Workpiece registration method, device, electronic device and storage medium

Abstract

The present application relates to a workpiece registration method, device, electronic device and storage medium. The method performs virtual photographing registration based on the three-dimensional model of the workpiece to be registered, obtains the first numerical number of top-view images and the attitude coordinate points corresponding to each top-view image according to the B top-view images at different rotation angles respectively collected at A target collection points in the virtual placement area, and registers the first numerical number of top-view images into the workpiece recognition model. Batch virtual recognition is performed according to the three-dimensional model randomly set in the virtual placement area to obtain the attitude coordinate points with recognition errors. Finally, based on the recognition camera, image collection and registration are performed on the workpiece to be registered placed on the passing track according to all the obtained attitude coordinate points. Since the collected images cover various preset attitudes of the workpiece to be registered and the attitudes with recognition errors of the workpiece recognition model, the attitudes of the workpiece to be registered can be fully collected, so that the recognition success rate after workpiece registration meets the recognition requirements.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular, to a workpiece registration method, apparatus, electronic device, and storage medium. Background Art

[0002] In the process of identifying a new model workpiece, in the face of a new model workpiece, it is first necessary to register the new model workpiece. The registration process is generally that workers randomly place the workpiece in various poses for photo registration. Usually, workers will select about 20 different angles for shooting and perform workpiece registration based on the taken photos. However, in the actual operation process, workers fail to fully present all kinds of placement poses of the workpiece, resulting in a large number of poses not being included in the registration scope. This problem of pose missing registration will have a serious impact on the subsequent actual in-line identification link, resulting in a low identification success rate, such as only about 90%, which cannot meet the identification requirements. Summary of the Invention

[0003] This application provides a workpiece registration method, apparatus, electronic device, and storage medium to solve the technical problem of how to fully collect the poses of the workpiece to be registered to meet the identification requirements after workpiece registration.

[0004] In a first aspect, this application provides a workpiece registration method, and the method includes:

[0005] Determine the effective placement area of the workpiece to be registered according to the shooting coverage of the identification camera and the track width of the in-line track;

[0006] Set the three-dimensional model of the workpiece to be registered on the horizontal plane where the virtual placement area determined based on the effective placement area is located according to a preset orientation; wherein, the contact points between the three-dimensional model and the horizontal plane include at least three points not on the same straight line;

[0007] Obtain A target collection points of the virtual placement area;

[0008] Respectively collect B top-down images at different rotation angles when the three-dimensional model is at each of the target collection points, and the pose coordinate points corresponding to each of the top-down images, to obtain the first numerical number of the top-down images; wherein, the first numerical number is equal to A multiplied by B; the pose coordinate points are used to represent the coordinates and angles of the three-dimensional model in the virtual placement area;

[0009] Register the first numerical number of the top-down images into the workpiece identification model, and perform identification tests on the three-dimensional model randomly set in the virtual placement area based on the workpiece identification model;

[0010] Obtain C pose coordinate points at which the workpiece identification model misidentifies the three-dimensional model under a preset number of test times;

[0011] Based on the recognition camera, image acquisition is performed on the workpiece to be registered placed on the passing track according to the second numerical value of the pose coordinate points, so as to register the workpiece to be registered according to the acquired image; wherein, the second numerical value is the sum of the first numerical value and C.

[0012] Optionally, obtaining A target acquisition points in the virtual placement area includes:

[0013] Determine the center point of the virtual placement area as the first acquisition point;

[0014] Determine a reference point from the target polygon formed by all the contact points;

[0015] Determine the target distance from the reference point to each vertex of the target polygon;

[0016] Based on the maximum distance among the target distances, determine four second acquisition points in the four corner areas of the virtual placement area; wherein, the second acquisition point is at a distance greater than or equal to the maximum distance from the edge of the virtual placement area;

[0017] Take the first acquisition point and the four second acquisition points as the A target acquisition points in the virtual placement area.

[0018] Optionally, determining a reference point from the target polygon formed by all the contact points includes:

[0019] Determine the highest point of the three-dimensional model based on the horizontal plane;

[0020] Take the projection point of the highest point in the target polygon as the reference point.

[0021] Optionally, based on the maximum distance among the target distances, determining four second acquisition points in the four corner areas of the virtual placement area includes:

[0022] Obtain the projection area of the three-dimensional model on the horizontal plane;

[0023] Determine the farthest distance between the reference point and all points in the projection area;

[0024] Based on the maximum value of the farthest distance and the maximum distance, determine four second acquisition points in the four corner areas of the virtual placement area; wherein, the second acquisition point is at a distance equal to the maximum value from the edge of the virtual placement area.

