Last head and toe removal method and medium

Abstract

The present application relates to a kind of shoe last head tail processing method and medium, method includes: obtaining the template file for the head tail processing of shoe last of multiple female last, for any one shoe last, based on the three-dimensional point cloud data of shoe last and pre-processing, based on the three-dimensional point cloud data of pre-processed shoe last, registration corresponding template file;Based on the local registration point cloud of template file after registration, template trajectory of template file and the point cloud of shoe last are trajectory registration, and trajectory point is traversed classification and obtains polishing point and fitting point;Based on the polishing point and fitting point, the removal of excess point and the smoothing processing of trajectory are carried out, to obtain smooth heel trajectory and toe trajectory, as the trajectory of processing.The matching precision of shoe last and female last in shoe last head tail processing is higher, the degree of automation of shoe last head tail processing is improved, and the degree of dependence on artificial is reduced.

Description

Technical Field

[0001] This invention relates to the field of shoe last processing technology, specifically to a method and medium for removing the toe and tail of a shoe last. Background Technology

[0002] A shoe last is a foot mold used to shape and refine the upper of a finished shoe. A master last is the final sample formed by grinding and shaping the shoe last, and is generally used as a template for producing shoe lasts.

[0003] In current shoe last processing, the trajectory of the master last template is generally applied directly to the actual product, and then polishing is done based on manual experience. The polishing precision is low, and the quality is unstable over a long period of time. In particular, manual polishing of irregularly shaped shoe lasts is more difficult, highly dependent on manual experience, and has high labor costs.

[0004] Although there are some automated processing methods for shoe lasts, the matching accuracy between shoe lasts and mother lasts is not high, resulting in a low degree of automation and a significant reliance on manual labor. Summary of the Invention

[0005] The technical problem to be solved by this application is to provide a method for removing the toe and tail of a shoe last, which can improve the matching accuracy between the shoe last and the mother last and reduce the dependence on manual labor.

[0006] In one aspect, one embodiment provides a method for removing the toe and tail of a shoe last, including:

[0007] Obtain multiple master last template files for shoe last toe and heel removal processing, wherein any template file includes shoe toe template trajectory, shoe heel template trajectory and point cloud data for local registration, and the multiple master lasts include master lasts of different shoe types and sizes;

[0008] For any shoe last, the 3D point cloud data of the shoe last is preprocessed;

[0009] Based on the preprocessed 3D point cloud data of the shoe last, the corresponding template file is registered;

[0010] Based on the locally registered point cloud of the registered template file, trajectory registration is performed between the template trajectory of the template file and the point cloud of the shoe last, including trajectory registration between the heel template trajectory of the template file and the heel point cloud of the shoe last, and trajectory registration between the toe template trajectory of the template file and the toe point cloud of the shoe last.

[0011] Based on the heel and toe trajectories after trajectory registration, the trajectory points are traversed and classified to obtain the polishing points and fitting points;

[0012] Based on the grinding points and bonding points, excess points are removed and the trajectory is smoothed to obtain a smooth heel trajectory and toe trajectory, which serve as the processing trajectory.

[0013] The registration of the corresponding template file based on the preprocessed 3D point cloud data of the shoe last includes:

[0014] Based on any template file, the iterative nearest algorithm is used to find the nearest adjacent shoe last point for each template point in the template file, forming a point pair. The distance between the two points in any point pair is calculated, and it is determined whether the distance is less than or equal to a preset third distance threshold. If so, the point pair is identified as the first valid point pair, thus obtaining all the first valid point pairs.

[0015] Based on any point in the template point cloud obtained from the first effective point pair, perform a spatial transformation to obtain the spatially transformed point cloud coordinates of that point. Based on the transformed point cloud coordinates, find the nearest adjacent shoe last point to obtain the second effective point pair, and thus obtain all the second effective point pairs.

[0016] Based on all the second valid point pairs obtained, the process of obtaining all first valid point pairs and second valid point pairs is repeated. Based on the distance convergence condition, the average distance of the final valid point pairs is obtained, and the average distance is used as a matching score.

[0017] The template file with the lowest matching score will be used as the template file for the current shoe last.

[0018] In a second aspect, one embodiment provides a computer-readable storage medium storing a program that can be loaded by a processor and executed as the shoe last removal method in any embodiment.

