Vamp side wall line registration method and device for robot glue application

Abstract

The application discloses a vamp side wall line registration method and device for robot glue application, and belongs to the technical field of vamp glue application. The method comprises the following steps: translating the center of a first vamp side wall line to the center of a second vamp side wall line, and performing rigid registration of the two vamp side wall lines by using an iterative closest point algorithm; projecting the two vamp side wall lines after rigid registration onto a two-dimensional plane to construct a first contour point set and a second contour point set; calculating a normal vector on each point in the first contour point set, and obtaining a corresponding point of each point in the first contour point set in the second contour point set based on the normal vector on each point; and mapping the corresponding point of each point in the first contour point set in the second contour point set to a three-dimensional space to realize registration of the two vamp side wall lines. The application can realize fast, efficient and stable flexible deformation registration.

Description

Technical Field

[0001] This invention relates to the field of shoe upper adhesive technology, specifically to a method and apparatus for aligning shoe upper sidewall lines for robotic adhesive application. Background Technology

[0002] Currently, in industrial settings, the glue application process for shoe uppers involves first manually marking lines (sidewall lines) and then manually applying the glue using these lines as a reference. However, due to the non-standard nature of shoes and the randomness of manual marking, the characteristics of the sidewall lines vary significantly between different shoe uppers. Workers rely on visual inspection to identify and compensate for these errors, resulting in low accuracy and efficiency in error identification.

[0003] With the development of machine vision technology, it has been widely used in image registration, including surface registration, point registration, and point cloud registration. For example, the invention patent application with application number 202011248876.X discloses a fast measurement algorithm for rail wear based on structured light. The steps are as follows: a) Constructing a contour through point cloud, establishing contour coordinate data, and locating the basic contour line; b) Accurate point cloud fitting of the basic contour mapping; c) Extracting the basic center of the middle rail waist M1 and the bottom rail waist M2, the intersection of the top straight line segment L1 and the bottom straight line segment L2, and their rail jaw points; d) Constructing coarse registration compatible with affine transformation deformation; e) Accurate registration based on the kd-ICP algorithm; f) Measuring the corresponding detection position after registration. By extracting the curvature of the reference invariant region and extracting key feature points to restore the deformed contour, the accuracy is improved while ensuring the registration speed.

[0004] Patent application number 201610288531.4 discloses a registration and navigation method based on 3D midpoint clouds of a skeleton. The implementation steps are: a) establishing a virtual coordinate system of X, Y, Z 3D midpoint clouds and a world coordinate system of X, Y, Z 3D midpoint clouds; b) dynamically and rapidly performing registration and navigation calculations in the virtual coordinate system and the world coordinate system of X, Y, Z 3D midpoint clouds, including acceleration for rigid registration and acceleration for flexible registration; c) correcting parameters such as navigation position and angle. This method enables robots to achieve dynamic registration and navigation of X, Y, Z 3D midpoint clouds, solving current problems such as unsatisfactory registration accuracy and excessive ray exposure, thus improving navigation accuracy and efficiency.

[0005] Patent application number 202010671178.4 discloses an automatic delineation method for organs at risk during multi-site whole-body radiotherapy. It employs a point cloud registration method to perform rigid registration on different patients, registering the delineated organs at risk from a large number of patients to obtain a prior knowledge graph of the probability distribution of organs at risk in each site. The point cloud registration method then maps this probability distribution onto the medical image to be processed, obtaining the prior probability distribution to guide the training of a convolutional neural network model. The point cloud registration technique used is the nearest-point iterative algorithm, which has advantages over traditional image grayscale-based registration techniques, being faster and less affected by noise.