[0025] Optionally, when collecting the three-dimensional model at each of the target collection points, B top-down images of the three-dimensional model at different rotation angles and the attitude coordinate points corresponding to each of the top-down images are collected, obtaining the first numerical number of the top-down images, including:

[0026] Obtain the height information of the recognition camera from the passing track;

[0027] When translating the three-dimensional model so that the reference point is at any one of the target collection points, based on the height information, collect B top-down images of the three-dimensional model at different rotation angles with the target line as the axis, and the attitude coordinate points corresponding to each of the top-down images; wherein, the target line is perpendicular to the horizontal plane and intersects the horizontal plane at the reference point;

[0028] When the collection of the B top-down images at the current target collection point is completed, translate the three-dimensional model to the next target collection point, and continue to execute the step of collecting B top-down images of the three-dimensional model at different rotation angles with the target line as the axis and the attitude coordinate points corresponding to each of the top-down images until the collection of all the target collection points is completed, obtaining the first numerical number of the top-down images.

[0029] Optionally, based on the recognition camera, image collection is performed on the workpiece to be registered placed on the passing track according to the second numerical number of the attitude coordinate points, so as to register the workpiece to be registered according to the collected images, including:

[0030] Input the second numerical number of the attitude coordinate points into the registration system;

[0031] Obtain the attitude virtual image of the workpiece to be registered generated by the registration system based on the attitude coordinate points;

[0032] Based on the recognition camera, collect the camera image of the workpiece to be registered placed on the passing track according to the attitude virtual image;

[0033] Register the workpiece to be registered according to the second numerical number of the camera images.

[0034] Optionally, set the three-dimensional model of the workpiece to be registered on the horizontal plane where the virtual placement area determined based on the placement valid area according to the preset orientation, including:

[0035] Obtain the three-dimensional model of the workpiece to be registered;

[0036] Determine the virtual placement area of the three-dimensional model based on the placement valid area;

[0037] Determine the riser position of the three-dimensional model;

[0038] Adjust the three-dimensional model to the orientation where the riser is far from the horizontal plane, and set the three-dimensional model on the virtual placement area.

[0039] In a second aspect, the present application provides a workpiece registration device, which includes:

[0040] A first determination module, configured to determine the effective placement area of the workpiece to be registered according to the shooting coverage of the recognition camera and the track width of the pass-through track;

[0041] An orientation adjustment module, configured to set the three-dimensional model of the workpiece to be registered on the horizontal plane where the virtual placement area is located according to a preset orientation; wherein, the contact points between the three-dimensional model and the horizontal plane include at least three points not on the same straight line;

[0042] A first acquisition module, configured to acquire A target acquisition points of the virtual placement area;

[0043] An attitude coordinate point acquisition module, configured to respectively acquire B top-down images at different rotation angles when the three-dimensional model is at each of the target acquisition points, and the attitude coordinate points corresponding to each of the top-down images, to obtain the first number of the top-down images; wherein, the first number is equal to A multiplied by B; the attitude coordinate points are used to represent the coordinates and angles of the three-dimensional model in the virtual placement area;

[0044] An identification test module, configured to register the first number of the top-down images to a workpiece identification model, and perform an identification test on the three-dimensional model randomly set in the virtual placement area based on the workpiece identification model;

[0045] A second acquisition module, configured to acquire C attitude coordinate points at which the workpiece identification model misidentifies the three-dimensional model under a preset number of test times;

[0046] An image acquisition module, configured to perform image acquisition on the workpiece to be registered placed on the pass-through track according to the second number of the attitude coordinate points based on the recognition camera, so as to register the workpiece to be registered according to the acquired images; wherein, the second number is the sum of the first number and C.

[0047] In a third aspect, the present application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete communication with each other through the communication bus;

[0048] The memory is used to store a computer program;

[0049] A processor, which is configured to implement the workpiece registration method according to any one of the embodiments of the first aspect when executing a program stored in a memory.

[0050] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the workpiece registration method according to any one of the embodiments of the first aspect is implemented.

[0051] The above technical solutions provided by the embodiments of the present application have the following advantages compared with the prior art: In the method provided by the embodiments of the present application, an effective placement area of a workpiece to be registered is determined according to the shooting coverage of an identification camera and the track width of a through-type track; a three-dimensional model of the workpiece to be registered is set on a horizontal plane where a virtual placement area determined based on the effective placement area is located according to a preset orientation, wherein at least three non-collinear points of contact between the three-dimensional model and the horizontal plane are included; A target acquisition points of the virtual placement area are obtained; B top-view images at different rotation angles and attitude coordinate points corresponding to each top-view image are respectively acquired when the three-dimensional model is at each of the target acquisition points, and the first numerical number of top-view images is obtained, wherein the first numerical number is equal to A multiplied by B; the attitude coordinate points are used to represent the coordinates and angles of the three-dimensional model in the virtual placement area; the first numerical number of top-view images are registered into a workpiece identification model, and the three-dimensional model randomly set in the virtual placement area is identified and tested based on the workpiece identification model; C attitude coordinate points at which the workpiece identification model misidentifies the three-dimensional model are obtained under a preset number of test times; the identification camera is used to collect images of the workpiece to be registered placed on the through-type track according to the second numerical number of attitude coordinate points respectively, so as to register the workpiece to be registered according to the collected images, wherein the second numerical number is the sum of the first numerical number and C. In this method, virtual photographing registration is performed on the three-dimensional model of the workpiece to be registered according to the actually identified scene, B top-view images at different rotation angles are respectively acquired according to A target acquisition points in the virtual placement area, the first numerical number of top-view images and attitude coordinate points corresponding to each top-view image are obtained, the first numerical number of top-view images are registered into the workpiece identification model, batch virtual identification tests are performed according to the three-dimensional model randomly set in the virtual placement area, and attitude coordinate points at which misidentification occurs are obtained. Finally, the identification camera is used to collect images and register the workpiece to be registered placed on the through-type track according to all the obtained attitude coordinate points. Since the collected images cover various preset postures of the workpiece to be registered and the postures at which the workpiece identification model misidentifies, the postures of the workpiece to be registered can be fully collected, so that the identification success rate after workpiece registration meets the identification requirements. Description of the Drawings