[0019] The beneficial effects of this invention are:

[0020] By finding effective point pairs between the master last and the shoe last, and through the continuous convergence and iteration process of the effective point pairs, the master last that is closest to the current shoe last is found. This makes the matching accuracy between the shoe last and the master last higher in the shoe last removal process in this application, thereby improving the automation level of the shoe last removal process and reducing the dependence on manual labor. Attached Figure Description

[0021] Figure 1 This is a schematic flowchart of a shoe last removal process according to an embodiment of this application;

[0022] Figure 2 This application Figure 1 A schematic diagram of the method flow of one embodiment of step S10;

[0023] Figure 3This application Figure 1 A schematic diagram of the method flow of one embodiment of step S30;

[0024] Figure 4 This application Figure 1 A schematic diagram of a method flow for one embodiment of step S40;

[0025] Figure 5 This application Figure 1 A schematic diagram of a method flow for one embodiment of step S50;

[0026] Figure 6 This application Figure 5 A schematic diagram of a method flow for one embodiment of step S501;

[0027] Figure 7 This is a schematic diagram illustrating the relationship between the heel template trajectory and the heel surface of the shoe last before trajectory adjustment in one embodiment.

[0028] Figure 8 Based on Figure 7 A schematic diagram of one embodiment of the trajectory adjustment shown;

[0029] Figure 9 This application Figure 5 A schematic diagram of the method flow for one embodiment of step S502. Detailed Implementation

[0030] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of this application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to this application are not shown or described in the specification. This is to avoid obscuring the core parts of this application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0031] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0032] The serial numbers assigned to components in this article, such as "first" and "second", are used only to distinguish the objects being described and have no sequential or technical meaning.

[0033] To facilitate the explanation of the inventive concept of this application, the following is a brief description of the shoe last processing technology.

[0034] Because of the differences between left and right feet, different sizes, and different shoe shapes, there are various models of mother lasts used for producing shoe lasts. These mother lasts themselves are not very different, and some mother lasts are even very similar. Therefore, it is quite difficult to match the corresponding mother last during the automatic processing of shoe lasts.

[0035] In the current shoe last removal process, the matching accuracy between the shoe last and the mother last is not high, resulting in a low degree of automation and a still large reliance on manual labor.

[0036] In view of this, and based on this, this application provides a method and medium for removing the toe and tail of a shoe last. By finding effective point pairs between a master last and a shoe last, and through a continuous convergence and iteration process of effective point pairs, the master last closest to the current shoe last is found. This scheme has higher matching accuracy, thereby improving the automation level of the shoe last removal process and reducing the reliance on manual labor.

[0037] One embodiment of this application provides a method for removing the toe and tail of a shoe last. Please refer to [the relevant documentation]. Figure 1 ,include:

[0038] Step S10: Obtain multiple template files for shoe last trimming. Each template file includes a toe template trajectory, a heel template trajectory, and point cloud data for local registration. The multiple templates include templates for different shoe types and sizes.

[0039] In one embodiment, please refer to Figure 2 Step S10 includes:

[0040] Step S101: For any given last, acquire its 3D point cloud data. The 3D point cloud data of the last includes the 3D point cloud data of the sole, the 3D point cloud data of the toe, and the 3D point cloud data of the heel.

[0041] In some embodiments, 3D point cloud data of the last can be acquired using a 3D sensor, or 3D point cloud data of the last can be generated based on an STL file. If the same shoe model has multiple sizes, the STL file can be scaled up to obtain multiple STL files for different shoe sizes, and thus obtain 3D point cloud data of the last for multiple shoe sizes.

[0042] In one embodiment, for any given 3D point cloud data of a mother last, the acquisition method includes: acquiring the original 3D point cloud data of the mother last; performing voxel filtering, radius culling, and Euclidean distance segmentation based on the original 3D point cloud data to reduce the amount of point cloud data; then performing edge burr removal processing to eliminate background noise and obtain a denoised point cloud; and finally using PCA analysis to transform the denoised point cloud to the local coordinate system of the shoe to obtain the 3D point cloud data of the mother last. Furthermore, it is always ensured that the x-axis is along the direction of the longest side of the shoe, the y-axis is along the direction of the second longest side of the shoe, and the z-axis is along the direction of the shortest side.

[0043] Step S102: Based on the three-dimensional point cloud data of the sole, extract the edge contour trajectory of the sole template as the sole reference trajectory; based on the three-dimensional point cloud data of the toe and the sole reference trajectory, extract the toe template trajectory; based on the three-dimensional point cloud data of the heel and the sole reference trajectory, extract the heel template trajectory.

[0044] In one embodiment, the edge contour trajectory of the sole template is extracted as a sole reference trajectory based on the three-dimensional point cloud data of the sole, including: extracting the sole point cloud based on the normal vector direction, and then extracting the edge contour trajectory of the sole template as a sole reference trajectory through the sole point cloud.

