Workpiece positioning method and positioning system based on multi-view vision

Abstract

The invention discloses a workpiece positioning method and system based on multi-view vision, and the method comprises the steps: collecting a first image of a to-be-detected workpiece through a first camera when the to-be-detected workpiece reaches a robot station, and collecting a second image of the to-be-detected workpiece through a plurality of second cameras; the workpiece positioning method comprises a coarse positioning process and a fine positioning process. The method comprises the following steps: roughly positioning a workpiece by using a first camera, and determining the approximate position of each part feature of the to-be-detected workpiece; then the workpieces are accurately positioned by a plurality of second cameras, and the accurate poses of the workpieces in the space are calculated by utilizing the corresponding relationship between the recognized 2D image feature points and the 3D workpiece contour feature template, so that the view of the cameras can effectively cover the workpieces of various sizes, and the workpieces can be accurately positioned under the large-view working condition; conditions are provided for operations such as accurate grabbing of workpieces in subsequent procedures, the production efficiency and stability are improved, and the production cost is reduced.

Description

Technical Field

[0001] The present invention relates to the field of vision guidance technology, and in particular to a workpiece positioning method and positioning system based on multi-eye vision. Background Art

[0002] At present, the new energy vehicle industry is showing a booming development trend, and automobile manufacturers are in urgent need of seeking automated ways to replace manual production in order to improve production efficiency and reduce labor costs.

[0003] At present, the degree of automation in the welding workshop of automobile production plants is relatively high, and many workstations use visual technology and robotic arms to complete various processes such as loading, welding, and gluing. However, in the sorting and framing process of the stamping workshop, manual sorting is still the main method in major automobile production plants.

[0004] In the existing automated positioning and guidance operation scheme in the stamping workshop, a single 2D camera is generally used to locate the position of the stamped parts. The parts stamped out from the stamping workshop are transported to the robot arm operation station via a conveyor belt. The photoelectric sensor triggers the 2D camera above the conveyor belt to take pictures, collect image data of the stamped parts on the conveyor belt, and then the visual positioning algorithm calculates the position of the stamped parts in the robot arm base coordinate system, guiding the robot arm to complete the grasping of the stamped parts. However, the existing single 2D camera positioning method has the following main disadvantages:

[0005] 1. There are many types of parts in the stamping workshop, and the size of stamping parts varies greatly. However, the visual shooting field of view is limited, and it is difficult to cover all types of parts of different sizes. Therefore, the 2D camera needs to have a large enough field of view. However, because the stamping parts need to be accurately loaded into the card slot of the material frame during subsequent framing, when grabbing the workpieces flowing out of the upstream production process, a certain grasping accuracy must be met to meet the subsequent framing requirements, and cameras with a large field of view are difficult to meet the grasping accuracy requirements. The visual positioning accuracy of the camera that meets the requirements of a large field of view is not high, resulting in grasping offset, which will affect the accuracy of subsequent stamping parts framing.

[0006] 2. The contour recognition algorithm for 2D images collected by 2D cameras is sensitive to image noise and interference, and is prone to misrecognition due to conveyor belt scratches, light, shadows, etc., reducing the capture accuracy and stability.

[0007] 3. Since stamped parts cannot generally be placed flat on a conveyor belt, they will present different placement postures on the horizontal conveyor belt, which will cause changes in the contours in the 2D image and lead to recognition deviations.

[0008] In summary, due to the low positioning and recognition accuracy of the existing single 2D camera workpiece positioning method, which leads to deviations in grasping, stamping workshops often need to spend more costs to produce material frames with higher precision to ensure that the robotic arm can stably place the grasped stamping parts accurately into the card slots of the material frame. Summary of the invention

[0009] The purpose of the present invention is to provide a workpiece positioning method and positioning system based on multi-eye vision to address the deficiencies of the above-mentioned prior art, so as to solve at least one of the above-mentioned prior art problems, effectively realize the camera field of view covering workpieces of various sizes, realize precise positioning of workpieces under large field of view conditions, and provide conditions for operations such as precise grasping of workpieces in subsequent processes.