[0052] The accompanying drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with the present application, and are used in conjunction with the specification to explain the principles of the present application.

[0053] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0054] One or more embodiments are exemplarily illustrated by the pictures in the corresponding accompanying drawings. These exemplary illustrations do not limit the embodiments. Elements with the same reference numerals in the drawings represent similar elements. Unless otherwise stated, the drawings in the figures do not constitute a scale limitation.

[0055] Figure 1 System architecture diagram of a workpiece registration method provided for an embodiment of the present application;

[0056] Figure 2 Flow schematic diagram of a workpiece registration method provided for an embodiment of the present application;

[0057] Figure 3 Schematic diagram of a workpiece to be registered;

[0058] Figure 4 Schematic diagram of determining a three-dimensional model reference point provided for an embodiment of the present application;

[0059] Figure 5 Schematic diagram of target acquisition points in a virtual placement area provided for an embodiment of the present application;

[0060] Figure 6 Structural schematic diagram of a workpiece registration device provided for an embodiment of the present application;

[0061] Figure 7 Structural schematic diagram of an electronic device provided for an embodiment of the present application. Detailed implementation manners

[0062] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts fall within the scope of protection of the present application.

[0063] The following disclosure provides many different embodiments or examples for implementing different structures of the present application. To simplify the disclosure of the present application, the components and settings of specific examples are described below. Of course, they are only examples and are not intended to limit the present application. In addition, the present application may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplification and clarity and does not itself indicate the relationship between the various embodiments and / or settings discussed.

[0064] To solve the technical problem in the prior art of how to fully collect the pose of a workpiece to be registered to meet the recognition requirements after workpiece registration, the present application provides a workpiece registration method, device, electronic device, and storage medium. When collecting images, it covers various preset poses of the workpiece to be registered and the poses where the workpiece recognition model makes mistakes, so that the pose of the workpiece to be registered can be fully collected, and the recognition success rate after workpiece registration meets the recognition requirements.

[0065] The first embodiment of the present application provides a workpiece registration method, which can be applied to a system architecture as Figure 1 shown. In this system architecture, it at least includes a data acquisition module 101 and a data processing module 102, and a communication connection is established between the data acquisition module 101 and the data processing module 102. The data acquisition module 101 can acquire data of the actual recognition scene, such as acquiring the shooting coverage of the recognition camera, the track width of the through-type track, the height of the recognition camera from the through-type track, etc., and can also acquire the three-dimensional model of the workpiece to be registered. The data processing module 102 can adjust the pose of the three-dimensional model, and collect B top-down images at different rotation angles of the three-dimensional model according to A target acquisition points in the virtual placement area to obtain the first numerical top-down images, and the pose coordinate points corresponding to each top-down image. Then register the first numerical top-down images into the workpiece recognition model, perform batch virtual recognition tests on the three-dimensional model randomly set in the virtual placement area to obtain the pose coordinate points where the recognition is incorrect, and output all the obtained pose coordinate points. Thus, it guides the user to collect and register images of the workpiece to be registered placed on the through-type track based on the recognition camera according to all the obtained pose coordinate points. Since the images collected cover various preset poses of the workpiece to be registered and the poses where the workpiece recognition model makes mistakes, the pose of the workpiece to be registered can be fully collected, and the recognition success rate after workpiece registration meets the recognition requirements.

[0066] Next, based on this system architecture, the workpiece registration method will be described in detail. As Figure 2 shown, the workpiece registration method includes:

[0067] Step 201, determine the effective placement area of the workpiece to be registered according to the shooting coverage of the recognition camera and the track width of the through-type track.

[0068] The workpiece to be registered is a new type of workpiece that has not been registered in the workpiece registration system. The identification camera is the camera that actually takes pictures of the workpiece to be registered to collect images, and it is also the camera for model identification after registration. The identification camera can be set above the passing track. The passing track can be a belt-type track. For example, the track width is 600 mm, the track length is half of the belt perimeter, the shooting coverage of the identification camera is 800 mm in diameter, and the height of the identification camera from the track is 550 mm. Then, the effective placement area of the workpiece to be registered can be determined according to the above parameters. For example, the effective placement area is a 500 mm (width) * 700 mm (length) area on the passing track under the identification camera. When the identification camera takes pictures, the workpiece to be registered needs to be in this effective placement area to ensure the accuracy of identification.

