Point cloud hole filling method, point cloud reconstruction method, system, device and medium

Abstract

The application discloses a point cloud hole filling method, a point cloud reconstruction method, a system, equipment and a medium, and relates to the technical field of data processing. The point cloud hole filling method comprises the following steps: acquiring a plurality of original point cloud data points of an object; determining a hole center, boundary point cloud data points and intermediate point cloud data points according to the plurality of original point cloud data points; determining a first newly-added point cloud data point according to the three-dimensional information corresponding to each boundary point cloud data point, the three-dimensional information corresponding to the hole center and a preset curvature radius; and completing point cloud hole filling according to the three-dimensional information corresponding to each original point cloud data point, the three-dimensional information corresponding to each first newly-added point cloud data point, the three-dimensional information corresponding to the hole center, the curvature radius and a preset point cloud hole filling end condition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a point cloud hole filling method, a point cloud reconstruction method, a system, a device and a medium. BACKGROUND

[0002] Point cloud refers to a mass of point set of object surface characteristics. In actual application process, due to the limitations of data acquisition device, manual operation and object structure complexity and other factors, the surface information of the object cannot be completely obtained, resulting in the point cloud data appearing holes or gaps, thereby affecting the subsequent processing such as three-dimensional reconstruction, finite element analysis and the like.

[0003] At present, the existing point cloud hole filling method based on grid is mainly used to solve the problem of point cloud hole. The method constructs a triangular mesh model from the point cloud, and then repairs the holes on the triangular mesh model. The point cloud filling method based on grid only needs to fill the point cloud of the grid defect part, and will not change the grid characteristics of other parts of the input model, but the method cannot well complete the filling for the holes with large curvature. SUMMARY

[0004] The technical problem to be solved by the present application is that the effect of point cloud hole filling by using the existing point cloud hole filling method is not good. In order to solve the technical problem, the present application provides a point cloud hole filling method, a point cloud reconstruction method, a system, a device and a medium. The present application can effectively repair the point cloud hole and densify the point cloud data by directly repairing the hole on the point cloud, increase the amount of point cloud data, and realize reverse engineering and three-dimensional reconstruction.

[0005] The technical solution of the present application to solve the above technical problem is as follows:

[0006] A point cloud hole filling method based on boundary point cloud, comprising:

[0007] Step S1, obtaining a plurality of original point cloud data points for an object, each of the original point cloud data points corresponding to three-dimensional information and image information, the three-dimensional information representing the three-dimensional coordinates of the original point cloud data point in a three-dimensional space, and the image information representing the three-dimensional color information of the original point cloud data point in the three-dimensional space;

[0008] Step S2, determining a hole center according to a plurality of original point cloud data points, the hole being a three-dimensional shape formed by the point cloud data points corresponding to the missing area of the object surface, and the hole center corresponding to three-dimensional information in the three-dimensional space; dividing each original point cloud data point into a boundary point cloud data point or an intermediate point cloud data point according to the three-dimensional information and image information corresponding to each boundary point cloud data point and the three-dimensional information corresponding to the hole center.

[0009] In step S3, according to the three-dimensional information corresponding to each of the boundary point cloud data points, the three-dimensional information corresponding to the hole center, and a preset curvature radius, a first new point cloud data point is determined, and each of the first new point cloud data points has corresponding three-dimensional information;

[0010] In step S4, according to the three-dimensional information corresponding to each of the original point cloud data points, the three-dimensional information corresponding to each of the first new point cloud data points, the three-dimensional information corresponding to the hole center, the curvature radius, and a preset point cloud hole filling end condition, the point cloud hole filling is completed.

[0011] The present application has the advantages that: by determining the boundary point cloud data points in the original point cloud data points, and continuously adding point cloud data points in the direction of the hole center according to the three-dimensional information of the boundary point cloud data points, a dense three-dimensional point cloud data point is generated, which is convenient for three-dimensional reconstruction in the later stage; the present application directly repairs the hole of the original point cloud data points, the point cloud data points are dense, and the present application has the advantages of high robustness, and does not need to reconstruct the topological relationship, has high storage and execution efficiency, and overcomes the problems of model repeated sampling and local feature loss when the point cloud hole is repaired by using the existing method.

[0012] On the basis of the above technical solution, the present application can also be improved as follows.

