A registration method, device and equipment for workpiece pose, and a storage medium

Abstract

The present application relates to a kind of registration method, device, equipment and storage medium for workpiece pose, the method comprises: the arc support line segment detection of any attitude of the image of workpiece to be registered is carried out, and the arc support group of workpiece to be registered is obtained;According to the arc support group, image ellipse candidate set is constructed, and image ellipse candidate set is processed to obtain image ellipse feature center point set;Determine the preset plane where the model ellipse feature center point set corresponding to the preset model of the workpiece to be registered is located, to indicate the model ellipse feature center point set with the normal vector of the preset plane;Based on preset matching algorithm, according to the image ellipse feature center point set and the model ellipse feature center point set, the workpiece to be registered is registered.The present application determines image ellipse feature center point set by the image of workpiece in any attitude, and iteratively registers with model ellipse feature center point set, and registration for workpiece in any pose is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of workpiece registration, in particular to a workpiece pose-oriented registration method, device, equipment and storage medium. BACKGROUND

[0002] In recent years, with the rapid development of computer technology and image processing technology, machine vision has been widely used in automobile manufacturing, food sorting, automatic processing and production, etc. In the industrial field, machine vision is usually used for visual measurement, visual guidance and visual detection. In modern visual detection automatic processing production, feature detection and model registration correction of workpieces are the basis for automatic processing.

[0003] In the field of feature detection and model registration correction, the existing technology mostly adopts the method of "eye on hand" robot end fixed plane structure light scanning, acquires the actual three-dimensional model of the target workpiece through the mobile scanning-splicing-reconstruction technology, and completes the registration through the large data amount of point cloud and the actual three-dimensional model.

[0004] However, when facing the processing of complex workpieces, the existing precision measurement-splicing technology algorithm is complex, and it is difficult to realize the efficient batch application of the visual measurement and robot processing integrated system. And for the large data amount of point cloud registration technology, under the condition that the actual pose of the workpiece is changed arbitrarily, it is easy to frequently fall into the local matching optimal condition, which affects the final pose correction accuracy and cannot accurately register the model. SUMMARY

[0005] Therefore, it is necessary to provide a workpiece pose-oriented registration method, device, equipment and storage medium to solve the problem that the existing workpiece model registration method affects the pose correction accuracy of the workpiece after changing the actual pose of the workpiece, and cannot accurately register the model.

[0006] To achieve the above technical purpose, the present application adopts the following technical scheme:

[0007] In a first aspect, the present application provides a workpiece pose-oriented registration method, comprising:

[0008] Detecting the arc support line segment of the image of the workpiece to be registered in any attitude to obtain an arc support group of the workpiece to be registered;

[0009] Constructing an image ellipse candidate set according to the arc support group, and processing the image ellipse candidate set to obtain an image ellipse feature center point set;

[0010] Determining a preset plane where a model ellipse feature center point set corresponding to a preset model of the workpiece to be registered is located, and representing the model ellipse feature center point set by a normal vector of the preset plane;

[0011] The workpiece to be registered is registered based on a preset matching algorithm according to the image ellipse feature center point set and the model ellipse feature center point set.

[0012] In some possible implementation manners, arc support line segment detection is performed on the workpiece image in any posture to be registered to obtain an arc support group of the workpiece to be registered, including: determining the arc support group of the workpiece to be registered according to sampling points of the image ellipse feature center point set.

[0013] In some possible implementation manners, the image ellipse candidate set is constructed according to the arc support group, and the image ellipse feature center point set is obtained by processing the image ellipse candidate set, including:

[0014] The initial image ellipse candidate set is constructed according to the arc support group based on a preset constraint condition, and the initial image ellipse candidate set is spatially clustered;

[0015] The initial image ellipse candidate set after spatial clustering is verified by setting a confidence function;

[0016] The initial image ellipse candidate set after verification is subjected to preset coordinate conversion to obtain the image ellipse feature center point set.

[0017] In some possible implementation manners, the preset constraint condition includes a same polarity constraint condition and a same region constraint condition; the initial image ellipse candidate set is constructed according to the arc support group based on the preset constraint condition, and the initial image ellipse candidate set is spatially clustered, including:

[0018] Based on the same polarity constraint condition, the polarity of the arc support group constituting the ellipse curve is determined;

[0019] Based on the same region constraint condition, it is determined whether the arc support group with the same polarity is located in a preset region.