[0069] Step 202: Set the three-dimensional model of the workpiece to be registered on the horizontal plane where the virtual placement area determined based on the effective placement area is located according to the preset orientation; among them, the contact points between the three-dimensional model and the horizontal plane include at least three points not on the same straight line.

[0070] As Figure 3 As shown in the schematic diagram of the workpiece to be registered, the preset orientation can be that the riser of the workpiece to be registered faces upward, which is convenient for the robot to grasp the workpiece on the passing track through the riser. When ensuring that the riser faces upward, set the three-dimensional model on the horizontal plane where the virtual placement area is located. To keep the three-dimensional model balanced, it needs to be in contact with the horizontal plane, and there are at least three contact points not on the same straight line (usually 3 to 5). It should be noted that setting the three-dimensional model on the horizontal plane of the virtual placement area here refers to setting in the software to simulate the real scene, rather than actually placing the real workpiece.

[0071] In one embodiment, setting the three-dimensional model of the workpiece to be registered on the horizontal plane where the virtual placement area determined based on the effective placement area is located according to the preset orientation includes: obtaining the three-dimensional model of the workpiece to be registered; determining the virtual placement area of the three-dimensional model based on the effective placement area; determining the riser position of the three-dimensional model; adjusting the three-dimensional model to the orientation with the riser position away from the horizontal plane, and setting the three-dimensional model on the virtual placement area.

[0072] In this embodiment, the three-dimensional model of the workpiece to be registered can be the design 3D drawing of the workpiece to be registered, or the three-dimensional model can be obtained by scanning the workpiece to be registered, without limitation. The virtual placement area can be set in combination with the effective placement area of the actual application scenario. For example, it can be set according to the same dimensions in the software to ensure that when the virtual three-dimensional model is in the virtual placement area, it is consistent with the placement effect of the actual workpiece to be registered in the effective placement area. After determining the riser position of the three-dimensional model, the three-dimensional model is adjusted to the orientation with the riser position away from the horizontal plane, that is, the three-dimensional model is set on the virtual placement area while keeping the riser direction upward. It should be understood that when collecting the top view image of the three-dimensional model subsequently, the collection position of the top view image is also determined according to the position of the recognition camera.

[0073] Step 203: Obtain A target collection points in the virtual placement area.

[0074] The target collection points can be the positions for collecting the top view image of the three-dimensional model. The quantity of A is not limited. For example, it can be 5. The 5 target collection points can include the center point and four points near the corners.

[0075] In one embodiment, obtaining A target collection points in the virtual placement area includes: determining the center point of the virtual placement area as the first collection point; determining a reference point from the target polygon formed by all the contact points; determining the target distance from the reference point to each vertex of the target polygon; based on the maximum distance among the target distances, determining four second collection points in the four corner regions of the virtual placement area; where the distance from the second collection point to the edge of the virtual placement area is greater than or equal to the maximum distance; and taking the first collection point and the four second collection points as the A target collection points in the virtual placement area.

[0076] In this embodiment, the center point of the virtual placement area can be determined as the first collection point. For example, if the virtual placement area is rectangular, the intersection of the two diagonals of the rectangle is the first collection point. When determining the second collection points, the shape of the three-dimensional model can be considered to ensure that when the three-dimensional model rotates around the second collection point, the three-dimensional model does not exceed the range of the virtual placement area. For example, a reference point is determined from the target polygon formed by the contact points. It should be understood that when there are three contact points, the target polygon is a triangle, and when there are four contact points, the target polygon is a quadrilateral. The reference point can be considered as the center position where the three-dimensional model rotates when taking pictures at different rotation angles, not necessarily the center point of the actual projection. For example, the projection point of the highest point of the three-dimensional model can also be used as the reference point, without limitation.

[0077] In one embodiment, determining a reference point from a target polygon formed by all contact points includes: determining the highest point of the 3D model based on the horizontal plane; using the projection point of the highest point in the target polygon as the reference point.

[0078] In this embodiment, the schematic diagram for determining the reference point of the 3D model is as Figure 4 , and the projection point of the highest point of the 3D model based on the horizontal plane in the target polygon can be used as the reference point. Among them, the outer rectangular area is the virtual placement area, X, Y, and Z are the contact points of the 3D model and the horizontal plane where the virtual placement area is located, XYZ encloses a triangle, H is the highest point of the 3D model based on the horizontal plane where the virtual placement area is located, D is the projection point of H in the virtual placement area, and D can be used as the reference point of the 3D model.