[0013] Further, in step S2, according to the three-dimensional information and image information corresponding to each of the boundary point cloud data points, and the three-dimensional information corresponding to the hole center, each of the original point cloud data points is divided into a boundary point cloud data point or an intermediate point cloud data point, which includes:

[0014] The hole center is projected into a two-dimensional space to obtain a center projection point, and the center projection point has corresponding two-dimensional coordinates in the two-dimensional space;

[0015] For each of the original point cloud data points, the original point cloud data point is projected into the two-dimensional space to obtain a two-dimensional projection data point corresponding to the original point cloud data point, and the two-dimensional projection data point has corresponding two-dimensional coordinates and two-dimensional color information in the two-dimensional space;

[0016] For each of the two-dimensional projection data points, according to the two-dimensional coordinates and two-dimensional color information corresponding to the two-dimensional projection data point, and the two-dimensional coordinates corresponding to the center projection point, the original point cloud data point corresponding to the two-dimensional projection data point is divided into a boundary point cloud data point or an intermediate point cloud data point.

[0017] The beneficial effect of the above further scheme is that: by projecting each original point cloud data point into a two-dimensional space to obtain a two-dimensional projection data point, and quickly dividing the boundary point cloud data points and the intermediate point cloud data points according to the two-dimensional coordinates and the two-dimensional color information of the two-dimensional projection data point, the calculation amount is small, and the topology structure formed by the original point cloud data points is not disturbed.

[0018] Further, the step of dividing, for each two-dimensional projection data point, the original point cloud data point corresponding to the two-dimensional projection data point into a boundary point cloud data point or an intermediate point cloud data point according to the two-dimensional coordinates and the two-dimensional color information of the two-dimensional projection data point and the two-dimensional coordinates of the center projection point comprises:

[0019] For each two-dimensional projection data point, the original point cloud data point corresponding to the two-dimensional projection data point is divided into a boundary point cloud data point or an intermediate point cloud data point according to the two-dimensional coordinates of the center projection point by using an image binarization method and a fixed threshold method.

[0020] Or,

[0021] For each two-dimensional projection data point, the original point cloud data point corresponding to the two-dimensional projection data point is divided into a boundary point cloud data point or an intermediate point cloud data point according to the two-dimensional coordinates of the center projection point by using an adaptive threshold method and a histogram threshold method.

[0022] The beneficial effect of the above further scheme is that: based on the fixed threshold method, the original point cloud data points are divided into boundary point cloud data points and intermediate point cloud data points, which has small calculation amount and high efficiency; considering the influence of uneven illumination or environmental light on the collection of the original point cloud data points, based on the adaptive threshold method, the original point cloud data points are divided into boundary point cloud data points and intermediate point cloud data points, which is flexible and easy to implement.

[0023] Further, the step S3 comprises:

[0024] For each boundary point cloud data point, a growth region of the boundary point cloud data point is constructed with the boundary point cloud data point as the center and the curvature radius as the radius of the boundary point cloud data point; and for two adjacent boundary point cloud data points, there is at least one intersection point between the growth regions corresponding to the two boundary point cloud data points.

[0025] For two adjacent boundary point cloud data points, if there is only one intersection point between the growth regions corresponding to the two boundary point cloud data points, the intersection point is determined as a first newly added point cloud data point.

[0026] For two adjacent boundary point cloud data points, if the number of intersection points between the growth areas corresponding to the two boundary point cloud data points is greater than 1, a first new point cloud data point is determined according to the distance between the three-dimensional information corresponding to each intersection point and the three-dimensional information corresponding to the hole center.

[0027] The beneficial effect of the further scheme is that the new point cloud data point is determined by determining the intersection point, which ensures that the point cloud hole filling is contracted towards the hole center, and avoids invalid filling.

[0028] Further, the step S4 includes:

[0029] Step S4.1, for each first new point cloud data point, a growth area of the first new point cloud data point is constructed with the first new point cloud data point as the center and the curvature radius as the radius of the first new point cloud data point.

[0030] For each first new point cloud data point, if there is only one intersection point between the growth area of the first new point cloud data point and the growth area of the adjacent original point cloud data point, the intersection point is determined as a second new point cloud data point, and if the number of intersection points between the growth area of the first new point cloud data point and the growth area of the adjacent original point cloud data point is greater than 1, a second new point cloud data point is determined according to the distance between the three-dimensional information corresponding to each intersection point and the three-dimensional information corresponding to the hole center.

[0031] Step S4.2, for each second new point cloud data point, the second new point cloud data point is taken as a first new point cloud data point, and step S4.1 is repeated until the point cloud hole filling end condition is met, and the point cloud hole filling is completed.

[0032] The beneficial effect of the further scheme is that the new point cloud data point is determined in a cyclic manner towards the hole center, which realizes fine and dense filling of the hole and avoids uneven hole filling.

[0033] Further, the point cloud hole filling end condition includes:

[0034] According to the three-dimensional information corresponding to each original point cloud data point and the curvature radius, it is determined that the point cloud data point cannot be added along the hole center direction any more;

[0035] Or,

[0036] According to the image information corresponding to each original point cloud data point, it is determined that the point cloud data point cannot be added along the hole center direction any more.

