CN116188683A

CN116188683A - Three-dimensional object bounding box determination method, three-dimensional object bounding box determination device, computer equipment and storage medium

Info

Publication number: CN116188683A
Application number: CN202211725903.7A
Authority: CN
Inventors: 周浩源
Original assignee: Shenzhen Lingyun Shixun Technology Co ltd
Current assignee: Shenzhen Lingyun Shixun Technology Co ltd
Priority date: 2022-12-29
Filing date: 2022-12-29
Publication date: 2023-05-30

Abstract

The invention discloses a method, a device, computer equipment and a storage medium for determining a three-dimensional object bounding box, wherein the method comprises the following steps: acquiring a triangularization result of a three-dimensional convex hull of the three-dimensional object; classifying the convex hull edge lines according to similar data among the convex hull edge lines to obtain a plurality of edge line clusters; wherein the edge line cluster corresponds to a designated edge line; combining the appointed side lines to obtain a plurality of side line pairs; and determining a candidate bounding box with the volume meeting the bounding box screening condition as a target bounding box of the three-dimensional object in a plurality of candidate bounding boxes corresponding to the plurality of local coordinate systems determined by the side line pairs. Therefore, the acquisition efficiency of the bounding box can be greatly improved while the volume of the finally acquired bounding box is ensured to be approximately the theoretical minimum.

Description

Three-dimensional object bounding box determination method, three-dimensional object bounding box determination device, computer equipment and storage medium

Technical Field

The present invention relates to the field of data processing technologies, and in particular, to a method and apparatus for determining a three-dimensional object bounding box, a computer device, and a storage medium.

Background

The three-dimensional vision sensor combined with the three-dimensional machine vision detection algorithm is widely applied to a plurality of fields such as industrial intelligent manufacturing, biomedical treatment, reverse engineering, virtual reality and the like. The bounding box algorithm is used as a basic processing operator of the three-dimensional machine vision detection algorithm, and plays an important role in the fields of three-dimensional image processing, pattern recognition, collision detection, mold parting design, mechanical control and the like.

In the related art, the size and the direction of the bounding box are determined according to the geometry of the object, so that the original object can be compactly fitted. However, since the bounding box algorithm takes a long time, it cannot be directly used in the fields of collision detection and the like. The efficiency of the bounding box algorithm is to be improved.

Disclosure of Invention

The present invention aims to solve at least one of the technical problems in the related art to some extent. Therefore, a first object of the present invention is to provide a method for determining a bounding box of a three-dimensional object, which can greatly improve the acquisition efficiency of the bounding box while ensuring that the volume of the finally acquired bounding box approximately reaches the theoretical minimum, and can be directly applied to the related field with high-speed and high-precision detection requirements on the three-dimensional object.

A second object of the present invention is to provide a three-dimensional object bounding box determination apparatus.

A third object of the invention is to propose a computer device.

A fourth object of the present invention is to propose a computer readable storage medium.

To achieve the above object, an embodiment of a first aspect of the present invention provides a method for determining a three-dimensional object bounding box, including: acquiring a triangularization result of a three-dimensional convex hull of the three-dimensional object; wherein the triangularization result comprises a plurality of triangular patches; the common edge of two adjacent triangular patches is a convex hull edge line of the three-dimensional convex hull; classifying the convex hull edge lines according to similar data among the convex hull edge lines to obtain a plurality of edge line clusters; wherein the edge line cluster corresponds to a designated edge line; combining the appointed side lines to obtain a plurality of side line pairs; and determining a candidate bounding box with the volume meeting the bounding box screening condition as a target bounding box of the three-dimensional object in a plurality of candidate bounding boxes corresponding to the plurality of local coordinate systems determined by the side line pairs.

According to one embodiment of the invention, the similarity data between the convex hull edge lines is determined by the included angle between the expression vectors of the convex hull edge lines; the two adjacent triangular patches comprise a first triangular patch and a second triangular patch; the generation mode of the expression vector of the convex hull edge comprises the following steps: determining a cross vector of the first normal vector and the second normal vector by using a first normal vector of the first triangular patch and a second normal vector of the second triangular patch; and generating a representation vector of the convex hull edge according to the first normal vector, the second normal vector and the cross multiplication vector.

According to one embodiment of the present invention, the classifying the plurality of convex hull edge lines according to the similarity data between the convex hull edge lines to obtain a plurality of edge line clusters includes: selecting one convex hull edge line from the convex hull edge lines as a reference edge line; traversing the convex hull edge lines to determine an included angle between the representative vector of each convex hull edge line and the representative vector of the reference edge line; determining convex hull edge lines corresponding to included angles smaller than a first angle threshold, wherein the reference edge lines belong to the same edge line cluster.

According to one embodiment of the present invention, the designated edge uses the reference edge to represent a convex hull edge included in the edge cluster; combining the specified edges to obtain a plurality of edge pairs, including: and combining the reference edges in pairs to obtain a plurality of edge pairs.

According to one embodiment of the present invention, the determining manner of the plurality of local coordinate systems includes: establishing a target equation through a plurality of edge line pairs based on the characteristic that normals of two adjacent surfaces of the bounding box are perpendicular to each other; determining a plurality of local coordinate systems according to a plurality of solution set vector pairs included in the solution set of the target equation; the solution set vector pair comprises a first vector and a second vector, wherein the first vector is a normal vector of a first surface of two adjacent surfaces of the bounding box, and the second vector is a normal vector of a second surface of the two adjacent surfaces of the bounding box.

According to one embodiment of the present invention, the determining a plurality of local coordinate systems according to a plurality of solution set vector pairs included in a solution set of the target equation includes: constructing a plurality of initial coordinate systems to be screened according to a plurality of solution set vector pairs included in the solution set; and screening a plurality of initial coordinate systems according to the similar data among the initial coordinate systems to obtain a plurality of local coordinate systems.

According to one embodiment of the present invention, the filtering the plurality of initial coordinate systems according to the similarity data between the initial coordinate systems to obtain the plurality of local coordinate systems includes: classifying the plurality of initial coordinate systems based on similar data among the initial coordinate systems to obtain a plurality of coordinate system clusters; the method comprises the steps of aiming at a first initial coordinate system and a second initial coordinate system in the coordinate system cluster, wherein the included angle between three coordinate axes in the first initial coordinate system and corresponding three coordinate axes in the second initial coordinate system is smaller than a second angle threshold value respectively; and reserving a specified coordinate system in the coordinate system cluster as the local coordinate system.

According to one embodiment of the invention, the pair of edges comprises a first edge and a second edge; the establishing a target equation through a plurality of edge line pairs based on the normal mutually perpendicular characteristics of two adjacent surfaces of the bounding box comprises the following steps: constructing a first normal vector to be solved of the first surface based on normal vectors of two triangular patches sharing the first side line; constructing a second normal vector to be solved of the second surface based on normal vectors of two triangular patches sharing the second side line; and establishing the target equation according to the normal mutually perpendicular characteristics of two adjacent surfaces of the bounding box, the first normal vector to be solved and the second normal vector to be solved.

According to one embodiment of the present invention, the constructing a first normal vector to be solved for the first surface based on normal vectors of two triangular patches sharing the first edge includes: determining a first pair of normal vectors based on normal vectors of two triangular patches sharing the first edge; constructing a first normal vector to be solved of the first surface by utilizing the first normal vector pair; the constructing a second normal vector to be solved of the second surface based on normal vectors of two triangular patches sharing the second edge includes: determining a second pair of normal vectors based on normal vectors of two triangular patches sharing the second edge; and constructing a second normal vector to be solved of the second surface by using the second normal vector pair.

According to one embodiment of the invention, the first face corresponds to a first weight factor and the second face corresponds to a second weight factor; the constructing a first normal vector to be solved of the first surface by using the first normal vector pair includes: constructing the first normal vector to be solved by using the first normal vector pair and the first weight factor; the constructing a second normal vector to be solved of the second surface by using the second normal vector pair includes: and constructing a second normal vector to be solved of the second surface by using the second normal vector pair and the second weight factor.

According to one embodiment of the invention, the solution set of the target equation is determined in a first weight value range of the first weight factor and in a second weight value range of the second weight factor.

