A method for quickly individualizing a sole model based on an upper model
By combining 3D scanning and 3D printing technologies with PCA and ICP algorithms for model registration and deformation processing, the problems of high cost and low efficiency in traditional shoe sole customization are solved, achieving efficient and low-cost personalized shoe sole customization and ensuring precise matching between the shoe sole and the upper model.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIAMEN ARITA WUWEI INTELLIGENT TECH CO LTD
- Filing Date
- 2023-05-31
- Publication Date
- 2026-08-04
AI Technical Summary
Traditional shoe sole manufacturing methods make it difficult to achieve personalized customization. Mold making is costly and difficult to match one-to-one, manual customization is expensive and inefficient, and existing CAD/CAM software design is costly and complex to operate, making it difficult to mass-produce.
3D scanning technology is used to obtain user foot information, PCA and ICP algorithms are used for model registration, and 3D printing technology is used to print personalized custom shoe sole models. Harmonic function and radial basis function algorithms are combined for deformation processing to ensure that the shoe sole model and the shoe upper model fit perfectly.
It enables efficient and low-cost personalized sole customization, improves production efficiency and accuracy, reduces labor and time costs, and ensures precise matching between sole and upper models.
Smart Images
Figure CN116533527B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of shoe sole personalization technology, specifically relating to a method for rapid personalization of shoe sole models based on shoe upper models. Background Technology
[0002] Currently, shoe sole manufacturing primarily relies on molds. The main process includes: 1. Material selection: different materials have different qualities and properties; 2. Mold design: using CAD software to design the sole mold; 3. Material processing: extruding the material into the mold for injection molding and pressing; 4. Finishing: improving surface quality through physical polishing and chemical soaking. However, different people have different foot characteristics, resulting in variations in the designed lasts and uppers, naturally leading to differences in soles produced for different individuals. This mass production method using molds results in high similarity and standardization. This leads to shoes of the same style having fixed shapes, differing only in size. This results in a reverse wearing pattern, where the feet adapt to the shoes. Research shows that long-term wear of shoes that do not fit the physiological needs of the feet can have serious adverse effects on physical health and development, even worsening existing symptoms and leading to serious consequences. Therefore, modern life has an increasing demand for personalized, customized soles.
[0003] Traditional personalized shoe soles rely on handcrafting, resulting in high labor costs and hindering large-scale commercial application. Currently, an increasing number of enterprises, institutions, and universities are integrating digitalization with the shoe industry, proposing various solutions for personalized shoe sole customization. A common approach is parametric sole design, which involves the following steps: 1. Using 3D scanning technology to obtain individual foot models and create last and upper models; 2. Using CAD / CAM software (such as Rhino and Solidworks) to assist manual design, or directly using programmatic parametric design, parametrically designing the sole model based on the last and upper models; 3. Using additive manufacturing technology to print the parametrically designed sole model. This personalized shoe sole manufacturing method fully integrates 3D scanning, model design, and 3D printing technologies. In intelligent manufacturing, product quality has become a key competitive factor for shoe companies, and personalized customization can help them stand out in the competition by shifting from a quantity-based production model to a quality-based production model.
[0004] Traditional mold-based sole manufacturing methods, due to the cost of mold production, make it difficult to achieve a one-to-one correspondence between the mold and the sole. Therefore, this method is only suitable for mass production and cannot meet the needs of personalized sole customization. Furthermore, the fixed production chain hinders industrial restructuring. Manual personalized sole customization is expensive due to high labor costs, making large-scale production difficult, inefficient, and prohibitively expensive for ordinary consumers, thus lacking commercial value. Parametric sole design, on the other hand, offers a novel production method that better adapts to different individual foot characteristics and allows for customized sole designs. However, the CAD / CAM software it relies on is expensive, manual design is inefficient, and its intelligence is limited. Furthermore, parametric design suffers from insufficient reliability and stability, primarily focusing on sole structure rather than shape, making it difficult to guarantee the shape of the produced soles. The biggest drawbacks of using CAD / CAM software to assist manual design are its high cost, low efficiency, complex operation, and high level of professional knowledge required from users. Summary of the Invention
[0005] In view of the shortcomings of the prior art, the purpose of this invention is to provide a method for rapid personalized customization of shoe sole models based on shoe upper models.
