Pose matching method of planar object and electronic device

Abstract

The application provides a pose matching method of a planar object, comprising: acquiring a point cloud of the planar object; extracting an edge point cloud based on the point cloud of the planar object; calculating edge vectors of the edge point cloud, wherein the edge vectors comprise a principal normal vector, a tangent vector or an edge normal vector; constructing a point pair feature vector set of any two points in the edge point cloud according to the principal normal vector, the tangent vector or the edge normal vector; and matching the point pair feature vector in the edge point cloud with a reference point pair feature vector set to obtain the pose of the planar object. The technical scheme of the application matches based on the edge normal, solves the problem that the matching of the planar object cannot be realized based on the face matching or the principal normal matching in the prior art, has high reliability and strong stability, and greatly improves the recognition accuracy and recognition efficiency of the planar object.

Description

Technical Field

[0001] This invention relates generally to the field of machine vision technology, and more particularly to a method for matching the pose of a planar object, and an electronic device for matching the pose of a planar object. Background Technology

[0002] With the development of industrial intelligence, the use of robots to replace manual labor in handling objects (such as industrial parts, boxes, etc.) is becoming increasingly common. When robots operate, they typically need to grasp objects, move them from one location to another, such as grasping objects from a conveyor belt and placing them on a pallet or in a cage, or grasping objects from a pallet and placing them on a conveyor belt or other pallets as required. Figures 1a to 1c As shown, most current matching methods are based on surfaces. For objects with rich surface features and large differences in shape, the matching stability is good. However, for planar objects (not shown in the figure), such as simple cubes or planar plates, the original point-pair features lose their discriminative power for different objects due to the lack of shape variation on the surface, resulting in matching failure.

[0003] The content of the background section is merely the technology known to the inventor and does not necessarily represent the prior art in this field. Summary of the Invention

[0004] To address one or more of the problems existing in the prior art, the present invention provides a pose matching method for a planar object, the pose matching method comprising:

[0005] Obtain the point cloud of the planar object;

[0006] Extract edge point clouds based on the point cloud of the planar object;

[0007] Calculate the edge vectors of the edge point cloud, wherein the edge vectors include principal normal vectors, tangent vectors, or edge normal vectors;

[0008] Based on the principal normal vector, tangent vector, or edge normal vector, construct a set of feature vectors for any two points in the edge point cloud; and

[0009] The pose of the planar object is obtained by matching the feature vectors of the point pairs in the edge point cloud with the set of feature vectors of the reference point pairs.

[0010] According to one aspect of the present invention, the step of acquiring the point cloud of the planar object includes: acquiring the point cloud of the planar object using a 3D camera.

[0011] According to one aspect of the present invention, the step of extracting edge point clouds from point clouds based on planar objects includes:

[0012] Search for neighboring points within the neighborhood of each point in the point cloud of the planar object;

[0013] Whether a point is an edge point is determined by each point and its neighboring points in the neighborhood.

[0014] According to one aspect of the invention, the step of determining whether each point is an edge point by means of each point and its neighboring points in its neighborhood includes:

[0015] Construct a vector for each point and its neighboring points in its neighborhood;

[0016] Determine the maximum and minimum values ​​of the angles between the vectors of each point and its neighboring points and the reference vector; and

[0017] Determine whether each point is an edge point based on the relationship between the maximum and minimum values ​​of the included angle and the threshold.

[0018] The reference vector is located in the plane of the planar object.

[0019] According to one aspect of the present invention, the step of extracting edge point clouds from point clouds based on planar objects further includes:

[0020] The point cloud of the planar object is projected onto a preset plane to obtain a two-dimensional projected image;

[0021] Extract the contour edges of the two-dimensional projection image; and

[0022] The outline edge of the two-dimensional projection image is projected into three-dimensional space to obtain the edge point cloud of the planar object.