[0037] The beneficial effect of adopting the further scheme is that: by determining the point cloud hole filling end condition, the complete point cloud hole filling is realized, and redundant calculation is avoided.

[0038] To solve the technical problem that the subsequent processing is affected due to the poor point cloud hole filling, the application further provides a point cloud reconstruction method based on boundary point cloud, comprising:

[0039] Obtaining a plurality of initial point cloud data points for a target to be reconstructed;

[0040] Taking the initial point cloud data points as original point cloud data points, performing the point cloud hole filling method based on boundary point cloud as described above to obtain target point cloud data points of the target to be reconstructed;

[0041] Triangulating the target point cloud data points to obtain a triangulation result;

[0042] According to the triangulation result, completing the point cloud reconstruction of the target to be reconstructed.

[0043] The beneficial effect of the application is that: by triangulating the target point cloud data points, the point cloud reconstruction result of the target to be reconstructed is obtained, which is convenient for subsequent related processing of the point cloud reconstruction result of the target to be reconstructed, such as three-dimensional measurement and reverse engineering application.

[0044] To solve the above technical problem, the embodiment further provides a point cloud hole filling system based on boundary point cloud, comprising:

[0045] A point cloud data point acquisition module is configured to acquire a plurality of original point cloud data points of an object, each of the original point cloud data points corresponding to three-dimensional information and image information, the three-dimensional information representing the three-dimensional coordinates of the original point cloud data point in a three-dimensional space, and the image information representing the three-dimensional color information of the original point cloud data point in the three-dimensional space;

[0046] A point cloud data point classification module is configured to determine a hole center according to the plurality of original point cloud data points, the hole being a three-dimensional shape formed by point cloud data points corresponding to a missing area of a surface of the object, the hole center corresponding to three-dimensional information in the three-dimensional space; and divide each of the original point cloud data points into a boundary point cloud data point or an intermediate point cloud data point according to the three-dimensional information and the image information corresponding to each of the boundary point cloud data points and the three-dimensional information corresponding to the hole center.

[0047] A first point cloud hole filling module is configured to determine a first newly added point cloud data point according to the three-dimensional information corresponding to each of the boundary point cloud data points, the three-dimensional information corresponding to the hole center, and a preset curvature radius, each of the first newly added point cloud data points corresponding to three-dimensional information.

[0048] The second point cloud hole filling module is used to complete point cloud hole filling based on the three-dimensional information corresponding to each of the original point cloud data points, the three-dimensional information corresponding to each of the first newly added point cloud data points, the three-dimensional information corresponding to the hole center, the radius of curvature, and the preset point cloud hole filling termination condition.

[0049] To address the aforementioned technical problems, this embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the point cloud hole filling method based on boundary point clouds as described above.

[0050] To address the aforementioned technical problems, this embodiment also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the point cloud hole filling method based on boundary point clouds as described above.

[0051] It should be noted that, in this invention, the collection of multiple original point cloud data points is a massive collection of points that constitute the surface characteristics of the target, i.e., a point cloud. Attached Figure Description

[0052] Figure 1 This is a flowchart illustrating the point cloud hole filling method based on boundary point clouds in this invention.

[0053] Figure 2 This is a schematic diagram illustrating the determination of the first newly added point cloud data point using boundary point cloud data points in this invention.

[0054] Figure 3 This is a schematic diagram of the point cloud contour lines formed by the newly added multi-layer point cloud data points in this invention.

[0055] Figure 4 This is a schematic diagram of the point cloud hole filling system based on boundary point clouds in this invention. Detailed Implementation

[0056] The principles and features of the present invention are described below. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0057] Example 1

[0058] like Figure 1 As shown, this embodiment provides a point cloud hole filling method based on boundary point clouds, including:

[0059] Step S1, obtaining a plurality of original point cloud data points corresponding to an object, each of the original point cloud data points corresponding to three-dimensional information and image information, the three-dimensional information representing three-dimensional coordinates of the original point cloud data point in a three-dimensional space, and the image information representing three-dimensional color information of the original point cloud data point in the three-dimensional space;

[0060] Step S2, determining a hole center according to a plurality of original point cloud data points, the hole being a three-dimensional shape formed by point cloud data points corresponding to a missing area of a surface of the object, the hole center corresponding to three-dimensional information in the three-dimensional space; dividing each of the original point cloud data points into a boundary point cloud data point or an intermediate point cloud data point according to the three-dimensional information and the image information corresponding to each of the boundary point cloud data points and the three-dimensional information corresponding to the hole center;

[0061] Step S3, determining a first new point cloud data point according to the three-dimensional information corresponding to each of the boundary point cloud data points, the three-dimensional information corresponding to the hole center, and a preset radius of curvature, each of the first new point cloud data points corresponding to three-dimensional information;

[0062] Step S4, completing point cloud hole filling according to the three-dimensional information corresponding to each of the original point cloud data points, the three-dimensional information corresponding to each of the first new point cloud data points, the three-dimensional information corresponding to the hole center, the radius of curvature, and a preset point cloud hole filling end condition.