[0006] Existing rigid registration techniques are insufficient for teaching applications involving adhesive application on industrial shoe lasts and soles. Most common flexible registration techniques employ deep learning or Gaussian mixture models. Deep learning-based flexible registration often requires extensive training to ensure accuracy, and its generalization ability is not guaranteed, making it difficult to meet the robust and accurate requirements of industrial applications. Gaussian mixture model-based flexible registration treats the neighborhood of each 3D point as a distribution of the Gaussian mixture model, iteratively completing the flexible matching process, which is often very time-consuming. These common flexible registration algorithms have very high time complexity and lack stability, often failing to meet the real-time, efficient, and stable application requirements of the footwear industry. Summary of the Invention

[0007] The technical problem to be solved by this invention is how to achieve efficient and stable registration of the shoe upper sidewall line.

[0008] The present invention solves the above-mentioned technical problems through the following technical means:

[0009] On one hand, embodiments of the present invention propose a method for registering the edge lines of a shoe upper for robotic adhesive application, the method comprising the following steps:

[0010] The center of the first shoe upper sidewall line is translated to the center of the second shoe upper sidewall line, and the iterative nearest point algorithm is used to perform rigid registration of the first shoe upper sidewall line and the second shoe upper sidewall line.

[0011] The rigidly registered first and second shoe upper sidewall lines are projected onto a two-dimensional plane to construct a first and second contour point set.

[0012] Calculate the normal vector at each point in the first contour point set, and obtain the corresponding point in the second contour point set for each point in the first contour point set based on the normal vector at each point.

[0013] Map each point in the first contour point set to its corresponding point in the second contour point set into three-dimensional space to achieve registration of the first shoe upper sidewall line and the second shoe upper sidewall line.

[0014] The shoe upper sidewall line registration process for robot glue application in this invention includes a rigid registration process and a flexible registration process. The rigid registration process is used to roughly register the outlines of the two shoe lasts together; the flexible registration process, based on this, performs flexible deformation registration. Furthermore, to address the high time consumption and low robustness of general flexible registration algorithms, a robust normal vector is constructed to quickly find corresponding points in the flexible registration, achieving fast, efficient, and stable flexible deformation registration.

[0015] Further, the rigid registration of the first shoe upper sidewall line and the second shoe upper sidewall line using the iterative nearest point algorithm includes:

[0016] (a) Let the first shoe upper sidewall line be a point set P = {p1, p2, ..., p...} n Let the second shoe upper edge wall line be a point set Q = {q1,q2,…,q}. n}, set k=1, P0=P, d0=0;

[0017] (b) Use a kd-tree to calculate the nearest point on point set Q for each point in point set P, denoted as point set C. k ={c1,c2,…,c n};

[0018] (c) Based on point set P and point set C k Calculate the rotation matrix R between point set P and point set Q. k Translation vector T k ;

[0019] (d) Using rotation matrix R k Translation vector T k Convert P0 to P k =R k P0+t k And calculate P k+1 With C k The root mean square error d between k ;

[0020] (e) If ||d k -d k-1 If || < τ, set k = k + 1 and jump to step (b) for the next iteration; otherwise, the algorithm terminates and outputs R. out =R k and T out =T k ;

[0021] (f) Using Rout and t out Transform the point set P and rigidly register it onto the point set Q.

[0022] Further, the step of calculating the normal vector at each point in the first contour point set, and obtaining the corresponding point in the second contour point set for each point in the first contour point set based on the normal vector at each point, includes:

[0023] Calculate the normal vector at each point in the first contour point set, and for each point in the first contour point set and its normal vector, determine a straight line;

[0024] Using kd-tree technology, the s nearest neighbor points to each point in the first contour point set are obtained on the second shoe upper sidewall line after rigid registration, and the distance from the neighbor points to the line is calculated and denoted as the distance set.

[0025] Obtain the index J corresponding to the minimum value in the distance set for each point in the first contour point set, and then obtain the corresponding point of that point in the second contour point set.

[0026] Further, calculating the normal vector at each point in the first contour point set includes:

[0027] Use kd-tree technology to obtain the k nearest points of each point in the first contour point set;

[0028] Calculate the covariance matrix corresponding to each point based on the k nearest points of each point;

[0029] Calculate the eigenvector corresponding to the smallest eigenvalue in the covariance matrix of each point in the first contour point set, and use it as the normal vector of that point.