[0045] In one embodiment, the toe template trajectory is extracted based on the three-dimensional point cloud data of the shoe toe and the reference trajectory of the shoe sole, including:

[0046] First, the point cloud at the bottom of the shoe toe is extracted, and the point cloud is sliced ​​along the x-axis. The slice distance is controlled by a preset trajectory spacing parameter, thus obtaining point cloud slices within a preset range along the x-axis. Then, based on the t-axis direction, the slices are divided into blocks according to a preset y-axis spacing, and the centroid of each small block is extracted as a trajectory point. Finally, each slice yields a trajectory, which is then surface-fitted and smoothed to obtain the trajectory at the bottom of the shoe toe.

[0047] Secondly, the point cloud of the side of the shoe tip is extracted, and the shoe tip reference trajectory is cut out based on the shoe sole reference trajectory as the initial trajectory of the side of the shoe tip. The initial trajectory is translated at a preset interval in the z-axis direction. Each translation yields a trajectory, thus obtaining the initial shoe tip trajectory surface.

[0048] Finally, the obtained trajectory at the bottom of the toe and the initial toe trajectory surface are projected onto the point cloud surface on the side of the toe, so that the initial trajectory surface fits the surface of the shoe last and is smoothed to obtain the final toe template trajectory.

[0049] In one embodiment, the heel template trajectory is extracted based on the three-dimensional point cloud data of the heel and the reference trajectory of the sole. This includes: extracting the point cloud of the side of the heel; cropping the sole reference trajectory of the heel portion based on the sole reference trajectory; for any cropped trajectory point, extracting adjacent trajectory points and fitting a straight line L1; and using points in the side point cloud whose distance to the straight line L1 is less than a preset first distance threshold as slice point clouds. This yields slice point clouds for each trajectory point. These slice point clouds are then divided into blocks in the Z direction, and the centroid of each small block is extracted as a trajectory point. A trajectory is obtained based on each slice. The trajectory is then surface-fitted and smoothed to obtain the final heel template trajectory.

[0050] The preset first distance threshold is set based on actual needs.

[0051] In one embodiment, the first distance threshold can be 2.5 mm.

[0052] Step S103: Based on the edge contour trajectory of the shoe sole template, extract point cloud data for local registration.

[0053] In one embodiment, step S103 includes: for any point in the three-dimensional point cloud of the shoe sole, based on the edge contour trajectory of the shoe sole template, using a contour enclosing method in the xy plane where the shoe sole is located, identifying the minimum distance between the arbitrary point and the edge contour trajectory; if the minimum distance is less than or equal to a preset second distance threshold, then the arbitrary point is taken as a point in the point cloud data of local registration.

[0054] The preset second distance threshold is set based on actual needs.

[0055] Step S104: The toe template trajectory, heel template trajectory, and point cloud data used for local registration are used as template files for the toe and heel removal processing of the shoe last, thereby obtaining multiple template files, which correspond to the master lasts of multiple shoe types.

[0056] Step S20: For any shoe last, perform preprocessing based on the 3D point cloud data of the shoe last.

[0057] In one embodiment, the method for acquiring the three-dimensional point cloud data of the shoe last can be the same as or different from the method for acquiring the three-dimensional point cloud data of the mother last.

[0058] In one embodiment, the preprocessing method may include: acquiring the original 3D point cloud data of the mother last; performing voxel filtering, radius culling, and Euclidean distance segmentation based on the original 3D point cloud data to reduce the amount of point cloud data; then performing edge burr removal processing to eliminate background noise and obtain a denoised point cloud; and then using PCA analysis to transform the denoised point cloud to the local coordinate system of the shoe to obtain the 3D point cloud data of the mother last. Furthermore, it is always ensured that the x-axis is along the direction of the longest side of the shoe, the y-axis is along the direction of the second longest side of the shoe, and the z-axis is along the direction of the shortest side.

[0059] Step S30: Based on the preprocessed 3D point cloud data of the shoe last, register the corresponding template file.

[0060] In one embodiment, please refer to Figure 3 Step S30 includes:

[0061] Step S301: Based on any template file, use the iterative nearest algorithm to find the nearest adjacent shoe last point for each template point in the template file, form a point pair, and calculate the distance between the two points in any point pair. Determine whether the distance is less than or equal to a preset third distance threshold. If so, identify the point pair as the first valid point pair, and thus obtain all the first valid point pairs.

[0062] The preset third distance threshold is set based on actual needs.

[0063] In one embodiment, the Iterative Nearest Point Registration (ICP) algorithm is used to match the 3D point cloud data in the template file with the 3D point cloud data of the shoe last. For each point in the template file, the nearest adjacent shoe last point is found to form a point pair. If the distance between the two points in the point pair is greater than a preset third distance threshold, it is identified as an invalid point and the invalid point does not participate in the subsequent calculation.