[0063] In the step S2, the dividing of each of the original point cloud data points into a boundary point cloud data point or an intermediate point cloud data point according to the three-dimensional information and the image information corresponding to each of the boundary point cloud data points and the three-dimensional information corresponding to the hole center, comprises:

[0064] Projecting the hole center into a two-dimensional space to obtain a center projection point, the center projection point corresponding to two-dimensional coordinates in the two-dimensional space;

[0065] For each of the original point cloud data points, projecting the original point cloud data point into the two-dimensional space to obtain a two-dimensional projection data point corresponding to the original point cloud data point, the two-dimensional projection data point corresponding to two-dimensional coordinates and two-dimensional color information in the two-dimensional space;

[0066] For each of the two-dimensional projection data points, dividing the original point cloud data point corresponding to the two-dimensional projection data point into a boundary point cloud data point or an intermediate point cloud data point by using a threshold method according to the two-dimensional coordinates and the two-dimensional color information corresponding to the two-dimensional projection data point and the two-dimensional coordinates corresponding to the center projection point.

[0067] In the present application, the obtained original point cloud data points are huge and scattered, the present application constructs a point cloud space including all the original point cloud data points obtained in the step S1, the point cloud space is a three-dimensional space, all the original point cloud data points obtained in the step S1 are mapped into the point cloud space, and then the three-dimensional point cloud space is projected onto a certain specific plane (i.e. the two-dimensional space), the boundary data points are extracted by using the two-dimensional image processing technology, and finally the determined boundary point cloud data points are mapped onto the point cloud space, so as to complete the hole recognition, i.e. to determine the hole center; in the present application, the three-dimensional information of the point cloud data points (including the original point cloud data points, the first newly added point cloud data points obtained through the step S3 and the second newly added point cloud data points obtained through the step S4) includes the horizontal coordinate, the vertical coordinate and the vertical coordinate of the point cloud data points in the three-dimensional space.

[0068] In the present embodiment, for each two-dimensional projection data point, according to the two-dimensional coordinates and two-dimensional color information corresponding to the two-dimensional projection data point and the two-dimensional coordinates corresponding to the center projection point, the original point cloud data point corresponding to the two-dimensional projection data point is divided into a boundary point cloud data point or an intermediate point cloud data point, which includes:

[0069] For each two-dimensional projection data point, according to the two-dimensional coordinates corresponding to the center projection point, the original point cloud data point corresponding to the two-dimensional projection data point is divided into a boundary point cloud data point or an intermediate point cloud data point by using the image binarization method and the fixed threshold method.

[0070] Specifically, for each two-dimensional projection data point, the two-dimensional projection data point is subjected to image binarization processing to obtain a binarized data point, if the gray value of the binarized data point is greater than a preset comparison threshold, the original point cloud data point corresponding to the binarized data point is determined as a boundary point cloud data point, and if the gray value of the binarized data point is less than or equal to the comparison threshold, the original point cloud data point corresponding to the binarized data point is determined as an intermediate point cloud data point.

[0071] The step S3 includes:

[0072] For each boundary point cloud data point, a growth region of the boundary point cloud data point is constructed with the boundary point cloud data point as the center and the curvature radius as the radius of the boundary point cloud data point, and there is at least one intersection point between the growth regions corresponding to two adjacent boundary point cloud data points.

[0073] If there is only one intersection point between the growth areas corresponding to the two adjacent boundary point cloud data points, the intersection point is determined as a first new point cloud data point;

[0074] If the number of intersection points between the growth areas corresponding to the two adjacent boundary point cloud data points is greater than 1, a first new point cloud data point is determined according to the distance between the three-dimensional information corresponding to each intersection point and the three-dimensional information corresponding to the hole center.

[0075] In the present application, the radius of curvature is determined according to the density of the original point cloud data point; in this embodiment, the growth area constructed by the radius of curvature is a circle for a planar point cloud hole, and in a curved surface or a three-dimensional point cloud hole, the growth area constructed by the radius of curvature can be set as a sphere.

[0076] By performing the step S3, the first new point cloud data point enclosing the point cloud hole is generated, and the inward growth of the first layer of point cloud data points is completed. As shown in Figure 2 P i (i=1, 2...n) represents the boundary point cloud data point, L i (i=1, 2...n) represents the growth area of the boundary point cloud data point, C i (i=1, 2...n) represents the first new point cloud data point, K i (i=1, 2...n) represents the growth area of the first new point cloud data point.