[0020] In some possible implementation manners, a preset plane in which a model ellipse feature center point set corresponding to a preset model of the workpiece to be registered is located is determined, and a normal vector of the preset plane is used to represent the model ellipse feature center point set, including:

[0021] An equation expression of the preset plane in which the model ellipse feature center point set is located is determined;

[0022] The equation of the preset plane is determined based on a minimum normal distance formula according to the equation expression of the preset plane;

[0023] The normal vector representing the model ellipse feature center point set is determined according to the equation of the preset plane.

[0024] In some possible implementation manners, the workpiece to be registered is registered based on a preset matching algorithm according to the image ellipse feature center point set and the model ellipse feature center point set.

[0025] The distance threshold and the registration error function are set to denoise the image ellipse feature center point set and the model ellipse feature center point set.

[0026] In some possible implementation manners, the workpiece to be registered is registered based on a preset matching algorithm according to the image ellipse feature center point set and the model ellipse feature center point set, including:

[0027] A rotation and translation matrix conversion expression of the image ellipse feature center point set to the model ellipse feature center point set is determined;

[0028] A correlation function of the image ellipse feature center point set and the model ellipse feature center point set is established based on the rotation and translation matrix conversion expression;

[0029] A cylinder rigid body transformation matrix of the image ellipse feature center point set and the model ellipse feature center point set is calculated through the correlation function until iterative convergence.

[0030] In a second aspect, the present application further provides a registration device for a workpiece pose, including:

[0031] A detection module is configured to detect an arc support line segment of an image of a workpiece to be registered in any attitude, and obtain an arc support group of the workpiece to be registered.

[0032] An image point set module is configured to construct an image ellipse candidate set according to the arc support group, and process the image ellipse candidate set to obtain an image ellipse feature center point set.

[0033] A model point set module is configured to determine a preset plane where a model ellipse feature center point set corresponding to a preset model of the workpiece to be registered is located, and represent the model ellipse feature center point set by a normal vector of the preset plane.

[0034] A registration module is configured to register the workpiece to be registered based on a preset matching algorithm according to the image ellipse feature center point set and the model ellipse feature center point set.

[0035] In a third aspect, the present application further provides an electronic device including a memory and a processor, wherein,

[0036] The memory is configured to store a program.

[0037] The processor is coupled with the memory, and is configured to execute the program stored in the memory to implement steps in the registration method for a workpiece pose in any of the above implementation manners.

[0038] In a fourth aspect, the present application further provides a computer readable storage medium for storing computer readable programs or instructions, which, when executed by a processor, can realize the steps in the workpiece pose-oriented registration method in any of the above implementation manners.

[0039] The beneficial effects of the above embodiments are that the workpiece pose-oriented registration method, device, equipment and storage medium provided by the present application can acquire a workpiece image in any attitude, detect an arc support line segment of the workpiece image in any attitude, thereby determining a set of image ellipse feature center points, extract a set of model ellipse feature center points by determining a normal vector of a preset plane, and finally register a workpiece to be registered according to the set of image ellipse feature center points and the set of model ellipse feature center points, so that the workpiece in any attitude can be corrected, and the efficiency and accuracy of registration of a workpiece model in any attitude are improved. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 A flowchart of an embodiment of the workpiece pose-oriented registration method provided by the present application;

[0041] Figure 2 An embodiment of the step S102 in the method; Figure 1 A flowchart of an embodiment of the step S102;

[0042] Figure 3 An embodiment of the step S103 in the method; Figure 1 A flowchart of an embodiment of the step S103;

[0043] Figure 4 An embodiment of the step S104 in the method; Figure 1 A flowchart of an embodiment of the step S104;

[0044] Figure 5 A structural diagram of an embodiment of the workpiece pose-oriented registration device provided by the present application;

[0045] Figure 6 A structural diagram of an electronic device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0046] The preferred embodiments of the present application will be described in detail below with reference to the accompanying drawings, which form a part of this application. The accompanying drawings and the embodiments together illustrate the principles of the present application, but are not intended to limit the scope of the present application.