[0079] After determining the reference point, the target distances from the reference point to each vertex of the target polygon can be obtained, and thus the maximum distance among the target distances can be determined. For example, the line segment lengths of XD, YD, and ZD can be determined, and the largest line segment length can be used as the maximum distance. The second collection point can be set at a position greater than or equal to the maximum distance from the edge of the virtual placement area. The schematic diagram of the target collection point in the virtual placement area is as Figure 5 , for example, the virtual placement area is a rectangular area with a length of 700 mm and a width of 500 mm, the maximum distance calculated among the target distances is 100 mm. Taking the lower left corner of the virtual placement area as the origin of the coordinate system as an example, the lower left corner is (0, 0), and the upper right corner is (700, 500). Then the first collection point is the center point of the virtual placement area, that is, (350, 250), and the second collection points at the other four corners are (100, 100), (600, 100), (100, 400), and (600, 400) respectively.

[0080] In one embodiment, based on the maximum distance among the target distances, four second collection points are determined in the four corner areas of the virtual placement area, including: obtaining the projection area of the 3D model on the horizontal plane; determining the farthest distance between the reference point and all points in the projection area; based on the maximum value of the farthest distance and the maximum distance, determining four second collection points in the four corner areas of the virtual placement area; where the second collection point is at a distance equal to the maximum value from the edge of the virtual placement area.

[0081] The actual shape of the workpiece to be registered may not be very regular. If only the maximum distance in the target distance is used to calculate the second acquisition point, there may still be some areas that will exceed the range of the virtual placement area when the workpiece rotates around the acquisition point. In this embodiment, the projection area of the three-dimensional model on the horizontal plane can be combined to determine the maximum distance from the reference point to all points in the projection area. According to the maximum value of the maximum distance and the maximum distance, four second acquisition points are determined in the four corner areas of the virtual placement area, which can ensure that the workpiece will not exceed the range of the virtual placement area no matter how it rotates, and ensure that the three-dimensional model can be completely covered during image acquisition.

[0082] Step 204: Respectively collect B top-down images of the three-dimensional model at different rotation angles when the three-dimensional model is at each target acquisition point, and the corresponding attitude coordinate points of each top-down image, to obtain the first numerical top-down images; where the first numerical value is equal to A multiplied by B; the attitude coordinate points are used to represent the coordinates and angles of the three-dimensional model in the virtual placement area.

[0083] When the three-dimensional model is at each target acquisition point, B top-down images of the three-dimensional model at different rotation angles can be collected. For example, at the first acquisition point, which is the center point of the virtual placement area, a top-down image can be collected every 45° rotation, for a total of 8 images. Then, the three-dimensional model is sequentially moved to the other four acquisition points, collecting a total of 5×8 = 40 top-down images (i.e., the first numerical value is 40 at this time). At the same time, the corresponding attitude coordinate points of each top-down image are recorded. The attitude coordinate points can include the coordinates and the rotation angle of the three-dimensional model. The attitude of the three-dimensional model corresponding to the first numerical top-down images can also be called the preset attitude of the three-dimensional model.

[0084] It should be noted that the 5 target acquisition points exemplified in the above embodiments are only for illustration. In actual applications, there may be other numbers of target acquisition points, or target acquisition points at other positions, without limitation. Similarly, the rotation angle at the same target acquisition point can also be rotated by other angles each time, such as rotating 30° each time (collecting 12 images at one acquisition point), rotating 22.5° each time (collecting 16 images at one acquisition point), or rotating 60° each time (collecting 6 images at one acquisition point), 90° (collecting 4 images at one acquisition point), etc., without limitation.

[0085] In one embodiment, when the 3D model is at each target acquisition point, B top-down images at different rotation angles are respectively acquired, along with the pose coordinate points corresponding to each top-down image, to obtain the first numerical number of top-down images, including: obtaining the height information of the recognition camera from the passing track; when the 3D model is translated to a reference point at any one of the target acquisition points, based on the height information, B top-down images of the 3D model at different rotation angles with the target line as the axis are acquired, along with the pose coordinate points corresponding to each top-down image; wherein, the target line is perpendicular to the horizontal plane and intersects the horizontal plane at the reference point; when the acquisition of the B top-down images at the current target acquisition point is completed, the 3D model is translated to the next target acquisition point, and the steps of acquiring B top-down images of the 3D model at different rotation angles with the target line as the axis and the pose coordinate points corresponding to each top-down image are continued until all target acquisition points are acquired, to obtain the first numerical number of top-down images.

[0086] In this embodiment, when acquiring the top-down images, the height information of the recognition camera from the passing track can be combined, and the top-down images are also acquired in the software according to this height information, so that the obtained top-down images are closer to the images actually taken by the recognition camera. When the 3D model rotates, it rotates at a preset angle with the target line perpendicular to the horizontal plane and intersecting the horizontal plane at the reference point as the axis. For example, one top-down image is acquired every 45° rotation, that is, 8 top-down images at different angles are acquired at each target acquisition point. A total of 40 top-down images and 40 pose coordinate points are obtained by acquiring 8 top-down images at each target acquisition point in sequence.

