiment of the present invention; as Figure 6 shown, RANSAC (Random Sample Consensus) is an iterative random algorithm that can identify inliers from the input point cloud data and fit a plane in the point cloud data.

[0056] In an optional implementation manner, the S3 includes:

[0057] S31. Determine the minimum number of points required for each fitting (for plane fitting, at least 3 non-collinear points), the maximum number of iterations, and the distance threshold;

[0058] S32. Randomly select the minimum number of points required for the current fitting from the outer contour surface point cloud of the extracted ship and denote them as fitting points; calculate a candidate plane based on all the fitting points;

[0059] For example: Randomly select 3 points from the outer contour surface point cloud of the ship and fit a candidate plane based on these 3 points;

[0060] The plane equation is: Ax + By + Cz + D = 0

[0061] The normal vector of the plane is (A, B, C), and D is the intercept of the plane equation. Substituting three points into the above plane equation, the normal vector (A, B, C) of the plane can be obtained, and then the candidate plane can be calculated.

[0062] S33. Calculate the distance from each point in the extracted outer contour surface point cloud of the ship to the candidate plane. If the distance from the current point to the candidate plane is less than the distance threshold, mark the current point as an inlier and count the number of inliers.

[0063] S34. Repeat S32 - S33 until the maximum number of iterations is reached; select the candidate plane with the largest number of inliers as the first fitting plane.

[0064] Each time, different three points are used for fitting, and the number of inliers for each fitting is recorded. Finally, select the candidate plane with the most inliers as the first fitting plane.

[0065] S35. Fit the inliers of the first fitting plane by the least squares method to obtain the normal of the medial surface of the original ship point cloud.

[0066] The principle of the least squares method for plane fitting is to find the optimal plane model by minimizing the sum of squared errors. Given a set of three - dimensional point cloud data, the goal is to fit a plane equation , such that the sum of the distances between the plane and the data points is minimized. For each point , its distance to the plane can be expressed as an error form . The goal of the least squares method is to minimize the sum of the squares of these errors, that is . By constructing a system of linear equations and using linear algebra methods such as matrix operations (usually the normal equation ), the plane parameters a, b, c, d can be solved. These parameters define the optimal plane, making the sum of the squared errors between the plane and the data points the smallest.

[0067] In the above steps, after the normal of the medial surface has been determined, only one reference point of the medial surface needs to be determined to completely define the medial surface.

[0068] S41. When the scanning of the outer contour surface point cloud of the ship is complete, perform a convex hull operation on the extracted outer contour surface point cloud of the ship to obtain a convex hull; find the two endpoints of the longest line segment in the convex hull, and find the endpoint that is farthest from the outer contour surface point cloud of the ship from the two endpoints, and use it as the reference point of the medial surface.

[0069] In an alternative embodiment Figure 7It is a schematic structural diagram of the convex hull provided by an embodiment of the present invention; as Figure 7 shown, this step is to find the bow point of the point cloud on the outer contour surface of the ship, and use the bow point as the reference point of the central plane.

[0070] The convex hull processing of the point cloud on the outer contour surface of the ship is a computational geometry algorithm, and its purpose is to find the smallest convex polyhedron that encloses a set of points. The convex hull refers to the smallest convex hull that can contain all the points in the point cloud dataset. For a given point cloud dataset, the convex hull can be regarded as a "shell" that tightly wraps the data points, similar to the shape formed after wrapping all the points with an elastic membrane. The application of the convex hull is to simplify complex data structures.

[0071] When performing convex hull processing, first, a point cloud dataset defined by three-dimensional coordinates (the point cloud on the outer contour surface of the ship) needs to be input. The algorithm selects an initial point from the point set and starts constructing a convex polyhedron that contains all the points. Through iteration, the algorithm selects appropriate points to expand and adjust the facets of the convex polyhedron, so that the shape of each new point keeps the facet convex. In each step, it is ensured that the surface of the polyhedron formed after adding the new point is still convex, that is, any line connecting the inner and outer points will not pass through the polyhedron.