[0047] In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly and specifically limited.

[0048] Reference to“an embodiment” herein means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase“in an embodiment” in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another. It is expressly understood that any of the embodiments described herein can be incorporated in a combination of embodiments.

[0049] The application provides a workpiece pose-oriented registration method, device, equipment and storage medium, which are described below respectively.

[0050] Please refer to Figure 1 , Figure 1 The flowchart of an embodiment of the workpiece pose-oriented registration method provided by the application, a specific embodiment of the application discloses a workpiece pose-oriented registration method, comprising:

[0051] S101, arc support line segment detection is performed on an arbitrary attitude of a workpiece image to be registered, to obtain an arc support group of the workpiece to be registered;

[0052] S102, an image ellipse candidate set is constructed according to the arc support group, and the image ellipse candidate set is processed to obtain an image ellipse feature center point set;

[0053] S103, a preset plane on which a model ellipse feature center point set corresponding to a preset model of the workpiece to be registered is located is determined, and the model ellipse feature center point set is represented by a normal vector of the preset plane;

[0054] S104, the workpiece to be registered is registered based on a preset matching algorithm according to the image ellipse feature center point set and the model ellipse feature center point set.

[0055] In the above embodiment, before the workpiece model is registered, the actual image of the workpiece model to be registered is first acquired, at this time, the image of the workpiece model to be registered acquired can be in an arbitrary attitude, and the arc support group of the workpiece to be registered is determined by performing arc support line segment detection on the workpiece image to be registered in an arbitrary attitude.

[0056] The image ellipse candidate set of the workpiece to be registered is constructed by taking the arc support group of the workpiece to be registered as a constraint object, and the image ellipse feature center point set is finally obtained by sequentially performing spatial clustering, verification and coordinate conversion on the image ellipse candidate set.

[0057] The preset plane is an optimal plane F on which the model ellipse feature center point set is located, which is solved by using a minimum normal distance formula, and the model ellipse feature center point set is represented by a normal vector of the optimal plane F.

[0058] The preset matching algorithm is an improved normal consistency iterative nearest point registration algorithm. It constructs a rotation and translation matrix transformation correlation function from the actual image elliptical feature center point set to the model elliptical feature center point set to complete the iterative registration.

[0059] Compared with existing technologies, this embodiment provides a workpiece pose-oriented registration method. The method includes: detecting arc support segments on an image of a workpiece to be registered in any pose to obtain an arc support group of the workpiece; constructing an image ellipse candidate set based on the arc support group, and processing the image ellipse candidate set to obtain an image ellipse feature center point set; determining a preset plane containing the model ellipse feature center point set corresponding to a preset model of the workpiece to be registered, and representing the model ellipse feature center point set with the normal vector of the preset plane; and registering the workpiece to be registered based on a preset matching algorithm, using the image ellipse feature center point set and the model ellipse feature center point set. This invention first acquires a workpiece image in any pose, then detects arc support segments on the workpiece image in any pose to determine the image ellipse feature center point set, then extracts the model ellipse feature center point set by determining the normal vector of the preset plane, and finally registers the workpiece to be registered based on the image ellipse feature center point set and the model ellipse feature center point set. This allows for correction of workpieces in any pose, improving the efficiency and accuracy of workpiece model registration in any pose.

[0060] In some embodiments of the present invention, arc support line segment detection is performed on the image of the workpiece to be registered in any posture to obtain the arc support group of the workpiece to be registered, including: determining the arc support group of the workpiece to be registered based on the sampling points of the image ellipse feature center point set.

[0061] In the above embodiment, assuming that the two ends of the circumscribed rectangle of the arc support line segment are M and N respectively, and the centroid is represented by O, the principal angle expression of the arc support line segment is as shown in the following formula (1):

[0062]

[0063] In the formula, A i Represents the set of center points of the ellipse in the image {W} i The sampling points are}, i = 1, 2, 3, ..., N. W (N W represents the point set {W i The total number of sampling points (}), Region represents the area where the arc support line segment is located, and level-line angel represents the contour features of the sampling points. This represents the principal angle direction vector of the arc-supported line segment.