[0087] Step 205: Register the first numerical number of top-down images to the workpiece recognition model, and perform recognition tests on the 3D models randomly set in the virtual placement area based on the workpiece recognition model.

[0088] For example, after obtaining 40 top-down images and 40 pose coordinate points, a simulation recognition test can be carried out in the software. The 40 top-down images can be registered to the workpiece recognition model, and a simulation recognition test can be performed in the workpiece recognition model. During the test, the 3D models can be randomly placed in the virtual placement area in batches at different angles and positions, and the 3D models are recognized by the workpiece recognition model. For example, it can be recognized 10,000 times continuously. It should be noted that when placing the 3D models, it can be limited to be placed within the area enclosed by four second acquisition points.

[0089] It should be noted that the workpiece recognition model can be a neural network model trained based on the 3D model, which is used to perform simulation recognition on the 3D models of workpieces. The training process of the neural network model will not be described in detail here.

[0090] Step 206: Obtain C pose coordinate points at which the workpiece recognition model misrecognizes the 3D model under a preset number of test times.

[0091] The preset number of test times can be 5000 times, 10000 times or other preset number of test times, without limitation. When misrecognition occurs, obtain the pose coordinate points of misrecognition. For example, if C is 10, then obtain the pose coordinate points of the positions and angles of the 3D model placed corresponding to these 10 misrecognitions. If an error occurs, it means that the image in this pose also needs to be registered during workpiece registration.

[0092] Step 207: Based on the recognition camera, perform image acquisition on the workpiece to be registered placed on the passing track according to the second numerical value of pose coordinate points respectively, so as to register the workpiece to be registered according to the acquired images; where the second numerical value is the sum of the first numerical value and C.

[0093] For example, if the first numerical value is 40 and C is 10, then the second numerical value is 50, that is, place the workpiece to be registered on the passing track according to these 50 pose coordinate points for image acquisition, and register the workpiece to be registered according to the acquired images.

[0094] This method performs virtual photographing registration on the 3D model of the workpiece to be registered according to the actual recognition scenario. Obtain the first numerical value of top-down images and the pose coordinate points corresponding to each top-down image according to B top-down images at different rotation angles respectively collected at A target collection points in the virtual placement area, and register the first numerical value of top-down images to the workpiece recognition model. Perform batch virtual recognition tests on the 3D model randomly set in the virtual placement area to obtain the misrecognized pose coordinate points. Finally, based on the recognition camera, perform image acquisition and registration on the workpiece to be registered placed on the passing track according to all the obtained pose coordinate points above. Since the acquired images cover various preset poses of the workpiece to be registered and the poses at which the workpiece recognition model misrecognizes, the poses of the workpiece to be registered can be fully collected, so that the recognition success rate after workpiece registration meets the recognition requirements.

[0095] In one embodiment, based on the recognition camera, performing image acquisition on the workpiece to be registered placed on the passing track according to the second numerical value of pose coordinate points respectively, so as to register the workpiece to be registered according to the acquired images includes: inputting the second numerical value of pose coordinate points into the registration system; obtaining the pose virtual image of the workpiece to be registered generated by the registration system based on the pose coordinate points; based on the recognition camera, collecting the camera images of the workpiece to be registered placed on the passing track according to the pose virtual image; registering the workpiece to be registered according to the second numerical value of camera images.

[0096] In this embodiment, taking the second numerical value including 50 pose coordinate points as an example, the 50 pose coordinate points are input into the registration system. The registration system will pop up the virtual images of the poses where the workpiece should be placed one by one, allowing people to place the workpiece according to the requirements. The recognition camera collects the camera images of the workpiece to be registered placed on the passing track according to the virtual images of the poses. The workpiece to be registered is registered based on the obtained 50 camera images. Since the 50 collected camera images cover various preset poses of the workpiece to be registered and the poses where the workpiece recognition model makes mistakes, the poses of the workpiece to be registered can be fully collected, so that the recognition success rate after workpiece registration meets the recognition requirements. After verification, using the workpiece registration method of the above embodiment of the present application, performing model recognition entry according to the simulated poses can increase the original recognition success rate of 90% to more than 99.9%, ensuring the workpiece recognition rate.

[0097] Based on the same inventive concept, the second embodiment of the present application provides a workpiece registration device, as Figure 6 , the device includes:

[0098] The first determination module 601 is used to determine the effective placement area of the workpiece to be registered according to the shooting coverage range of the recognition camera and the track width of the passing track;

[0099] The orientation adjustment module 602 is used to set the three-dimensional model of the workpiece to be registered on the horizontal plane where the virtual placement area determined based on the effective placement area is located according to a preset orientation; wherein, the contact points between the three-dimensional model and the horizontal plane include at least three points not on the same straight line;

[0100] The first acquisition module 603 is used to acquire A target acquisition points of the virtual placement area;