[0023] According to one aspect of the present invention, the step of calculating the edge vector of the edge point cloud includes: calculating the principal normal vector of the plane containing the edge point cloud using a principal component analysis algorithm; wherein the principal normal vector corresponds to the shortest axis of the principal component.

[0024] According to one aspect of the present invention, the step of calculating the edge vector of the edge point cloud further includes: calculating the tangent vector of the edge point cloud using a principal component analysis algorithm; wherein the tangent vector corresponds to the longest axis of the principal component.

[0025] According to one aspect of the present invention, the step of calculating the edge vector of the edge point cloud further includes: calculating the edge normal vector of the edge point cloud; wherein the edge normal vector points outward along the plane in which the edge point cloud is located.

[0026] According to one aspect of the invention, the step of calculating the edge normal vector of the edge point cloud includes:

[0027] Calculate the principal normal vector of the plane containing the edge point cloud;

[0028] Calculate the tangent vector of the edge point cloud;

[0029] Calculate the vector between the center point of the plane containing the edge point cloud and the points in the edge point cloud; and

[0030] The edge normal vector is determined based on the principal normal vector, the tangent vector, and the vector between the center point of the plane containing the edge point cloud and the points in the edge point cloud.

[0031] According to one aspect of the present invention, the step of constructing a set of point pair feature vectors for any two points in the edge point cloud based on the principal normal vector, tangent vector, or edge normal vector includes:

[0032] Construct a point pair vector from any two points in the edge point cloud;

[0033] Calculate the magnitude of the point-to-vector pair;

[0034] Determine the first angle between the principal normal vector, tangent vector, or edge normal vector of the first point among the arbitrary two points and the point-pair vector;

[0035] Determine the second angle between the principal normal vector, tangent vector, or edge normal vector of the second point among the arbitrary two points and the point-pair vector;

[0036] Determine the third angle between the principal normal vector, tangent vector, or edge normal vector of the first point and the edge vector of the second point; and

[0037] The set of feature vectors for the point pairs is constructed using the magnitude of the point-pair vectors, the first included angle, the second included angle, and the third included angle.

[0038] According to one aspect of the present invention, there exist multiple sets of reference point pairs feature vectors, each set of reference point pairs feature vectors having a corresponding pose, and the step of matching the set of point pairs feature vectors in the edge point cloud with the set of reference point pairs feature vectors includes:

[0039] The transformation relationship between the set of point pairs feature vectors in the edge point cloud and each set of reference point pairs feature vectors is determined, the set of reference point pairs feature vectors that matches the set of point pairs feature vectors is determined, and the pose of the planar object is determined based on the pose of the set of matching reference point pairs feature vectors.

[0040] This invention also relates to an electronic device for pose matching of a planar object, comprising:

[0041] A point cloud acquisition unit is configured to acquire the point cloud of the planar object;

[0042] An edge point cloud extraction unit, communicatively connected to the point cloud acquisition unit, is configured to extract edge point clouds based on the point cloud of the planar object; and

[0043] The processing unit is communicatively connected to the point cloud acquisition unit and the edge point cloud acquisition unit, and is configured to calculate the edge vectors of the edge point cloud, wherein the edge vectors include principal normal vectors, tangent vectors, or edge normal vectors; and construct point pair feature vectors of any two points in the edge point cloud based on the principal normal vectors, tangent vectors, or edge normal vectors; and match the point pair feature vectors in the edge point cloud with the reference point pair feature vectors to obtain the pose of the planar object.

[0044] The present invention also relates to a computer-readable storage medium including computer-executable instructions stored thereon, which, when executed by a processor, implement the pose matching method as described above.

