Workpiece identification method, system and equipment, medium and program product

By selecting points close to the seed point in point cloud segmentation and segmenting them according to the angle between the normal vectors, the problem of low accuracy in point cloud segmentation of traditional Euclidean clustering algorithm is solved, realizing efficient and accurate identification and detail capture of workpieces, which is suitable for workpiece identification in industrial scenarios.

CN119048829APending Publication Date: 2024-11-29BEIJING INST OF TECH
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411172741.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-26
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

Traditional Euclidean clustering segmentation algorithms suffer from low segmentation accuracy and insufficient algorithm stability in point cloud segmentation, making it difficult to accurately identify workpieces, especially under complex working conditions.

Method used

By acquiring raw point cloud data, seed points for the workpiece are determined and a point cloud queue is established. Points close to the seed points are filtered using the Euclidean distance threshold. The point cloud is segmented based on the angle between the normal vector of the workpiece to be inspected and the seed point. An empty cluster set is established to filter points that meet the angle threshold and remove non-compliant points, thereby achieving accurate identification of the workpiece.

Benefits of technology

It improves the efficiency and accuracy of point cloud data processing, ensures high-precision identification and detail capture of workpieces, enhances sensitivity to complex geometries and subtle features, and avoids increasing computational complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048829A_ABST
    Figure CN119048829A_ABST
Patent Text Reader

Abstract

The invention provides a workpiece identification method, system and device, a medium and a program product, and the method comprises the steps: obtaining original point cloud data which comprises the point cloud data of at least one workpiece; determining point cloud data serving as workpiece seed points from the original point cloud data; based on the original point cloud data, a point cloud queue corresponding to the workpiece seed points is established, the point cloud queue comprises the workpiece seed points and workpiece to-be-inspected points, and the Euclidean distance between the workpiece to-be-inspected points and the workpiece seed points is smaller than an Euclidean distance threshold value; determining a point cloud segmentation result of the workpiece according to an included angle between the normal vector of the to-be-detected point of the workpiece and the normal vector of the workpiece seed point; and identifying the workpiece based on the cloud segmentation result of the workpiece. The problem of low segmentation precision of a traditional European clustering algorithm is effectively improved, the workpiece point cloud in the industrial scene is accurately segmented to identify the workpiece, the efficiency of point cloud data processing is effectively improved, and high-precision identification and detail capture of the workpiece are also realized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present application relate to point cloud segmentation and its technical field, and in particular, to a workpiece identification method, system, device, medium and program product. Background Art

[0002] Robots with binocular vision often use template matching to identify the workpiece to be processed. This method requires segmenting the original point cloud data to obtain a separate target point cloud. The cluster segmentation algorithm divides point clouds with similar attributes into one category to achieve segmentation. Since it has a relatively high segmentation accuracy while having a low algorithm complexity, cluster segmentation has high universality.

[0003] However, the original point cloud is unstructured and dense near and sparse far. The traditional Euclidean clustering algorithm directly acting on the 3D point cloud has the problem of low segmentation accuracy. Accurately segmenting point cloud targets has always been a difficult problem in target detection.

[0004] Although there are a lot of research on the related technologies of Euclidean clustering segmentation algorithm, they often have the following shortcomings:

[0005] ⑴. The traditional Euclidean clustering segmentation algorithm is prone to under-segmentation and over-segmentation problems due to the difficulty in setting the distance threshold, and there is little room for improving the segmentation accuracy;

[0006] ⑵. The traditional Euclidean clustering segmentation algorithm is difficult to ensure the stability of the algorithm when dealing with complex working conditions, such as stacked workpieces. Summary of the invention

[0007] The embodiments of the present application provide a workpiece identification method, system, device, medium and program product to solve the problem that the existing segmentation method has low precision and leads to inaccurate workpiece identification.

[0008] In order to solve the above technical problems, this application is implemented as follows:

[0009] In a first aspect, an embodiment of the present application provides a workpiece identification method, comprising:

[0010] Acquire original point cloud data, wherein the original point cloud data includes point cloud data of at least one workpiece;

[0011] Determine point cloud data serving as a workpiece seed point from the original point cloud data;

[0012] Based on the original point cloud data, a point cloud queue corresponding to the workpiece seed point is established, wherein the point cloud queue includes the workpiece seed point and the workpiece to be inspected point, and the Euclidean distance between the workpiece to be inspected point and the workpiece seed point is less than a Euclidean distance threshold;

[0013] Determining a point cloud segmentation result of the workpiece according to an angle between a normal vector of the workpiece to be inspected point and a normal vector of a seed point of the workpiece;

[0014] Based on the cloud segmentation result of the artifact, the artifact is identified.