[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for rapid personalized customization of shoe sole models based on shoe upper models, comprising the following steps:
[0007] S1. Using 3D scanning technology, scan the outline of the user's foot that needs personalized custom soles to obtain the user's foot information, and make the corresponding shoe last model and the shoe upper model that matches the shoe last model accordingly.
[0008] S2. Select the original sole model; based on the length, width and height of the upper model, perform a step-by-step registration process on the original sole model to obtain an intermediate sole model that is effectively registered with the upper model; based on the shape characteristics of the upper model, perform adaptive deformation processing on the intermediate sole model to obtain the final sole model.
[0009] S3. Using 3D printing technology, the final shoe sole model is printed.
[0010] Preferably, in step S2, the specific method for performing scaled-down registration processing on the original sole model is as follows: In the world coordinate system, fix the spatial position of the upper model, and manually adjust the length, width, and height directions of the original sole model to be approximately the same as those of the upper model, with the top of the original sole model and the bottom of the upper model roughly facing each other; use a feature extraction method to extract feature A from the upper model and the corresponding feature B from the sole model, then use the PCA algorithm to calculate the principal directions of feature A and feature B, and determine the sign of the principal directions based on the x, y, and z axes of the world coordinate system. Next, calculate the offset distance of the vertex data center points of feature A and feature B in the width direction. Using the principal directions and their positive and negative values of feature A and feature B, and the offset distance of the vertex data center points in the width direction, perform pose transformation of the original sole model so that the length, width and height directions of the original sole model are exactly the same as those of the upper model, and the top of the original sole model is completely facing the bottom of the upper model. Finally, based on the projection distance of the vertex data center points in the length direction, perform tiered registration processing on the original sole model to obtain an intermediate sole model that is effectively registered with the upper model.
[0011] Preferably, the PCA algorithm calculates as follows:
[0012]
[0013]
[0014] Where cov is the covariance, x i Let x be the x-coordinate component of vertex i, and let x be the mean of the x-coordinate components of the vertex dataset; D is the covariance matrix of the vertex dataset; calculate the covariance between all vertex coordinate components x, y, z, and solve the covariance matrix through eigenvalue decomposition or singular value decomposition to obtain the main directions of the length, width, and height of the vertex dataset; calculate the mean of the feature point set to obtain the vertex data center point.
[0015] Preferably, when performing pose transformation of the original sole model, an ICP registration algorithm is added as an auxiliary method.
[0016] Preferably, the ICP registration algorithm is calculated as follows:
[0017]
[0018] Where E is the distance error between the source and target point sets under the R,t transformation, R is the rotation matrix, t is the translation matrix, and p i Let the target point be i, q i Let i be the source point corresponding to the target point i; calculate R and t under the minimum distance error to obtain the transformation matrix, and then perform the position transformation.
[0019] Preferably, in step S2, the specific method for adaptive deformation processing of the intermediate shoe sole model is as follows: select control points and deformation points of the intermediate shoe sole model, calculate the displacement of the control points, use the deformation algorithm to calculate the displacement of the deformation points in reverse, and thus calculate the deformed shoe sole model based on the displacement of the control points and the displacement of the deformation points, so that the features of the final shoe sole model meet the surface feature requirements of the shoe upper model.
[0020] Preferably, the deformation algorithm includes a harmonic function deformation algorithm and a radial basis function interpolation algorithm;
[0021] Preferably, the harmonic function transformation algorithm satisfies the Laplace equation:
[0022]
[0023]
[0024]
[0025] in, Let Δ denote the second-order differential, Δ be the Laplace operator, and δ be the second-order differential. i For the discretized Laplace operator, v i v j Let N(i) ∈ {j | (i,j) ∈ E} be the set of vertices, and E be the set of edges, ω. ij For weights, w ij w ik These are the weights of other edges within the adjacent range;
[0026] The radial basis function interpolation algorithm is defined as follows:
[0027]
[0028] Where G is the radial basis function between control points, and A x A y A z D represents the weighting coefficients for the displacement of the control points in the x, y, and z directions, respectively. x D y D z These represent the displacements of the control points in the x, y, and z directions, respectively. After calculating the weighting coefficients, the displacements of the deformation points are inversely calculated based on the radial basis functions between the deformation points.