[0045] The technical solution of this invention is based on edge normal matching, which solves the problem that the existing technology based on surface matching or principal normal matching cannot achieve the matching of planar objects. It has high reliability and strong stability, and greatly improves the accuracy and efficiency of planar object recognition. Attached Figure Description

[0046] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0047] Figure 1a A schematic diagram of an object with rich surface features and large variations in shape is shown;

[0048] Figure 1b and Figure 1c Schematic diagrams of existing surface matching and principal normal vector matching are shown respectively;

[0049] Figure 2 A flowchart of a pose matching method for a planar object according to an embodiment of the present invention is shown;

[0050] Figure 3 A schematic diagram of a point cloud of a planar object according to an embodiment of the present invention is shown;

[0051] Figure 4a The diagram shows the effect of directly extracting the contour edges of a point cloud from a point cloud of a planar object according to an embodiment of the present invention.

[0052] Figure 4b A flowchart illustrating the extraction of the contour edges of a point cloud according to a preferred embodiment of the present invention is shown;

[0053] Figure 4cA schematic diagram illustrating the determination of whether a point in a point cloud of a planar object is an edge point according to a preferred embodiment of the present invention is shown.

[0054] Figure 4d A rendering of a 3D point cloud contour obtained according to a preferred embodiment of the present invention is shown.

[0055] Figure 4e A flowchart illustrating a 3D point cloud contour obtained from a 2D image contour according to a preferred embodiment of the present invention is shown;

[0056] Figure 5a A schematic diagram illustrating the calculation of edge vectors in an edge point cloud according to a preferred embodiment of the present invention is shown;

[0057] Figure 5b A schematic diagram illustrating the calculation of the tangent vector according to a preferred embodiment of the present invention is shown;

[0058] Figure 6a A schematic diagram illustrating the construction of a set of feature vectors for any two points in an edge point cloud according to a preferred embodiment of the present invention is shown.

[0059] Figure 6b A schematic diagram illustrating the construction of a hash table according to a preferred embodiment of the present invention is shown; and

[0060] Figure 7 A schematic diagram of an electronic device for locating the pose of a planar object according to an embodiment of the present invention is shown. Detailed Implementation

[0061] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0062] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0063] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection, an electrical connection, or a connection that allows for communication; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0064] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0065] The following disclosure provides many different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the invention. Furthermore, reference numerals and / or letters may be repeated in different examples; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. In addition, examples of various specific processes and materials are provided in this invention, but those skilled in the art will recognize the application of other processes and / or the use of other materials.

[0066] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0067] This invention provides a pose matching method for planar objects. Since the normal directions of each side are different, the stability of the matching is greatly improved. The method is described in detail below.

[0068] Figure 2 A flowchart of a pose matching method 10 for a planar object according to an embodiment of the present invention is shown, as follows: Figure 2 As shown, the pose matching method 10 includes steps S11 to S15. The steps of the pose matching method 10 are described in detail below.

[0069] In step S11, the point cloud of the planar object is obtained. Figure 3 A schematic diagram of a point cloud of a planar object according to an embodiment of the present invention is shown. According to a preferred embodiment of the present invention, the point cloud of the planar object can be acquired using a 3D camera. Alternatively, other 3D scanning devices such as LiDAR (2D / 3D), RGB binoculars, 3D structured light cameras, and ToF cameras (time-of-flight cameras) can also be used to acquire the point cloud of the planar object. The present invention does not limit the specific method of acquiring the point cloud of the planar object. Furthermore, the present invention does not limit the specific type of the planar object, which includes, but is not limited to, objects such as cubes and planar panels.

[0070] In step S12, edge point cloud is extracted based on the point cloud of the planar object. It should be understood that the so-called extraction of edge point cloud based on the point cloud of the planar object actually means extracting the contour edges of the point cloud of the planar object.

[0071] According to one embodiment of the present invention, the contour edges of a point cloud can be directly extracted from the point cloud of a planar object. However, this method is slow and the edges are prone to having stray point clouds. Figure 4a This illustrates such a situation.