To achieve the above object, a second aspect of the present invention provides a three-dimensional object bounding box determining apparatus, including: the triangularization result acquisition module is used for acquiring the triangularization result of the three-dimensional convex hull of the three-dimensional object; wherein the triangularization result comprises a plurality of triangular patches; the common edge of two adjacent triangular patches is a convex hull edge line of the three-dimensional convex hull; the classification processing module is used for classifying the convex hull edge lines according to the similar data among the convex hull edge lines to obtain a plurality of edge line clusters; wherein the edge line cluster corresponds to a designated edge line; the combination module is used for combining the appointed side lines to obtain a plurality of side line pairs; and the determining module is used for determining the candidate bounding box with the volume meeting the bounding box screening condition as the target bounding box of the three-dimensional object in a plurality of candidate bounding boxes corresponding to a plurality of local coordinate systems determined by a plurality of edge pairs.

To achieve the above object, an embodiment of a third aspect of the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the method according to any one of the preceding embodiments when the processor executes the computer program.

To achieve the above object, a fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps of the method according to any one of the preceding embodiments.

According to the embodiments provided by the invention, the vertex model representing the complete three-dimensional object is obtained by adopting the convex hull algorithm and the convex hull triangularization algorithm, and the data volume processed by the algorithm can be reduced while the boundary characteristics of the three-dimensional object are maintained, so that the efficiency of obtaining the bounding box is improved; by carrying out similarity screening on the convex hull edge line and carrying out similarity screening on the local coordinate system, the repeated calculation amount of the similar bounding box can be avoided, and the overall efficiency of the bounding box determining method is greatly improved; and a target equation is established by combining the inherent geometric characteristics of the minimum volume bounding box so as to solve the local coordinate system, and the accuracy of the local coordinate system is ensured.

Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.

Drawings

Fig. 1 is a flowchart of a method for determining a three-dimensional object bounding box according to an embodiment of the present disclosure.

Fig. 2 is a flow chart illustrating a generation manner of a representation vector of a convex hull edge according to an embodiment of the present disclosure.

Fig. 3 is a schematic flow chart of a classification process for multiple convex hull edge lines according to an embodiment of the present disclosure.

Fig. 4a is a flow chart illustrating a method for determining a plurality of local coordinate systems according to an embodiment of the present disclosure.

Fig. 4b is a schematic view of a normal vector distribution of one face of a bounding box according to an embodiment of the present disclosure.

Fig. 5 is a schematic flow chart of determining a plurality of local coordinate systems according to one embodiment of the present disclosure.

Fig. 6 is a schematic flow chart of screening a plurality of initial coordinate systems according to one embodiment of the present disclosure.

FIG. 7a is a schematic flow chart of establishing a target equation by a plurality of edge pairs according to one embodiment of the present disclosure.

FIG. 7b is a schematic flow chart of establishing a target equation through a plurality of edge pairs according to one embodiment of the present disclosure.

Fig. 8 is a flowchart of a method for determining a three-dimensional object bounding box according to an embodiment of the present disclosure.

Fig. 9 is a block diagram of a three-dimensional object bounding box determination apparatus provided according to an embodiment of the present specification.

Fig. 10 is a block diagram of a computer device according to one embodiment of the present disclosure.

Detailed Description

Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention.

At present, the three-dimensional vision sensor combined with the three-dimensional machine vision detection algorithm is widely applied to a plurality of fields such as industrial intelligent manufacturing, biomedical science, reverse engineering, virtual reality and the like, and has the characteristics of high detection precision, high detection efficiency, simplicity and convenience in operation and deployment and the like. The bounding box algorithm is used as a basic processing operator of the three-dimensional machine vision detection algorithm, and plays an important role in the fields of three-dimensional image processing, pattern recognition, collision detection, mold parting design, mechanical control and the like of a three-dimensional object. In the field of industrial intelligent manufacturing, the requirement of high detection precision and high efficiency on the detection of a three-dimensional object can be met, so that an efficient, stable and high-precision bounding box algorithm is particularly important.

In the related art, the most widely applied OBB (Oriented Bounding Box, directed bounding box) algorithm determines the size and direction of the bounding box according to the geometry of the object itself, and can perform a compact fit to the original object. For a given set of three-dimensional object spatial coordinate points, how to efficiently and accurately obtain its smallest directed bounding box has been the focus of research in the related art.

The optimal direction is generally found by different algorithms based on the spatial distribution of all vertices of the three-dimensional object to determine several principal axes of the directional bounding box. The common bounding box calculation method mainly comprises the following two steps:

(1) A Principal Component Analysis (PCA) -based method acquires a covariance matrix based on a set of spatial coordinate points of sampling points of an input three-dimensional object to calculate a feature vector, estimates a direction of a maximum spread distribution in the set of spatial coordinate points, and takes it as a principal axis direction of a bounding box. The algorithm flow of the method is simpler, but for three-dimensional objects with uneven spatial coordinate distribution of sampling points or large difference of spatial relations in spatial coordinate point concentration, the method is difficult to accurately acquire the minimum directional bounding box of the three-dimensional object.

(2) Sampling on a unit hemisphere of a three-dimensional object space coordinate point set, and acquiring candidate directions of a plurality of bounding box spindles through violent search; secondly, traversing and searching all candidate directions by combining algorithms such as genetic algorithm or particle swarm optimization, and searching minimum volume bounding boxes in a plurality of candidate directions by combining projection and two-dimensional minimum bounding rectangles; the bounding boxes in all candidate directions are then compared, with the bounding box with the smallest volume being the smallest directional bounding box of the three-dimensional object. The method has the defects of difficult adjustment of genetic algorithm parameters, difficult selection of initial direction candidates for violent search, low algorithm efficiency and the like.

In order to improve the calculation efficiency and calculation accuracy of the minimum directional bounding box, it is necessary to propose a three-dimensional object bounding box determination method, apparatus, computer device and storage medium. Firstly, acquiring a three-dimensional convex hull of a three-dimensional object space coordinate point set and triangulating convex hull vertexes to obtain a vertex model of the three-dimensional object, and greatly reducing the data volume to be processed while keeping the boundary characteristics of the three-dimensional object, so that the calculation efficiency of a bounding box is improved; secondly, screening convex hull edge lines of the three-dimensional convex hull through a similarity screening strategy of the convex hull edge lines so as to reduce the number of the convex hull edge lines for acquiring the candidate directions of the principal axes of the bounding boxes; then, according to the geometric characteristics that two mutually perpendicular surfaces in the minimum directional bounding box are respectively overlapped with two side lines of a three-dimensional convex hull of the three-dimensional object, establishing a target equation and solving the target equation to obtain a local coordinate system of a plurality of bounding box main shaft candidate directions; then, combining a similarity screening principle of the local coordinate systems, and screening the local coordinate systems in all candidate directions to further reduce the number of candidate local coordinate systems for calculating the minimum directional bounding box and improve the calculation efficiency of the bounding box; and finally, traversing in the candidate bounding boxes calculated based on the local coordinate system, and searching for a directional bounding box with the smallest volume to serve as the smallest directional bounding box of the three-dimensional object. Compared with the related art, the method can improve the calculation efficiency of the bounding box while ensuring that the volume of the finally obtained bounding box approximately reaches the theoretical minimum, and can be directly applied to the related field with high-speed and high-precision detection requirements on the three-dimensional object.

The embodiment of the present specification provides a three-dimensional object bounding box determination method, which may include the following steps with reference to fig. 1.

S110, acquiring a triangularization result of a three-dimensional convex hull of the three-dimensional object. Wherein the triangularization result comprises a plurality of triangular patches; the common edge of two adjacent triangular surface patches is a convex hull edge line of the three-dimensional convex hull.

The three-dimensional object can be a three-dimensional object, the surface of the three-dimensional object corresponds to a point data set, and the point data set can be calculated by utilizing a three-dimensional convex hull calculation algorithm to obtain a three-dimensional convex hull of the three-dimensional object. In some embodiments, the three-dimensional object may also be a point data set of a three-dimensional object surface, i.e. a three-dimensional point cloud. The three-dimensional convex hull may be obtained based on a three-dimensional point cloud. The triangulation result may be based on the vertices of a three-dimensional convex hull. The convex hull edge lines may be lines between vertices of a three-dimensional convex hull.

Specifically, a convex hull calculation method may be used to obtain a three-dimensional convex hull corresponding to a three-dimensional point cloud of a three-dimensional object according to a spatial relationship or the like of point data in the three-dimensional point cloud. And then triangulating the vertex point set of the three-dimensional convex hull by using a triangulating algorithm with the vertex of the three-dimensional convex hull as the point set to obtain a triangulating result of the three-dimensional convex hull.