[0064] Please see Figure 2 , Figure 2 for Figure 1A flowchart illustrating an embodiment of step S102. In some embodiments of the present invention, an image ellipse candidate set is constructed based on the arc support group, and the image ellipse candidate set is processed to obtain an image ellipse feature center point set, including:

[0065] S201. Based on preset constraints, construct an initial image ellipse candidate set according to the arc support group, and perform spatial clustering on the initial image ellipse candidate set;

[0066] S202. Validate the initial image ellipse candidate set after spatial clustering using a pre-set confidence function;

[0067] S203. Perform a preset coordinate transformation on the verified initial image ellipse candidate set to obtain the image ellipse feature center point set.

[0068] In the above embodiments, the same polarity constraint is: determine the polarity of the "arc support group" that makes up the elliptic curve of the image. If the internal area of ​​the arc support group is brighter or darker than the outer area, it means that the polarity is positive or negative. Usually, the polarity of the elliptic curve of the same image is the same.

[0069] The same constraint in the region is: to determine whether two paired "arc support groups" are located in mutually valid regions, the specific judgment condition satisfies formula (2):

[0070]

[0071] In the formula, i and j represent two different "arc support groups". Taking i as an example, the t of the "arc support group" i Starting point is The termination point is The midpoint is P i and P j The initial arc support segment is The terminating arc support segment is The normalized vector representing the arc support direction L; express A normalized vector rotated 90° clockwise; when the polarity of the candidate ellipse set {t1, t2} is positive, ply is +1, otherwise ply is -1; γ represents the distance threshold.

[0072] The pre-set reliability function includes five constraints: goodness index, shape index, position index, gradient index, and weighted index. The calculation formulas for the five constraints—goodness index CI, shape index HI, position index PI, gradient index PI, and weighted index AI—are shown in formula (3).

[0073]

[0074] (Formula 3)

[0075] The confidence function U of the image ellipse candidate set constructed by the above indicators e As shown in formula (4) below, the verification is completed by removing elliptic curves with low confidence through thresholding.

[0076]

[0077] In the formula, SI(e) represents the arc support segment of ellipse e, and θ represents the coverage area of ​​the ellipse's connected angle. a and b represent the sizes of the major and minor axes of the ellipse, P i R represents an edge point. i Represents the sampling point, ζ i R represents the i-th sampling point. i The eccentric angle.

[0078] The preset coordinates are converted to center point coordinates. That is, the target elliptical image coordinates are transformed to the robot's base coordinate system using a multi-view stereo vision algorithm, resulting in the actual acquired set of elliptical feature center points {W}. i The expression for the multi-view stereo vision algorithm is shown in formula (5) below:

[0079]

[0080] In the formula, the coordinates of the center point of the target ellipse are (X1, Y1, Z1). This represents the transformation matrix from the coordinates of the multi-camera group to the robot's base coordinates. This represents the matrix showing the relationship between the center point of the actual image ellipse and the main camera Cl, (x o y o , z o ) represents the three-dimensional coordinates of the center point of the image ellipse after transformation to the robot's base coordinate system, and a, b, c, t, f, u, v represent the intrinsic and extrinsic parameters of the matrix transformation, which can be obtained through singular value decomposition.

[0081] In some embodiments of the present invention, the preset constraints include polarity-same constraints and region-same constraints; based on the preset constraints, an initial image ellipse candidate set is constructed according to the arc support group, and spatial clustering is performed on the initial image ellipse candidate set, including:

[0082] Determine the polarity of the arc support group that makes up the elliptic curve based on the constraint condition of the same polarity.

[0083] Based on the same regional constraint condition, determine whether the arc support group with the same polarity is located in the preset region.

[0084] Please see Figure 3 , Figure 3 for Figure 1A flowchart illustrating an embodiment of step S103 of the present invention shows that, in some embodiments of the present invention, a preset plane is determined to be located where the set of feature center points of the model ellipse corresponding to the preset model of the workpiece to be registered is located, and the set of feature center points of the model ellipse is represented by the normal vector of the preset plane, including:

[0085] S301. Determine the equation expression of the preset plane where the set of feature center points of the model ellipse lies;

[0086] S302. Based on the minimum normal distance formula, determine the equation of the preset plane according to the equation expression of the preset plane;

[0087] S303. Determine the normal vector representing the set of feature center points of the ellipse model according to the preset plane equation.