[0072] According to a preferred embodiment of the present invention, the contour edges of a point cloud can be extracted in the following manner. For example... Figure 4b As shown, in this embodiment, step S12 includes sub-steps S121 to S122. In sub-step S121, neighboring points within the neighborhood of each point in the point cloud of the planar object are searched. In sub-step S122, it is determined whether each point is an edge point based on each point and its neighboring points within its neighborhood. In some preferred embodiments, a vector can be constructed for each point and its neighboring points within its neighborhood; the maximum or minimum value of the angle between the vector of each point and its neighboring points and a reference vector can be determined; and the relationship between the maximum or minimum value of the angle and a threshold value can be used to determine whether each point is an edge point; wherein the reference vector lies within the plane of the planar object. Figure 4c A schematic diagram illustrating the determination of whether a point in the point cloud of a planar object is an edge point according to a preferred embodiment of the present invention is shown, such as... Figure 4c For example, taking point A in the point cloud of a planar object, the orthogonal vectors u and v of the point set in the neighborhood of point A can be obtained through the normal of point A. The neighborhood can be a region enclosed by a circle of a preset size. The neighboring points in the neighborhood are, for example, point B, point C, point D, point E, and point F. All the neighboring points in the neighborhood (e.g., points B, C, D, E, and F) are connected to point A (e.g., point A) to form vectors P1, P2, P3, P4, and P5. The angles between vectors P1, P2, P3, P4, and P5 and the reference vector are calculated and sorted. The maximum angle θ among the angles is determined, and the relationship between the maximum angle θ and a threshold is determined. If the maximum angle θ is greater than the threshold, the point is considered an edge point and is retained; otherwise, if the maximum angle θ is not greater than the threshold, the point is considered not an edge point and is discarded. By traversing all points in the point cloud of a planar object in this way, it can be determined whether each point is an edge point. The set of all edge points is the edge point cloud of the planar object. An image of the edge point cloud can be found for reference. Figure 4d By comparison Figure 4a and Figure 4d As can be seen, the point cloud contour edges extracted using this embodiment are clearer and have less noise. It should be noted that this embodiment is for illustrative purposes only and does not constitute a limitation of the invention. The neighborhood can be an area enclosed by a circle of a preset size, or by other shapes (e.g., rectangles, triangles, etc.) of a preset size. Furthermore, the invention does not limit the specific number of adjacent points within the neighborhood. In addition to determining whether a point is an edge point through the relationship between the maximum angle and a threshold, it can also be determined through the relationship between the minimum angle and a threshold, depending on the actual situation.

[0073] According to another preferred embodiment of the present invention, 3D point cloud contours can also be obtained from 2D image contours. It should be understood that the key to obtaining 3D point cloud contours from 2D image contours is that the 3D camera knows the transformation relationships from 3D points to 2D points and from 2D points to 3D points. For example... Figure 4e As shown, in this embodiment, step S12 includes sub-steps S125 to S127. In sub-step S125, the point cloud of the planar object is projected onto a preset plane to obtain a two-dimensional projected image. In specific implementations, for example, perspective projection can be used to project the 3D point cloud of the planar object onto a two-dimensional preset plane to obtain a 2D projected image of the 3D point cloud. In sub-step S126, the contour edges of the two-dimensional projected image are extracted. In sub-step S127, the contour edges of the two-dimensional projected image are projected into three-dimensional space to obtain the edge point cloud of the planar object. Specifically, for example, perspective projection can be used to project the contour edges of the two-dimensional projected image into three-dimensional space to obtain the edge point cloud. The effect diagram of the edge point cloud can be referred to... Figure 4d Obtaining 3D point cloud contours from 2D image contours is clearer and faster.

[0074] In step S13, the edge vectors of the edge point cloud are calculated, wherein the edge vectors include principal normal vectors, tangent vectors, or edge normal vectors.

[0075] According to a preferred embodiment of the present invention, the principal normal vector of the plane containing the edge point cloud can be calculated using the principal component analysis (PCA) algorithm; wherein the principal normal vector corresponds to the shortest axis of the principal component.