[0015] Optionally, determining the point cloud segmentation result of the workpiece according to the angle between the normal vector of the to-be-inspected point of the workpiece and the normal vector of the seed point of the workpiece includes:

[0016] Establishing an empty cluster set, wherein the empty cluster set is used to store points that meet the normal vector angle threshold;

[0017] If the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point is less than the normal vector angle threshold, the workpiece to be inspected point is placed in the empty cluster set;

[0018] If the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point is greater than or equal to the normal vector angle threshold, the workpiece to be inspected point is eliminated.

[0019] Optionally, establishing a point cloud queue corresponding to the workpiece seed point based on the original point cloud data includes:

[0020] Establishing a topological relationship between the original point cloud data;

[0021] Based on the topological relationship, traverse the Euclidean distance between the points in the original point cloud data that have not entered the point cloud queue and the workpiece seed point;

[0022] The point whose Euclidean distance is less than the Euclidean distance threshold is used as the workpiece to be inspected point;

[0023] Based on the workpiece seed point and the workpiece point to be inspected, a point cloud queue corresponding to the workpiece seed point is established.

[0024] Optionally, before determining the point cloud segmentation result according to the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point, the method includes:

[0025] Based on the points in the point cloud queue, establishing a spherical neighborhood set of the workpiece seed point;

[0026] Determine, based on the points in the spherical neighborhood set, a local plane in the least squares sense of the points in the spherical neighborhood set;

[0027] Based on the local plane, a normal vector of the workpiece to be inspected point and a normal vector of the workpiece seed point are determined.

[0028] Optionally, determining the normal vector of the workpiece to-be-inspected point and the normal vector of the workpiece seed point based on the local plane includes:

[0029] determining the centroid of the local plane;

[0030] Determining a covariance matrix based on the centroid of the local plane;

[0031] Determining the normal vector of the workpiece to be inspected point according to the covariance matrix;

[0032] Perform eigenvalue decomposition on the covariance matrix, and use the eigenvector corresponding to the minimum eigenvalue as the normal vector of the workpiece seed point.

[0033] Optionally, also include:

[0034] Based on the inner product of the normal vectors of two adjacent points in the point cloud queue, the normal vectors of the points in the point cloud queue are oriented, and the orientation includes:

[0035] Set the normal vector direction;

[0036] If the inner product of the normal vectors of two adjacent points is less than 0, the direction of the normal vector of the opposite point is rotated 180 degrees, wherein the opposite point is a point opposite to the direction of the normal vector.

[0037] In a second aspect, an embodiment of the present application provides a workpiece identification system, including:

[0038] An acquisition module, used for acquiring original point cloud data, wherein the original point cloud data includes point cloud data of at least one workpiece;

[0039] A determination module, used to determine point cloud data serving as a workpiece seed point from the original point cloud data;

[0040] A distance module is used to establish a point cloud queue corresponding to the workpiece seed point based on the original point cloud data, wherein the point cloud queue includes the workpiece seed point and the workpiece to be inspected point, and the Euclidean distance between the workpiece to be inspected point and the workpiece seed point is less than a Euclidean distance threshold;

[0041] An angle module, used to determine the point cloud segmentation result of the workpiece according to the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point;

[0042] The recognition module is used to recognize the artifact based on the cloud segmentation result of the artifact.

[0043] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, the steps of the workpiece identification method described in the first aspect are implemented.

[0044] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the workpiece identification method described in the first aspect above are implemented.

[0045] In a fifth aspect, a computer program product is provided, comprising computer instructions, which, when executed by a processor, implement the steps of the workpiece identification method as described in the first aspect.