[0010] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0011] A workpiece positioning method based on multi-eye vision is characterized in that when a workpiece to be measured arrives at a robot station, a first camera collects a first image of the workpiece to be measured, and a plurality of second cameras collect a second image of the workpiece to be measured; the workpiece positioning method includes a coarse positioning process and a fine positioning process, wherein:

[0012] The coarse positioning process includes:

[0013] Calibrate the transformation matrix T from the camera coordinate system of the first camera to the robot base coordinate system h And the transformation matrix from the camera coordinate system of each second camera to the robot base coordinate system Wherein, i represents the i-th second camera;

[0014] Based on the first image and a pre-made workpiece contour feature template, the rough pose of the workpiece to be measured in the robot arm base coordinate system is identified and located;

[0015] Based on the transformation matrix The rough pose of the workpiece to be measured is transmitted to a plurality of second cameras respectively to obtain the region of interest of the workpiece template feature in each second image;

[0016] The precise positioning process includes:

[0017] Based on the second image, obtain the corresponding coordinate m of the jth feature on the workpiece to be measured in the camera coordinate system of the i-th second camera ij ;

[0018] Based on the transformation matrix m ij and the position M of the jth feature in the second image captured by the i-th second camera in the workpiece contour feature template ij , obtain the transformation T from the workpiece to be measured to the camera coordinate system of the i-th second camera i (Ri ,t i );

[0019] Integrated conversion of multiple second cameras i (R i ,t i ) to obtain the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system.

[0020] As a preferred embodiment, the workpiece contour feature template pre-production process includes:

[0021] Collect characteristic parameters of the middle area of ​​the workpiece template;

[0022] The collected feature parameters are saved as feature vectors and made into workpiece contour feature templates.

[0023] As a preferred method, the transformation T of the workpiece to be measured to the camera coordinate system of the i-th second camera is obtained based on the least squares method. i (R i ,t i ), where the objective function of the least squares method is:

[0024]

[0025] Where N represents the total number of second cameras, j represents the jth feature, and M i represents the number of features under the i-th second camera, R i represents the rotation matrix between the workpiece to be measured and the i-th second camera, t i represents the translation matrix between the workpiece to be measured and the i-th second camera, 1 z =[0,0,1] is a normalized vector.

[0026] As a preferred method, the conversion T of multiple second cameras is integrated i (R i ,t i ), obtaining the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system includes:

[0027] For multiple second cameras, we get

[0028] Transform the above formula into

[0029] Minimize the previous formula to obtain the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system;

[0030] Among them, R represents the rotation matrix of the workpiece to be measured in the robot arm base coordinate system, and t represents the translation matrix of the workpiece to be measured in the robot arm base coordinate system.

[0031] As a preferred approach, minimize The position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system includes:

[0032] The formula Transformed to

[0033]

[0034] Transform the previous formula into a quadratic formula;

[0035] Solve the minimum value of the quadratic formula to obtain the position and posture T(R, t) of the workpiece to be measured in the robot base coordinate system.

[0036] As a preferred method, the quadratic formula obtained by transformation includes:

[0037]

[0038] in, I3 represents the third-order identity matrix;

[0039] When solving the minimum value of the quadratic formula, the rotation matrix R is constrained to be an orthogonal matrix.

[0040] As a preferred method, a sequential quadratic optimization method is used to solve the minimum value of the quadratic formula.

[0041] As a preferred embodiment, the method of solving the minimum value of the quadratic formula by using the sequential quadratic optimization method includes:

[0042] Using Taylor expansion and Lagrange multiplier method, the nonlinear quadratic optimization problem is transformed into a linear quadratic optimization problem.

[0043] The minimum value of the quadratic formula is obtained by cyclically iterating the calculation according to the sequential quadratic optimization method.

[0044] As a preferred embodiment, the method of identifying and locating the rough pose of the workpiece to be measured in the robot arm base coordinate system based on the first image and the pre-made workpiece contour feature template includes:

[0045] Based on the workpiece contour feature template, identifying feature information of the workpiece to be measured in the first image;

[0046] Calculating the offset of the workpiece to be measured in the camera coordinate system of the first camera based on the feature information of the workpiece to be measured in the first image and the workpiece contour feature template;

[0047] The offset information is converted to the robot base coordinate system to identify the position of the workpiece to be measured in the robot base coordinate system.