[0077] The step S4 includes:

[0078] Step S4.1, for each first new point cloud data point, taking the first new point cloud data point as the center, taking the radius of curvature as the radius of the first new point cloud data point, and constructing the growth area of the first new point cloud data point;

[0079] For each first new point cloud data point, if there is only one intersection point between the growth area of the first new point cloud data point and the growth area of the adjacent original point cloud data point, the intersection point is determined as a second new point cloud data point, and if the number of intersection points between the growth area of the first new point cloud data point and the growth area of the adjacent original point cloud data point is greater than 1, a second new point cloud data point is determined according to the distance between the three-dimensional information corresponding to each intersection point and the three-dimensional information corresponding to the hole center.

[0080] Step S4.2, for each of the second new point cloud data points, repeating step S4.1 with the second new point cloud data point as the first new point cloud data point until the point cloud hole filling end condition is met, completing the point cloud hole filling.

[0081] The present application generates a plurality of new point cloud data points that close the point cloud holes by performing the step S4, and completes the in-growth of the multi-layer point cloud data points.

[0082] In this embodiment, the point cloud hole filling end condition includes:

[0083] According to the three-dimensional information corresponding to each of the original point cloud data points and the radius of curvature, it is determined that the point cloud data points cannot be added along the direction of the hole center.

[0084] Specifically, for each of the first new point cloud data points, a growth region T of the first new point cloud data point is constructed according to the radius of curvature, and if there is no intersection between the growth region T and the growth region of the adjacent original point cloud data point, the point cloud hole filling is ended; wherein for each of the first new point cloud data points, whether the first new point cloud data point is adjacent to each of the original point cloud data points is determined according to a preset distance threshold, if the distance between the first new point cloud data point and the original point cloud data point is greater than the distance threshold, it is determined that the first new point cloud data point is not adjacent to the original point cloud data point, and if the distance between the first new point cloud data point and the original point cloud data point is less than or equal to the distance threshold, it is determined that the first new point cloud data point is adjacent to the original point cloud data point.

[0085] As shown in Figure 3 The filling of the point cloud hole is gradually completed by gradually compressing and filling from the boundary point cloud data point to the center of the point cloud hole, and after each layer of point cloud data points is determined, the contour line can be formed by connecting the adjacent point cloud data points of this layer.

[0086] Embodiment two

[0087] Based on the above-mentioned embodiment one, for each of the two-dimensional projection data points, according to the two-dimensional coordinates and two-dimensional color information corresponding to the two-dimensional projection data points and the two-dimensional coordinates corresponding to the center projection point, the original point cloud data points corresponding to the two-dimensional projection data points are divided into boundary point cloud data points or intermediate point cloud data points, including:

[0088] For each of the two-dimensional projection data points, according to the two-dimensional coordinates corresponding to the center projection point, the adaptive threshold method and the histogram threshold method are used to divide the original point cloud data points corresponding to the two-dimensional projection data points into boundary point cloud data points or intermediate point cloud data points.

[0089] Embodiment three

[0090] On the basis of the above-mentioned embodiment one or two, the hole filling end condition of the point cloud comprises:

[0091] According to the image information corresponding to each of the original point cloud data points, it is determined that the point cloud data points cannot be added along the hole center direction.

[0092] Specifically, the two-dimensional projection data points corresponding to each of the original point cloud data points used to determine the first newly added point cloud data point or the second newly added point cloud data point are adjusted to the same color in the two-dimensional space, and if all the two-dimensional projection data points in the two-dimensional space are the same color, it is determined that the point cloud data points cannot be added along the hole center direction.

[0093] Embodiment four

[0094] To solve the technical problem that the subsequent processing is affected due to poor point cloud hole filling, the embodiment provides a point cloud reconstruction method based on boundary point cloud, comprising:

[0095] Obtaining a plurality of initial point cloud data points for a target to be reconstructed;

[0096] Taking the initial point cloud data points as original point cloud data points, performing the point cloud hole filling method based on boundary point cloud as described in any one of embodiments one to three to obtain target point cloud data points of the target to be reconstructed;

[0097] Triangulating the target point cloud data points to obtain a triangulation result;

[0098] According to the triangulation result, completing point cloud reconstruction of the target to be reconstructed.