[0029] The matrix expression form corresponding to formula (7) is as follows:
[0030]
[0031]
[0032]
[0033]
[0034] Among them, g ij Let x be the radial basis function values corresponding to control points i and j, where i,j∈n. n y n z n Let x, y, and z be the components of control point n; C1, C2, C3, and C4 are the weighting coefficients of control point n in the x, y, and z directions, respectively, and represent the polynomial correlation coefficients. These represent the displacements of control point n in the x, y, and z directions, respectively; the basis function g(x) uses Wendland's C2 function:
[0035]
[0036] Where x = ||r i -r j || / d,||r i -r j || represents the Euclidean distance between vertices i and j, and d represents the radius of action of the radial basis function.
[0037] Preferably, the deformed sole model is sealed using a hole-filling algorithm to obtain a closed sole model. The difference between the closed sole model and the upper model is calculated using a Boolean operation to obtain the final sole model that perfectly matches the upper model.
[0038] Preferably, the deformation strategy of the deformation algorithm includes deformation in the height direction of the sole, deformation in the horizontal plane of the sole's length and width, and deformation in local areas of the sole.
[0039] Preferably, the deformation parameters of the deformation algorithm include sampling distance, sampling cosine, and shoe upper gap; wherein the sampling distance is set to 10mm, the sampling cosine is set to 0.75, and the shoe upper gap is set to 0mm.
[0040] Preferably, the original sole model includes an inner cavity.
[0041] Compared with the prior art, the present invention has the following beneficial effects:
[0042] (1) This method is highly intelligent, does not require high professional knowledge from operators, is easy to operate, and has simple steps, which greatly reduces labor costs, time costs, and production costs, and greatly improves efficiency.
[0043] (2) This method is mainly for shoe sole models, with a more precise scope of application, stronger targeting, high efficiency, good effect and strong reliability in personalized shoe sole customization;
[0044] (3) This method ensures that the basic structural features and basic texture features of the original sole model remain unchanged, with good surface quality and high accuracy; (4) The final sole model obtained by this method can perfectly adapt to the contour features of the upper model, truly realizing the personalized customization of the sole shape features.
[0045] (5) This method can automatically achieve effective matching between the sole model and the upper model, so that the 3D printed sole can match the upper more accurately, which is convenient for shoe production. Attached Figure Description
[0046] Figure 1 This is a flowchart illustrating Embodiment 1 of the present invention.
[0047] Figure 2 This is a flowchart illustrating Embodiment 2 of the present invention.
[0048] Figure 3 This is the front view of the shoe upper model and the original shoe sole model facing each other in Case 1.
[0049] Figure 4 This is a rear view of the shoe upper model and the original shoe sole model facing each other in Case 1.
[0050] Figure 5 This is a left view of the shoe upper model and the original shoe sole model facing each other in Case 1.
[0051] Figure 6 This is a right-hand view of the shoe upper model and the original shoe sole model from Case 1, facing each other.
[0052] Figure 7 This is a top view of the shoe upper model and the original shoe sole model facing each other in Case 1.
[0053] Figure 8 This is a bottom view of the shoe upper model and the original shoe sole model from Case 1, facing each other.
[0054] Figure 9 This is a 3D view of the shoe upper model and the original shoe sole model facing each other in Case 1.
[0055] Figure 10 This is the front view that matches the upper model and the final sole model of Case 1.
[0056] Figure 11 This is a rear view showing the combination of the upper model and the final sole model from Case 1.
[0057] Figure 12 The left view shows the combination of the upper model and the final sole model from Case 1.
[0058] Figure 13 The right view shows the combination of the upper model and the final sole model from Case 1.
[0059] Figure 14 This is a top view showing the combination of the upper model and the final sole model from Case 1.
[0060] Figure 15 This is a bottom view showing the upper and final sole models of Case 1.
[0061] Figure 16 This is a 3D model that combines the upper and final sole models from Case 1.
[0062] Figure 17 For example, the cross-sectional features between the upper model and the midsole model in Case 1. Figure 1 .
[0063] Figure 18 For example, the cross-sectional features between the upper model and the midsole model in Case 1. Figure 2 .
[0064] Figure 19 For the cross-sectional features between the upper model and the final sole model in Case 1 Figure 1 .
[0065] Figure 20 For the cross-sectional features between the upper model and the final sole model in Case 1 Figure 2 .
[0066] Figure 21 For the cross-sectional features between the upper model and the final sole model in Case 1 Figure 3 .
[0067] Figure 22 For the cross-sectional features between the upper model and the final sole model in Case 1 Figure 4 .