[0030] Furthermore, this invention also proposes a shoe upper edge line registration device for robot glue application, the device comprising:

[0031] The rigid registration module is used to translate the center of the first shoe upper sidewall line to the center of the second shoe upper sidewall line, and to perform rigid registration of the first shoe upper sidewall line and the second shoe upper sidewall line using the iterative nearest point algorithm.

[0032] The projection module is used to project the rigidly registered first and second shoe upper sidewall lines onto a two-dimensional plane to construct a first contour point set and a second contour point set.

[0033] The flexible registration module is used to calculate the normal vector of each point in the first contour point set, and obtain the corresponding point of each point in the first contour point set in the second contour point set based on the normal vector of each point.

[0034] The mapping module is used to map the corresponding point of each point in the first contour point set to the second contour point set into three-dimensional space, thereby achieving the registration of the first shoe upper sidewall line and the second shoe upper sidewall line.

[0035] Furthermore, the rigid registration module includes:

[0036] The translation unit is used to translate the center of the first shoe upper sidewall line to the center of the second shoe upper sidewall line;

[0037] Iteration units are used to perform the following steps:

[0038] (a) Let the first shoe upper sidewall line be a point set P = {p1, p2, ..., p...} n Let the second shoe upper edge wall line be a point set Q = {q1,q2,…,q}. n}, set k=1, P0=P, d0=0;

[0039] (b) Use a kd-tree to calculate the nearest point on point set Q for each point in point set P, denoted as point set C. k ={c1,c2,…,c n};

[0040] (c) Based on point set P and point set C k Calculate the rotation matrix R between point set P and point set Q. k Translation vector T k ;

[0041] (d) Using rotation matrix R k Translation vector T k Convert P0 to P k =R k P0+t k And calculate P k+1 With C k The root mean square error d between k ;

[0042] (e) If ||d k -d k-1 If || < τ, set k = k + 1 and jump to step (b) for the next iteration; otherwise, the algorithm terminates and outputs R. out =R k and T out =T k ;

[0043] (f) Using R out and t out Transform the point set P and rigidly register it onto the point set Q.

[0044] Furthermore, the flexible registration module includes:

[0045] The normal vector calculation unit is used to calculate the normal vector at each point in the first contour point set, and to determine a straight line for each point in the first contour point set and its normal vector.

[0046] The distance set construction unit is used to obtain the s nearest neighbor points to each point in the first contour point set on the second shoe upper sidewall line after rigid registration using kd tree technology, and calculate the distance from the neighbor points to the line, which is denoted as the distance set;

[0047] The corresponding point lookup unit is used to obtain the index number J corresponding to the minimum value in the distance set corresponding to each point in the first contour point set, and to obtain the corresponding point of that point in the second contour point set.

[0048] Furthermore, the normal vector calculation unit is used for:

[0049] Use kd-tree technology to obtain the k nearest points of each point in the first contour point set;

[0050] Calculate the covariance matrix corresponding to each point based on the k nearest points of each point;

[0051] Calculate the eigenvector corresponding to the smallest eigenvalue in the covariance matrix of each point in the first contour point set, and use it as the normal vector of that point.

[0052] Furthermore, this invention also proposes a shoe last outline gluing robot, the device including a memory and a processor; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the method described above.

[0053] Furthermore, the present invention also proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0054] The advantages of this invention are:

[0055] (1) The registration process for the shoe upper sidewall lines for robot glue application in this invention includes a rigid registration process and a flexible registration process. The rigid registration process is used to roughly register the outlines of the two shoe lasts together; the flexible registration process, based on this, implements flexible deformation registration. In order to solve the problems of high time consumption and low robustness of general flexible registration algorithms, a robust normal vector is constructed to quickly find the corresponding points in the flexible registration, thereby achieving fast, efficient and stable flexible deformation registration.