[0099] Embodiment five

[0100] Based on the same principle as the point cloud hole filling method based on boundary point cloud in the above-mentioned embodiment one, the embodiment provides a point cloud hole filling system based on boundary point cloud, as shown in Figure 4 The system comprises:

[0101] A point cloud data point acquisition module is configured to acquire a plurality of original point cloud data points for an object, each of the original point cloud data points corresponding to three-dimensional information and image information, the three-dimensional information representing three-dimensional coordinates of the original point cloud data points in a three-dimensional space, and the image information representing three-dimensional color information of the original point cloud data points in the three-dimensional space;

[0102] a point cloud data point classification module configured to determine a hole center according to a plurality of the original point cloud data points, the hole being a three-dimensional shape formed by point cloud data points corresponding to a missing area of the object surface, the hole center corresponding to three-dimensional information in the three-dimensional space; and divide each of the original point cloud data points into a boundary point cloud data point or an intermediate point cloud data point according to the three-dimensional information and the image information corresponding to each of the boundary point cloud data points and the three-dimensional information corresponding to the hole center.

[0103] a first point cloud hole filling module configured to determine a first new point cloud data point according to the three-dimensional information corresponding to each of the boundary point cloud data points, the three-dimensional information corresponding to the hole center, and a preset radius of curvature, each of the first new point cloud data points corresponding to three-dimensional information;

[0104] a second point cloud hole filling module configured to complete point cloud hole filling according to the three-dimensional information corresponding to each of the original point cloud data points, the three-dimensional information corresponding to each of the first new point cloud data points, the three-dimensional information corresponding to the hole center, the radius of curvature, and a preset point cloud hole filling end condition.

[0105] The point cloud hole filling end condition includes:

[0106] determining that a new point cloud data point cannot be added in the direction of the hole center according to the three-dimensional information corresponding to each of the original point cloud data points and the radius of curvature;

[0107] or,

[0108] determining that a new point cloud data point cannot be added in the direction of the hole center according to the image information corresponding to each of the original point cloud data points.

[0109] The point cloud data point classification module includes:

[0110] a first processing unit configured to project the hole center into a two-dimensional space to obtain a center projection point, the center projection point corresponding to two-dimensional coordinates in the two-dimensional space;

[0111] a second processing unit configured to project each of the original point cloud data points into the two-dimensional space to obtain a two-dimensional projection data point corresponding to the original point cloud data point, the two-dimensional projection data point corresponding to two-dimensional coordinates and two-dimensional color information in the two-dimensional space;

[0112] The third processing unit is configured to, for each two-dimensional projection data point, divide the original point cloud data point corresponding to the two-dimensional projection data point into a boundary point cloud data point or an intermediate point cloud data point according to the two-dimensional coordinate and the two-dimensional color information corresponding to the two-dimensional projection data point and the two-dimensional coordinate corresponding to the center projection point.

[0113] The third processing unit is specifically configured to:

[0114] For each two-dimensional projection data point, the original point cloud data point corresponding to the two-dimensional projection data point is divided into a boundary point cloud data point or an intermediate point cloud data point according to the two-dimensional coordinate corresponding to the center projection point by using an image binarization method and a fixed threshold method.

[0115] Or,

[0116] For each two-dimensional projection data point, the original point cloud data point corresponding to the two-dimensional projection data point is divided into a boundary point cloud data point or an intermediate point cloud data point according to the two-dimensional coordinate corresponding to the center projection point by using an adaptive threshold method and a histogram threshold method.

[0117] The first point cloud hole filling module comprises:

[0118] The fourth processing unit is configured to, for each boundary point cloud data point, construct a growth region of the boundary point cloud data point with the boundary point cloud data point as the center and the curvature radius as the radius of the boundary point cloud data point; and for two adjacent boundary point cloud data points, there is at least one intersection point between the growth regions of the two boundary point cloud data points.

[0119] The fifth processing unit is configured to, for two adjacent boundary point cloud data points, if there is only one intersection point between the growth regions of the two boundary point cloud data points, determine the intersection point as a first new point cloud data point.

[0120] The sixth processing unit is configured to, for two adjacent boundary point cloud data points, if the number of intersection points between the growth regions of the two boundary point cloud data points is greater than 1, determine the intersection point close to the hole as a first new point cloud data point.

[0121] The second point cloud hole filling module comprises:

[0122] The seventh processing unit is configured to, for each first new point cloud data point, construct a growth region of the first new point cloud data point with the first new point cloud data point as the center and the curvature radius as the radius of the first new point cloud data point.

[0123] an eighth processing unit, configured to, for each of the first new point cloud data points, if there is only one intersection point between the growth region of the first new point cloud data point and the growth region of the adjacent original point cloud data point, determining the intersection point as a second new point cloud data point, and if the number of intersection points between the growth region of the first new point cloud data point and the growth region of the adjacent original point cloud data point is greater than 1, determining the second new point cloud data point according to the distance between the three-dimensional information corresponding to each of the intersection points and the three-dimensional information corresponding to the hole center;

[0124] a ninth processing unit, configured to, for each of the second new point cloud data points, taking the second new point cloud data point as a first new point cloud data point, repeatedly executing the seventh processing unit until the point cloud hole filling end condition is met, and completing the point cloud hole filling.