[0064] Step S302: Based on any point in the template point cloud obtained from the first valid point pair, perform a spatial transformation to obtain the spatially transformed point cloud coordinates of that arbitrary point. Based on the transformed point cloud coordinates, find the nearest adjacent shoe last point to obtain the second valid point pair, and thus obtain all the second valid point pairs.

[0065] The purpose of the registration algorithm is to minimize the distance between two points in a point pair. In one embodiment, a 4x4 spatial transformation matrix is ​​obtained using the least squares method. This spatial transformation matrix includes rotation and translation transformations. Applying this spatial transformation matrix to the template point cloud yields the latest coordinates of any point in the template point cloud. Based on the obtained latest coordinates, the nearest adjacent shoe last points are found again, resulting in the corresponding second valid point pair.

[0066] Step S303: Based on all the obtained second valid point pairs, iterate through all the first valid point pairs to the second valid point pairs. Based on the distance convergence condition, obtain the average distance of the finally obtained valid point pairs and use the average distance as a matching score.

[0067] Based on all the obtained second valid point pairs, step S301 can be executed again to obtain the latest first valid point pair, and then step S302 can be executed to obtain the latest second valid point pair. In this way, steps S301 and S302 can be continuously repeated. When the distance between the two points in a point pair tends to stabilize and no longer decreases, that is, when the distance convergence condition is reached, the most useful template point cloud coordinates and the corresponding spatial transformation matrix can be obtained. At this time, the average distance between the two points in the final registration point pair can be used as a matching score. The larger the average distance, the larger the matching score. The larger the matching score, the greater the difference between the mother last template and the current shoe last.

[0068] Step S304: Use the template file with the lowest matching score as the template file to be registered with the current shoe last.

[0069] In this way, the size and left / right foot parameters corresponding to the master last that is matched with the current shoe last can be obtained, and the template file of the corresponding size and left / right foot can be loaded for subsequent trajectory calculation.

[0070] In this way, by finding effective point pairs between the master last and the shoe last, and through the continuous convergence and iteration process of the effective point pairs, the master last that is closest to the current shoe last can be found, resulting in higher matching accuracy, improved automation of the shoe last toe-and-tail removal process, and reduced reliance on manual labor.

[0071] Step S40: Based on the locally registered point cloud of the registered template file, perform trajectory registration between the template trajectory of the template file and the point cloud of the shoe last. This includes trajectory registration between the heel template trajectory of the template file and the heel point cloud of the shoe last, and trajectory registration between the toe template trajectory of the template file and the toe point cloud of the shoe last.

[0072] In one embodiment, please refer to Figure 4 For trajectory registration of either the heel template trajectory of the template file or the heel point cloud trajectory of the shoe last, or the toe template trajectory of the template file or the toe point cloud trajectory of the shoe last, it can include:

[0073] Step S401: Based on the locally registered point cloud of the registered template file, perform coarse registration of the overall point cloud with a preset fourth distance threshold to obtain the third valid point pair.

[0074] The fourth distance threshold can be set according to actual needs. In one embodiment, the fourth distance threshold can be 3.5mm.

[0075] In coarse registration, a relatively large fourth distance threshold needs to be set to avoid failing to obtain valid point pairs. However, a large fourth distance threshold will include some interfering point pairs. After coarse registration, some interfering points are removed. However, since the distance between the template point cloud and the shoe last point cloud is still relatively large after coarse registration, the distance threshold needs to be continuously reduced to achieve fine registration.

[0076] Step S402: Based on any point in the template point cloud obtained from the third valid point pair, perform a spatial transformation to obtain the spatially transformed point cloud coordinates of that point. Based on the transformed point cloud coordinates, identify the nearest adjacent shoe last point to obtain the fourth valid point pair, and thus obtain all fourth valid point pairs. In this way, the template point cloud and the shoe last point cloud are closer together.

[0077] Step S403: For any fourth valid point pair, calculate the distance between the two points in the pair, and determine whether the distance is less than or equal to a preset fifth distance threshold. If so, identify the point pair as a fifth valid point pair, thus obtaining all fifth valid point pairs. The fifth distance threshold is less than the fourth distance threshold. In this way, more accurate valid point pairs are obtained from the transformed template point cloud and shoe last point cloud in step S402.

[0078] Step S404: Based on all the obtained fourth valid point pairs, the process of obtaining the fourth valid point pairs to the fifth valid point pairs is repeated. In each round of the loop, the fifth distance threshold is updated so that the updated fifth distance threshold is less than the previous fifth distance threshold, until the fifth distance threshold meets the preset convergence condition. The spatial transformation matrix corresponding to the fourth valid point pairs obtained in the last round of the loop is obtained, thereby transforming the toe and heel trajectories of the template to the corresponding positions of the toe and heel of the shoe last, respectively.