In some embodiments, the open source third party library Qhull includes a mature three-dimensional convex hull calculation algorithm and a trigonometric algorithm, and the three-dimensional convex hull and the trigonometric result thereof can be obtained through the open source third party library Qhull. By way of example, taking a three-dimensional point cloud of a three-dimensional object as input, firstly, a three-dimensional convex hull of the three-dimensional point cloud can be obtained through a three-dimensional convex hull calculation algorithm of a third party library Qhull, and secondly, a vertex point set of the three-dimensional convex hull can be taken as input, and a triangulation result of the vertex of the three-dimensional convex hull can be obtained through a triangulation algorithm of the third party library Qhull.

It can be appreciated that after triangulating the vertices of the three-dimensional convex hull, a plurality of triangular patches can be obtained, wherein the vertices of the triangular patches are the vertices of the three-dimensional convex hull. Therefore, the triangular surface patches can embody the connection relation between the vertexes of the three-dimensional convex hulls, and the common edge of two adjacent triangular surface patches is the convex hull edge line of the three-dimensional convex hulls. It should be noted that QHull is an open source program software for researching and solving the convex hull problem and generating a convex hull form.

S120, classifying the convex hull edge lines according to similar data among the convex hull edge lines to obtain a plurality of edge line clusters. Wherein the edge cluster corresponds to a designated edge.

The similarity data may be used to measure similarity between any two convex hull edge lines, such as whether any two convex hull edge lines are oriented close to each other. The specified edge may be a convex hull edge obtained from the edge cluster for computing a bounding box candidate direction, and may represent similar spatial characteristics of the convex hull edge in the corresponding edge cluster.

It will be appreciated that for a plurality of edge clusters, the similarity of convex hull edge lines in each edge cluster is higher and the similarity of convex hull edge lines between edge clusters is lower. The appointed edge lines which can represent the convex hull edge line characteristics in the corresponding edge line clusters are obtained from the edge line clusters, and the number of the convex hull edge lines with higher similarity can be reduced, so that the number of the convex hull edge lines used for calculating and obtaining the candidate directions of the principal axes of the bounding boxes is reduced, and the efficiency of an algorithm can be effectively improved.

In some embodiments, the similarity between any two convex hull edge lines may be determined based on the included angle and orientation of two adjacent triangular patches of the convex hull edge lines. Further, the included angle and the orientation of the triangular patches may be determined by the orientation of the normal vector corresponding to the triangular patches. Illustratively, two adjacent triangular patches with convex hull edge line L1 as a common edge are P11 and P12 respectively, the normal vector corresponding to the triangular patch P11 is N11, and the normal vector corresponding to the triangular patch P12 is N12; two adjacent triangular patches taking the convex hull edge line L2 as a common edge are respectively P21 and P22, the normal vector corresponding to the triangular patch P21 is N21, and the normal vector corresponding to the triangular patch P22 is N22. The normal vector N11 and N12 are used as the normal line pair of the adjacent triangular surface patches of the convex hull edge line L1, the normal vector N21 and N22 are used as the normal line pair of the adjacent triangular surface patches of the convex hull edge line L2, and the orientation similarity of the normal line pair of the adjacent triangular surface patches respectively corresponding to the convex hull edge line L1 and the convex hull edge line L2 is used as similar data to judge the similarity between the convex hull edge line L1 and the convex hull edge line L2. If the orientation similarity of the normal pairs of the adjacent triangular patches corresponding to the convex hull edge line L1 and the convex hull edge line L2 is large, the convex hull edge line L1 and the convex hull edge line L2 can be classified into the same edge line cluster, and the convex hull edge line L1 and/or the convex hull edge line L2 can be used as the appointed edge line of the corresponding edge line cluster; if the orientation similarity of the normal pairs of the adjacent triangular patches to which the convex hull edge line L1 and the convex hull edge line L2 respectively correspond is small, the convex hull edge line L1 and the convex hull edge line L2 may be classified into different edge line clusters, and the convex hull edge line L1 and the convex hull edge line L2 may respectively serve as designated edge lines of the corresponding edge line clusters.

In some embodiments, a weighted average calculation may be performed on the normal pairs of two adjacent triangular patches of the convex hull edge according to a weighted average rule, resulting in a corresponding weighted average vector. And determining similar data based on the included angle between the weighted average vectors, and judging the similarity between the convex hull edge lines.

S130, combining the specified edges to obtain a plurality of edge pairs.

Specifically, arbitrary two-edge line combinations may be performed on the designated edges obtained from each edge line cluster, and each of the obtained combinations of the designated edges may be used as one edge line pair, so that a plurality of edge line pairs may be obtained. Illustratively, the plurality of edge clusters may include edge cluster C1, edge cluster C2, edge cluster C3. If the designated edge corresponding to the edge cluster C1 is LC1, the designated edge corresponding to the edge cluster C2 is LC2, and the designated edge corresponding to the edge cluster C1 is LC2, the designated edge LC1, the designated edge LC2, and the designated edge LC3 are combined in pairs to obtain an edge pair including LC1 and LC2, an edge pair including LC1 and LC3, and an edge pair including LC2 and LC 3.

And S140, determining the candidate bounding box with the volume meeting the bounding box screening condition as a target bounding box of the three-dimensional object in a plurality of candidate bounding boxes corresponding to the plurality of local coordinate systems determined by the plurality of edge pairs.

The local coordinate system may be a coordinate system in a bounding box principal axis candidate direction corresponding to the three-dimensional convex hull. The target bounding box may be a minimum volume bounding box.

In some cases, two mutually perpendicular adjacent faces of the minimum volume bounding box corresponding to the three-dimensional object are respectively overlapped with two convex hull edge lines of the three-dimensional convex hull corresponding to the three-dimensional object. Thus, the normals of two adjacent faces of the minimum volume bounding box may be represented by the normals of two adjacent triangular patches of convex hull edge lines that coincide with the corresponding faces, respectively. It will be appreciated that from the normals of the two adjacent faces of the minimum volume bounding box, and the cross-product vector of those normals, the coordinate system in the principal axis candidate direction of the minimum volume bounding box can be determined. Therefore, the coordinate system in the principal axis candidate direction of the minimum volume bounding box can be determined as the local coordinate system from the convex hull edge lines respectively overlapping with the adjacent faces.

Specifically, according to the plurality of edge pairs obtained above, it may be assumed that the two convex hull edge lines in each edge pair respectively coincide with two adjacent faces of the minimum volume bounding box, so that a corresponding plurality of local coordinate systems may be determined according to the plurality of edge pairs. On the determined local coordinate systems, candidate bounding boxes corresponding to the local coordinate systems can be obtained through calculation through a bounding box algorithm, and the candidate bounding boxes with volumes meeting the bounding box screening conditions can be determined to be target bounding boxes of the three-dimensional object. For example, the volumes of the candidate bounding boxes corresponding to the respective local coordinate systems may be calculated and compared, and the candidate bounding box in which the volume is smallest is determined as the smallest volume bounding box of the three-dimensional object and finally output.

In the above embodiment, by performing similarity screening on the convex hull edge lines of the three-dimensional convex hull, the number of the convex hull edge lines for acquiring the local coordinate system in the principal axis candidate direction of the bounding box is greatly reduced, and the efficiency of acquiring the local coordinate system subsequently can be effectively improved.

In some embodiments, the similarity data between convex hull edge lines is determined by the angles between the representative vectors of the convex hull edge lines; the two adjacent triangular panels include a first triangular panel and a second triangular panel. Referring to fig. 2, the generation of the representation vector of the convex hull edge may include the following steps.

S210, determining a cross vector of the first normal vector and the second normal vector by using the first normal vector of the first triangular patch and the second normal vector of the second triangular patch.

The expression vector may be a weighted average vector obtained by calculating normal vectors corresponding to two adjacent triangular patches of the convex hull edge line according to a weighted average rule of the vectors, and is used for expressing included angles and orientations of the two adjacent triangular patches of the convex hull edge line.

In some cases, there may be rotation symmetry or the like between two adjacent triangular panels corresponding to two convex hull edge lines, so that a representation vector of a convex hull edge line may be determined according to normal vectors corresponding to two adjacent triangular panels of the convex hull edge line, so that the representation vector can be used to represent an included angle and an orientation of two adjacent triangular panels of the convex hull edge line.

S220, generating a representation vector of the convex hull edge line according to the first normal vector, the second normal vector and the cross multiplication vector.