[0046] In the present application, original point cloud data is obtained, and the original point cloud data includes point cloud data of at least one workpiece; point cloud data serving as a workpiece seed point is determined from the original point cloud data; based on the original point cloud data, a point cloud queue corresponding to the workpiece seed point is established, wherein the point cloud queue includes the workpiece seed point and a workpiece point to be inspected, and the Euclidean distance between the workpiece point to be inspected and the workpiece seed point is less than a Euclidean distance threshold; based on the angle between the normal vector of the workpiece point to be inspected and the normal vector of the workpiece seed point, a point cloud segmentation result of the workpiece is determined; and based on the cloud segmentation result of the workpiece, the workpiece is identified. The present application obtains the original point cloud data in sequence, determines the workpiece seed points, and establishes a point cloud queue including the seed points and the points to be inspected, uses the Euclidean distance constraint to screen the points close to the seed points, and then determines the point cloud segmentation results according to the angles between these points and the normal vectors of the seed points, so as to accurately identify the workpiece. This can effectively improve the problem of low segmentation accuracy in traditional Euclidean clustering algorithms, and achieve accurate segmentation of workpiece point clouds in industrial scenes to identify the workpieces. This not only effectively improves the efficiency of point cloud data processing, but also ensures high-precision recognition and detail capture of the workpiece, avoids increasing the computational complexity, and enhances the sensitivity to complex geometric shapes and subtle features, making the overall processing more reliable and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0048] Figure 1 is a flow chart of a workpiece identification method provided in an embodiment of the present application;

[0049] Figure 2 is a flow chart of another workpiece identification method provided in an embodiment of the present application;

[0050] Figure 3 It is a schematic diagram of the whole process of a workpiece identification method provided in an embodiment of the present application;

[0051] Figure 4 It is a structural schematic diagram of a workpiece identification system provided in an embodiment of the present application;

[0052] Figure 5 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0054] Please refer to Figure 1 , the present application embodiment provides a workpiece identification method, comprising:

[0055] Step 11: Acquire original point cloud data, wherein the original point cloud data includes point cloud data of at least one workpiece;

[0056] Optionally, point cloud data represents the surface of an object or the structure of an environment through a series of discrete points in a three-dimensional space, and these points usually include three-dimensional coordinates, and sometimes also include additional information such as color, normal direction, intensity value, etc. In this application, the original point cloud data including the point cloud data of the workpiece can be collected by a depth camera. In addition to the point cloud data of the workpiece, the original point cloud data may also include point cloud data of other unrelated items in the environment, and may include multiple workpieces, so it is necessary to identify the workpiece through the following embodiments in this application.

[0057] Step 12: determining point cloud data serving as a workpiece seed point from the original point cloud data;

[0058] Optionally, the workpiece seed point can be selected according to actual needs, and this application does not impose any restrictions. Specifically, the topological relationship between the scattered point clouds can be established through the data structure of a k-dimensional tree (kd-tree), and then the seed point p can be selected.

[0059] Step 13: Based on the original point cloud data, a point cloud queue corresponding to the workpiece seed point is established, wherein the point cloud queue includes the workpiece seed point and the workpiece to be inspected point, and the Euclidean distance between the workpiece to be inspected point and the workpiece seed point is less than a Euclidean distance threshold;

[0060] Optionally, the number of seed points can be 1 or more, depending on whether all the original point cloud data are classified into a corresponding point cloud queue. That is to say, when clustering, all the original point cloud data can be classified one by one, and new seed points can be selected for unclassified points. The above steps 12 and 13 are repeated until all points are classified into a point cloud queue and the clustering process ends. Optionally, the next unclassified point after the classification of the previous point cloud queue can be directly used as a new seed point.

[0061] Step 14: determining a point cloud segmentation result of the workpiece according to an angle between a normal vector of the workpiece to be inspected point and a normal vector of a seed point of the workpiece;

[0062] Step 15: Identify the artifact based on the cloud segmentation result of the artifact.