[0072] Find the two endpoints of the longest line segment in the convex hull, and find the endpoint that is farthest from the point cloud on the outer contour surface of the ship from the two endpoints. This endpoint is the bow point of the point cloud on the outer contour surface of the ship, and it is used as the reference point of the central plane.

[0073] S42. When the original ship point cloud contains symmetric feature information, find the midpoint of the line connecting any two points on the symmetric plane by using the symmetric feature information, and use it as the reference point of the central plane;

[0074] In an optional embodiment, Figure 8 It is a schematic structural diagram of the original ship point cloud containing symmetric feature information provided by an embodiment of the present invention; as Figure 8 shown, the S5 includes:

[0075] S421. Remove the point cloud on the outer contour surface of the ship and the points near the point cloud on the outer contour surface of the ship from the original ship point cloud to obtain the target symmetric point cloud;

[0076] Figure 9 It is a schematic structural diagram of the target symmetric point cloud provided by an embodiment of the present invention, as Figure 9 shown. Through the above process, the point cloud other than the point cloud on the outer contour surface of the ship can be retained, that is, the target symmetric point cloud.

[0077] S422. Use the RANSAC algorithm to fit two second fitting planes to the target symmetric point cloud respectively; calculate the center points of the two second fitting planes;

[0078] That is, the target symmetric point cloud consists of two point clouds, and the two point clouds are symmetric in shape. The RANSAC algorithm is used to fit a second fitting plane for each point cloud respectively, obtaining two second fitting planes; the center point of each second fitting plane is calculated.

[0079] S423. Connect the center points of the two second fitting planes, find the midpoint of the connection line, and use the midpoint of the connection line as the reference point of the mid-axis plane.

[0080] S5. For each case, generate a mid-axis plane according to the normal line of the mid-axis plane and the reference point of the mid-axis plane; remove the point cloud data on the back side of the mid-axis plane.

[0081] In an alternative embodiment, two mid-axis planes based on two methods are generated in the above manner. For each case: remove the point cloud data located on the back side of the mid-axis plane.

[0082] The removal of the point cloud data on the back side of the mid-axis plane includes:

[0083] Calculate the point cloud data on both sides of the mid-axis plane, determine the side with the smaller quantity as the back side of the mid-axis plane; remove the point cloud data on the back side of the mid-axis plane.

[0084] S6. For each case, perform a symmetry operation on the point cloud data in front of the mid-axis plane according to the mid-axis plane to obtain a mirrored ship point cloud.

[0085] S7. Compare the similarity between the two mirrored ship point clouds and the original ship point cloud to select the mirrored ship point cloud with the highest similarity as the target point cloud.

[0086] Specifically, taking one mirrored ship point cloud as an example:

[0087] Calculate the Chamfer Distance between the current mirrored ship point cloud and the original ship point cloud. The calculation formula is as follows:

[0088] ;

[0089] Where is the current mirrored ship point cloud, is the original ship point cloud, is the calculation point set the average distance from each point in to the nearest point in the point set , is the calculation point set the average distance from each point in to the nearest point in the point set Add these two average distances to obtain the Chamfer Distance between the two point sets. The smaller this distance is, the more similar the two point sets are.

[0090] Through the above method, the mirror ship point cloud with the highest similarity is selected as the target point cloud.

[0091] Figure 2 It is a schematic structural diagram of a ship point cloud dorsal complementing system provided by an embodiment of the present invention; as Figure 2 shown, the system includes:

[0092] A filtering unit 201, configured to perform radius filtering preprocessing on the original ship point cloud to obtain a noise-reduced ship point cloud;

[0093] A clustering unit 202, configured to screen the noise-reduced ship point cloud according to the Z value, and cluster the screened noise-reduced ship point cloud using the DBSCAN clustering algorithm to extract the class with the largest number of point clouds as the outer contour surface point cloud of the ship;

[0094] A fitting unit 203, configured to fit the extracted outer contour surface point cloud of the ship using the RANSAC algorithm to obtain a first fitting plane and the inliers of the first fitting plane; fit the inliers of the first fitting plane by the least squares method to obtain the normal of the central axis plane of the original ship point cloud;