[0088] In the above embodiments, it is assumed that the model elliptical feature center point set {V} j The optimal plane for all points in} is F, and the equation of the optimal plane is expressed as formula (6). Solving for the parameters a, b, c, d yields the point set {V}. j The corresponding optimal plane F.

[0089]

[0090] In the formula, cosα, cosβ, and cosγ are the direction cosines of the normal vector at the point (X, Y, Z) on the optimal plane, |p| is the normal distance from the origin to the optimal plane, and d is the distance from the point set {V}. j The normal distance from all points in the plane to the optimal plane.

[0091] Satisfying the model elliptical feature center point set {V j The optimal plane is the one where the normal distance d from all points to the hypothetical optimal plane F is minimized. Based on this, the error function f is constructed as shown in formula (7), where τ represents the error modulus, and N... V Represents the point set {V j The total number of sampling points, j = 1, 2, 3, ..., N V .

[0092]

[0093] Substituting formula (7) into formula (6), the point set {V j The mean value of all points in the three-dimensional direction is represented as: Then the minimum normal distance Taking partial derivatives with respect to parameters a, b, c, and d yields formula (8):

[0094]

[0095] Rearranging formula (8) into matrix form, as shown in formula (9), the eigenvector corresponding to the smallest eigenvalue of the matrix is ​​the normal vector of the optimal plane, and at the same time, it satisfies the solution of a2+b2+c2=1.

[0096]

[0097] In some embodiments of the present invention, based on a preset matching algorithm, registration is performed on the workpiece to be registered according to the image elliptical feature center point set and the model elliptical feature center point set, before which the following is included:

[0098] Denoising is performed on the image elliptical feature center point set and the model elliptical feature center point set by setting a distance threshold and a registration error function.

[0099] In the above embodiments, based on the obtained optimal plane normal vector, in order to avoid the iteration process getting trapped in local optima, a distance threshold and registration error function are set between the image elliptical feature center point set and the model elliptical feature center point set, and points whose corresponding Euclidean distance does not meet the threshold are regarded as noise points and removed.

[0100] The Euclidean distance calculation formula is shown in formula (10) below, with the distance threshold η and registration error function R set. MS The registration error R obtained by the (k-1)th iteration is... MS Set as the distance threshold for the k-th iteration to achieve dynamic adjustment.

[0101]

[0102] In the formula, R k-1 and t k-1 Let represent the rotation and translation matrices of the (k-1)th iteration, and η represent the distance threshold of the kth iteration. During the iteration process, points whose Euclidean distance does not meet the threshold are considered noise points and are removed.

[0103] Please see Figure 4 , Figure 4 for Figure 1 A flowchart illustrating an embodiment of step S104. In some embodiments of the present invention, based on a preset matching algorithm, registration is performed on the workpiece to be registered according to the image elliptical feature center point set and the model elliptical feature center point set, including:

[0104] S401. Determine the transformation expression of the rotation and translation matrix from the set of elliptical feature center points in the image to the set of elliptical feature center points in the model;

[0105] S402. Based on the transformation expression of the rotation and translation matrix, establish the correlation function between the image elliptical feature center point set and the model elliptical feature center point set;

[0106] S403. Calculate the rigid body transformation matrix of the image elliptical feature center point set and the model elliptical feature center point set using the correlation function until the iteration converges.

[0107] In the above embodiments, the image ellipse feature center point set {W i} to the model elliptical feature center point set {V j The transformation expression of the rotation and translation matrix of} is shown in formula (11):

[0108]

[0109] In the formula, and R represents the rotation matrix, t represents the translation matrix, and N... W and N V Represent the set {W} i} and {V j The number of registration points to be registered in}

[0110] Based on the mathematical method of inverse solution, given matrix transformation R k-1 and t k-1 and the elliptic point set {W i} Perform R k-1 W i +t k-1 Transformation, establishing the correlation function q between two sets. k (i), as shown in formula (12):

[0111]

[0112] In the formula, i = 1, 2, ..., N W k-1 indicates that the iteration has reached the (k-1)th step.