[0076] According to a preferred embodiment of the present invention, the tangent vector of the edge point cloud can be calculated using Principal Component Analysis (PCA); wherein the tangent vector corresponds to the longest axis of the principal component. Figure 5b A schematic diagram for calculating the tangent vector is shown.

[0077] According to a preferred embodiment of the present invention, the edge normal vector of the edge point cloud can be calculated by the cross product of the principal normal vector and the tangent vector, wherein the edge normal vector points outward along the plane containing the edge point cloud. Figure 5a As shown, the normal directions of each edge are different, which greatly improves the stability of planar object matching.

[0078] According to a preferred embodiment of the present invention, the step of calculating the edge normal vector of the edge point cloud includes: calculating the principal normal vector of the plane containing the edge point cloud; calculating the tangent vector of the edge point cloud; calculating the vector between the center point of the plane containing the edge point cloud and the points in the edge point cloud; and determining the edge normal vector based on the principal normal vector, the tangent vector, and the vector between the center point of the plane containing the edge point cloud and the points in the edge point cloud.

[0079] In step S14, a set of point pair feature vectors for any two points in the edge point cloud is constructed based on the principal normal vector, tangent vector, or edge normal vector.

[0080] Suppose there are two points m1 and m2, and their normal vectors (principal normal vector, tangent vector, or edge normal vector) are n1 and n2, respectively. A point-to-vector pair can be constructed using any two points in the edge point cloud, such as m1 and m2.<m1,m2> In some preferred embodiments, the modulus F1 of the point-pair vector can be calculated; the principal normal vector, tangent vector, or edge normal vector n1 of the first point (e.g., m1) can be determined in relation to the point-pair vector.<m1,m2> The first included angle F2; determine the principal normal vector, tangent vector, or edge normal vector n2 of the second point (e.g., m1, m2) and the point pair vector.<m1,m2> The second included angle F3; the third included angle F4 between the principal normal vector, tangent vector, or edge normal vector n1 of the first point (e.g., m1) and the edge vector n2 of the second point; the feature vector set of the point pair is constructed using the modulus F1, the first included angle F2, the second included angle F3, and the third included angle F4. That is, the point pair (m1, m2) has one distance and three angular features.<F1,F2,F3,F4> ,refer to Figure 6a .

[0081] In step S15, the feature vectors of point pairs in the edge point cloud are matched with the set of feature vectors of reference point pairs to obtain the pose of the planar object.

[0082] According to a preferred embodiment of the present invention, there are multiple sets of reference point pairs feature vectors, each set of reference point pairs feature vectors having a corresponding pose. By determining the transformation relationship between the set of point pairs feature vectors in the edge point cloud and each set of reference point pairs feature vectors, a set of reference point pairs feature vectors that matches the set of point pairs feature vectors can be determined. Based on the pose of the matched set of reference point pairs feature vectors, the pose of the planar object can be determined.

[0083] In some specific embodiments, a reference point pair (mr, mi) ∈ M with similar eigenvectors to the point pair (sr, si) ∈ S in the edge point cloud can be selected; mr is moved to the origin of the local coordinate system using the transformation matrix Tm → g, and the reference point cloud is rotated so that its normal axis coincides with the x-axis of the local coordinate system; similarly, the same operation can be performed on the edge point cloud using Ts → g; finally, a point mi in the reference point cloud is rotated Rx(α) around the x-axis and registered with si. The transformation from the reference point cloud to the edge point cloud can be defined as:

[0084]

[0085] In some preferred embodiments, a hash table can be constructed to store the feature vectors of point pairs in the edge point cloud with corresponding poses together with the feature vectors of reference point pairs. That is, the key of the hash table is the feature vector F, and the value is the set A of point pairs with the same feature vector. (Refer to...) Figure 6b .