In some embodiments, the weighted average rule of the vectors may be vector addition. Illustratively, the normal vectors of all triangular patches are calculated according to the triangularization result of the three-dimensional convex hull of the three-dimensional object obtained as described above. The first normal vector n1 of the adjacent first triangular patch 1 and the second normal vector n2 of the second triangular patch 2 are subjected to a cross multiplication operation, and a cross multiplication vector n3 of the first normal vector n1 and the second normal vector n2 can be obtained. Vector addition is performed on the first normal vector n1, the second normal vector n2 and the cross vector n3, so that a representation vector of the corresponding convex hull edge is n=n1+n2+n3.

Based on the included angles between the representative vectors of the convex hull edge lines, the similarity between the convex hull edge lines can be determined. For example, an angle threshold value representing the included angle between vectors may be set as a basis for judging similarity or dissimilarity between convex hull edge lines. When the included angle between the two expression vectors is in the range of the angle threshold value, the two corresponding convex hull edge lines are similar; when the included angle between the two expression vectors is out of the angle threshold range, the two corresponding convex hull edge lines are dissimilar.

In some embodiments, referring to fig. 3, classifying the plurality of convex hull edge lines according to similar data between the convex hull edge lines to obtain a plurality of edge line clusters may include the following steps.

S310, selecting one convex hull edge line from a plurality of convex hull edge lines as a reference edge line.

S320, traversing a plurality of convex hull edge lines to determine the included angle between the representation vector of each convex hull edge line and the representation vector of the reference edge line.

S330, determining convex hull edge lines corresponding to the included angle smaller than the first angle threshold and the reference edge lines belong to the same edge line cluster.

Wherein the reference edge may be used as a central element for constructing the edge cluster.

Illustratively, the normal lines of all triangular patches and the expression vector of the convex hull edge line are calculated according to the triangularization result of the three-dimensional convex hull obtained by the above. Setting a first angle threshold, and setting a classification rule for classifying convex hull edge lines according to the first angle threshold:

and selecting one convex hull edge line as a reference edge line in the triangularization result of the three-dimensional convex hull, taking the reference edge line as a central element for constructing an edge line cluster, traversing the rest convex hull edge lines, and respectively calculating the included angle between the representation vector of the rest convex hull edge line and the representation vector of the reference edge line. According to the classification rule, the convex hull edge corresponding to the included angle smaller than the first angle threshold can be placed into the edge cluster where the reference edge is located, and the convex hull edge corresponding to the included angle which does not meet the condition smaller than the first angle threshold is used as the remaining convex hull edge to participate in the next traversal process. And when the convex hull edge lines corresponding to the included angles which meet the condition of being smaller than the first angle threshold value are not found in the rest convex hull edge lines, ending the traversing process and starting the next traversing. The traversal process is repeated until the convex hull edge lines are classified into corresponding edge line clusters. And respectively taking out the appointed side lines from the finally obtained side line clusters, and participating in the subsequent process of determining the local coordinate system.

It should be noted that, the first angle threshold may be set according to an accuracy requirement of calculating the bounding box. In some embodiments, the first angle threshold may be valued in the range of 1 ° to 10 °.

In some embodiments, the designated edge uses a reference edge to represent a convex hull edge that is included in the edge cluster. Combining the plurality of designated edges to obtain a plurality of edge pairs may include: and combining the reference edges in pairs to obtain a plurality of edge pairs.

Specifically, in the process of performing traversal and classification processing on convex hull edge lines, the reference edge lines selected in each traversal can be used as appointed edge lines, all the reference edge lines are summarized, and a convex hull edge line set after similarity screening can be obtained. And combining the reference edges in the set in pairs to obtain a plurality of edge pairs for subsequently establishing a target equation, and obtaining a local coordinate system for constructing the candidate bounding box by solving the target equation.

In some embodiments, the reference coordinate system selected by each traversal may be reserved as a designated edge, and other convex hull edges in the edge cluster except for the reference edge are removed. And summarizing all the reference edges to obtain a convex hull edge set subjected to similarity screening.

In some embodiments, referring to FIG. 4a, the manner in which the plurality of local coordinate systems are determined may include the following steps.

S410, establishing a target equation through a plurality of edge line pairs based on the characteristic that normals of two adjacent surfaces of the bounding box are perpendicular to each other.

It will be appreciated that the adjacent faces of the bounding box are perpendicular to each other, and therefore, the normals of the adjacent faces of the bounding box are perpendicular to each other, and the product of the normals of the adjacent faces of the bounding box is 0. According to the geometric characteristics of the minimum volume bounding box, two adjacent surfaces of the minimum volume bounding box corresponding to the three-dimensional object are respectively overlapped with two convex hull edge lines of the three-dimensional convex hull corresponding to the three-dimensional object, so that the normal vector of the two adjacent surfaces of the minimum volume bounding box can be respectively linearly represented by the normal vector of two adjacent triangular patches of the corresponding convex hull edge lines overlapped with the normal vector. Assuming that the two convex hull edge lines in each edge line pair are respectively coincident with the adjacent two faces of the bounding box, the normal vector of the adjacent two faces of the bounding box can be respectively represented linearly by the normal vector of the adjacent triangular patches corresponding to the convex hull edge lines in the edge line pairs. According to the normal vector product of two adjacent surfaces of the bounding box is 0, a corresponding target equation can be established.

Illustratively, the convex hull borderline L is centered with the borderline ₁ The normal vector of the first surface of the overlapped bounding box is N, and the convex hull edge line L ₁ Corresponding to adjacent triangular surface patches P ₁₀ And triangular surface patch P ₁₁ Triangular patch P ₁₀ Corresponding to normal vector N ₀ Triangular patch P ₁₁ Corresponding to normal vector N ₁ . The normal vector of the first face of the bounding box may be expressed as: n=a ₁ N ₀ +b ₁ N ₁ Wherein a is ₁ And b ₁ Is a weight factor.

Convex hull edge line L centered with the edge line ₂ The normal vector of the second surface of the overlapped bounding box is M, and the convex hull edge line L ₂ Corresponding to adjacent triangular surface patches P ₂₀ And triangular surface patch P ₂₁ Triangular patch P ₂₀ Corresponding to normal vector M ₀ Triangular patch P ₂₁ Corresponding to normal vector M ₁ . The normal vector of the second face of the bounding box can be expressed as: m=a ₂ M ₀ +b ₂ M ₁ Wherein a is ₂ And b ₂ Is a weight factor.

From the product of the normal vector of the first face and the second face being 0, the objective equation can be established as follows:

F＝N·M

＝a ₁ a ₂ N ₀ ·M ₀ +b ₁ a ₂ N ₁ ·M ₀ +a ₁ b ₂ N ₀ ·M ₁ +b ₁ b ₂ N ₁ ·M ₁

＝a ₁ a ₂ f ₀₀ +b ₁ a ₂ f ₁₀ +a ₁ b ₂ f ₀₁ +b ₁ b ₂ f ₁₁

＝0

wherein f ₀₀ ＝N ₀ ·M ₀ ，f ₁₀ ＝N ₁ ·M ₀ ,f ₀₁ ＝N ₀ ·M ₁ ，f ₁₁ ＝N ₁ ·M ₁ 。

S420, determining a plurality of local coordinate systems according to a plurality of solution set vector pairs included in the solution set of the target equation.

The solution vector pair comprises a first vector and a second vector, wherein the first vector is a normal vector of a first surface of two adjacent surfaces of the bounding box, and the second vector is a normal vector of a second surface of the two adjacent surfaces of the bounding box.

It will be appreciated that, according to the geometric characteristics of the bounding box, the triangular patches corresponding to the convex hull edge are all contained within the bounding box, so that the geometric characteristics of the normal vectors corresponding to the first and second faces of the bounding box can be obtained. Illustratively, referring to FIG. 4b, a first face of the bounding box presents at most a convex hull edge line L ₁ Triangular surface patch P of (2) ₁₀ Or triangular surface patch P ₁₁ Limit cases of coincidence. From this, the normal vector N is equal to the normal vector N ₀ From the normal vector N ₁ Within an angle range of (2), and the normal vector N is present at most with the normal vector N ₀ Or normal vector N ₁ And (3) overlapping. Thus, the weight factor a ₁ And b ₁ Sum is 1, weight factor a ₂ And b ₂ And also is 1, and a ₁ 、b ₁ 、a ₂ 、b ₂ The value range of (2) is [0,1 ]]. Thus, the problem of obtaining the minimum volume bounding box translates into the problem of finding a solution set that satisfies the target equation f=0, it being understood that the solution set includes the weight factors (a ₁ ,b ₁ ) And (a) ₂ ,b ₂ ) Is a combination of all possible pairs of values.