[0079] In one embodiment, in step S404, the process from step S402 to step S403 is repeated.

[0080] In one embodiment, the convergence condition is set based on requirements. In another embodiment, the convergence condition is that the updated fifth distance threshold is less than or equal to 0.5 mm.

[0081] In one embodiment, the fifth distance threshold updated in each iteration is 0.5 mm smaller than the previous fifth distance threshold, and the fifth distance threshold is 0.5 mm smaller than the fourth distance threshold. Registration is completed after six iterations.

[0082] The registration process described above can be either the registration of the heel template trajectory from the template file with the trajectory of the heel point cloud from the shoe last, or the registration of the toe template trajectory from the template file with the trajectory of the toe point cloud from the shoe last. If it is the registration of the heel template trajectory from the template file with the trajectory of the heel point cloud from the shoe last, after multiple rounds of registration, a final 4x4 spatial transformation matrix T for the heel point cloud trajectory registration can be obtained. Since the heel trajectory and the registered heel point cloud maintain the same orientation, matrix T is multiplied by the heel trajectory coordinates through matrix multiplication, thus performing the same spatial transformation on the heel trajectory and the registered point cloud. If the process involves registering the toe template trajectory of the template file with the toe point cloud trajectory of the shoe last, after multiple rounds of registration, a final 4x4 spatial transformation matrix T can be obtained. Since the toe trajectory and the registered toe point cloud maintain the same orientation, matrix T is multiplied by the toe trajectory coordinates to perform the same spatial transformation on the toe trajectory and the registered point cloud. In this way, the spatial transformation matrices for the heel and toe are calculated separately, thus transforming the heel and toe trajectories of the template to their corresponding positions.

[0083] The above-described gradually converging trajectory registration process improves the accuracy of trajectory registration, reduces reliance on manual labor, and enhances the automation level of shoe last processing.

[0084] Step S50: Based on the heel trajectory and toe trajectory after trajectory registration, the trajectory points are traversed and classified to obtain the polishing points and fitting points.

[0085] In one embodiment, please refer to Figure 5 Step S50 includes:

[0086] Step S501: Traverse any trajectory in the heel trajectory and toe trajectory. For any trajectory, obtain the endpoints and midpoints of the two ends of the trajectory. Based on these endpoints and midpoints, adjust each trajectory point from the endpoints toward the midpoint with a gradually decreasing adjustment range, so that the trajectory points move closer to the surface of the shoe last.

[0087] In one embodiment, please refer to Figure 6 Step S501 includes:

[0088] Please refer to step S5011. Figure 7 For any trajectory, obtain the endpoints p1 and p2 at both ends of the trajectory and the midpoint p3 at both ends. Based on any one of the endpoints, identify multiple points in the point cloud on the shoe last surface that are less than or equal to the preset sixth distance threshold, or identify multiple points in the point cloud on the shoe last surface that are closest to the endpoint. Then, fit the multiple points to obtain a three-dimensional surface, so as to obtain two three-dimensional surfaces corresponding to the two endpoints.

[0089] In step S5011, in one embodiment, the multiple points obtained can be based on a preset sixth distance threshold, that is, identifying points in the point cloud on the shoe last surface that are less than or equal to the preset sixth distance threshold at any given endpoint, thereby obtaining multiple points. Alternatively, they can be based on a preset point count threshold, that is, identifying the multiple points in the point cloud on the shoe last surface that are closest to any given endpoint. In one embodiment, the point count threshold can be 20. This would identify the 20 points in the point cloud on the shoe last surface that are closest to any given endpoint. The point count threshold is set according to actual needs.

[0090] In one embodiment, KD tree nearest neighbor search is used to identify the multiple points in the point cloud on the shoe last surface that are closest to the endpoint.

[0091] After obtaining multiple points, a three-dimensional surface can be obtained by fitting the data based on these points.

[0092] Step S5012: Calculate the distance from any one of the two endpoints to its corresponding three-dimensional surface, so as to obtain the distance d1 from endpoint p1 to its corresponding three-dimensional surface and the distance d2 from endpoint p2 to its corresponding three-dimensional surface.

[0093] In step S5013, based on distance d1, the trajectory points between endpoint p1 and midpoint p3 are adjusted so that endpoint p1 fits against the surface of the shoe last. The adjustment range of the trajectory points closer to the midpoint p3 is smaller, and the adjustment range of p3 is 0. Based on distance d2, the trajectory points between endpoint p2 and midpoint p3 are adjusted so that endpoint p2 fits against the surface of the shoe last. The adjustment range of the trajectory points closer to the midpoint p3 is smaller, and the adjustment range of p3 is 0.