[0068] Figure 23 For the cross-sectional features between the upper model and the final sole model in Case 1 Figure 5 .
[0069] Figure 24 For the cross-sectional features between the upper model and the final sole model in Case 1 Figure 6 .
[0070] Figure 25 This is a 3D view of the original shoe sole model in Case 2.
[0071] Figure 26 This is a 3D model that combines the upper and final sole models of Case 2.
[0072] Figure 27 For the cross-sectional features between the upper model and the midsole model in Case 2 Figure 1 .
[0073] Figure 28 For the cross-sectional features between the upper model and the midsole model in Case 2 Figure 2 .
[0074] Figure 29 For the cross-sectional features between the upper model and the final sole model in Case 2 Figure 1 .
[0075] Figure 30 For the cross-sectional features between the upper model and the final sole model in Case 2 Figure 2 .
[0076] Figure 31 This is a 3D view of the original shoe sole model in Case 3.
[0077] Figure 32 This is a 3D model that combines the upper and final sole models of Case 3.
[0078] Figure 33 For the cross-sectional features between the upper model and the midsole model in Case 3 Figure 1 .
[0079] Figure 34 For the cross-sectional features between the upper model and the midsole model in Case 3 Figure 2 .
[0080] Figure 35 For the cross-sectional features between the upper model and the final sole model in Case 3 Figure 1 .
[0081] Figure 36 For the cross-sectional features between the upper model and the final sole model in Case 3 Figure 2 . Detailed Implementation
[0082] To make the above features and advantages of the present invention more apparent and understandable, specific embodiments are described below in conjunction with the accompanying drawings for detailed explanation.
[0083] Example 1: As Figure 1 As shown, a method for rapid personalized customization of shoe sole models based on shoe upper models includes the following steps:
[0084] S1. Using 3D scanning technology, scan the outline of the user's foot that needs personalized custom soles to obtain the user's foot information, and make the corresponding shoe last model and the shoe upper model that matches the shoe last model accordingly.
[0085] S2. Select the original sole model; based on the length, width and height of the upper model, perform a step-by-step registration process on the original sole model to obtain an intermediate sole model that is effectively registered with the upper model; based on the shape characteristics of the upper model, perform adaptive deformation processing on the intermediate sole model to obtain the final sole model.
[0086] S3. Using 3D printing technology, the final shoe sole model is printed.
[0087] In this embodiment, in step S2, after selecting the original sole model, the internal area of the original sole model can be selected. Of course, the original sole model may already include the internal area of the original sole model.
[0088] In this embodiment, the specific method for performing scaled-down registration processing on the original sole model in step S2 is as follows: In the world coordinate system, since the upper model and the original sole model are at different positions, the spatial position of the upper model needs to be fixed, and the length, width, and height directions of the original sole model should be manually adjusted to be approximately the same as those of the upper model, with the top of the original sole model and the bottom of the upper model roughly facing each other. Since the upper model and the original sole model are two different models, traditional registration algorithms such as ICP, FPFH, and NDT cannot be directly used (these algorithms are applicable to the registration of the same model). Therefore, this embodiment uses a feature extraction method to extract feature A of the upper model and the corresponding feature B of the sole model, and then utilizes... The PCA algorithm calculates the principal directions of features A and B, and determines the sign of the principal directions based on the x, y, and z axes of the world coordinate system. It then calculates the offset distance of the vertex data centers of features A and B in the width direction. Using the principal directions and their signs, along with the offset distance of the vertex data centers in the width direction, the pose transformation of the original sole model is performed, ensuring that the length, width, and height directions of the original sole model are identical to those of the upper model, and that the top of the original sole model and the bottom of the upper model are completely aligned. Finally, based on the projection distance of the vertex data centers in the length direction, the original sole model undergoes a step-by-step registration process to obtain an intermediate sole model that is effectively registered with the upper model.
[0089] In this embodiment, the PCA algorithm calculates as follows:
[0090]
[0091]
[0092] Where cov is the covariance, x iLet x be the x-coordinate component of vertex i, and let x be the mean of the x-coordinate components of the vertex dataset; D is the covariance matrix of the vertex dataset; calculate the covariance between all vertex coordinate components x, y, z, and solve the covariance matrix through eigenvalue decomposition or singular value decomposition to obtain the main directions of the length, width, and height of the vertex dataset; calculate the mean of the feature point set to obtain the vertex data center point.