Illustratively, the value range based on the weight factors is according to f ₀₀ 、f ₁₀ 、f ₀₁ 、f ₁₁ The values of (2) are positive number, negative number or zero, and the like, and the target equation can be discussed and solved according to 81 conditions. It will be appreciated that the solution set of the objective equation for each case includes infinity (a ₁ ,b ₁ ) And (a) ₂ ,b ₂ ) Therefore, the number of samples for sampling the solution set of each case needs to be set to obtain a partial solution set thereof as a candidate solution set in the corresponding case.

Solving the target equation, acquiring a candidate solution set according to the sampling number, and then obtaining (a) in the candidate solution set ₁ ,b ₁ ) And (a) ₂ ,b ₂ ) Substituting the corresponding values of the target equation into the linear expression of the normal N and the normal M to obtain a plurality of structures of the normal vector N and the normal vector MThe resulting solution sets vector pairs are used to represent pairs of normal vector combinations of two adjacent faces of the bounding boxes, respectively. And respectively carrying out cross multiplication calculation on the solution set vector pairs to obtain corresponding cross multiplication result vectors of the solution set vector pairs. According to each solution vector pair and the corresponding cross product vector, a direction vector combination formed by three mutually perpendicular vectors can be respectively obtained. Based on the combination of the plurality of direction vectors, a plurality of corresponding local coordinate systems can be constructed. In some embodiments, the origin of the local coordinate system may be established at the origin of the world coordinate system.

Further, bounding boxes corresponding to the multiple local coordinate systems can be obtained through a bounding box algorithm based on the multiple local coordinate systems obtained through construction, the bounding boxes are used as multiple candidate bounding boxes, and the length, width and height of each candidate bounding box can be determined and used for calculating the volume of the corresponding candidate bounding box. The candidate bounding box with the smallest volume may be selected from the plurality of candidate bounding boxes as the smallest volume bounding box of the three-dimensional object to be finally output.

The number of samples may be set according to the accuracy requirement for calculating the bounding box.

In some embodiments, referring to fig. 5, determining a plurality of local coordinate systems from a plurality of solution set vector pairs included in a solution set of the target equation may include the following steps.

S510, constructing a plurality of initial coordinate systems to be screened according to a plurality of solution set vector pairs included in the solution set.

S520, screening the plurality of initial coordinate systems according to the similar data among the initial coordinate systems to obtain a plurality of local coordinate systems.

Wherein the similarity data can be used to measure the similarity between any two initial coordinate systems.

In some cases, there may be some similarity between the corresponding plurality of local coordinate systems constructed from the plurality of solution set vector pairs, i.e., there may be some local coordinate systems with directions oriented similarly. It can be understood that bounding boxes calculated based on local coordinate systems with similar direction orientation are similar, so that a plurality of local coordinate systems constructed according to a plurality of solution vector pairs can be used as an initial coordinate system, and screening is performed on the initial coordinate system to remove coordinate systems with larger direction orientation similarity, so that the number of bounding boxes to be calculated is reduced, and the overall efficiency of an algorithm is improved.

Specifically, after a plurality of initial coordinate systems are constructed, the similarity between the initial coordinate systems can be judged according to the directions of three mutually perpendicular direction vectors used for constructing the corresponding initial coordinate systems, and coordinate systems with larger similarity are removed, so that local coordinate systems with smaller similarity between the initial coordinate systems are finally obtained, and the screening of the initial coordinate systems is completed.

Illustratively, for the initial coordinate system C1, three mutually perpendicular direction vectors for constructing the corresponding initial coordinate system are X1, Y1, and Z1, respectively; for the initial coordinate system C2, three mutually perpendicular direction vectors for constructing the corresponding initial coordinate system are X2, Y2 and Z2, respectively. The vector included angles of the direction vectors X1 and X2, the vector included angles of Y1 and Y2, and the vector included angles of Z1 and Z2 can be calculated respectively, so as to determine the similarity of the direction orientations of the initial coordinate system C1 and the initial coordinate system C2. If the vector included angle between X1 and X2, the vector included angle between Y1 and Y2 and the vector included angle between Z1 and Z2 are smaller, the initial coordinate system C1 is similar to the initial coordinate system C2, and a local coordinate system which is finally needed can be selected from the initial coordinate system C1; if the vector included angle between X1 and X2, the vector included angle between Y1 and Y2, and the vector included angle between Z1 and Z2 are larger, the initial coordinate system C1 and the initial coordinate system C2 are dissimilar, and both the initial coordinate system C1 and the initial coordinate system C2 can be used as the final local coordinate system.

In some embodiments, referring to fig. 6, the steps of screening the plurality of initial coordinate systems according to the similarity data between the initial coordinate systems to obtain a plurality of local coordinate systems may include the following steps.

S610, classifying the plurality of initial coordinate systems based on similar data among the initial coordinate systems to obtain a plurality of coordinate system clusters.

The method comprises the steps of aiming at a first initial coordinate system and a second initial coordinate system in a coordinate system cluster, wherein the included angle between three coordinate axes in the first initial coordinate system and corresponding three coordinate axes in the second initial coordinate system is smaller than a second angle threshold value respectively. The similar data may be determined according to angles between three coordinate axes of any initial coordinate system and three coordinate axes corresponding to any other initial coordinate system.

It will be appreciated that the initial coordinate system is constructed based on three mutually perpendicular direction vectors, the three coordinate axes of which coincide with the three direction vectors. Therefore, the calculation of the included angle between the three coordinate axes in the first initial coordinate system and the corresponding three coordinate axes in the second initial coordinate system can be realized by calculating the included angle between the three direction vectors in the first initial coordinate system and the corresponding three direction vectors in the second initial coordinate system.

Illustratively, a second angle threshold is set, and a classification rule for classifying the initial coordinates may be set according to the second angle threshold:

according to the constructed multiple initial coordinate systems, selecting one initial coordinate system as a reference coordinate system, taking the reference coordinate system as a central element of a coordinate system cluster, traversing in the rest initial coordinate systems, and respectively calculating included angle angles between three direction vectors corresponding to the rest initial coordinate systems and the three direction vectors of the reference coordinate system. According to the classification rule, an initial coordinate system corresponding to the included angle which meets the condition that the included angle among the three direction vectors is smaller than the second angle threshold value can be put into a coordinate system cluster where the reference coordinate system is located, and an initial coordinate system corresponding to the included angle which does not meet the condition that the included angle among the three direction vectors is smaller than the second angle threshold value is taken as the rest initial coordinate system to participate in the next traversal process. And when the initial coordinate system corresponding to the included angle which does not meet the condition that the included angles among the three direction vectors are smaller than the second angle threshold value in the rest initial coordinate systems is not available, ending the traversing process and starting the next traversing. The traversal process is repeated until the initial coordinate systems are classified into the corresponding coordinate system clusters.

It should be noted that the second angle threshold may be set according to the accuracy requirement of calculating the bounding box. In some embodiments, the second angle threshold may be valued in the range of 1 ° to 10 °.

S620, reserving a designated coordinate system in the coordinate system cluster as a local coordinate system.

The designated coordinate system may be a coordinate system obtained from a cluster of coordinate systems and used for calculating and obtaining a bounding box, and may represent similar spatial features of an initial coordinate system in the corresponding cluster of coordinate systems.

Specifically, the specified coordinate systems can be respectively taken out from the plurality of finally obtained coordinate system clusters, and the specified coordinate systems participate in the subsequent process of calculating the bounding box.

In some embodiments, during the process of traversing and classifying the initial coordinate system, the reference coordinate system selected by each traversal may be reserved as a designated coordinate system, and the initial coordinate systems except the reference coordinate system in the coordinate system cluster are removed. And summarizing all the reference coordinate systems to obtain a plurality of local coordinate systems subjected to similarity screening.

Further, bounding boxes corresponding to the multiple local coordinate systems can be obtained through a bounding box algorithm based on the multiple screened local coordinate systems respectively to serve as multiple candidate bounding boxes, and the length, width and height of each candidate bounding box can be determined and used for calculating the volume of the corresponding candidate bounding box. The candidate bounding box with the smallest volume may be selected from the plurality of candidate bounding boxes as the smallest volume bounding box of the three-dimensional object to be finally output.