[0094] Please refer to the adjusted trajectory. Figure 8 After adjustment, the trajectory surface is made to fit as closely as possible to the surface of the shoe last, which facilitates subsequent processing.

[0095] Step S502: Based on each adjusted trajectory point, identify the trajectory points inside the shoe last point cloud and the trajectory points outside the shoe last point cloud, and obtain the polishing points and bonding points.

[0096] In one embodiment, please refer to Figure 9 Step S502 includes:

[0097] Step S5021: For any adjusted trajectory point, identify multiple points in the point cloud on the shoe last surface that are less than or equal to a preset seventh distance threshold, or identify multiple points in the point cloud on the shoe last surface that are closest to the trajectory point, and fit the multiple points to obtain a three-dimensional surface.

[0098] In step S502 1, in one embodiment, the multiple points obtained can be based on a preset seventh distance threshold, that is, identifying points in the point cloud on the shoe last surface that are less than or equal to the preset seventh distance threshold from any trajectory point, thereby obtaining multiple points. Alternatively, they can be based on a preset point count threshold, that is, identifying the multiple points in the point cloud on the shoe last surface that are closest to any trajectory point. In one embodiment, the point count threshold can be 20. This would identify the 20 points in the point cloud on the shoe last surface that are closest to any trajectory point. The seventh distance threshold and the point count threshold are set according to actual needs.

[0099] Step S5022: Calculate the distance from any trajectory point to its corresponding three-dimensional surface, and obtain the projection point of the three-dimensional surface.

[0100] Step S5023: Obtain the first vector from any trajectory point to the origin of the coordinate system, obtain the second vector from any trajectory point to its projection point, and multiply the first and second vectors to obtain the product. If the product is less than m, the arbitrary trajectory point is identified as an internal point; if the product is greater than or equal to m, the arbitrary trajectory point is identified as an external point and marked as a fitting point. For internal points, if the distance from the internal point to its corresponding three-dimensional surface is greater than a preset difference threshold, the internal point is marked as a polishing point; otherwise, the internal point is marked as a fitting point. The difference threshold is determined based on the difference between the shoe last and the template.

[0101] This improves the fitting accuracy between the template trajectory and the shoe last, reduces the reliance on manual labor, and increases the automation level of shoe last processing.

[0102] Step S60: Based on the grinding points and bonding points, remove excess points and smooth the trajectory to obtain a smooth heel trajectory and toe trajectory, which serve as the processing trajectory.

[0103] For the polishing points, the original coordinates remain unchanged. For the bonding points, the nearest point cloud of the shoe last surface can be found by the KD tree nearest neighbor search method and projected onto the surface so that the trajectory fits the surface, thereby ensuring a smooth transition between the reprocessed surface and the original surface.

[0104] In one embodiment, since the area of ​​trajectory points in the template is larger than the actual area that needs to be polished, that is, some of the outermost areas do not need to be polished, a polishing range can be set, all the mating points are traversed, the KD tree is used to search for the nearest polishing point to the current mating point, and the distance d3 between the two points is calculated. The mating points whose distance exceeds the polishing range are discarded, thus avoiding meaningless polishing and improving polishing efficiency.

[0105] In one embodiment, each trajectory is smoothed individually. A lifting parameter L2 can be set, and the contact point is lifted based on the lifting parameter L2 and a distance d3. This raises the contact point from its original position against the shoe last surface by a height of L2*d3, moving it away from the shoe last surface, thus avoiding over-cutting in the contact area. Then, a moving least squares (MLS) smoothing is performed on all trajectory points to make the trajectory surface smooth and flat.

[0106] In one embodiment, the process further includes: checking the heel and toe tracks to determine if any abnormal areas exist; if so, issuing an alarm. If the heel track height is higher than the heel height of the shoe last, the track is extended to ensure complete polishing.

[0107] In one embodiment, the method further includes traversing all trajectory points, obtaining the neighboring points of the current trajectory point through the nearest neighbor method, fitting the local surface of the trajectory using the neighboring points, calculating the normal vector direction of the local surface, and thus determining the grinding posture.

[0108] Finally, the final polishing trajectory and normal vector are output.

[0109] One embodiment of this application provides a computer-readable storage medium storing a program, the stored program including a method for removing the toe and tail of a shoe last as described in any of the above embodiments, which can be loaded by a processor and processed.