[0086] In some preferred embodiments, for a reference point sr in the edge point cloud, a point pair (sr, si) is formed with all other points in the edge point cloud. For each point pair, Fs(sr, si) is calculated. Fs is used as the key of a hash table to search the hash table and find the feature vector Fm and point pair (mr, mi) of the reference point cloud that are similar to Fs. For each matching reference point pair (mr, mi), the rotation angle α can be calculated using the following formula.

[0087]

[0088] For α and discrete α in a two-dimensional array 1,2,..,n The corresponding position votes (+1); after all calculations are completed, we can obtain the mr and α corresponding to the maximum votes of the reference point sr, that is, the optimal local coordinates (mr, α).

[0089] The above provides a detailed description of the planar object pose matching method 10 of the present invention. The technical solution of the present invention is based on edge normal matching, which solves the problem that the existing technology based on surface matching or principal normal matching cannot achieve the matching of planar objects. It has high reliability and strong stability, and greatly improves the accuracy and efficiency of planar object recognition.

[0090] In addition, the present invention also relates to an electronic device 200 for pose matching of planar objects, such as... Figure 7 As shown, the pose matching electronic device 200 includes a point cloud acquisition unit 210, an edge point cloud extraction unit 220, and a processing unit 230. The point cloud acquisition unit 210 is configured to acquire the point cloud of the planar object. The edge point cloud extraction unit 220 is communicatively connected to the point cloud acquisition unit 210 and configured to extract edge point clouds based on the point cloud of the planar object. The processing unit 230 is communicatively connected to the point cloud acquisition unit 210 and the edge point cloud acquisition unit 220 and configured to calculate the edge vectors of the edge point cloud, wherein the edge vectors include principal normal vectors, tangent vectors, or edge normal vectors; and construct point pair feature vectors of any two points in the edge point cloud based on the principal normal vectors, tangent vectors, or edge normal vectors; and match the point pair feature vectors in the edge point cloud with the reference point pair feature vectors to obtain the pose of the planar object.

[0091] The present invention also relates to a computer-readable storage medium including computer-executable instructions stored thereon, which, when executed by a processor, implement the pose matching method 10 as described above.

[0092] In some preferred embodiments, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used or combined with an instruction execution system, apparatus, or device. The computer-readable storage medium can take any combination of one or more computer-readable media. The computer-readable storage medium can be, for example, but not limited to, electrical, magnetic, optical, or semiconductor forms or devices, and more specific examples (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer hard disk, a hard disk, random access memory (RAM), non-volatile random access memory (NVRAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0093] The processing unit can be a central processing unit (CPU), a microcontroller unit (MCU), or other general-purpose processors, digital signal processors (DSPs), or application-specific integrated circuits (ASICs).

[0094] The present invention does not limit the use of (Application Special Integrated Circuit, ASIC), off-the-shelf programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., but depends on the specific circumstances.

[0095] It should be noted that this specification provides the operational steps of the methods described in the embodiments or diagrams, but based on conventional or non-inventive labor, more or fewer operational steps may be included. The order of steps listed in the embodiments is merely one possible execution order among many steps and does not represent the only execution order. In actual system or device products, the methods shown in the embodiments or flowcharts can be executed sequentially or in parallel.

[0096] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for pose matching of a planar object, comprising: obtaining a point cloud of the planar object; extracting an edge point cloud based on the point cloud of the planar object; calculating edge vectors of the edge point cloud, wherein the edge vectors comprise a principal normal vector, a tangent vector or an edge normal vector; constructing a set of point pair feature vectors of any two points in the edge point cloud according to the principal normal vector, the tangent vector or the edge normal vector; and matching the point pair feature vectors in the edge point cloud with a set of reference point pair feature vectors to obtain a pose of the planar object; wherein the calculating of the edge vectors of the edge point cloud comprises: calculating a principal normal vector of a plane on which the edge point cloud is located by using a principal component analysis algorithm; the principal normal vector corresponds to a shortest axis of the principal component; calculating a tangent vector of the edge point cloud by using the principal component analysis algorithm; the tangent vector corresponds to a longest axis of the principal component; and calculating an edge normal vector of the edge point cloud, comprising: calculating a vector of a center point of the plane on which the edge point cloud is located and a point in the edge point cloud; determining the edge normal vector based on the principal normal vector, the tangent vector and the vector of the center point of the plane on which the edge point cloud is located and the point in the edge point cloud; the edge normal vector points to the outside along the plane on which the edge point cloud is located. The point cloud of the planar object is obtained by a 3D camera.