[0056] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0057] Figure 1 This is a flowchart of the shoe upper sidewall line registration method for robot glue application in the first embodiment of the present invention;

[0058] Figure 2 This is a schematic diagram of the shoe upper sidewall line registration method for robot glue application in the first embodiment of the present invention;

[0059] Figure 3 This is a schematic diagram of the rigid registration principle of the shoe upper sidewall line in the first embodiment of the present invention;

[0060] Figure 4 This is a schematic diagram of the flexible registration principle of the shoe upper sidewall line in the first embodiment of the present invention;

[0061] Figure 5 This is a structural diagram of the shoe upper sidewall line registration device for robot glue application in the second embodiment of the present invention. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] like Figures 1 to 2 As shown, one embodiment of the present invention proposes a method for registering the edge wall lines of a shoe upper for robot glue application, including the following steps:

[0064] S10. The center of the first shoe upper sidewall line is translated to the center of the second shoe upper sidewall line, and the rigid registration of the first shoe upper sidewall line and the second shoe upper sidewall line is performed by using the iterative nearest point algorithm.

[0065] S20. Project the rigidly registered first and second shoe upper sidewall lines onto a two-dimensional plane to construct a first contour point set and a second contour point set.

[0066] S30. Calculate the normal vector on each point in the first contour point set, and obtain the corresponding point in the second contour point set based on the normal vector on each point.

[0067] S40. Map each point in the first contour point set to its corresponding point in the second contour point set to three-dimensional space to achieve registration of the first shoe upper sidewall line and the second shoe upper sidewall line.

[0068] It should be noted that rigid registration means that during the registration process, the objects to be registered are treated as rigid objects, and only rigid rotation and translation can be performed between the two objects.

[0069] Flexible registration refers to the registration process in which the two objects to be registered are considered to be flexible and variable. The two objects can not only have rigid motion, but also include motion that changes their shape.

[0070] Considering the significant deviation between the two contour point sets to be registered during the registration of the shoe upper sidewall lines, and the fact that their shapes do not perfectly match, this embodiment addresses this unique characteristic by first employing rigid registration to ensure their centers coincide and their orientations are aligned. Then, flexible registration is used to register the two point sets whose shapes do not perfectly match. Taking into account the flattened shape of the shoe last contour line in the xy plane, the three-dimensional contour line point set is projected onto the xy two-dimensional plane to form a two-dimensional point set for rapid flexible registration, reducing algorithm complexity. Furthermore, to address the high time consumption and low robustness of general flexible registration algorithms, robust normal vectors are constructed to quickly find corresponding points in the flexible registration, achieving fast, efficient, and stable flexible deformation registration.

[0071] Compared to Scheme 1 described in the invention patent application with application number 201610288531.4, Scheme 1's rigid registration first involves establishing a Gaussian mixture model in one point cloud and then performing ICP iterative registration using the model's points and another point cloud dataset. This method aims to improve robustness to noise interference; however, it has high time complexity. In this embodiment, where noise interference is minimal, the approach is to first translate the centers of the two point sets together and then use conventional point-to-point ICP for efficient registration.

[0072] Furthermore, the flexible registration in Scheme 1 is performed directly in three-dimensional space and involves complex differentiation and optimization algorithms, resulting in high time complexity. This scheme utilizes the flattened xy-plane characteristic of the shoe last's outline to project the point set onto the two-dimensional xy-plane for fast flexible registration based on normal vectors. Then, it is mapped back to the original three-dimensional space to complete the flexible registration of the three-dimensional point cloud.

[0073] In one embodiment, in step S10, the first shoe upper sidewall line and the second shoe upper sidewall line are first moved to approximately the same position, for example, the center of the first shoe upper sidewall line is translated to the center of the second shoe upper sidewall line, providing an initial position for rigid registration. Then, the standard Iterative Closest Point (ICP) algorithm is used for precise rigid registration, including the following steps:

[0074] (a) Let the first shoe upper sidewall line be a point set P = {p1, p2, ..., p...} n Let the second shoe upper edge wall line be a point set Q = {q1,q2,…,q}. n}, set k=1, P0=P, d0=0, k is the number of iterations, P and 0 are the initial values ​​of the P0 and d0 iteration algorithms respectively, P0 is the first shoe upper sidewall line, and d0 is the root mean square error;