[0125] Embodiment six

[0126] To solve the technical problems in the prior art, an electronic device is provided in the embodiment, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the point cloud hole filling method based on boundary point cloud according to any one of the embodiments one to three when executing the computer program.

[0127] Embodiment seven

[0128] To solve the technical problems in the prior art, a computer readable storage medium is provided in the embodiment, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the point cloud hole filling method based on boundary point cloud according to any one of the embodiments one to three.

[0129] In the description of the present application, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application.

[0130] In addition, the terms "first", "second", etc. are used only for the purpose of description, and should not be understood as indicating or implying relative importance or a specific number of the technical features indicated. Therefore, the features defined as "first", "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "a plurality of" is at least two, for example, two, three, etc., unless otherwise explicitly and specifically limited.

[0131] In the description of the present application, the description referring to the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Also, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the present specification and the features of the different embodiments or examples, without contradiction.

[0132] Although the embodiments of the present application have been shown and described above, it is understood that the above-described embodiments are exemplary and should not be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-described embodiments within the scope of the present application.

Claims

1. A method for hole filling of a point cloud based on a boundary point cloud, characterized in that, The method comprises the following steps: S1, obtaining a plurality of original point cloud data points corresponding to an object, each of the original point cloud data points corresponding to three-dimensional information and image information, the three-dimensional information representing three-dimensional coordinates of the original point cloud data point in a three-dimensional space, and the image information representing three-dimensional color information of the original point cloud data point in the three-dimensional space; S2, determining a hole center according to a plurality of original point cloud data points, the hole being a three-dimensional shape formed by point cloud data points corresponding to a missing area of a surface of the object, the hole center corresponding to three-dimensional information in the three-dimensional space; According to the three-dimensional information and the image information corresponding to each of the boundary point cloud data points and the three-dimensional information corresponding to the hole center, each of the original point cloud data points is divided into a boundary point cloud data point or an intermediate point cloud data point; S3, determining a first new point cloud data point according to the three-dimensional information corresponding to each of the boundary point cloud data points, the three-dimensional information corresponding to the hole center and a preset curvature radius, each of the first new point cloud data points corresponding to three-dimensional information, and the specific steps comprising: For each of the boundary point cloud data points, a growth region of the boundary point cloud data point is constructed with the boundary point cloud data point as the center and the curvature radius as the radius of the boundary point cloud data point; for two adjacent boundary point cloud data points, there is at least one intersection point between the growth regions corresponding to the two boundary point cloud data points; For two adjacent boundary point cloud data points, if there is only one intersection point between the growth regions corresponding to the two boundary point cloud data points, the intersection point is determined as a first new point cloud data point; For two adjacent boundary point cloud data points, if the number of intersection points between the growth regions corresponding to the two boundary point cloud data points is greater than 1, a first new point cloud data point is determined according to the distance between the three-dimensional information corresponding to each of the intersection points and the three-dimensional information corresponding to the hole center; S4, completing point cloud hole filling according to the three-dimensional information corresponding to each of the original point cloud data points, the three-dimensional information corresponding to each of the first new point cloud data points, the three-dimensional information corresponding to the hole center, the curvature radius and a preset point cloud hole filling end condition.

2. The method of claim 1, wherein, In the step S2, according to the three-dimensional information and the image information corresponding to each of the boundary point cloud data points and the three-dimensional information corresponding to the hole center, each of the original point cloud data points is divided into a boundary point cloud data point or an intermediate point cloud data point, which comprises: Projecting the hole center into a two-dimensional space to obtain a center projection point, the center projection point corresponding to two-dimensional coordinates in the two-dimensional space; For each of the original point cloud data points, the original point cloud data point is projected into the two-dimensional space to obtain a two-dimensional projection data point corresponding to the original point cloud data point, the two-dimensional projection data point corresponding to two-dimensional coordinates and two-dimensional color information in the two-dimensional space; For each of the two-dimensional projection data points, according to the two-dimensional coordinates and two-dimensional color information corresponding to the two-dimensional projection data points and the two-dimensional coordinates corresponding to the center projection point, the original point cloud data points corresponding to the two-dimensional projection data points are divided into boundary point cloud data points or intermediate point cloud data points.