[0048] Based on the same inventive concept, the present invention also provides a workpiece positioning system based on multi-eye vision, which is characterized in that the workpiece positioning system is used to implement the workpiece positioning method.

[0049] Compared with the prior art, the present invention firstly utilizes the first camera to roughly locate the workpiece to determine the approximate position of the features of each part of the workpiece to be measured; then, multiple second cameras are used to accurately locate the workpiece, and the correspondence between the identified 2D image feature points and the 3D workpiece contour feature template is used to calculate the accurate position and posture of the workpiece in space, thereby effectively realizing that the camera field of view covers workpieces of various sizes, and realizing accurate positioning of workpieces under large field of view conditions, providing conditions for operations such as accurate grasping of workpieces in subsequent processes, improving production efficiency and stability, and reducing production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Schematic diagram of the positional relationship between each camera and the workpiece to be measured in the present invention.

[0051] Among them, A is the workpiece to be measured, C1 is the first camera, and C2 and C3 are the second cameras. DETAILED DESCRIPTION

[0052] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described below in conjunction with the drawings in the embodiment. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0053] The present invention provides a high-precision workpiece positioning method and positioning system based on multi-eye vision, including a first camera and a plurality of second cameras. Figure 1 As shown, the first camera C1 is located in the middle area of ​​the plurality of second cameras C2 and C3, and the first camera captures the image of the middle area of ​​the workpiece A to be measured. Figure 1 Only two second cameras are shown in the figure. In other implementations, the number of second cameras may be greater.

[0054] In the present invention, the first camera and the plurality of second cameras are preferably 2D cameras. The first camera has a larger field of view but lower precision, and the second camera has a smaller field of view but higher precision.

[0055] First, the workpiece is roughly positioned using a first camera with a larger field of view, and the contour of the part of the workpiece to be measured is recognized to determine the approximate position of the features of each part of the workpiece to be measured.

[0056] Then, multiple second cameras with small fields of view but higher accuracy are used to accurately locate the workpiece. The correspondence between the identified 2D image feature points and the 3D workpiece contour feature template is used to locate the local features of the workpiece to be measured and calculate its posture. The accurate position and posture of the workpiece in space are calculated, realizing high-precision positioning and identification of workpieces of various sizes to be measured.

[0057] Taking the workpiece to be tested as a stamping part as an example, the present invention can improve the positioning accuracy of the stamping part, thereby realizing high-precision positioning guidance, improving the positioning and grasping accuracy of subsequent processes, and meeting the subsequent stamping part framing requirements. The present invention reduces the requirements for the accuracy of the material frame slot by improving the visual accuracy in the grasping stage, thereby reducing the cost of customizing the material frame for automobile factories and improving production efficiency and stability.

[0058] The first aspect of the embodiment of the present invention provides a workpiece positioning method based on multi-eye vision, when the workpiece to be measured is transported to the robot station through a conveyor belt, the photoelectric sensor triggers the camera to take pictures at the same time. The first camera captures the first image of the workpiece to be measured, and multiple second cameras capture the second image of the workpiece to be measured; the workpiece positioning method includes a coarse positioning process and a fine positioning process.

[0059] The coarse positioning process includes:

[0060] Hand-eye calibration, calibrate the transformation matrix T from the camera coordinate system of the first camera to the robot base coordinate system h And the transformation matrix from the camera coordinate system of each second camera to the robot base coordinate system Wherein, i represents the i-th second camera;

[0061] Based on the first image and a pre-made workpiece contour feature template, the rough pose of the workpiece to be measured in the robot arm base coordinate system is identified and located;

[0062] Based on the transformation matrix The rough pose of the workpiece to be measured is transmitted to multiple second cameras respectively to obtain the workpiece template feature in the area of ​​interest of each second image.

[0063] Taking stamping parts as an example, since the size of stamping parts varies greatly, ranging from 1 meter to 3.7 meters, if the traditional overall contour recognition solution is used, it is difficult to find a suitable camera that can meet the recognition requirements of stamping parts in such a large range of sizes. Therefore, the present invention adopts a local recognition method to locate the local contour features of the stamping parts under a camera with a larger field of view (the first camera).