[0113] Furthermore, calculate the set of elliptical feature center points {W} in the image. k} and the model elliptical feature center point set {V k The rigid body transformation matrix (R) of} k , t k The following formula (13) is as follows:

[0114]

[0115] Repeat the calculation of the function expression above until the two sets of points converge iteratively, and the transformation matrix between the target point pairs RW+t=V can be obtained, which means that the registration of the workpiece model is achieved.

[0116] To better implement the workpiece pose-oriented registration method in the embodiments of the present invention, based on the workpiece pose-oriented registration method, please refer to the corresponding... Figure 5 , Figure 5This is a schematic diagram of an embodiment of the workpiece pose-oriented registration device provided by the present invention. The embodiment of the present invention provides a workpiece pose-oriented registration device 500, comprising:

[0117] The detection module 510 is used to detect the arc support line segments of the image of the workpiece to be registered in any posture, so as to obtain the arc support group of the workpiece to be registered.

[0118] The image point set module 520 is used to construct an image ellipse candidate set based on the arc support group, and to process the image ellipse candidate set to obtain the image ellipse feature center point set.

[0119] The model point set module 530 is used to determine the preset plane where the preset model ellipse feature center point set corresponding to the preset model of the workpiece to be registered is located, and the model ellipse feature center point set is represented by the normal vector of the preset plane.

[0120] The registration module 540 is used to register the workpiece to be registered based on the image ellipse feature center point set and the model ellipse feature center point set according to the preset matching algorithm.

[0121] It should be noted that the device 500 provided in the above embodiments can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be found in the corresponding content in the above method embodiments, and will not be repeated here.

[0122] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Based on the above-described registration method for workpiece pose, the present invention also provides a registration device for workpiece pose, which can be a computing device such as a mobile terminal, desktop computer, laptop, handheld computer, or server. The registration device for workpiece pose includes a processor 610, a memory 620, and a display 630. Figure 6 Only some components of the electronic device are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0123] In some embodiments, memory 620 may be an internal storage unit of the workpiece pose-oriented registration device, such as a hard disk or memory of the workpiece pose-oriented registration device. In other embodiments, memory 620 may be an external storage device of the workpiece pose-oriented registration device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the workpiece pose-oriented registration device. Furthermore, memory 620 may include both internal and external storage units of the workpiece pose-oriented registration device. Memory 620 is used to store application software and various types of data installed on the workpiece pose-oriented registration device, such as the program code for the workpiece pose-oriented registration device. Memory 620 may also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 620 stores a workpiece pose-oriented registration program 640, which can be executed by the processor 610 to implement the workpiece pose-oriented registration method of the various embodiments of this application.

[0124] In some embodiments, processor 610 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 620 or process data, such as executing a registration method oriented towards workpiece pose.

[0125] In some embodiments, display 630 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 630 is used to display information from the workpiece pose-oriented registration device and to display a user interface for visualization. Components 610-630 of the workpiece pose-oriented registration device communicate with each other via a system bus.

[0126] In one embodiment, when the processor 610 executes the workpiece pose-oriented registration program 640 in the memory 620, the steps in the workpiece pose-oriented registration method described above are implemented.

[0127] This embodiment also provides a computer-readable storage medium storing a workpiece pose-oriented registration program, which, when executed by a processor, performs the following steps:

[0128] Arc support line segment detection is performed on the image of the workpiece to be registered in any posture to obtain the arc support group of the workpiece to be registered;

[0129] Based on the arc support group, construct the image ellipse candidate set, and process the image ellipse candidate set to obtain the image ellipse feature center point set;

[0130] Determine the preset plane where the set of feature center points of the model ellipse corresponding to the preset model of the workpiece to be registered is located, and represent the set of feature center points of the model ellipse with the normal vector of the preset plane;

[0131] Based on a preset matching algorithm, the workpiece to be registered is registered according to the set of elliptical feature center points in the image and the set of elliptical feature center points in the model.