[0101] The pose coordinate point acquisition module 604 is used to respectively acquire B top-down images at different rotation angles when the three-dimensional model is at each of the target acquisition points, and the pose coordinate points corresponding to each of the top-down images, to obtain the first numerical value of the top-down images; wherein, the first numerical value is equal to A multiplied by B; the pose coordinate points are used to represent the coordinates and angles of the three-dimensional model in the virtual placement area;

[0102] The recognition test module 605 is used to register the first numerical value of the top-down images into the workpiece recognition model, and perform recognition tests on the three-dimensional model randomly set in the virtual placement area based on the workpiece recognition model;

[0103] The second acquisition module 606 is used to acquire C pose coordinate points where the workpiece recognition model makes mistakes in recognizing the three-dimensional model under a preset number of test times;

[0104] An image acquisition module 607 is configured to perform image acquisition on the workpiece to be registered placed on the passing track according to the second number of the pose coordinate points based on the recognition camera, so as to register the workpiece to be registered according to the acquired image; wherein, the second number is the sum of the first number and C.

[0105] Based on the three-dimensional model of the workpiece to be registered, the device performs virtual photographing registration according to the actually recognized scene, obtains the first number of top-view images and the pose coordinate points corresponding to each top-view image by respectively collecting B top-view images at different rotation angles according to A target acquisition points in the virtual placement area, and registers the first number of top-view images into the workpiece recognition model. Then, batch virtual recognition tests are performed according to the three-dimensional model randomly set in the virtual placement area to obtain the pose coordinate points with recognition errors. Finally, based on the recognition camera, image acquisition and registration are performed on the workpiece to be registered placed on the passing track according to all the obtained pose coordinate points. Since the acquired images cover various preset poses of the workpiece to be registered and the poses with recognition errors of the workpiece recognition model, the poses of the workpiece to be registered can be fully collected, so that the recognition success rate after workpiece registration meets the recognition requirements.

[0106] As Figure 7 shown, an embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114. Among them, the processor 111, the communication interface 112, and the memory 113 complete mutual communication through the communication bus 114.

[0107] The memory 113 is used to store a computer program.

[0108] In an embodiment of the present application, when the processor 111 is used to execute the program stored on the memory 113, it implements the workpiece registration method provided by any one of the foregoing method embodiments.

[0109] The communication bus mentioned in the above terminal may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity, only a thick line is used in the figure to represent it, but it does not mean that there is only one bus or one type of bus.

[0110] The communication interface is used for communication between the above terminal and other devices.

[0111] The memory may include a Random Access Memory (RAM), or may also include a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located far from the aforementioned processor.

[0112] The aforementioned processor may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0113] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the workpiece registration method provided in any one of the foregoing method embodiments.

[0114] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0115] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the related technology, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0116] It should be understood that the terms used herein are for the purpose of describing particular example embodiments only and are not intended to be limiting. Unless the context clearly dictates otherwise, the singular forms "a", "an", and "the" as used herein may also include the plural forms. The terms "comprising", "including", "containing", and "having" are inclusive and thus specify the presence of the stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not to be construed as necessarily requiring them to be performed in the particular order described or illustrated, unless the order of performance is explicitly stated. It should also be understood that additional or alternative steps may be used.

[0117] It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present application and are not intended to limit the present application. In the description, the suffixes such as "module", "component", or "unit" used to denote elements are only for the convenience of describing the present application and have no specific meaning in themselves. Therefore, "module", "component", or "unit" may be used interchangeably.

[0118] The above are only the specific embodiments of the present application, enabling those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features claimed herein.

Claims

1. A workpiece registration method, characterized in that, The method includes: Determining the effective placement area of the workpiece to be registered according to the shooting coverage of the recognition camera and the track width of the through-type track; Setting the three-dimensional model of the workpiece to be registered on the horizontal plane where the virtual placement area determined based on the effective placement area is located according to a preset orientation; wherein, the contact points between the three-dimensional model and the horizontal plane include at least three points not on the same straight line; Obtaining A target acquisition points in the virtual placement area; Respectively collecting B top-down images at different rotation angles when the three-dimensional model is at each of the target acquisition points, and the attitude coordinate points corresponding to each of the top-down images, to obtain the first numerical number of the top-down images; wherein, the first numerical number is equal to A multiplied by B; the attitude coordinate points are used to represent the coordinates and angles of the three-dimensional model in the virtual placement area; Registering the first numerical number of the top-down images into the workpiece recognition model, and performing recognition tests on the three-dimensional model randomly set in the virtual placement area based on the workpiece recognition model; Obtaining C attitude coordinate points at which the workpiece recognition model misidentifies the three-dimensional model under a preset number of test times; Performing image acquisition on the workpiece to be registered placed on the through-type track according to the second numerical number of the attitude coordinate points by the recognition camera, so as to register the workpiece to be registered according to the acquired images; wherein, the second numerical number is the sum of the first numerical number and C.