[0075] (b) Use a kd-tree to calculate the nearest point on point set Q for each point in point set P, denoted as point set C. k ={c1,c2,…,c n};

[0076] (c) Based on point set P and point set C k Calculate the rotation matrix R between point set P and point set Q. k Translation vector T k ;

[0077] (d) Using rotation matrix R k Translation vector T k Convert P0 to P k =R k P0+t k And calculate P k+1 With C k The root mean square error d between k ;

[0078] (e) If ||d k -d k-1 If || < τ, set k = k + 1 and jump to step (b) for the next iteration; otherwise, the algorithm terminates and outputs R. out =R k and T out =T k ;

[0079] (f) Using R out and T out The transformation formula for the point set P is P′=R. out P+T out ,P′={p′1,p′2,…,p′ nLet} be the points transformed from P. Then, the point set P is rigidly registered onto the point set Q. The rigid registration effect is as follows: Figure 3 As shown, Figure 3 -(a) is a schematic diagram of the sidewall lines of the two shoe uppers before rigid registration. Figure 3 -(b) is a schematic diagram of the sidewall lines of the two shoe uppers after rigid registration.

[0080] In one embodiment, step S30 includes the following steps:

[0081] S31. Calculate the normal vector at each point in the first contour point set, and determine a straight line for each point in the first contour point set and its normal vector.

[0082] In one embodiment, step S31 includes the following steps:

[0083] Use kd-tree technology to obtain the k nearest points of each point in the first contour point set;

[0084] Calculate the covariance matrix corresponding to each point based on the k nearest points of each point;

[0085] Calculate the eigenvector corresponding to the smallest eigenvalue in the covariance matrix of each point in the first contour point set, and use it as the normal vector of that point.

[0086] Specifically, the calculation process for the normal vector corresponding to each point in the first contour point set is as follows:

[0087] Calculate the normal vector at each point in the first contour point set P″. For each point p″ in the first contour point set P″... i Use a kd-tree to obtain point p″ from the first contour point set P″. i Let Ap be the k nearest points. i ={p″ i1 ,p″ i2 ,…,p″ ik}

[0088] Calculate Ad i The covariance matrix Cov(Ad) i ):

[0089]

[0090] In the formula, For Ad i The center point is used to calculate Cov(Ad). i The eigenvector corresponding to the smallest eigenvalue in the matrix is ​​denoted as v. i , will v i As point p″ i The normal vector on.

[0091] More preferably, in this embodiment, when calculating the covariance matrix, a kd-tree is used to obtain a neighborhood point set within a certain radius sphere from the first contour point set P″, denoted as Ap. i ={p″ i1 ,p″ i2 ,…,p″ ik}, used to calculate Ad i The covariance matrix.

[0092] Compared to using the k nearest points to calculate the normal vector, using a sphere neighborhood with a fixed radius (ranging from about 1.5 to 3 cm) to calculate the normal vector can avoid changes in the size range of the normal vector calculation caused by inconsistent density of the shoe last outline.

[0093] Based on the above method, the normal vector at each point in P″ is calculated and denoted as V={v1,v2,…,v n}

[0094] S32. Using kd-tree technology, obtain the s nearest neighbor points to each point in the first contour point set on the second shoe upper sidewall line after rigid registration, and calculate the distance from the neighbor points to the line, denoted as the distance set;

[0095] S33. Obtain the index number J corresponding to the minimum value in the distance set corresponding to each point in the first contour point set, and obtain the corresponding point of that point in the second contour point set.

[0096] It should be noted that, considering that the shape of the shoe upper sidewall line is roughly flat, Q and the rigidly registered P′ are projected onto the XY two-dimensional plane to construct a two-dimensional contour line. The first contour point set and the second contour point set obtained after projection are denoted as P″ and Q′, respectively, and flexible registration processing is performed on the two-dimensional plane.