[0093] In this embodiment, to further improve the model registration accuracy, an ICP registration algorithm can be added to assist in the pose transformation of the original shoe sole model.
[0094] In this embodiment, the ICP registration algorithm is calculated as follows:
[0095]
[0096] Where E is the distance error between the source and target point sets under the R,t transformation, R is the rotation matrix, t is the translation matrix, and p i Let the target point be i, q i Let i be the source point corresponding to the target point i; calculate R and t under the minimum distance error to obtain the transformation matrix, and then perform the position transformation.
[0097] In this embodiment, the specific method for adaptive deformation processing of the intermediate shoe sole model in step S2 is as follows: select control points and deformation points of the intermediate shoe sole model, calculate the displacement of the control points, use the deformation algorithm to calculate the displacement of the deformation points in reverse, and thus calculate the deformed shoe sole model based on the displacement of the control points and the displacement of the deformation points, so that the features of the final shoe sole model meet the surface feature requirements of the shoe upper model.
[0098] In this embodiment, the deformation algorithm includes a harmonic function deformation algorithm and a radial basis function (RBF) interpolation algorithm;
[0099] In this embodiment, the harmonic function transformation algorithm satisfies the Laplace equation:
[0100]
[0101]
[0102]
[0103] in, Let Δ denote the second-order differential, Δ be the Laplace operator, and δ be the second-order differential. i For the discretized Laplace operator, v i v j Let N(i) ∈ {j | (i,j) ∈ E} be the set of vertices, and E be the set of edges, ω. ij For weights, w ij wik These are the weights of other edges within the adjacent range.
[0104] In this embodiment, the radial basis function interpolation algorithm is defined as follows:
[0105]
[0106] Where G is the radial basis function between control points, and A x A y A z D represents the weighting coefficients for the displacement of the control points in the x, y, and z directions, respectively. x D y D z These represent the displacements of the control points in the x, y, and z directions, respectively. After calculating the weighting coefficients, the displacements of the deformation points are inversely calculated based on the radial basis functions between the deformation points.
[0107] The matrix expression form corresponding to formula (7) is as follows:
[0108]
[0109]
[0110]
[0111]
[0112] Among them, g ij Let x be the radial basis function values corresponding to control points i and j, where i,j∈n. n y n z n Let x, y, and z be the components of control point n; C1, C2, C3, and C4 are the weighting coefficients of control point n in the x, y, and z directions, respectively, and represent the polynomial correlation coefficients. These represent the displacements of control point n in the x, y, and z directions, respectively; the basis function g(x) uses Wendland's C2 function:
[0113]
[0114] Where x = ||r i -r j || / d,||r i -r j || represents the Euclidean distance between vertices i and j, and d represents the radius of action of the radial basis function.
[0115] In this embodiment, the deformed sole model is sealed using a hole-filling algorithm to obtain a closed sole model. The difference between the closed sole model and the upper model is then calculated using a Boolean operation to obtain the final sole model that perfectly matches the upper model. The hole-filling algorithm can be a 3D model hole-filling algorithm based on OpenMesh, a hole-filling algorithm based on the Libigl library, or other similar algorithms.
[0116] In this embodiment, the deformation strategy of the deformation algorithm includes deformation in the height direction of the sole, deformation in the horizontal plane of the sole length and width, and deformation in the local area of the sole. Based on the coordination between the shape features of the intermediate sole model and the upper model, the shape features of the intermediate sole model are deformed step by step so that the features of the deformed sole model meet the surface feature requirements of the upper model.
[0117] In this embodiment, the deformation parameters of the deformation algorithm include sampling distance, sampling cosine, and shoe upper gap. The sampling distance controls the number of sampling points (including control points and deformation points). A larger sampling distance results in fewer sampling points, faster calculation speed, but also greater calculation error, and vice versa. The sampling cosine is the dot product of the normal vector of the model patch direction and the normal vector of the shoe upper / sole height direction. A smaller sampling cosine value results in a larger extracted feature area of the shoe upper and sole models, and vice versa. This parameter directly affects the deformation effect. The shoe upper gap is used to offset the shoe upper model to increase or decrease the overall thickness of the sole model. The sampling distance is preferably, but not limited to, 10mm, and the sampling cosine is preferably, but not limited to, 0.75. The shoe upper gap is set to 0mm by default. By setting reasonable values for the sampling distance, sampling cosine, and shoe upper gap, faster deformation processing and better deformation effects can be achieved.