[0063] In the present application, original point cloud data is obtained, and the original point cloud data includes point cloud data of at least one workpiece; point cloud data serving as a workpiece seed point is determined from the original point cloud data; based on the original point cloud data, a point cloud queue corresponding to the workpiece seed point is established, wherein the point cloud queue includes the workpiece seed point and a workpiece point to be inspected, and the Euclidean distance between the workpiece point to be inspected and the workpiece seed point is less than a Euclidean distance threshold; based on the angle between the normal vector of the workpiece point to be inspected and the normal vector of the workpiece seed point, a point cloud segmentation result of the workpiece is determined; and based on the cloud segmentation result of the workpiece, the workpiece is identified. The present application obtains the original point cloud data in sequence, determines the workpiece seed points, and establishes a point cloud queue including the seed points and the points to be inspected, uses the Euclidean distance constraint to screen the points close to the seed points, and then determines the point cloud segmentation results according to the angles between these points and the normal vectors of the seed points, so as to accurately identify the workpiece. This can effectively improve the problem of low segmentation accuracy in traditional Euclidean clustering algorithms, and achieve accurate segmentation of workpiece point clouds in industrial scenes to identify the workpieces. This not only effectively improves the efficiency of point cloud data processing, but also ensures high-precision recognition and detail capture of the workpiece, avoids increasing the computational complexity, and enhances the sensitivity to complex geometric shapes and subtle features, making the overall processing more reliable and accurate.

[0064] Optionally, before determining the point cloud data of the workpiece seed point, the original point cloud data may be preprocessed, and downsampling may be achieved using a voxelized grid method to reduce the amount of point cloud data, reduce subsequent calculations, and speed up processing.

[0065] Specifically, the filtering operators in the Point Cloud Library (PCL) can be combined to reduce outliers and noise points in the data and streamline the data volume.

[0066] Optionally, establishing a point cloud queue corresponding to the workpiece seed point based on the original point cloud data includes:

[0067] Establishing a topological relationship between the original point cloud data;

[0068] Based on the topological relationship, traverse the Euclidean distance between the points in the original point cloud data that have not entered the point cloud queue and the workpiece seed point;

[0069] The point whose Euclidean distance is less than the Euclidean distance threshold is used as the workpiece to be inspected point;

[0070] Based on the workpiece seed point and the workpiece point to be inspected, a point cloud queue corresponding to the workpiece seed point is established.

[0071] Optionally, a k-dimensional tree (kd-tree) data structure can be used to establish the topological relationship between scattered point clouds, and the Euclidean distance d between all unclassified points (points that have not entered the point cloud queue in the original point cloud data) and the workpiece seed point can be traversed to set the Euclidean distance threshold d min , will meet d <d min The workpiece to be inspected points are stored in the point cloud queue Q. Specifically, the calculation formula of the Euclidean distance d is as follows:

[0072]

[0073] Among them, x and y are the two points for which the Euclidean distance needs to be calculated, (x1, x2...x n ) and (y1, y2…y n ) are the coordinates of x and y in n-dimensional space respectively.

[0074] If the distance d is less than the set Euclidean distance threshold d min , it will be put into the point cloud queue Q.

[0075] Optionally, in the embodiment of the present application, by establishing the topological relationship between the original point cloud data and calculating the Euclidean distance, the relevant points near the workpiece seed point can be efficiently identified and added to the point cloud queue. This ensures that only points that are close enough to the workpiece seed point are selected, which helps to improve the accuracy and effectiveness of point cloud processing, while reducing the interference of irrelevant points and improving the overall inspection and analysis efficiency.

[0076] Optionally, before determining the point cloud segmentation result according to the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point, the method includes:

[0077] Based on the points in the point cloud queue, establishing a spherical neighborhood set of the workpiece seed point;

[0078] Determine, based on the points in the spherical neighborhood set, a local plane in the least squares sense of the points in the spherical neighborhood set;

[0079] Based on the local plane, a normal vector of the workpiece to be inspected point and a normal vector of the workpiece seed point are determined.

[0080] Optionally, a spherical neighborhood set S with point p as the center and d as the radius in query Q is created based on kd-tree.

[0081] Optionally, determining the normal vector of the workpiece to-be-inspected point and the normal vector of the workpiece seed point based on the local plane includes:

[0082] determining the centroid of the local plane;

[0083] Determining a covariance matrix based on the centroid of the local plane;

[0084] Determining the normal vector of the workpiece to be inspected point according to the covariance matrix;

[0085] Perform eigenvalue decomposition on the covariance matrix, and use the eigenvector corresponding to the minimum eigenvalue as the normal vector of the workpiece seed point.