[0097] Specifically, for a point p″ in P″ i and its normal vector v i A straight line can be determined, denoted as l. i Then, using kd-tree techniques, we obtain the distance from p in Q. i The nearest s points are denoted as Aq. i ={q i1 ,q i2 ,…,q is}; Calculate the distances from these neighboring points to the line, denoted as D. i ={d i1 ,d i2 ,…,d is};Get D i The index number J corresponding to the minimum value is shown in the following formula.

[0098]

[0099] Point q corresponding to index number J iJ That is, point p″ i The corresponding point in Q′.

[0100] Based on the above process, the corresponding point of each point in P″ on Q′ is obtained, and then these obtained corresponding points are mapped to three-dimensional space, that is, the corresponding points of all points in P′ on Q are obtained. The flexible registration effect diagram is shown below. Figure 4 As shown, Figure 4 -(a) is the position diagram of the contour line before flexible registration, 4-(b) is the diagram of calculating the normal vector and finding the mapping points, and 4-(c) is the position diagram of the contour line after flexible registration.

[0101] In this embodiment, rigid registration is used to roughly register the outlines of the two shoe lasts together. Then, flexible deformation registration is performed on the outlines after rigid registration. However, in order to solve the problem of high time consumption of general flexible registration algorithms, this embodiment constructs a robust normal vector to quickly find the corresponding points in the flexible registration, thereby achieving fast and efficient flexible deformation registration.

[0102] In addition, such as Figure 5 As shown, one embodiment of the present invention also proposes a shoe upper edge line registration device for robot glue application, the device comprising:

[0103] The rigid registration module 10 is used to translate the center of the first shoe upper sidewall line to the center of the second shoe upper sidewall line, and to perform rigid registration of the first shoe upper sidewall line and the second shoe upper sidewall line using the iterative nearest point algorithm.

[0104] Projection module 20 is used to project the rigidly registered first and second shoe upper sidewall lines onto a two-dimensional plane to construct a first contour point set and a second contour point set.

[0105] The flexible registration module 30 is used to calculate the normal vector on each point in the first contour point set, and obtain the corresponding point in the second contour point set based on the normal vector on each point.

[0106] The mapping module 40 is used to map the corresponding point of each point in the first contour point set to the second contour point set into three-dimensional space, thereby achieving the registration of the first shoe upper sidewall line and the second shoe upper sidewall line.

[0107] In this embodiment, the registration process for the shoe upper sidewall lines for robot glue application includes a rigid registration process and a flexible registration process. The rigid registration process is used to roughly register the outlines of the two shoe lasts together; the flexible registration process, based on this, performs flexible deformation registration. Furthermore, to address the high time consumption and low robustness of general flexible registration algorithms, a robust normal vector is constructed to quickly find corresponding points in the flexible registration, achieving fast, efficient, and stable flexible deformation registration.

[0108] In one embodiment, the rigid registration module 10 includes:

[0109] A translation unit is used to translate the center of the first shoe upper sidewall line to the center of the second shoe upper sidewall line;

[0110] Iteration units are used to perform the following steps:

[0111] (a) Let the first shoe upper sidewall line be a point set P = {p1, p2, ..., p...} n Let the second shoe upper edge wall line be a point set Q = {q1,q2,…,q}. n}, set k=1, P0=P, d0=0;

[0112] (b) Use a kd-tree to calculate the nearest point on point set Q for each point in point set P, denoted as point set C. k ={c1,c2,…,c n};

[0113] (c) Based on point set P and point set C k Calculate the rotation matrix R between point set P and point set Q. k Translation vector T k ;

[0114] (d) Using rotation matrix R k Translation vector T k Convert P0 to P k =R k P0+t k And calculate P k+1 With C k The root mean square error d between k ;

[0115] (e) If ||d k -d k-1 If || < τ, set k = k + 1 and jump to step (b) for the next iteration; otherwise, the algorithm terminates and outputs R. out =R k and T out =T k ;

[0116] (f) Using R out and tout Transform the point set P and rigidly register it onto the point set Q.