[0095] A first reference point obtaining unit 2041, configured to, when the scanning of the outer contour surface point cloud of the ship is complete, perform a convex hull process on the extracted outer contour surface point cloud of the ship to obtain a convex hull; find the two endpoints of the longest line segment in the convex hull, and find the endpoint farthest from the outer contour surface point cloud of the ship from the two endpoints, and use it as the reference point of the central axis plane;

[0096] A second reference point obtaining unit 2042, configured to, when the original ship point cloud contains symmetric feature information, find the midpoint of the connection line of any two points on the symmetric plane by using the symmetric feature information, and use it as the reference point of the central axis plane;

[0097] A removing unit 205, configured to, for each case, generate a central axis plane according to the normal of the central axis plane and the reference point of the central axis plane; remove the point cloud data on the dorsal side of the central axis plane;

[0098] A symmetric unit 206, configured to, for each case, perform a symmetric operation on the point cloud data on the front side of the central axis plane according to the central axis plane to obtain a mirror ship point cloud;

[0099] A similarity comparison unit 207, configured to compare the similarity of the two mirror ship point clouds with the original ship point cloud to select the mirror ship point cloud with the highest similarity as the target point cloud.

[0100] In an optional embodiment, the clustering unit 202 includes:

[0101] A screening subunit 2021, configured to sort all points in the noise-reduced ship point cloud from smallest to largest according to the Z value, and extract the points in the top preset percentage to obtain the screened noise-reduced ship point cloud;

[0102] A clustering subunit 2022, configured to cluster the screened noise-reduced ship point cloud using the DBSCAN clustering algorithm to obtain multiple clusters; and use the cluster with the largest number of points in all clusters as the point cloud of the outer contour surface of the ship.

[0103] In an optional embodiment, the fitting unit 203 includes:

[0104] A determination subunit 2031, configured to determine the minimum number of points, the maximum number of iterations, and the distance threshold required for each fitting;

[0105] A first fitting subunit 2032, configured to randomly select the minimum number of points required for the current fitting from the extracted point cloud of the outer contour surface of the ship, and denote them as fitting points; calculate a candidate plane based on all the fitting points;

[0106] A marking subunit 2033, configured to calculate the distance from each point in the extracted point cloud of the outer contour surface of the ship to the candidate plane. If the distance from the current point to the candidate plane is less than the distance threshold, mark the current point as an inlier and count the number of inliers;

[0107] A repetition subunit 2034, configured to repeat the first fitting subunit 2032 and the marking subunit 2033 until the maximum number of iterations is reached; select the candidate plane with the largest number of inliers as the first fitting plane;

[0108] A second fitting subunit 2035, configured to fit the inliers of the first fitting plane by the least squares method to obtain the normal of the central axis plane of the original ship point cloud.

[0109] In an optional embodiment, the reference point second acquisition unit 2042 includes:

[0110] A removal subunit 20421, configured to remove the point cloud of the outer contour surface of the ship and the points near the point cloud of the outer contour surface of the ship from the original ship point cloud to obtain a target symmetric point cloud;

[0111] A third fitting subunit 20422, configured to respectively fit two second fitting planes for the target symmetric point cloud using the RANSAC algorithm; calculate the center points of the two second fitting planes;

[0112] A connection subunit 20423, configured to connect the center points of the two second fitting planes, find the midpoint of the connection line, and use the midpoint of the connection line as the reference point of the central axis plane.

[0113] In an alternative embodiment, the removal of the point cloud data on the dorsal side of the central plane includes:

[0114] Calculating the point cloud data on both sides of the central plane, determining the side with the smaller quantity as the dorsal side of the central plane; removing the point cloud data on the dorsal side of the central plane.