[0086] Optionally, based on the points in the spherical neighborhood set, it is determined that the sampling surface of the point cloud is smooth everywhere, and the local neighborhood of any point can be fitted by a plane. Therefore, for each seed point p in the point cloud, the k nearest neighboring points are searched, and the normal vector of the local plane P fitted by the k nearest points is calculated. The centroid of the normal vector represented by this plane.

[0087] Centroid The calculation formula is:

[0088]

[0089] in, is the normal vector of plane P, d is the distance from p to the origin of the coordinate system, p i is a point in the spherical neighborhood set.

[0090] Further, according to the above, the centroid is obtained Determine the covariance matrix M, the calculation formula of M is:

[0091]

[0092] Among them, p0 is the center of mass

[0093] Then perform eigenvalue decomposition on the covariance matrix M to obtain the eigenvalues ​​of M, and calculate the normal vectors of p and each point cloud in its neighborhood S, where the eigenvector corresponding to the minimum eigenvalue of M is the normal vector of the seed point p.

[0094] Optionally, the method for calculating the normal vector in the above embodiment can improve the accuracy of normal vector calculation, help to more accurately describe the geometric characteristics of the point cloud surface, and enhance subsequent analysis and processing effects.

[0095] Optionally, determining the point cloud segmentation result of the workpiece according to the angle between the normal vector of the to-be-inspected point of the workpiece and the normal vector of the seed point of the workpiece includes:

[0096] Establishing an empty cluster set, wherein the empty cluster set is used to store points that meet the normal vector angle threshold;

[0097] If the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point is less than the normal vector angle threshold, the workpiece to be inspected point is placed in the empty cluster set;

[0098] If the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point is greater than or equal to the normal vector angle threshold, the workpiece to be inspected point is eliminated.

[0099] Optionally, set the normal vector angle threshold α′ and create an empty cluster set C to store the normal vector angle threshold; calculate the normal vector angle between p and any point in its neighborhood S. If α<α′ is satisfied, put the point in C, otherwise remove the point. Please refer to Figure 3 For each seed point p, the kd-tree nearest neighbor search is first used to determine whether the normal vector satisfies the set threshold, and then decide whether to delete the neighboring point or put it into the set.

[0100] Optionally, segmentation is a key step in point cloud data processing. The embodiment of the present application establishes an empty cluster set and screens points that meet the normal vector angle threshold, incorporates the points that meet the requirements into the cluster set, and removes those that do not meet the requirements. On the basis of the Euclidean clustering segmentation algorithm, it can further classify based on the normal vector to enhance the accuracy of point cloud segmentation. At the same time, it reduces noise points and error points, improves the recognition accuracy of workpiece surface features, and thus optimizes the quality of subsequent processing and analysis. The point cloud segmentation results of the workpiece determined by this application can provide important information for subsequent feature extraction, and are the premise and basis for realizing object recognition in this application.

[0101] Optionally, after calculating the above normal vector, it is also necessary to orient the previously calculated normal vector, including:

[0102] Based on the inner product of the normal vectors of two adjacent points in the point cloud queue, the normal vectors of the points in the point cloud queue are oriented, and the orientation includes:

[0103] Set the normal vector direction;

[0104] If the inner product of the normal vectors of two adjacent points is less than 0, the direction of the normal vector of the opposite point is rotated 180 degrees, wherein the opposite point is a point opposite to the direction of the normal vector.

[0105] Optional, assuming For two adjacent points x i 、x j If the normal vectors are in the same direction, then If the inner product is negative, it means that the normal vector of one of the points needs to be flipped. Therefore, first set a normal vector direction for a point in the point cloud, and then traverse all other points. If the normal vector of the current point is set to is the next point to be traversed. If Then Flip, otherwise remain unchanged.

[0106] Optionally, in an embodiment of the present application, the normal vector of adjacent points in the point cloud queue is oriented based on the inner product of the normal vectors, thereby ensuring the consistency of the normal vector and reducing the calculation error caused by inconsistent normal vector directions, improving the overall consistency and quality of the point cloud data, and making subsequent results more accurate and reliable.

[0107] Optionally, after obtaining the normal vector of each point, the normal vector angle is calculated. and They are two normal vectors, and the calculation formula is as follows:

[0108]

[0109] Convert the resulting angle into degrees or radians to express the final result.