2. The method according to claim 1, characterized in that, Obtaining A target acquisition points in the virtual placement area includes: Determining the center point of the virtual placement area as the first acquisition point; Determining a reference point from the target polygon formed by all the contact points; Determining the target distance from the reference point to each vertex of the target polygon; Based on the maximum distance among the target distances, determining four second acquisition points in the four corner regions of the virtual placement area; wherein, the second acquisition points are at a distance greater than or equal to the maximum distance from the edge of the virtual placement area; Taking the first acquisition point and the four second acquisition points as the A target acquisition points in the virtual placement area.

3. The method according to claim 2, wherein Determining a reference point from the target polygon formed by all the contact points includes: Determining the highest point of the three-dimensional model based on the horizontal plane; Taking the projection point of the highest point in the target polygon as the reference point.

4. The method according to claim 2, characterized in that Based on the maximum distance among the target distances, determining four second acquisition points in the four corner regions of the virtual placement area includes: Obtaining the projection area of the three-dimensional model on the horizontal plane; Determining the farthest distance between the reference point and all points in the projection area; Based on the maximum value among the farthest distance and the maximum distance, determining four second acquisition points in the four corner regions of the virtual placement area; wherein, the second acquisition points are at a distance equal to the maximum value from the edge of the virtual placement area.

5. The method according to claim 2, wherein Respectively collecting B top-down images at different rotation angles when the three-dimensional model is at each of the target acquisition points, and the attitude coordinate points corresponding to each of the top-down images, to obtain the first numerical number of the top-down images, includes: Obtaining the height information of the recognition camera from the through-type track; When translating the 3D model so that the reference point is at any one of the target acquisition points, based on the height information, collect B top-down images of the 3D model at different rotation angles with the target line as the axis, and the pose coordinate points corresponding to each top-down image; wherein, the target line is perpendicular to the horizontal plane and intersects the horizontal plane at the reference point. When the collection of the B top-down images at the current target acquisition point is completed, translate the 3D model to the next target acquisition point, and continue to execute the step of collecting B top-down images of the 3D model at different rotation angles with the target line as the axis, and the pose coordinate points corresponding to each top-down image, until the collection of all target acquisition points is completed, obtaining the first numerical number of top-down images.

6. The method according to claim 1, characterized in that Based on the identification camera, perform image acquisition on the workpiece to be registered placed on the passing track according to the second numerical number of pose coordinate points, so as to register the workpiece to be registered according to the acquired images, including: Input the second numerical number of pose coordinate points into the registration system; Obtain the pose virtual image of the workpiece to be registered generated by the registration system based on the pose coordinate points; Based on the identification camera, collect the camera image of the workpiece to be registered placed on the passing track according to the pose virtual image; Register the workpiece to be registered according to the second numerical number of camera images.

7. The method according to claim 1, characterized in that Set the 3D model of the workpiece to be registered on the horizontal plane where the virtual placement area determined based on the placement valid area is located according to a preset orientation, including: Obtain the 3D model of the workpiece to be registered; Determine the virtual placement area of the 3D model based on the placement valid area; Determine the riser position of the 3D model; Adjust the 3D model to the orientation where the riser position is away from the horizontal plane, and set the 3D model on the virtual placement area.

8. A workpiece registration device, characterized in that, The device includes: The first determination module is used to determine the placement valid area of the workpiece to be registered according to the shooting coverage range of the identification camera and the track width of the passing track; The orientation adjustment module is used to set the 3D model of the workpiece to be registered on the horizontal plane where the virtual placement area determined based on the placement valid area is located according to a preset orientation; wherein, the contact points between the 3D model and the horizontal plane include at least three points not on the same straight line; The first acquisition module is used to acquire A target acquisition points of the virtual placement area; The pose coordinate point acquisition module is used to respectively acquire B top-down images of the 3D model at different rotation angles and the pose coordinate points corresponding to each top-down image when the 3D model is at each target acquisition point, obtaining the first numerical number of top-down images; wherein, the first numerical number is equal to A multiplied by B; the pose coordinate points are used to represent the coordinates and angles of the 3D model in the virtual placement area; The identification test module is used to register the first numerical number of top-down images into the workpiece identification model, and perform identification test on the 3D model randomly set in the virtual placement area based on the workpiece identification model. A second acquisition module, configured to acquire C pose coordinate points at which the workpiece recognition model makes recognition errors for the three-dimensional model under a preset number of test times; An image acquisition module, configured to perform image acquisition on the workpiece to be registered placed on the passing track according to the second numerical number of the pose coordinate points respectively based on the recognition camera, so as to register the workpiece to be registered according to the acquired images; wherein, the second numerical number is the sum of the first numerical number and C.

9. An electronic device, characterized in that, It includes a processor, a communication interface, a memory and a communication bus. Among them, the processor, the communication interface and the memory complete mutual communication through the communication bus; The memory is used for storing computer programs; The processor is configured to implement the workpiece registration method according to any one of claims 1-7 when executing the programs stored on the memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program, when executed by the processor, implements the workpiece registration method according to any one of claims 1-7.

Images