[0118] The following section describes how the method of this invention was used to process three different original sole and upper models to obtain three case studies.
[0119] like Figures 3-24 Case 1 shown, in which Figures 3-9 This shows the spatial relative positions of the first original sole model and the first upper model. Figures 10-16 This shows the fit between the first final sole model and the first upper model obtained after the initial original sole model has undergone progressive registration and adaptive deformation. Figures 17-18 The cross-sectional features between the first intermediate sole model and the first upper model without adaptive deformation are shown. Figures 19-24 The cross-sectional features between the first final sole model and the first upper model, obtained after adaptive deformation of the first intermediate sole model, are shown.
[0120] like Figures 25-30 Case 2 shown, in which Figure 25The second original sole model was shown. Figure 26 This shows the fit between the second final sole model and the second upper model. Figures 27-28 The cross-sectional features between the second intermediate sole model and the second upper model, which exhibit no adaptive deformation, are shown. Figures 29-30 The cross-sectional features between the second final sole model and the second upper model, obtained after adaptive deformation of the second intermediate sole model, are shown.
[0121] like Figures 31-36 Case 3 is shown, in which Figure 31 The third original sole model was shown. Figure 32 The image shows the fit between the third final sole model and the third upper model. Figures 33-34 The cross-sectional features between the third intermediate sole model and the third upper model, which exhibit no adaptive deformation, are shown. Figures 35-36 The cross-sectional features between the third final sole model and the third upper model, obtained after adaptive deformation of the third intermediate sole model, are shown.
[0122] From Case 1 Figures 19-24 and Figures 17-18 In contrast, Case 2 Figures 29-30 and Figures 27-28 In comparison, Case 3 Figures 35-36 and Figures 33-34 In comparison, it can be seen that the sole model customized by the method of the present invention ensures that the basic structural and textural features of the original sole model remain unchanged, with good surface quality and high accuracy; it can perfectly adapt to the contour features of the upper model, and truly realize the personalized customization of the sole shape features.
[0123] Example 2: As Figure 2 As shown, a method for rapid personalized customization of shoe sole models based on shoe upper models differs from Embodiment 1 in that the original shoe sole model includes an inner cavity.
[0124] In this embodiment, in step S2, after selecting the original sole model (or the internal area of the original sole model), the inner cavity model is then selected, and the inner cavity model is assembled onto the original sole model. After selecting the inner cavity model, the internal area of the inner cavity model can then be selected. Of course, the inner cavity model may already contain the internal area of the inner cavity model.
[0125] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, and alterations made by those skilled in the art to the above embodiments without departing from the technical solution of the present invention, based on the technical essence of the present invention, should be within the scope of the present invention.
Claims
1. A method for rapid personalized customization of shoe sole models based on shoe upper models, characterized in that, Includes the following steps: S1. Using 3D scanning technology, scan the outline of the user's foot that needs personalized custom soles to obtain the user's foot information, and make the corresponding shoe last model and the shoe upper model that matches the shoe last model accordingly. S2. Select the original sole model; based on the length, width and height of the upper model, perform a step-by-step registration process on the original sole model to obtain an intermediate sole model that is effectively registered with the upper model; based on the shape characteristics of the upper model, perform adaptive deformation processing on the intermediate sole model to obtain the final sole model. The specific method for performing scaled-down registration on the original sole model is as follows: In the world coordinate system, fix the spatial position of the upper model and manually adjust the length, width, and height directions of the original sole model to be approximately the same as those of the upper model, with the top of the original sole model and the bottom of the upper model roughly facing each other. Use feature extraction methods to extract feature A from the upper model and the corresponding feature B from the sole model. Then, use the PCA algorithm to calculate the principal directions of feature A and feature B, and determine the sign of the principal directions based on the x, y, and z axes of the world coordinate system. Finally, calculate the feature... The offset distance of the vertex data center points of feature A and feature B in the width direction is used to perform pose transformation of the original sole model. This makes the length, width and height directions of the original sole model exactly the same as those of the upper model, and the top of the original sole model and the bottom of the upper model are completely facing each other. Finally, based on the projection distance of the vertex data center points in the length direction, the original sole model is subjected to tiered registration processing to obtain an intermediate sole model that is effectively registered with the upper model. S3. Using 3D printing technology, the final shoe sole model is printed.