[0110] Please refer to Figure 2 In general, the embodiments of the present application may also include: first, obtaining point cloud data and performing downsampling; second, using the kd-tree data structure to establish the topological relationship between scattered point clouds, setting the Euclidean distance threshold, calculating the Euclidean distance, and storing the points that meet the threshold in a queue; finally, performing segmentation based on the improved Euclidean clustering segmentation algorithm based on the angle.

[0111] Please refer to Figure 4The present application embodiment further provides a workpiece identification system 20, including:

[0112] An acquisition module 21, used to acquire original point cloud data, wherein the original point cloud data includes point cloud data of at least one workpiece;

[0113] A determination module 22, used to determine point cloud data serving as a workpiece seed point from the original point cloud data;

[0114] A distance module 23 is used to establish a point cloud queue corresponding to the workpiece seed point based on the original point cloud data, wherein the point cloud queue includes the workpiece seed point and the workpiece to be inspected point, and the Euclidean distance between the workpiece to be inspected point and the workpiece seed point is less than a Euclidean distance threshold;

[0115] An angle module 24, used to determine the point cloud segmentation result of the workpiece according to the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point;

[0116] The identification module 25 is used to identify the workpiece based on the cloud segmentation result of the workpiece.

[0117] Optionally, the angle module 24 includes:

[0118] An angle submodule, used to establish an empty cluster set, wherein the empty cluster set is used to store points that meet the normal vector angle threshold;

[0119] If the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point is less than the normal vector angle threshold, the workpiece to be inspected point is placed in the empty cluster set;

[0120] If the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point is greater than or equal to the normal vector angle threshold, the workpiece to be inspected point is eliminated.

[0121] Optionally, the distance module 23 includes:

[0122] A distance submodule, used to establish a topological relationship between the original point cloud data;

[0123] Based on the topological relationship, traverse the Euclidean distance between the points in the original point cloud data that have not entered the point cloud queue and the workpiece seed point;

[0124] The point whose Euclidean distance is less than the Euclidean distance threshold is used as the workpiece to be inspected point;

[0125] Based on the workpiece seed point and the workpiece point to be inspected, a point cloud queue corresponding to the workpiece seed point is established.

[0126] Optionally, the workpiece identification system 20 further includes:

[0127] A plane module, used for establishing a spherical neighborhood set of the workpiece seed point based on the points in the point cloud queue;

[0128] Determine, based on the points in the spherical neighborhood set, a local plane in the least squares sense of the points in the spherical neighborhood set;

[0129] Based on the local plane, a normal vector of the workpiece to be inspected point and a normal vector of the workpiece seed point are determined.

[0130] Optionally, the plane module includes:

[0131] A normal vector module, for determining the centroid of the local plane;

[0132] Determining a covariance matrix based on the centroid of the local plane;

[0133] Determining the normal vector of the workpiece to be inspected point according to the covariance matrix;

[0134] Perform eigenvalue decomposition on the covariance matrix, and use the eigenvector corresponding to the minimum eigenvalue as the normal vector of the workpiece seed point.

[0135] Optionally, the workpiece identification system 20 further includes:

[0136] An orientation module is used to orient the normal vectors of points in the point cloud queue based on the inner product of the normal vectors of two adjacent points in the point cloud queue, wherein the orientation includes:

[0137] Set the normal vector direction;

[0138] If the inner product of the normal vectors of two adjacent points is less than 0, the direction of the normal vector of the opposite point is rotated 180 degrees, wherein the opposite point is a point opposite to the direction of the normal vector.

[0139] The workpiece recognition system provided by the embodiment of the present application can achieve Figure 1 The various processes implemented by the method embodiment and achieving the same technical effect are not described here to avoid repetition.

[0140] The present application embodiment provides an electronic device 30, see Figure 5 As shown, Figure 5 This is a principle block diagram of an electronic device 30 according to an embodiment of the present application, including a processor 31, a memory 32, and a program or instruction stored in the memory 32 and executable on the processor 31. When the program or instruction is executed by the processor, the steps in any one of the workpiece identification methods of the present application are implemented.

[0141] An embodiment of the present application provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of an embodiment of an artifact identification method such as any one of the above-mentioned items are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0142] Computer readable media include permanent and non-permanent, removable and non-removable media, and can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0143] The present application also provides a computer program product, including computer instructions, which, when executed by a processor, implement the above Figure 1 Each process of any one of the workpiece identification method embodiments shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0144] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0145] It should be noted that the collection, collection, updating, analysis, processing, use, transmission, storage and other aspects of user personal information involved in the technical solution of this disclosure are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information and network security.