[0115] Advantages of the present invention:

[0116] The present invention provides a method and system for completing the dorsal side of a ship's point cloud. Among them, through the mirror driving technology and the full utilization of the inherent symmetry of the ship, the method can complete the dorsal side data of the ship's point cloud under limited sample conditions, reducing the impact of data loss on analysis and application. Through two strategies for determining the central plane, it respectively processes the cases of complete scanning and cases with symmetric feature information, ensuring accurate completion effects in different situations and improving the overall completion accuracy. The method does not rely on a large amount of training data, but realizes the completion of the point cloud through geometric features and symmetry, and is suitable for application under the condition of limited data acquisition, reducing the cost and complexity of collecting and maintaining a large-scale data set. Through the methods of symmetry driving and geometric analysis, compared with methods such as deep learning, the method is more efficient in terms of computational resource requirements and is suitable for use in resource-constrained environments.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for completing the back side of a ship point cloud, characterized in that: include: S1, performing radius filtering preprocessing on the original ship point cloud to obtain a de-noised ship point cloud; S2, filtering the noise reduction ship point cloud according to the Z value, and clustering the filtered noise reduction ship point cloud using the DBSCAN clustering algorithm to extract the class with the largest number of point clouds as the outer contour surface point cloud of the ship; S3, fitting the extracted outer contour surface point cloud of the ship using the RANSAC algorithm to obtain a first fitting plane and the inner points of the first fitting plane; fitting the inner points of the first fitting plane using the least squares method to obtain the normal line of the medial axis surface of the original ship point cloud; S41. When the scan of the outer contour surface point cloud of the ship is complete, the extracted outer contour surface point cloud of the ship is subjected to convex hull processing to obtain a convex hull; two endpoints of the longest line segment in the convex hull are found, and the endpoint farthest from the outer contour surface point cloud of the ship is found from the two endpoints, and the endpoint is used as the reference point of the medial axis surface; S42, when the original ship point cloud contains symmetry feature information, the midpoint of the line connecting any two points on the symmetry plane is calculated by using the symmetry feature information, and the midpoint is used as the reference point of the median axis plane; S5. For each case, generate a medial axis surface according to the normal line of the medial axis surface and the reference point of the medial axis surface; remove the point cloud data on the back side of the medial axis surface; S6. For each case, the point cloud data on the front side of the mid-axis plane is symmetrically operated according to the mid-axis plane to obtain a mirror image ship point cloud; S7. Compare the two mirror image ship point clouds with the original ship point cloud for similarity, so as to select the mirror image ship point cloud with the highest similarity as the target point cloud.

2. The method according to claim 1, characterized in that The S2 includes: S21, sorting all points in the noise reduction ship point cloud from small to large according to the Z value, and extracting a preset percentage of points to obtain a filtered noise reduction ship point cloud; S22, clustering the filtered noise-reduced ship point cloud using the DBSCAN clustering algorithm to obtain multiple classes; and taking the class with the largest number of point clouds among all classes as the outer contour surface point cloud of the ship.

3. The method according to claim 1, characterized in that The S3 includes: S31, determining the minimum number of points, maximum number of iterations and distance threshold required for each fitting; S32, randomly selecting the minimum number of points required for the current fitting in the extracted point cloud of the outer contour surface of the ship, and recording them as fitting points; and calculating a candidate plane according to all the fitting points; S33, calculating the distance between each point in the extracted point cloud of the outer contour surface of the ship and the candidate plane, if the distance between the current point and the candidate plane is less than the distance threshold, the current point is recorded as an inner point, and the number of inner points is counted; S34, repeat S32 to S33 until the maximum number of iterations is reached; select the candidate plane with the largest number of internal points as the first fitting plane; S35. Fit the inner points of the first fitting plane by the least square method to obtain the normal line of the medial axis surface of the original ship point cloud.

4. The method according to claim 1, characterized in that: The S42 includes: S421, removing the outer contour surface point cloud of the ship and points near the outer contour surface point cloud of the ship from the original ship point cloud to obtain a target symmetrical point cloud; S422, using the RANSAC algorithm to fit the target symmetrical point cloud to two second fitting planes respectively; and calculating the center points of the two second fitting planes; S423, connecting the center points of the two second fitting planes, finding the midpoint of the connecting line, and using the midpoint of the connecting line as the reference point of the medial axis surface.

5. The method according to claim 1, characterized in that The step of removing the point cloud data on the back side of the medial axis surface comprises: The point cloud data on both sides of the mid-axis surface are calculated, and the side with a smaller number of point cloud data is determined as the back side of the mid-axis surface; and the point cloud data on the back side of the mid-axis surface are removed.