3. The method of claim 2, wherein, The step of dividing, for each of the two-dimensional projection data points, the original point cloud data points corresponding to the two-dimensional projection data points into boundary point cloud data points or intermediate point cloud data points according to the two-dimensional coordinates and two-dimensional color information corresponding to the two-dimensional projection data points and the two-dimensional coordinates corresponding to the center projection point comprises: For each of the two-dimensional projection data points, according to the two-dimensional coordinates corresponding to the center projection point, the original point cloud data points corresponding to the two-dimensional projection data points are divided into boundary point cloud data points or intermediate point cloud data points by using an image binarization method and a fixed threshold method; Or, For each of the two-dimensional projection data points, according to the two-dimensional coordinates corresponding to the center projection point, the original point cloud data points corresponding to the two-dimensional projection data points are divided into boundary point cloud data points or intermediate point cloud data points by using an adaptive threshold method and a histogram threshold method.

4. The method of claim 1, wherein, The step S4 comprises: Step S4.1, for each of the first new point cloud data points, taking the first new point cloud data point as the center, and taking the curvature radius as the radius of the first new point cloud data point, a growth region of the first new point cloud data point is constructed; For each of the first new point cloud data points, if there is only one intersection point between the growth region of the first new point cloud data point and the growth region of the adjacent original point cloud data point, the intersection point is determined as a second new point cloud data point, and if the number of intersection points between the growth region of the first new point cloud data point and the growth region of the adjacent original point cloud data point is greater than 1, a second new point cloud data point is determined according to the distance between the three-dimensional information corresponding to each of the intersection points and the three-dimensional information corresponding to the hole center; Step S4.2, for each of the second new point cloud data points, taking the second new point cloud data point as a first new point cloud data point, repeating step S4.1 until the point cloud hole filling end condition is met, and completing the point cloud hole filling.

5. The method according to any one of claims 1 to 4, characterized in that, The point cloud hole filling end condition comprises: According to the three-dimensional information corresponding to each of the original point cloud data points and the curvature radius, it is determined that the point cloud data point cannot be continuously added along the direction of the hole center; Or, According to the image information corresponding to each of the original point cloud data points, it is determined that the point cloud data point cannot be continuously added along the direction of the hole center. 6.A point cloud reconstruction method based on boundary point cloud, characterized in that, Comprise: Obtaining a plurality of initial point cloud data points for a target to be reconstructed; Taking the initial point cloud data points as original point cloud data points, performing the point cloud hole filling method based on boundary point clouds according to any one of claims 1 to 5 to obtain target point cloud data points of the target to be reconstructed; Triangulating the target point cloud data points to obtain a triangulation result; According to the triangulation result, completing the point cloud reconstruction of the target to be reconstructed.

7. A point cloud hole filling system based on boundary point cloud, characterized in that, Comprise: The point cloud data point obtaining module is configured to obtain a plurality of original point cloud data points for an object, each of the original point cloud data points corresponding to three-dimensional information and image information, the three-dimensional information representing a three-dimensional coordinate of the original point cloud data point in a three-dimensional space, and the image information representing three-dimensional color information of the original point cloud data point in the three-dimensional space; The point cloud data point classification module is configured to determine a hole center according to the plurality of original point cloud data points, the hole being a three-dimensional shape formed by point cloud data points corresponding to a missing area of a surface of the object, and the hole center corresponding to three-dimensional information in the three-dimensional space. Each of the original point cloud data points is divided into a boundary point cloud data point or an intermediate point cloud data point according to the three-dimensional information and the image information corresponding to each of the boundary point cloud data points and the three-dimensional information corresponding to the hole center. The first point cloud hole filling module is configured to determine a first newly added point cloud data point according to the three-dimensional information corresponding to each of the boundary point cloud data points, the three-dimensional information corresponding to the hole center, and a preset curvature radius, each of the first newly added point cloud data points corresponding to three-dimensional information, specifically: for each of the boundary point cloud data points, a growth region of the boundary point cloud data point is constructed with the boundary point cloud data point as a center and the curvature radius as a radius of the boundary point cloud data point; for two adjacent boundary point cloud data points, there is at least one intersection point between the growth regions corresponding to the two boundary point cloud data points; for two adjacent boundary point cloud data points, if there is only one intersection point between the growth regions corresponding to the two boundary point cloud data points, the intersection point is determined as a first newly added point cloud data point; for two adjacent boundary point cloud data points, if the number of intersection points between the growth regions corresponding to the two boundary point cloud data points is greater than 1, a first newly added point cloud data point is determined according to a distance between the three-dimensional information corresponding to each of the intersection points and the three-dimensional information corresponding to the hole center. The second point cloud hole filling module is configured to complete point cloud hole filling according to the three-dimensional information corresponding to each of the original point cloud data points, the three-dimensional information corresponding to each of the first newly added point cloud data points, the three-dimensional information corresponding to the hole center, the curvature radius, and a preset point cloud hole filling end condition.

8. An electronic device, comprising: The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the boundary point cloud-based point cloud hole filling method according to any one of claims 1 to 5.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the boundary point cloud-based point cloud hole filling method according to any one of claims 1 to 5.

Images