[0132] In summary, this embodiment provides a registration method, apparatus, device, and storage medium for workpiece pose orientation. The method includes: detecting arc support segments in an image of a workpiece to be registered in any pose to obtain an arc support group of the workpiece; constructing an image ellipse candidate set based on the arc support group, and processing the image ellipse candidate set to obtain an image ellipse feature center point set; determining a preset plane containing the model ellipse feature center point set corresponding to a preset model of the workpiece to be registered, and representing the model ellipse feature center point set with the normal vector of the preset plane; and registering the workpiece to be registered based on a preset matching algorithm, according to the image ellipse feature center point set and the model ellipse feature center point set. This invention first acquires an image of a workpiece in any pose, then performs arc support line segment detection on the workpiece image in any pose to determine the set of elliptical feature center points in the image. Then, it extracts the set of elliptical feature center points in the model by determining the normal vector of a preset plane. Finally, it registers the workpiece to be registered based on the set of elliptical feature center points in the image and the set of elliptical feature center points in the model. This invention can correct workpieces in any pose, improving the efficiency and accuracy of workpiece model registration in any pose.

[0133] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A registration method for workpiece pose orientation, characterized in that, include: Arc support line segment detection is performed on the image of the workpiece to be registered in any posture to obtain the arc support group of the workpiece to be registered; Constructing an image ellipse candidate set based on the arc support group, and processing the image ellipse candidate set to obtain an image ellipse feature center point set, includes: constructing an initial image ellipse candidate set based on the arc support group according to the same polarity constraint and the same region constraint, and performing spatial clustering on the initial image ellipse candidate set, wherein, based on the same polarity constraint, the polarity of the arc support group that makes up the ellipse curve is determined, and based on the same region constraint, it is determined whether the arc support group with the same polarity is located in a preset region; verifying the initial image ellipse candidate set after spatial clustering through a preset confidence function; and performing a preset coordinate transformation on the verified initial image ellipse candidate set to obtain the image ellipse feature center point set; Determine the preset plane where the set of feature center points of the model ellipse corresponding to the preset model of the workpiece to be registered is located, and represent the set of feature center points of the model ellipse with the normal vector of the preset plane; Denoise the image elliptical feature center point set and the model elliptical feature center point set by setting a distance threshold and a registration error function; Based on a preset matching algorithm, the workpiece to be registered is registered according to the image elliptical feature center point set and the model elliptical feature center point set, including: determining the rotation and translation matrix transformation expression from the image elliptical feature center point set to the model elliptical feature center point set; establishing a correlation function between the image elliptical feature center point set and the model elliptical feature center point set based on the rotation and translation matrix transformation expression; and calculating the cylinder rigid body transformation matrix of the image elliptical feature center point set and the model elliptical feature center point set through the correlation function until iterative convergence.

2. The registration method for workpiece pose according to claim 1, characterized in that, The step of detecting arc support segments in an image of a workpiece to be registered in any pose to obtain the arc support group of the workpiece to be registered includes: determining the arc support group of the workpiece to be registered based on the sampling points of the image ellipse feature center point set.

3. The registration method for workpiece pose according to claim 1, characterized in that, The step of determining the preset plane containing the set of model elliptical feature center points corresponding to the preset model of the workpiece to be registered, and representing the set of model elliptical feature center points with the normal vector of the preset plane, includes: Determine the equation expression for the preset plane containing the set of elliptical feature center points of the model; Based on the minimum normal distance formula, the equation of the preset plane is determined according to the equation expression of the preset plane; The normal vector representing the set of feature center points of the ellipse in the model is determined based on the preset plane equation.

4. A registration device oriented towards workpiece pose, used to implement the registration method oriented towards workpiece pose as described in any one of claims 1 to 3, characterized in that, include: The detection module is used to detect the arc support line segments of the image of the workpiece to be registered in any posture, so as to obtain the arc support group of the workpiece to be registered. The image point set module is used to construct an image ellipse candidate set based on the arc support group, and process the image ellipse candidate set to obtain the image ellipse feature center point set; The model point set module is used to determine the preset plane where the preset model ellipse feature center point set corresponding to the preset model of the workpiece to be registered is located, and the model ellipse feature center point set is represented by the normal vector of the preset plane. The registration module is used to register the workpiece to be registered based on a preset matching algorithm, according to the set of elliptical feature center points of the image and the set of elliptical feature center points of the model.

5. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the workpiece pose-oriented registration method according to any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, are capable of implementing the steps in the workpiece pose-oriented registration method described in any one of claims 1 to 3.

Images