In some embodiments, in order to improve efficiency of acquiring bounding boxes, a transformation relationship between a local coordinate system and a world coordinate system may be utilized to transform a three-dimensional object under the local coordinate system and the local coordinate system together into the world coordinate system, and a candidate bounding box corresponding to the transformed local coordinate system is acquired through a bounding box algorithm. According to the method, the plurality of local coordinate systems obtained after screening are traversed in sequence, candidate bounding boxes corresponding to the converted local coordinate systems can be obtained respectively, and the volumes of the corresponding candidate bounding boxes are calculated. And selecting a candidate bounding box with the smallest volume from the plurality of candidate bounding boxes, and converting the bounding box back to the corresponding local coordinate system according to the transformation relation between the local coordinate system corresponding to the bounding box and the world coordinate system to be used as the smallest volume bounding box of the finally output three-dimensional object. In some embodiments, the bounding box algorithm may be an axially parallel bounding box algorithm.

In the above embodiment, by performing similarity screening on the initial coordinate system, the number of local coordinate systems for subsequently acquiring the candidate bounding boxes is greatly reduced, the repeated calculation amount of the similar bounding boxes can be avoided, the number of the candidate bounding boxes is reduced, and the overall efficiency of the bounding box determining method is greatly improved.

In some embodiments, the pair of edges includes a first edge and a second edge. Referring to fig. 7a, the establishment of the objective equation by a plurality of edge pairs based on the normal mutually perpendicular characteristics of the adjacent two faces of the bounding box may include the following steps.

S710, constructing a first normal vector to be solved of the first surface based on normal vectors of two triangular patches sharing the first side line.

Specifically, the two triangular panels sharing the first edge are triangular panels F ₁₀ And triangular dough sheet F ₁₁ Triangular dough sheet F ₁₀ Corresponding to normal vector N ₀ Triangular dough sheet F ₁₁ Corresponding to normal vector N ₁ . By means of normal vector N ₀ And normal vector N ₁ A first vector N of normals to be solved for the first surface is constructed. Illustratively, n=a ₁ N ₀ +b ₁ N ₁ Wherein a is ₁ And b ₁ Is a weight factor.

S720, constructing a second normal vector to be solved of the second surface based on normal vectors of two triangular patches sharing the second side line.

Specifically, the two triangular panels sharing the first edge are triangular panels F ₂₀ And triangular dough sheet F ₂₁ Triangular dough sheet F ₂₀ Corresponding to normal vector M ₀ Triangular dough sheet F ₂₁ Corresponding to normal vector M ₁ . Using normal vector M ₀ And normal vector M ₁ And constructing a second normal vector M to be solved of the second surface. Illustratively, m=a ₂ M ₀ +b ₂ M ₁ Wherein a is ₂ And b ₂ Is a weight factor.

And S730, establishing a target equation according to the mutually perpendicular characteristics of normals of two adjacent surfaces of the bounding box, the first normal vector to be solved and the second normal vector to be solved.

Specifically, according to the characteristic that the normals of two adjacent surfaces of the bounding box are perpendicular to each other, it is known that the product of the normal vector of the first surface and the normal vector of the second surface of the bounding box is 0, and therefore, the target equation f=n·m=0 can be established by using the first normal vector to be solved N and the second normal vector to be solved M.

In the embodiment, the objective equation is established by combining the inherent geometric characteristics of the bounding box with the minimum volume, and the objective equation is solved to obtain the local coordinate system of the candidate bounding box, so that the accuracy and precision of the finally obtained bounding box can be ensured to be higher, and the theoretical minimum can be approximately achieved.

In some embodiments, referring to fig. 7b, constructing a first to-be-solved normal vector of a first face based on normal vectors of two triangular patches sharing a first edge may include the steps of:

s712, determining a first normal vector pair based on normal vectors of two triangular patches sharing a first side line; and constructing a first normal vector to be solved of the first surface by using the first normal vector pair.

Wherein the first face corresponds to a first weight factor, specifically, a first normal vector to be solved may be constructed using the first normal vector pair and the first weight factor. The first weight factor may be used to represent an angle of the first normal vector to be solved within an included angle range of the first normal vector pair.

Specifically, the two triangular panels sharing the first edge are triangular panels F ₁₀ And triangular dough sheet F ₁₁ Triangular dough sheet F ₁₀ Corresponding to normal vector N ₀ Triangular dough sheet F ₁₁ Corresponding to normal vector N ₁ . By means of normal vector N ₀ And normal vector N ₁ A first pair of normal vectors is constructed. Based on the geometric characteristics of the first normal vector to be solved N, a weight factor s may be set for the first normal vector pair. The constructed first normal vector N to be solved can be expressed linearly as: n= (1-s) N ₀ +sN ₁ Wherein the weight factor s E [0,1 ]]。

Constructing a second to-be-solved normal vector of the second face based on normal vectors of two triangular patches sharing the second edge may include the steps of:

s722, determining a second normal vector pair based on normal vectors of two triangular patches sharing a second side line; and constructing a second normal vector to be solved of the second surface by using the second normal vector pair.

Wherein the second face corresponds to a second weight factor. Specifically, a second normal vector to be solved for the second surface may be constructed using the second normal vector pair and the second weight factor. The second weight factor can be used to represent the angle of the second normal vector to be solved within the range of the included angles of the second normal vector pair

Specifically, the two triangular panels sharing the second edge line are triangular panels F ₂₀ And triangular dough sheet F ₂₁ Triangular dough sheet F ₂₀ Corresponding to normal vector M ₀ Triangular dough sheet F ₂₁ Corresponding to normal vector M ₁ . Using normal vector M ₁ And normal vector M ₂ A second pair of normal vectors is constructed. Based on the geometric characteristics of the second normal vector M to be solved, a weight factor t may be set for the second normal vector pair. The constructed second to-be-solved normal vector M can be expressed linearly as: m= (1-t) M ₀ +tM ₁ Wherein the weight factor t E [0,1 ]]。

Further, according to the product of the first to-be-solved normal vector and the second to-be-solved normal vector being 0, a target equation can be established as follows:

F(s,t)＝N·M

＝(1-s)(1-t)N ₀ ·M ₀ +s(1-t)N ₁ ·M ₀ +(1-s)tN ₀ ·M ₁ +stN ₁ ·M ₁

＝(1-s)(1-t)f ₀₀ +s(1-t)f ₁₀ +(1-s)tf ₀₁ +stf ₁₁

＝0

value range [0,1 ] based on s and t]According to f ₀₀ 、f ₁₀ 、f ₀₁ 、f ₁₁ The values of (2) are positive number, negative number or zero, and the like, and the target equation can be discussed and solved according to 81 conditions.

Illustratively, when f ₀₀ 、f ₁₀ 、f ₀₁ 、f ₁₁ The target equation F (s, t) =0 is constant, so s and t are in the value interval [0,1]All values in the solution are solutions of the target equation, and a partial solution set can be obtained from the corresponding solution set according to the sampling number to serve as a candidate solution set.

Illustratively, when f ₀₀ Is a negative number, f ₁₀ Is zero, f ₀₁ Is positive in number, f ₁₁ When zero, the objective equation is F (s, t) = (1-s) (1-t) F ₀₀ +(1-s)tf ₀₁ ＝(1-s)(f ₀₀ -tf ₀₀ +tf ₀₁ ). When s=1, the target equation F (s, t) =0 is constant, where t may be the value interval [0,1]Any value within; when (when)

When the target equation F (s, t) =0 is constant, s can be the value interval [0,1]Any value within. Thus (S)>

The values of the two line segments corresponding to s=1 are solutions of the target equation, and a partial solution set can be obtained from the corresponding solution set according to the sampling number to serve as a candidate solution set.

Illustratively, when f ₀₀ Is zero, f ₁₀ Is positive in number, f ₀₁ Is a negative number, f ₁₁ When zero, the objective equation is F (s, t) =s (1-t) F ₁₀ +(1-s)tf ₀₁ . When s=0, t=0, the target equation F (s, t) =0 is constant; when s=1, t=1, the objective equation F (s, t) =0 is constant. Because ofThe values of s and t on the line segment formed by taking (0, 0) and (1, 1) as endpoints meet the condition that the value of the target equation is zero, any value on the line segment is the solution of the target equation, and a partial solution set can be obtained from the corresponding solution set according to the sampling number to be used as a candidate solution set.