[0064] The local contour matching method in the rough positioning stage can roughly locate the position of the workpiece to be measured in the first camera coordinate system, but because the first camera has a large field of view and a low resolution, it is difficult to achieve a high positioning accuracy. In addition, since the 2D camera cannot calculate the 3D posture of the stamping parts on the conveyor belt, some stamping parts on the conveyor belt will be tilted due to their own shape, which makes it difficult for the robot to accurately grasp the stamping parts when sucking the stamping parts in the horizontal direction, resulting in grasping failure or grasping with a large deviation, affecting the success rate of subsequent framing in the precision-free material frame. Therefore, after the present invention obtains the deviation of the workpiece in the horizontal direction through rough positioning, this deviation data is used to perform the next step of fine positioning.

[0065] The precise positioning process includes:

[0066] Based on the second image, obtain the corresponding coordinate m of the jth feature on the workpiece to be measured in the camera coordinate system of the i-th second camera ij (Normalized coordinates, i.e. coordinate values ​​when Z=1);

[0067] Based on the transformation matrix m ij and the position M of the jth feature in the second image captured by the i-th second camera in the workpiece contour feature template ij , obtain the transformation T from the workpiece to be measured to the camera coordinate system of the i-th second camera i (R i ,t i );

[0068] Integrated conversion of multiple second cameras i (R i ,t i ) to obtain the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system.

[0069] The workpiece contour feature template pre-production process preferably includes but is not limited to:

[0070] Collecting contour feature parameters of the middle area of ​​the workpiece template;

[0071] The collected feature parameters are saved as feature vectors and made into workpiece contour feature templates.

[0072] Among them, obvious features such as round holes and corner points in the middle area of ​​the workpiece template can be used as contour features, and the feature parameters of these obvious features are collected when making the workpiece contour feature template.

[0073] The present invention preferably but not limited to obtain the transformation T of the workpiece to be measured to the camera coordinate system of the i-th second camera based on the least squares method i (R i ,t i), where the objective function of the least squares method is:

[0074]

[0075] Where N represents the total number of second cameras, j represents the jth feature, and M i represents the number of features under the i-th second camera, R i represents the rotation matrix between the workpiece to be measured and the i-th second camera, t i represents the translation matrix between the workpiece to be measured and the i-th second camera, 1 z =[0,0,1] is a normalized vector.

[0076] Integrated conversion of multiple second cameras i (R i ,t i ), obtaining the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system preferably includes but is not limited to:

[0077] For multiple second cameras, we get

[0078] Transform the above formula into

[0079] Further transformed into

[0080] Minimize the previous formula to obtain the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system;

[0081] Among them, R represents the rotation matrix of the workpiece to be measured in the robot arm base coordinate system, and t represents the translation matrix of the workpiece to be measured in the robot arm base coordinate system.

[0082] minimize The position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system is preferably obtained, but not limited to, including:

[0083] The formula Transformed to

[0084]

[0085] Transform the previous formula into a quadratic formula;

[0086] Solve the minimum value of the quadratic formula to obtain the position and posture T(R, t) of the workpiece to be measured in the robot base coordinate system.

[0087] The quadratic formula obtained by transformation preferably includes but is not limited to:

[0088]

[0089] in, I3 represents the third-order identity matrix;

[0090] When solving the minimum value of the quadratic form formula, the constraint rotation matrix R is an orthogonal matrix. Since R in the above formula is a rotation transformation matrix, it is necessary to bring constraints on R when solving the minimum value of this quadratic form, so as to ensure that the optimization solution process of this constrained quadratic form can obtain R that meets the specifications.

[0091] The present invention preferably, but not limited to, adopts a sequential quadratic optimization method to solve the minimum value of the quadratic formula.

[0092] To solve the nonlinear constraints The method of using the sequential quadratic optimization method to solve the minimum value of the quadratic formula preferably includes but is not limited to:

[0093] Using Taylor expansion and Lagrange multiplier method, the nonlinear quadratic optimization problem is transformed into a linear quadratic optimization problem.