[0110] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0111] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. A method for processing shoe lasts by removing the toe and tail, characterized in that, include: Obtain multiple master last template files for shoe last toe and heel removal processing, wherein any template file includes shoe toe template trajectory, shoe heel template trajectory and point cloud data for local registration, and the multiple master lasts include master lasts of different shoe types and sizes; For any shoe last, preprocess the 3D point cloud data based on the shoe last; Based on the preprocessed 3D point cloud data of the shoe last, the corresponding template file is registered; Based on the locally registered point cloud of the registered template file, trajectory registration is performed between the template trajectory of the template file and the point cloud of the shoe last, including trajectory registration between the heel template trajectory of the template file and the heel point cloud of the shoe last, and trajectory registration between the toe template trajectory of the template file and the toe point cloud of the shoe last. Based on the heel and toe trajectories after trajectory registration, the trajectory points are traversed and classified to obtain the polishing points and fitting points; Based on the grinding points and bonding points, excess points are removed and the trajectory is smoothed to obtain a smooth heel trajectory and toe trajectory, which serve as the processing trajectory. The registration of the corresponding template file based on the preprocessed 3D point cloud data of the shoe last includes: Based on any template file, the iterative nearest algorithm is used to find the nearest adjacent shoe last point for each template point in the template file, forming a point pair. The distance between the two points in any point pair is calculated, and it is determined whether the distance is less than or equal to a preset third distance threshold. If so, the point pair is identified as the first valid point pair, thus obtaining all the first valid point pairs. Based on any point in the template point cloud obtained from the first effective point pair, perform a spatial transformation to obtain the spatially transformed point cloud coordinates of that point. Based on the transformed point cloud coordinates, find the nearest adjacent shoe last point to obtain the second effective point pair, and thus obtain all the second effective point pairs. Based on all the second valid point pairs obtained, the process of obtaining all first valid point pairs and second valid point pairs is repeated. Based on the distance convergence condition, the average distance of the final valid point pairs is obtained, and the average distance is used as a matching score. The template file with the lowest matching score will be used as the template file for the current shoe last.

2. The method for removing the toe and tail of a shoe last as described in claim 1, characterized in that, The aforementioned template file for obtaining multiple mother lasts for shoe last toe-removal and hem-removal processing includes: For any given last, obtain the three-dimensional point cloud data of the last, which includes the three-dimensional point cloud data of the sole, the three-dimensional point cloud data of the toe, and the three-dimensional point cloud data of the heel. Based on the 3D point cloud data of the sole, the edge contour trajectory of the sole template is extracted as the sole reference trajectory; based on the 3D point cloud data of the toe and the sole reference trajectory, the toe template trajectory is extracted; based on the 3D point cloud data of the heel and the sole reference trajectory, the heel template trajectory is extracted. Based on the edge contour trajectory of the shoe sole template, point cloud data for local registration is extracted; The toe template trajectory, heel template trajectory, and point cloud data used for local registration are used as template files for the toe and heel removal processing of the shoe last, thereby obtaining multiple template files, which correspond to the master lasts of multiple shoe types.

3. The method for removing the toe and tail of a shoe last as described in claim 2, characterized in that, The extraction of point cloud data for local registration based on the edge contour trajectory of the shoe sole template includes: For any point in the three-dimensional point cloud of the shoe sole, based on the edge contour trajectory of the shoe sole template, the minimum distance between the point and the edge contour trajectory is identified in the plane where the shoe sole is located using a contour enclosing method. If the minimum distance is less than or equal to a preset second distance threshold, then the point is taken as a point in the point cloud data of local registration.

4. The method for removing the toe and tail of a shoe last as described in claim 1, characterized in that, The step of performing a spatial transformation on any point in the template point cloud obtained based on the first valid point pair to obtain the spatially transformed point cloud coordinates of that arbitrary point includes: A 4x4 spatial transformation matrix is ​​obtained using the least squares method. This spatial transformation matrix includes rotation and translation transformations. The spatial transformation matrix is ​​then used to transform the template point cloud, resulting in the spatially transformed point cloud coordinates of any point in the template point cloud.

5. The method for removing the toe and tail of a shoe last as described in claim 1, characterized in that, The aforementioned local registration point cloud based on the registered template file performs trajectory registration between the template trajectory of the template file and the point cloud of the shoe last, including: Based on the locally registered point cloud of the registered template file, the overall point cloud is coarsely registered using a preset fourth distance threshold to obtain the third effective point pair; Based on any point in the template point cloud obtained from the third valid point pair, perform a spatial transformation to obtain the spatially transformed point cloud coordinates of that arbitrary point. Based on the transformed point cloud coordinates, identify the nearest adjacent shoe last point to obtain the fourth valid point pair, and thus obtain all the fourth valid point pairs. For any fourth valid point pair, calculate the distance between the two points in the pair, and determine whether the distance is less than or equal to a preset fifth distance threshold. If so, identify the point pair as a fifth valid point pair, and thus obtain all fifth valid point pairs; where the fifth distance threshold is less than the fourth distance threshold. Based on all the obtained fourth valid point pairs, the process of obtaining the fourth valid point pair to the fifth valid point pair is repeated. In each round of the loop, the fifth distance threshold is updated so that the updated fifth distance threshold is less than the original fifth distance threshold, until the fifth distance threshold meets the preset convergence condition. The spatial transformation matrix corresponding to the fourth valid point pair obtained in the last round of the loop is obtained, thereby transforming the toe and heel trajectories of the template to the corresponding positions of the toe and heel of the shoe last, respectively.