Illustratively, when f ₀₀ Is a negative number, f ₁₀ Is positive in number, f ₀₁ Is zero, f ₁₁ When positive, the target equation is F (s, t) = (1-s) (1-t) F ₀₀ +s(1-t)f ₁₀ +stf ₁₁ . When s=0, t=1, the objective equation F (s, t) =0 is constant; when (when)

When t=0, the target equation F (s, t) =0 is constant. Thus, with (0, 1) and +.>

The values of s and t on a line segment formed by the endpoints meet the condition that the value of the target equation is zero, any value on the line segment is a solution of the target equation, and a partial solution set can be obtained from the corresponding solution set according to the sampling number to serve as a candidate solution set.

It should be noted that, the objective equation solving process in the remaining 77 cases is similar to the above process, and detailed descriptions thereof are omitted here.

Solving the target equation, after obtaining a candidate solution set according to the sampling number, substituting the corresponding values of (s, t) in the candidate solution set into the linear expression of the first normal vector N to be solved and the second normal vector M to be solved, so as to obtain a plurality of solution set vector pairs consisting of the first normal vector N to be solved and the second normal vector M to be solved, which are included in the solution set of the target equation.

In some embodiments, the first face corresponds to a first weight factor and the second face corresponds to a second weight factor. Constructing a first to-be-solved normal vector of the first surface by using the first normal vector pair may include: and constructing a first normal vector to be solved by using the first normal vector pair and the first weight factor.

Constructing a second to-be-solved normal vector of the second surface using the second normal vector pair may include: and constructing a second normal vector to be solved of the second surface by using the second normal vector pair and the second weight factor.

In some embodiments, in order to facilitate obtaining a corresponding local coordinate system according to the first to-be-solved normal vector and the second to-be-solved normal vector obtained by solving, the first to-be-solved normal vector and the second to-be-solved normal vector may be unit vectors. Specifically, using the first normal vector pair and the first weight factor, the constructed first normal vector N to be solved may be expressed as:

s∈[0,1]the method comprises the steps of carrying out a first treatment on the surface of the The constructed second normal vector M to be solved can be expressed as: />

t∈[0,1]。

In some embodiments, the solution set of the target equation is determined at a first weight value range of the first weight factor, at a second weight value range of the second weight factor.

Specifically, according to the objective equation, the problem of acquiring the minimum volume bounding box can be converted into a problem of finding a solution set satisfying the objective equation F (s, t) =0, in which all possible pairs of values of the weighting factors (s, t) are included. Since the first surface of the bounding box has a limit that at most coincides with one of the two triangular patches sharing the first edge, and the second surface has a limit that at most coincides with one of the two triangular patches sharing the second edge, the normal vector of the first surface is within the angle range of the first normal vector pair, and the normal vector of the second surface is within the angle range of the second normal vector pair. Because the weight factors can be used for representing angles of the normal vector to be solved within the included angle range of the corresponding normal vector pair, the corresponding first weight factors have a first weight value range, and the second weight factors have a second weight value range. It will be appreciated that the solution set of the target equation needs to be determined in a first weight range of the first weight factor and in a second weight range of the second weight factor.

The present embodiment provides a three-dimensional object bounding box determination method, which may include the following steps, with reference to fig. 8.

S810, acquiring a triangularization result of a three-dimensional convex hull of the three-dimensional object. Wherein the triangularization result comprises a plurality of triangular patches; the common edge of two adjacent triangular surface patches is a convex hull edge line of the three-dimensional convex hull.

S820, classifying the convex hull edge lines according to similar data among the convex hull edge lines to obtain a plurality of edge line clusters. Wherein the edge cluster corresponds to a designated edge.

S830, combining the specified edges to obtain a plurality of edge pairs.

S840, establishing a target equation through a plurality of edge line pairs based on the characteristic that normals of two adjacent surfaces of the bounding box are perpendicular to each other.

S850, constructing a plurality of initial coordinate systems to be screened according to a plurality of solution set vector pairs included in the solution set of the target equation. The solution vector pair comprises a first vector and a second vector, wherein the first vector is a normal vector of a first surface of two adjacent surfaces of the bounding box, and the second vector is a normal vector of a second surface of the two adjacent surfaces of the bounding box.

S860, screening the plurality of initial coordinate systems according to the similar data among the initial coordinate systems to obtain a plurality of local coordinate systems.

S870, determining a candidate bounding box, of which the volume satisfies the bounding box screening condition, as a target bounding box of the three-dimensional object, from among the plurality of candidate bounding boxes corresponding to the plurality of local coordinate systems.

By way of example, taking a point cloud of a three-dimensional object as input, firstly, a three-dimensional convex hull of the point cloud can be obtained through a three-dimensional convex hull calculation algorithm of a third party library Qhull, and secondly, a vertex point set of the three-dimensional convex hull can be taken as input, and a triangulation result of the vertex of the three-dimensional convex hull can be obtained through a triangulation algorithm of the third party library Qhull.

And calculating normal vectors of all triangular patches according to the obtained triangulation result, and calculating cross vectors of two normal vectors corresponding to two adjacent triangular patches sharing one convex hull edge line. According to the two normal vectors corresponding to the two adjacent triangular patches and the corresponding cross vector, the representation vector of the convex hull edge line can be calculated. Setting a first angle threshold, selecting a reference edge line from convex hull edge lines, traversing the residual convex hull edge lines, judging the similarity between the residual convex hull edge lines and the reference edge lines according to the included angle between the representing vector of the residual convex hull edge lines and the representing vector of the reference edge line, and placing the convex hull edge line corresponding to the included angle smaller than the first angle threshold into an edge line cluster where the reference edge line is located. And repeating the traversal process to finally obtain a plurality of classified edge clusters. Summarizing the reference edge lines in each edge line cluster as appointed edge lines to obtain convex hull edge lines subjected to similarity screening. And combining the convex hull edge lines obtained after screening in pairs to obtain a plurality of edge line pairs.

According to the geometric characteristic that two adjacent faces of the bounding box with the minimum volume corresponding to the three-dimensional object are respectively overlapped with two convex hull edge lines of the three-dimensional convex hull corresponding to the three-dimensional object, assuming that the two convex hull edge lines in each edge line pair are respectively overlapped with two adjacent faces of the bounding box, the normal vector of the two adjacent faces of the bounding box can be respectively represented in a linear mode through the normal vector and the weight factor of the adjacent triangular patches corresponding to the convex hull edge lines in the edge line pairs. According to the fact that the product of normal vectors of two adjacent surfaces of the bounding box is 0, a corresponding target equation can be established and obtained, and the problem of obtaining the minimum volume bounding box is converted into the problem of searching a solution set meeting the condition that the target equation is equal to 0.

Solving the target equation can obtain a value set of weight factors for representing normal vectors of two adjacent faces of the bounding box, so that a solution set vector pair of the normal vectors of the two adjacent faces of the bounding box can be obtained. And respectively carrying out cross multiplication calculation on the solution set vector pairs to obtain corresponding cross multiplication result vectors of the solution set vector pairs. According to each solution vector pair and the corresponding cross product vector, a direction vector combination formed by three mutually perpendicular vectors can be respectively obtained. Based on the combination of the plurality of direction vectors, a corresponding plurality of initial coordinate systems can be constructed.

Setting a second angle threshold, selecting an initial coordinate system as a reference coordinate system according to the constructed multiple initial coordinate systems, traversing the rest initial coordinate systems by taking the reference coordinate system as a central element of a coordinate system cluster, and respectively calculating included angle angles between three direction vectors corresponding to the rest initial coordinate systems and three direction vectors of the reference coordinate system. And placing the initial coordinate system corresponding to the included angle smaller than the second angle threshold value into the coordinate system cluster where the reference coordinate system is located. And repeating the traversal process to finally obtain a plurality of classified coordinate system clusters. Summarizing the reference coordinate system in each coordinate system cluster as a designated coordinate system to obtain a plurality of local coordinate systems subjected to similarity screening.

And acquiring bounding boxes corresponding to the multiple local coordinate systems through a bounding box algorithm based on the multiple screened local coordinate systems respectively to serve as multiple candidate bounding boxes, and determining the length, width and height of each candidate bounding box to calculate the volume of the corresponding candidate bounding box. The candidate bounding box having the smallest volume may be selected from the plurality of candidate bounding boxes as a target bounding box of the finally output three-dimensional object.

The present embodiment provides a three-dimensional object bounding box determination apparatus, referring to fig. 9, the three-dimensional object bounding box determination apparatus 900 may include: the triangularization result acquisition module 910, the classification processing module 920, the combining module 930, and the determining module 940.

The triangularization result obtaining module 910 is configured to obtain a triangularization result of the three-dimensional convex hull of the three-dimensional object. Wherein the triangularization result comprises a plurality of triangular patches; the common edge of two adjacent triangular surface patches is a convex hull edge line of the three-dimensional convex hull.