[0117] In one embodiment, the flexible registration module 30 includes:

[0118] The normal vector calculation unit is used to calculate the normal vector at each point in the first contour point set, and to determine a straight line for each point in the first contour point set and its normal vector.

[0119] The distance set construction unit is used to obtain the s nearest neighbor points to each point in the first contour point set on the second shoe upper sidewall line after rigid registration using kd tree technology, and calculate the distance from the neighbor points to the line, which is denoted as the distance set;

[0120] The corresponding point lookup unit is used to obtain the index number J corresponding to the minimum value in the distance set corresponding to each point in the first contour point set, and to obtain the corresponding point of that point in the second contour point set.

[0121] In one embodiment, the normal vector calculation unit is used for:

[0122] Use kd-tree technology to obtain the k nearest points of each point in the first contour point set;

[0123] Calculate the covariance matrix corresponding to each point based on the k nearest points of each point;

[0124] Calculate the eigenvector corresponding to the smallest eigenvalue in the covariance matrix of each point in the first contour point set, and use it as the normal vector of that point.

[0125] It should be noted that other embodiments or implementation methods of the shoe upper edge line registration device for robot glue application described in this invention can refer to the above-described method embodiments, and will not be repeated here.

[0126] Furthermore, this invention also proposes a shoe last outline gluing robot, the robot including a memory and a processor; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the method described in the above embodiments.

[0127] Furthermore, this invention also proposes a computer-readable storage medium storing a computer program thereon, characterized in that, when the computer program is executed by a processor, it implements the method described in the above embodiments.

[0128] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0129] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0130] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0131] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0132] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for registering the edge lines of a shoe upper for robotic adhesive application, characterized in that, The method includes: The center of the first shoe upper sidewall line is translated to the center of the second shoe upper sidewall line, and the iterative nearest point algorithm is used to perform rigid registration of the first shoe upper sidewall line and the second shoe upper sidewall line. The rigidly registered first and second shoe upper sidewall lines are projected onto a two-dimensional plane to construct a first and second contour point set. Calculate the normal vector at each point in the first contour point set, and obtain the corresponding point in the second contour point set for each point in the first contour point set based on the normal vector at each point. Map each point in the first contour point set to its corresponding point in the second contour point set into three-dimensional space to achieve registration of the first shoe upper sidewall line and the second shoe upper sidewall line.

2. The method for registering shoe upper edge lines for robot adhesive application as described in claim 1, characterized in that, The rigid registration of the first and second shoe upper sidewall lines using the iterative nearest point algorithm includes: (a) Let the first shoe upper sidewall line be a point set P = {p1, p2, ..., p...} n Let the second shoe upper edge wall line be a point set Q = {q1,q2,…,q}. n }, set k=1, P0=P, d0=0; (b) Use a kd-tree to calculate the nearest point on point set Q for each point in point set P, denoted as point set C. k ={c1,c2,…,c n }; (c) Based on point set P and point set C k Calculate the rotation matrix R between point set P and point set Q. k Translation vector T k ; (d) Using rotation matrix R k Translation vector T k Convert P0 to P k =R k P0+t k And calculate P k+1 With C k The root mean square error d between k ; (e) If ||d k -d k-1 If || < τ, set k = k + 1 and jump to step (b) for the next iteration; otherwise, the algorithm terminates and outputs R. out =R k and T out =T k ; (f) Using R out and t out Transform the point set P and rigidly register it onto the point set Q.

3. The method for registering shoe upper edge lines for robot adhesive application as described in claim 1, characterized in that, The step of calculating the normal vector at each point in the first contour point set and obtaining the corresponding point in the second contour point set for each point in the first contour point set based on the normal vector at each point includes: Calculate the normal vector at each point in the first contour point set, and for each point in the first contour point set and its normal vector, determine a straight line; Using kd-tree technology, the s nearest neighbor points to each point in the first contour point set are obtained on the second shoe upper sidewall line after rigid registration, and the distance from the neighbor points to the line is calculated and denoted as the distance set. Obtain the index J corresponding to the minimum value in the distance set for each point in the first contour point set, and then obtain the corresponding point of that point in the second contour point set.