6. A ship point cloud back side completion system, characterized in that: include: A filtering unit is used to perform radius filtering preprocessing on the original ship point cloud to obtain a de-noised ship point cloud; A clustering unit is used to filter the noise reduction ship point cloud according to the Z value, and cluster the filtered noise reduction ship point cloud using the DBSCAN clustering algorithm to extract the class with the largest number of point clouds as the outer contour surface point cloud of the ship; The fitting unit is used to fit the extracted outer contour surface point cloud of the ship using the RANSAC algorithm to obtain a first fitting plane and the inner points of the first fitting plane; the inner points of the first fitting plane are fitted by the least square method to obtain the normal line of the medial axis surface of the original ship point cloud; The first reference point acquisition unit is used to perform convex hull processing on the extracted point cloud of the outer contour surface of the ship to obtain a convex hull when the scan of the point cloud of the outer contour surface of the ship is complete; find the two endpoints of the longest line segment in the convex hull, and find the endpoint farthest from the outer contour surface point cloud of the ship from the two endpoints, and use it as the reference point of the medial axis surface; The second reference point acquisition unit is used to find the midpoint of the line connecting any two points on the symmetry plane by using the symmetry feature information when the original ship point cloud contains symmetry feature information, and use it as the reference point of the median axis plane; A removal unit is used to generate a medial axis surface according to the normal line of the medial axis surface and the reference point of the medial axis surface for each case; and remove the point cloud data on the back side of the medial axis surface; A symmetry unit, for each case, performing a symmetric operation on the point cloud data on the front side of the mid-axis plane according to the mid-axis plane to obtain a mirror image ship point cloud; The similarity comparison unit is used to compare the two mirror image ship point clouds with the original ship point cloud to select the mirror image ship point cloud with the highest similarity as the target point cloud.

7. The system according to claim 6, characterized in that The clustering unit comprises: A screening subunit is used to sort all points in the noise reduction ship point cloud from small to large according to the Z value, and extract a preset percentage of points to obtain a screened noise reduction ship point cloud; The clustering subunit is used to cluster the filtered noise-reduced ship point cloud using the DBSCAN clustering algorithm to obtain multiple classes; the class with the largest number of point clouds among all classes is used as the outer contour surface point cloud of the ship.

8. The system according to claim 7, characterized in that The fitting unit comprises: Determine the subunit, which is used to determine the minimum number of points, the maximum number of iterations and the distance threshold required for each fitting; The first fitting subunit is used to randomly select the minimum number of points required for the current fitting in the extracted point cloud of the outer contour surface of the ship and record them as fitting points; and calculate a candidate plane according to all the fitting points; A marking subunit is used to calculate the distance between each point in the extracted point cloud of the outer contour surface of the ship and the candidate plane. If the distance between the current point and the candidate plane is less than the distance threshold, the current point is recorded as an inner point, and the number of inner points is counted; A repeating subunit is used to repeat the first fitting subunit and the marking subunit until the maximum number of iterations is reached; and the candidate plane with the largest number of inliers is selected as the first fitting plane; The second fitting subunit is used to fit the inner points of the first fitting plane by the least square method to obtain the normal line of the medial axis surface of the original ship point cloud.

9. The system according to claim 6, characterized in that The reference point second acquisition unit comprises: A removal subunit is used to remove the outer contour surface point cloud of the ship and points near the outer contour surface point cloud of the ship from the original ship point cloud to obtain a target symmetrical point cloud; The third fitting subunit is used to fit the target symmetrical point cloud into two second fitting planes respectively by using the RANSAC algorithm; and calculate the center points of the two second fitting planes; The connecting line subunit is used to connect the center points of the two second fitting planes and find the midpoint of the connecting line, and use the midpoint of the connecting line as the reference point of the medial axis surface.

10. The system according to claim 6, characterized in that The step of removing the point cloud data on the back side of the medial axis surface comprises: The point cloud data on both sides of the mid-axis surface are calculated, and the side with a smaller number of point cloud data is determined as the back side of the mid-axis surface; and the point cloud data on the back side of the mid-axis surface are removed.

Images