2. The method for rapid personalized customization of sole models based on shoe upper models according to claim 1, characterized in that, The PCA algorithm calculates the following: (1) (2) Where cov is the covariance, x i Let x be the x-coordinate component of vertex i. Let be the mean of the x-coordinate components of the vertex dataset; D is the covariance matrix of the vertex dataset; calculate the covariance between all vertex coordinate components x, y, z, and solve the covariance matrix through eigenvalue decomposition or singular value decomposition to obtain the main directions of the length, width, and height of the vertex dataset; calculate the mean of the feature point set to obtain the vertex data center point.
3. The method for rapid personalized customization of shoe sole models based on shoe upper models according to claim 1, characterized in that, When performing pose transformation of the original sole model, an ICP registration algorithm is added as an aid. The ICP registration algorithm is calculated as follows: (3) Where E is the distance error between the source and target point sets under the R, t transformation, R is the rotation matrix, t is the translation matrix, and p i Let the target point be i, q i Let i be the source point corresponding to the target point i; calculate R and t under the minimum distance error to obtain the transformation matrix, and then perform the position transformation.
4. The method for rapid personalized customization of sole models based on shoe upper models according to claim 1, characterized in that, In step S2, the specific method for adaptive deformation processing of the intermediate shoe sole model is as follows: select the control points and deformation points of the intermediate shoe sole model, calculate the displacement of the control points, use the deformation algorithm to calculate the displacement of the deformation points in reverse, and thus calculate the deformed shoe sole model based on the displacement of the control points and the displacement of the deformation points, so that the features of the final shoe sole model meet the surface feature requirements of the shoe upper model.
5. The method for rapid personalized customization of sole models based on shoe upper models according to claim 4, characterized in that, The deformation algorithm includes a harmonic function deformation algorithm and a radial basis function interpolation algorithm; The harmonic function transformation algorithm satisfies the Laplace equation: (4) (5) (6) in, Let Δ denote the second-order differential, Δ be the Laplace operator, and δ be the second-order differential. i For the discretized Laplace operator, v i v j Let N(i) ∈ {j | (i,j) ∈ E} be the set of vertices, and E be the set of edges, ω. ij For weights, w ij w ik These are the weights of other edges within the adjacent range; The radial basis function interpolation algorithm is defined as follows: (7) Where G is the radial basis function between control points, and A x A y A z D represents the weighting coefficients for the displacement of the control points in the x, y, and z directions, respectively. x D y D z These represent the displacements of the control points in the x, y, and z directions, respectively. After calculating the weighting coefficients, the displacements of the deformation points are inversely calculated based on the radial basis functions between the deformation points. The matrix expression form corresponding to formula (7) is as follows: (8) (9) (10) (11) Among them, g ij Let x be the radial basis function values corresponding to control points i and j, where i,j∈n. n y n z n Let x, y, and z be the components of control point n; , , C1, C2, C3, and C4 are the weighting coefficients of control point n in the x, y, and z directions, respectively, and represent the polynomial correlation coefficients. , , These represent the displacements of control point n in the x, y, and z directions, respectively; the basis function g(x) uses Wendland's C2 function: (12) Where x = ||r i -r j || / d,||r i -r j || represents the Euclidean distance between vertices i and j, and d represents the radius of action of the radial basis function.
6. The method for rapid personalized customization of shoe sole models based on shoe upper models according to claim 4, characterized in that, The deformed sole model is sealed using a hole-filling algorithm to obtain a closed sole model. The difference between the closed sole model and the upper model is calculated using a Boolean operation to obtain the final sole model that perfectly matches the upper model.
7. The method for rapid personalized customization of sole models based on shoe upper models according to claim 4, characterized in that, The deformation strategy of the deformation algorithm includes deformation in the height direction of the sole, deformation in the horizontal plane of the sole's length and width, and deformation in local areas of the sole.
8. The method for rapid personalized customization of sole models based on shoe upper models according to claim 4, characterized in that, The deformation parameters of the deformation algorithm include sampling distance, sampling cosine, and shoe upper gap; wherein, the sampling distance is set to 10 mm, the sampling cosine is set to 0.75, and the shoe upper gap is set to 0 mm.
9. The method for rapid personalized customization of shoe sole models based on shoe upper models according to claim 1, characterized in that: The original shoe sole model includes an inner cavity.