[0094] According to the sequential quadratic optimization method (SQP) iterative calculation, we know that the calculation obtains δ 2 The optimal solution of the quadratic formula is obtained. According to the above formula and the SQP solution algorithm, we can calculate the position transformation of the stamping workpiece to be grasped in the robot base coordinate system, so that we can accurately guide the robot arm to complete the accurate grasping of various types of stamping workpieces.

[0095] The method of identifying and locating the rough pose of the workpiece to be measured in the robot arm base coordinate system based on the first image and the pre-made workpiece contour feature template preferably includes but is not limited to:

[0096] Based on the workpiece contour feature template, identifying feature information of the workpiece to be measured in the first image;

[0097] Calculating the offset of the workpiece to be measured in the camera coordinate system of the first camera based on the feature information of the workpiece to be measured in the first image and the workpiece contour feature template;

[0098] The offset information is converted to the robot base coordinate system to identify the position of the workpiece to be measured in the robot base coordinate system.

[0099] The present invention proposes to use the SQP algorithm to solve the pose transformation under the mapping relationship between the feature points of the image taken by multiple 2D cameras and the feature points of the workpiece contour feature template, so as to calculate the accurate pose of the workpiece to be measured in the image camera coordinate system space composed of multiple 2D cameras. After testing, the positioning accuracy of the present invention can reach +-0.5mm, which meets the accuracy requirements of the subsequent framing of stamping parts. At the same time, the SQP algorithm is highly efficient and can complete the calculation within 1s. The time efficiency can also meet the process requirements of framing in the stamping workshop.

[0100] In order to obtain the 6D pose of the workpiece to be measured, it is necessary to identify the positions of multiple feature points of the workpiece to be measured (the number of feature points is ≥ 4), and use the spatial position data of these feature points in the workpiece contour feature template, combined with the calculation of the position of the workpiece to be measured in the camera coordinate system space, to obtain the pose Rt of the workpiece to be measured in the robot base coordinate system. The present invention innovatively transforms the multi-eye vision positioning problem into a quadratic optimization problem with specific constraints, and uses optimization technology to accurately obtain the accurate pose of the workpiece to be grasped when there are only 4 corresponding feature points. By innovatively using this method, the stability and accuracy of the system can be improved.

[0101] Taking the workpiece to be tested as a stamping part as an example, the present invention uses multiple 2D cameras to complete the calculation of the 6D pose of stamping parts of different sizes, and uses this to guide the robot arm to complete the grasping, which can meet the needs of high-precision framing of stamping parts in the stamping workshop. The workpiece positioning method of the present invention is not limited to stamping parts, but is also applicable to other workpieces that require positioning guidance.

[0102] A second aspect of an embodiment of the present invention provides a workpiece positioning system based on multi-eye vision, and the workpiece positioning system is used to implement the workpiece positioning method.

[0103] In summary, the present invention is mainly divided into two stages: coarse positioning and fine positioning. The rough positioning stage identifies the approximate posture of the workpiece to be measured, and transmits the offset of the workpiece to be measured under the first camera with a large field of view to multiple second cameras with small fields of view and high precision. The multiple second cameras identify the position of the feature points of the workpiece to be measured in the feature point area preset for the workpiece to be measured according to the offset data of the workpiece to be measured, and calculate the 6D posture of the entire workpiece to be measured through the 2D positions of multiple feature points of the multiple second cameras to achieve fine positioning.

[0104] The present invention combines coarse positioning with fine positioning to achieve calculation of the 6D pose of a workpiece in a scene with a large range of size variation, thereby improving positioning accuracy and operating efficiency and reducing costs.