[0146] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0147] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a service classification device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0148] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A workpiece identification method, characterized in that: include: Acquire original point cloud data, wherein the original point cloud data includes point cloud data of at least one workpiece; Determine point cloud data serving as a workpiece seed point from the original point cloud data; Based on the original point cloud data, a point cloud queue corresponding to the workpiece seed point is established, wherein the point cloud queue includes the workpiece seed point and the workpiece to be inspected point, and the Euclidean distance between the workpiece to be inspected point and the workpiece seed point is less than a Euclidean distance threshold; Determining a point cloud segmentation result of the workpiece according to an angle between a normal vector of the workpiece to be inspected point and a normal vector of a seed point of the workpiece; Based on the cloud segmentation result of the artifact, the artifact is identified.

2. The method according to claim 1, characterized in that Determining the point cloud segmentation result of the workpiece according to the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point includes: Establishing an empty cluster set, wherein the empty cluster set is used to store points that meet the normal vector angle threshold; If the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point is less than the normal vector angle threshold, the workpiece to be inspected point is placed in the empty cluster set; If the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point is greater than or equal to the normal vector angle threshold, the workpiece to be inspected point is eliminated.

3. The method according to claim 1, characterized in that The step of establishing a point cloud queue corresponding to the workpiece seed point based on the original point cloud data comprises: Establishing a topological relationship between the original point cloud data; Based on the topological relationship, traverse the Euclidean distance between the points in the original point cloud data that have not entered the point cloud queue and the workpiece seed point; The point whose Euclidean distance is less than the Euclidean distance threshold is used as the workpiece to be inspected point; Based on the workpiece seed point and the workpiece point to be inspected, a point cloud queue corresponding to the workpiece seed point is established.

4. The method according to claim 1, characterized in that: The step of determining the point cloud segmentation result according to the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point comprises: Based on the points in the point cloud queue, establishing a spherical neighborhood set of the workpiece seed point; Determine, based on the points in the spherical neighborhood set, a local plane in the least squares sense of the points in the spherical neighborhood set; Based on the local plane, a normal vector of the workpiece to be inspected point and a normal vector of the workpiece seed point are determined.

5. The method according to claim 4, characterized in that The determining, based on the local plane, the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point comprises: determining the centroid of the local plane; Determining a covariance matrix based on the centroid of the local plane; Determining the normal vector of the workpiece to be inspected point according to the covariance matrix; Perform eigenvalue decomposition on the covariance matrix, and use the eigenvector corresponding to the minimum eigenvalue as the normal vector of the workpiece seed point.

6. The method according to claim 1, characterized in that Also includes: Based on the inner product of the normal vectors of two adjacent points in the point cloud queue, the normal vectors of the points in the point cloud queue are oriented, and the orientation includes: Set the normal vector direction; If the inner product of the normal vectors of two adjacent points is less than 0, the direction of the normal vector of the opposite point is rotated 180 degrees, wherein the opposite point is a point opposite to the direction of the normal vector.

7. A workpiece identification system, characterized in that: include: An acquisition module, used for acquiring original point cloud data, wherein the original point cloud data includes point cloud data of at least one workpiece; A determination module, used to determine point cloud data serving as a workpiece seed point from the original point cloud data; A distance module is used to establish a point cloud queue corresponding to the workpiece seed point based on the original point cloud data, wherein the point cloud queue includes the workpiece seed point and the workpiece to be inspected point, and the Euclidean distance between the workpiece to be inspected point and the workpiece seed point is less than a Euclidean distance threshold; An angle module, used to determine the point cloud segmentation result of the workpiece according to the angle between the normal vector of the workpiece to be inspected point and the normal vector of the workpiece seed point; The recognition module is used to recognize the artifact based on the cloud segmentation result of the artifact.

8. An electronic device, characterized in that: include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, the steps of the workpiece identification method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that: A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the workpiece identification method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of the workpiece identification method according to any one of claims 1 to 6.

Citation Information

Cited By

  • Data matching identification method, identification device, readable storage medium and magnetic stimulator

    CN121366185A