6. The method for removing the toe and tail of a shoe last as described in claim 1, characterized in that, The process of classifying and traversing the trajectory points based on the registered heel and toe trajectories to obtain polishing points and fitting points includes: Traverse any trajectory in the heel trajectory and toe trajectory. For any trajectory, obtain the endpoints and midpoints of the two ends of the trajectory. Based on these endpoints and midpoints, adjust each trajectory point from the endpoints toward the midpoint with gradually decreasing adjustment range, so that the trajectory points move closer to the surface of the shoe last. Based on each adjusted trajectory point, the trajectory points inside and outside the shoe last point cloud are identified, and the polishing points and bonding points are obtained.

7. The method for removing the toe and tail of a shoe last as described in claim 6, characterized in that, The process of traversing any trajectory in the heel and toe tracks involves, for any trajectory, obtaining the endpoints and midpoints of both ends of the trajectory, and based on these endpoints and midpoints, adjusting each trajectory point from the endpoints towards the midpoint with gradually decreasing amplitude, so that the trajectory points move closer to the shoe last surface. This includes: For any trajectory, obtain the endpoints p1 and p2 at both ends of the trajectory and the midpoint p3 at both ends. Based on any one of the endpoints, identify multiple points in the point cloud on the shoe last surface that are less than or equal to a preset sixth distance threshold, or identify multiple points in the point cloud on the shoe last surface that are closest to the endpoint. Then, fit the multiple points to obtain a three-dimensional surface, so as to obtain two three-dimensional surfaces corresponding to the two endpoints. Calculate the distance from any one of the two endpoints to its corresponding three-dimensional surface to obtain the distance d1 from endpoint p1 to its corresponding three-dimensional surface and the distance d2 from endpoint p2 to its corresponding three-dimensional surface. Based on distance d1, the trajectory points between endpoint p1 and midpoint p3 are adjusted so that endpoint p1 fits against the surface of the shoe last. The closer the trajectory points are to the midpoint p3, the smaller the adjustment range, and the adjustment range of p3 is 0. Based on distance d2, the trajectory points between endpoint p2 and midpoint p3 are adjusted so that endpoint p2 fits against the surface of the shoe last. The closer the trajectory points are to the midpoint p3, the smaller the adjustment range, and the adjustment range of p3 is 0.

8. The method for removing the toe and tail of a shoe last as described in claim 6, characterized in that, The process of identifying trajectory points inside and outside the shoe last point cloud based on each adjusted trajectory point, and obtaining polishing points and bonding points, includes: For any adjusted trajectory point, identify multiple points in the point cloud on the shoe last surface that are less than or equal to a preset seventh distance threshold, or identify multiple points in the point cloud on the shoe last surface that are closest to the trajectory point, and fit the multiple points to obtain a three-dimensional surface. Calculate the distance from any trajectory point to its corresponding three-dimensional surface, and obtain the projection point of the three-dimensional surface; Obtain the first vector from any trajectory point to the origin, and the second vector from any trajectory point to its projection point. Multiply the first and second vectors to obtain the product. If the product is less than m, the trajectory point is identified as an internal point. If the product is greater than or equal to m, the trajectory point is identified as an external point and marked as a fitting point. For internal points, if the distance from the internal point to its corresponding 3D surface is greater than a preset difference threshold, the internal point is marked as a polishing point. Otherwise, the internal point is marked as a fitting point. The difference threshold is determined based on the difference between the shoe last and the template.

9. The method for removing the toe and tail of a shoe last as described in claim 6, characterized in that, The processing method further includes: Traverse all trajectory points, obtain the neighboring points of the current trajectory point through the nearest neighbor method, fit the local surface of the trajectory using the neighboring points, calculate the normal vector direction of the local surface, thereby determining the grinding posture, and output the final grinding trajectory and normal vector.

10. A computer-readable storage medium, characterized in that, The medium stores a program that can be loaded by a processor and executed as described in any one of claims 1 to 9.

Images