[0105] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0106] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0107] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0108] In the embodiments provided by the present invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the above-described method and system embodiments are only illustrative, for example, the division of the modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0109] The units described as separate components may or may not be physically separated, that is, they may be located in one place, or they may be distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0110] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0111] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0112] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A workpiece positioning method based on multi-eye vision, characterized in that: When the workpiece to be measured arrives at the robot station, the first camera captures the first image of the workpiece to be measured, and the plurality of second cameras capture the second image of the workpiece to be measured; the workpiece positioning method includes a coarse positioning process and a fine positioning process, wherein: The coarse positioning process includes: Calibrate the transformation matrix T from the camera coordinate system of the first camera to the robot base coordinate system h And the transformation matrix from the camera coordinate system of each second camera to the robot base coordinate system Wherein, i represents the i-th second camera; Based on the first image and a pre-made workpiece contour feature template, identify and locate the rough pose of the workpiece to be measured in the robot arm base coordinate system; Based on the transformation matrix The rough pose of the workpiece to be measured is transmitted to a plurality of second cameras respectively to obtain the region of interest of the workpiece template feature in each second image; The precise positioning process includes: Based on the second image, obtain the corresponding coordinate m of the jth feature on the workpiece to be measured in the camera coordinate system of the i-th second camera. ij ; Based on the transformation matrix m ij and the position M of the jth feature in the second image captured by the i-th second camera in the workpiece contour feature template ij , obtain the transformation T from the workpiece to be measured to the camera coordinate system of the i-th second camera i (R i ,t i ); Integrated conversion of multiple second cameras i (R i ,t i ) to obtain the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system.

2. The workpiece positioning method based on multi-eye vision according to claim 1 is characterized in that: The workpiece contour feature template pre-production process includes: Collect characteristic parameters of the middle area of ​​the workpiece template; The collected feature parameters are saved as feature vectors and made into workpiece contour feature templates.

3. The workpiece positioning method based on multi-eye vision according to claim 1 or 2, characterized in that: Based on the least squares method, the transformation T of the workpiece to be measured to the camera coordinate system of the i-th second camera is obtained. i (R i ,t i ), where the objective function of the least squares method is: Where N represents the total number of second cameras, j represents the jth feature, and M i represents the number of features under the i-th second camera, R i represents the rotation matrix between the workpiece to be measured and the i-th second camera, t i represents the translation matrix between the workpiece to be measured and the i-th second camera, 1 z =[0,0,1] is a normalized vector.

4. The workpiece positioning method based on multi-eye vision according to claim 3 is characterized in that: Integrated conversion of multiple second cameras i (R i ,t i ), obtaining the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system includes: For multiple second cameras, we get Transform the above formula into Minimize the previous formula to obtain the position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system; Among them, R represents the rotation matrix of the workpiece to be measured in the robot arm base coordinate system, and t represents the translation matrix of the workpiece to be measured in the robot arm base coordinate system.

5. The workpiece positioning method based on multi-eye vision according to claim 4 is characterized in that: minimize The position and posture T(R,t) of the workpiece to be measured in the robot base coordinate system includes: The formula Transformed to Transform the previous formula into a quadratic formula; Solve the minimum value of the quadratic formula to obtain the position and posture T(R, t) of the workpiece to be measured in the robot base coordinate system.

6. The workpiece positioning method based on multi-eye vision according to claim 5 is characterized in that: The quadratic formulas obtained by transformation include: in, I3 represents the third-order identity matrix; When solving the minimum value of the quadratic formula, the rotation matrix R is constrained to be an orthogonal matrix.

7. The workpiece positioning method based on multi-eye vision according to claim 5 is characterized in that: The sequential quadratic optimization method is used to solve the minimum value of the quadratic formula.

8. The workpiece positioning method based on multi-eye vision according to claim 7 is characterized in that: The method of using the sequential quadratic optimization method to solve the minimum value of the quadratic formula includes: Using Taylor expansion and Lagrange multiplier method, the nonlinear quadratic optimization problem is transformed into a linear quadratic optimization problem. The minimum value of the quadratic formula is obtained by cyclically iterating the calculation according to the sequential quadratic optimization method.

9. The workpiece positioning method based on multi-eye vision according to claim 2, characterized in that: The method of identifying and locating the rough pose of the workpiece to be measured in the robot arm base coordinate system based on the first image and the pre-made workpiece contour feature template includes: Based on the workpiece contour feature template, identifying feature information of the workpiece to be measured in the first image; Calculating the offset of the workpiece to be measured in the camera coordinate system of the first camera based on the feature information of the workpiece to be measured in the first image and the workpiece contour feature template; The offset information is converted to the robot base coordinate system to identify the position of the workpiece to be measured in the robot base coordinate system.

10. A workpiece positioning system based on multi-eye vision, characterized in that: The workpiece positioning system is used to implement the workpiece positioning method according to any one of claims 1 to 9.

Images