The classification processing module 920 is configured to perform classification processing on the plurality of convex hull edge lines according to similar data between the convex hull edge lines, so as to obtain a plurality of edge line clusters. Wherein the edge cluster corresponds to a designated edge.

And the combining module 930 is configured to combine the plurality of designated edges to obtain a plurality of edge pairs.

The determining module 940 is configured to determine, as a target bounding box of the three-dimensional object, a candidate bounding box whose volume satisfies a bounding box screening condition, from among a plurality of candidate bounding boxes corresponding to the plurality of local coordinate systems determined by the plurality of edge pairs.

For the specific definition of the three-dimensional object bounding box determining apparatus, reference may be made to the definition of the three-dimensional object bounding box determining method hereinabove, and the description thereof will not be repeated. The respective modules in the above-described three-dimensional object bounding box determination apparatus may be implemented in whole or in part by software, hardware, or a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

The present description further provides a computer device, referring to fig. 10, where the computer device 1000 includes a memory 1010, a processor 1020, and a computer program 1030 stored in the memory 1010 and executable on the processor 1020, and when the processor 1020 executes the computer program 1030, the method for determining a three-dimensional object bounding box according to any one of the foregoing embodiments is implemented.

The present specification embodiment also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the three-dimensional object bounding box determination method of any of the preceding embodiments.

It should be noted that the logic and/or steps represented in the flowcharts or otherwise described herein, for example, may be considered as a ordered listing of executable instructions for implementing logical functions, and may be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer readable medium may even be paper or other suitable medium on which the program is printed, as the program may be electronically captured, via, for instance, optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.

It is to be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, may be implemented using any one or combination of the following techniques, as is well known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application specific integrated circuits having suitable combinational logic gates, programmable Gate Arrays (PGAs), field Programmable Gate Arrays (FPGAs), and the like.

In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In the description of the present invention, the meaning of "plurality" means at least two, for example, two, three, etc., unless specifically defined otherwise.

In the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," "secured," and the like are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; either directly or indirectly, through intermediaries, or both, may be in communication with each other or in interaction with each other, unless expressly defined otherwise. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.

While embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the invention, and that variations, modifications, alternatives and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the invention.

Claims

1. A method of determining a three-dimensional object bounding box, the method comprising:

acquiring a triangularization result of a three-dimensional convex hull of the three-dimensional object; wherein the triangularization result comprises a plurality of triangular patches; the common edge of two adjacent triangular patches is a convex hull edge line of the three-dimensional convex hull;

classifying the convex hull edge lines according to similar data among the convex hull edge lines to obtain a plurality of edge line clusters; wherein the edge line cluster corresponds to a designated edge line;

combining the appointed side lines to obtain a plurality of side line pairs;

and determining a candidate bounding box with the volume meeting the bounding box screening condition as a target bounding box of the three-dimensional object in a plurality of candidate bounding boxes corresponding to the plurality of local coordinate systems determined by the side line pairs.

2. The method of claim 1, wherein the similarity data between the convex hull edge lines is determined by the angles between the representative vectors of the convex hull edge lines; the two adjacent triangular patches comprise a first triangular patch and a second triangular patch; the generation mode of the expression vector of the convex hull edge comprises the following steps:

determining a cross vector of the first normal vector and the second normal vector by using a first normal vector of the first triangular patch and a second normal vector of the second triangular patch;

And generating a representation vector of the convex hull edge according to the first normal vector, the second normal vector and the cross multiplication vector.

3. The method according to claim 1, wherein the classifying the plurality of convex hull edge lines according to the similar data between the convex hull edge lines to obtain a plurality of edge line clusters includes:

selecting one convex hull edge line from the convex hull edge lines as a reference edge line;

traversing the convex hull edge lines to determine an included angle between the representative vector of each convex hull edge line and the representative vector of the reference edge line;

determining convex hull edge lines corresponding to included angles smaller than a first angle threshold, wherein the reference edge lines belong to the same edge line cluster.

4. A method according to claim 3, wherein the designated edge uses the reference edge to represent convex hull edges comprised by the edge clusters; combining the specified edges to obtain a plurality of edge pairs, including:

and combining the reference edges in pairs to obtain a plurality of edge pairs.

5. The method of claim 1, wherein the determining the plurality of local coordinate systems comprises:

Establishing a target equation through a plurality of edge line pairs based on the characteristic that normals of two adjacent surfaces of the bounding box are perpendicular to each other;

determining a plurality of local coordinate systems according to a plurality of solution set vector pairs included in the solution set of the target equation; the solution set vector pair comprises a first vector and a second vector, wherein the first vector is a normal vector of a first surface of two adjacent surfaces of the bounding box, and the second vector is a normal vector of a second surface of the two adjacent surfaces of the bounding box.

6. The method of claim 5, wherein determining a plurality of local coordinate systems from a plurality of solution set vector pairs included in a solution set of the target equation comprises:

constructing a plurality of initial coordinate systems to be screened according to a plurality of solution set vector pairs included in the solution set;

and screening a plurality of initial coordinate systems according to the similar data among the initial coordinate systems to obtain a plurality of local coordinate systems.

7. The method of claim 6, wherein the filtering the plurality of initial coordinate systems according to the similarity data between the initial coordinate systems to obtain the plurality of local coordinate systems comprises:

Classifying the plurality of initial coordinate systems based on similar data among the initial coordinate systems to obtain a plurality of coordinate system clusters; the method comprises the steps of aiming at a first initial coordinate system and a second initial coordinate system in the coordinate system cluster, wherein the included angle between three coordinate axes in the first initial coordinate system and corresponding three coordinate axes in the second initial coordinate system is smaller than a second angle threshold value respectively;

and reserving a specified coordinate system in the coordinate system cluster as the local coordinate system.

8. The method of claim 5, wherein the pair of edges comprises a first edge and a second edge; the establishing a target equation through a plurality of edge line pairs based on the normal mutually perpendicular characteristics of two adjacent surfaces of the bounding box comprises the following steps:

constructing a first normal vector to be solved of the first surface based on normal vectors of two triangular patches sharing the first side line;

constructing a second normal vector to be solved of the second surface based on normal vectors of two triangular patches sharing the second side line;

and establishing the target equation according to the normal mutually perpendicular characteristics of two adjacent surfaces of the bounding box, the first normal vector to be solved and the second normal vector to be solved.

9. The method of claim 8, wherein constructing a first to-be-solved normal vector for the first face based on normal vectors for two triangular patches sharing the first edge comprises:

determining a first pair of normal vectors based on normal vectors of two triangular patches sharing the first edge; constructing a first normal vector to be solved of the first surface by utilizing the first normal vector pair;

the constructing a second normal vector to be solved of the second surface based on normal vectors of two triangular patches sharing the second edge includes:

determining a second pair of normal vectors based on normal vectors of two triangular patches sharing the second edge; and constructing a second normal vector to be solved of the second surface by using the second normal vector pair.

10. The method of claim 9, wherein the first one corresponds to a first weight factor and the second one corresponds to a second weight factor; the constructing a first normal vector to be solved of the first surface by using the first normal vector pair includes:

constructing the first normal vector to be solved by using the first normal vector pair and the first weight factor;

The constructing a second normal vector to be solved of the second surface by using the second normal vector pair includes:

and constructing a second normal vector to be solved of the second surface by using the second normal vector pair and the second weight factor.

11. The method of claim 10, wherein the solution set of the target equation is determined at a first weight value range of the first weight factor and at a second weight value range of the second weight factor.

12. A three-dimensional object bounding box determination apparatus, the apparatus comprising:

the triangularization result acquisition module is used for acquiring the triangularization result of the three-dimensional convex hull of the three-dimensional object; wherein the triangularization result comprises a plurality of triangular patches; the common edge of two adjacent triangular patches is a convex hull edge line of the three-dimensional convex hull;

the classification processing module is used for classifying the convex hull edge lines according to the similar data among the convex hull edge lines to obtain a plurality of edge line clusters; wherein the edge line cluster corresponds to a designated edge line;

the combination module is used for combining the appointed side lines to obtain a plurality of side line pairs;

And the determining module is used for determining the candidate bounding box with the volume meeting the bounding box screening condition as the target bounding box of the three-dimensional object in a plurality of candidate bounding boxes corresponding to a plurality of local coordinate systems determined by a plurality of edge pairs.

13. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 11 when the computer program is executed.

14. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 11.