4. The method for registering shoe upper edge lines for robot adhesive application as described in claim 1 or 3, characterized in that, The calculation of the normal vector at each point in the first contour point set includes: Use kd-tree technology to obtain the k nearest points of each point in the first contour point set; Calculate the covariance matrix corresponding to each point based on the k nearest points of each point; Calculate the eigenvector corresponding to the smallest eigenvalue in the covariance matrix of each point in the first contour point set, and use it as the normal vector of that point.

5. A shoe upper edge line registration device for robotic glue application, characterized in that, The device includes: The rigid registration module is used to translate the center of the first shoe upper sidewall line to the center of the second shoe upper sidewall line, and to perform rigid registration of the first shoe upper sidewall line and the second shoe upper sidewall line using the iterative nearest point algorithm. The projection module is used to project the rigidly registered first and second shoe upper sidewall lines onto a two-dimensional plane to construct a first contour point set and a second contour point set. The flexible registration module is used to calculate the normal vector of each point in the first contour point set, and obtain the corresponding point of each point in the first contour point set in the second contour point set based on the normal vector of each point. The mapping module is used to map the corresponding point of each point in the first contour point set to the second contour point set into three-dimensional space, thereby achieving the registration of the first shoe upper sidewall line and the second shoe upper sidewall line.

6. The shoe upper edge line registration device for robot glue application as described in claim 5, characterized in that, The rigid registration module includes: A translation unit is used to translate the center of the first shoe upper sidewall line to the center of the second shoe upper sidewall line; Iteration units are used to perform the following steps: (a) Let the first shoe upper sidewall line be a point set P = {p1, p2, ..., p...} n Let the second shoe upper edge wall line be a point set Q = {q1,q2,…,q}. n }, set k=1, P0=P, d0=0; (b) Use a kd-tree to calculate the nearest point on point set Q for each point in point set P, denoted as point set C. k ={c1,c2,…,c n }; (c) Based on point set P and point set C k Calculate the rotation matrix R between point set P and point set Q. k Translation vector T k ; (d) Using rotation matrix R k Translation vector T k Convert P0 to P k =R k P0+t k And calculate P k+1 With C k The root mean square error d between k ; (e) If ||d k -d k-1 If || < τ, set k = k + 1 and jump to step (b) for the next iteration; otherwise, the algorithm terminates and outputs R. out =R k and T out =T k ; (f) Using R out and t out Transform the point set P and rigidly register it onto the point set Q.

7. The shoe upper edge line registration device for robot glue application as described in claim 5, characterized in that, The flexible registration module includes: The normal vector calculation unit is used to calculate the normal vector at each point in the first contour point set, and to determine a straight line for each point in the first contour point set and its normal vector. The distance set construction unit is used to obtain the s nearest neighbor points to each point in the first contour point set on the second shoe upper sidewall line after rigid registration using kd tree technology, and calculate the distance from the neighbor points to the line, which is denoted as the distance set; The corresponding point lookup unit is used to obtain the index number J corresponding to the minimum value in the distance set corresponding to each point in the first contour point set, and to obtain the corresponding point of that point in the second contour point set.

8. The shoe upper edge line registration device for robot glue application as described in claim 7, characterized in that, The normal vector calculation unit is used for: Use kd-tree technology to obtain the k nearest points of each point in the first contour point set; Calculate the covariance matrix corresponding to each point based on the k nearest points of each point; Calculate the eigenvector corresponding to the smallest eigenvalue in the covariance matrix of each point in the first contour point set, and use it as the normal vector of that point.

9. A robot for applying glue to the outline of a shoe last, characterized in that, The robot includes a memory and a processor; wherein the processor runs a program corresponding to the executable program code stored in the memory to implement the method as described in any one of claims 1-4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-4.

Images