2. The pose matching method of claim 1, wherein the step of acquiring a point cloud of a planar object comprises: 3.The method for pose matching of a planar object according to claim 1, wherein the step of extracting an edge point cloud based on the point cloud of the planar object comprises: searching for neighboring points within a neighborhood of each point in the point cloud of the planar object; determining whether the each point is an edge point by the each point and the neighboring points within the neighborhood of the each point. 4.The method for pose matching of a planar object according to claim 3, wherein the step of determining whether the each point is an edge point by the each point and the neighboring points within the neighborhood of the each point comprises: constructing vectors of the each point and the neighboring points within the neighborhood of the each point; determining extremum values of angles between the vectors of the each point and the neighboring points and a reference vector; and determining whether the each point is an edge point based on a size relationship between the extremum values of the angles and a threshold value; wherein the reference vector is located in a plane of the planar object. 5.The method for pose matching of a planar object according to claim 1, wherein the step of extracting an edge point cloud based on the point cloud of the planar object further comprises: projecting the point cloud of the planar object to a preset plane to obtain a two-dimensional projection image; extracting a contour edge of the two-dimensional projection image; and projecting the contour edge of the two-dimensional projection image to a three-dimensional space to obtain the edge point cloud of the planar object. 6.The method for pose matching of a planar object according to any one of claims 1-5, wherein the step of constructing a set of point pair feature vectors of any two points in the edge point cloud according to the principal normal vector, the tangent vector or the edge normal vector comprises: constructing a point pair vector with the any two points in the edge point cloud; calculating a module of the point pair vector; determining a first angle between a principal normal vector, a tangent vector or an edge normal vector of a first point of the any two points and the point pair vector; determining a second angle between a principal normal vector, a tangent vector or an edge normal vector of a second point of the any two points and the point pair vector. ​ ​ ​ determining a third angle between the principal normal vector, tangent vector or edge normal vector of the first point and an edge vector of the second point; and constructing the point pair feature vector set by the modulus of the point pair vector, the first angle, the second angle and the third angle. 7.The pose matching method of claim 6, wherein there are a plurality of reference point pair feature vector sets, each of which has a corresponding pose, and the step of matching the point pair feature vector set in the edge point cloud with the reference point pair feature vector sets comprises: determining a transformation relationship between the point pair feature vector set in the edge point cloud and each of the reference point pair feature vector sets, determining a reference point pair feature vector set matched with the point pair feature vector set, and determining the pose of the planar object according to the pose of the matched reference point pair feature vector set. 8.A pose matching electronic device of a planar object capable of performing the pose matching method of any one of claims 1-7, comprising: a point cloud acquisition unit configured to acquire a point cloud of the planar object; an edge point cloud extraction unit communicatively connected to the point cloud acquisition unit and configured to extract an edge point cloud based on the point cloud of the planar object; and a processing unit communicatively connected to the point cloud acquisition unit and the edge point cloud acquisition unit and configured to calculate an edge vector of the edge point cloud, wherein the edge vector comprises a principal normal vector, tangent vector or edge normal vector, construct a point pair feature vector of any two points in the edge point cloud according to the principal normal vector, tangent vector or edge normal vector, and match the point pair feature vector in the edge point cloud with a reference point pair feature vector to obtain the pose of the planar object. 9.A computer readable storage medium comprising computer executable instructions stored thereon, which when executed by a processor implement the pose matching method of any one of claims 1-7.

Images