A 3D printing path planning method and system

Abstract

The application provides a 3D printing path planning method and system, and belongs to the field of 3D printing. The method comprises the following steps: performing multi-scale geometric feature extraction on a three-dimensional model to be printed to obtain a layered feature mapping set; performing self-adaptive voxelization decomposition on the three-dimensional model to be printed based on the layered feature mapping set to obtain a non-uniform voxel grid structure; performing direction field calculation on the non-uniform voxel grid structure to obtain a printing direction tensor field; generating a spiral trajectory based on the printing direction tensor field and the non-uniform voxel grid structure to obtain a continuous spiral printing path; and controlling a preset printing nozzle to perform material deposition printing based on the continuous spiral printing path. The application effectively solves the technical problems that the existing method is prone to uneven filling, excessive support structure and surface precision reduction.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of 3D printing, and particularly relates to a 3D printing path planning method and system. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute the prior art.

[0003] Under the background of rapid development of 3D printing technology, path planning as a key link affecting printing quality and efficiency is attracting more and more researchers' attention. Traditional path planning methods are usually based on uniform slicing and fixed scanning strategy, which is difficult to adapt to the printing needs of complex geometric structure models. Especially when dealing with models with multi-scale features, high curvature or overhanging structures, existing methods are prone to uneven filling, excessive support structure and surface precision decline, which limits the application expansion of 3D printing in high-precision manufacturing field.

[0004] In recent years, although some researches have tried to introduce adaptive voxelization and direction field optimization to improve the intelligence and continuity of the printing path, most of the methods still rely on experience setting parameters, and lack the ability of deep perception of model geometric features. In addition, when dealing with non-uniform material distribution or multi-nozzle collaborative printing scene, existing algorithms often ignore the collaborative optimization between local details and overall structure, resulting in defects such as stress concentration and insufficient interlayer bonding force in the printing process. Therefore, how to realize efficient and high-quality printing path planning for complex three-dimensional models is still a key challenge in current research. SUMMARY

[0005] In order to overcome the above technical problems of the prior art, the present application provides a 3D printing path planning method and system, which extracts multi-scale geometric features from a three-dimensional model to be printed to obtain a hierarchical feature mapping set, and then performs adaptive voxelization decomposition to obtain a non-uniform voxel grid structure, and then calculates a direction field to obtain a printing direction tensor field, and then generates a continuous spiral printing path, and finally controls the nozzle to print, solving the technical problems of uneven filling, excessive support structure and surface precision decline in existing methods.

[0006] To achieve the above purpose, one or more embodiments of the present application provide the following technical solutions:

[0007] The first aspect of the present application provides a 3D printing path planning method;

[0008] A 3D printing path planning method, comprising:

[0009] Multi-scale geometric feature extraction is performed on a three-dimensional model to be printed to obtain a hierarchical feature mapping set;

[0010] Based on the hierarchical feature mapping set, the 3D model to be printed is adaptively decomposed into voxels to obtain a non-uniform voxel mesh structure.

[0011] The orientation field of the non-uniform voxel mesh structure is calculated to obtain the printing orientation tensor field;

[0012] Based on the printing direction tensor field, a spiral trajectory is generated for the non-uniform voxel mesh structure to obtain a continuous spiral printing path.

[0013] Material deposition printing is performed using a preset printhead controlled by the continuous spiral printing path.

[0014] As a further technical solution, the step of extracting multi-scale geometric features from the 3D model to be printed to obtain a hierarchical feature mapping set includes:

[0015] The shape tensor decomposition of the three-dimensional model to be printed is performed to obtain the feature component field, and the spectral analysis of the feature component field is performed to obtain the multi-scale characterization spectrum.

[0016] The multi-scale representation spectrum is hierarchically mapped by manifold embedding to obtain a feature layer map, and connectivity is calculated based on the feature layer map to obtain the inter-layer connection matrix.

[0017] Structural feature analysis is performed on the interlayer connection matrix to obtain a hierarchical feature mapping set.

[0018] As a further technical solution, the adaptive voxelization decomposition of the 3D model to be printed based on the hierarchical feature mapping set to obtain a non-uniform voxel mesh structure includes:

[0019] Geometric density calculation is performed on the hierarchical feature mapping set to obtain the feature density distribution field, and gradient tensor analysis is performed on the feature density distribution field to obtain the voxel scale mapping matrix.

[0020] The voxel scale mapping matrix is ​​spatially partitioned using an octree partitioning method to obtain a multi-level voxel hierarchical tree, and boundary refinement is performed based on the multi-level voxel hierarchical tree to obtain an adaptive mesh topology.

[0021] Anisotropy analysis is performed on the adaptive mesh topology to obtain the deformation feature tensor field, and mesh deformation is calculated based on the deformation feature tensor field to obtain the deformable mesh element set.

[0022] Based on the deformable mesh unit set, the three-dimensional model to be printed is reconstructed using voxels to obtain a non-uniform voxel mesh structure.

[0023] As a further technical solution, the step of spatially partitioning the voxel scale mapping matrix using the octree segmentation method to obtain a multi-level voxel hierarchical tree includes:

[0024] The voxel scale mapping matrix is ​​decomposed into eigenvalues ​​to obtain a set of eigenvectors, and the eigenvectors are aligned with coordinate axes to obtain a set of principal direction components.

[0025] The main direction component set is hierarchically divided by spatial bisection to obtain a recursive subspace set, and boundary nodes are extracted based on the recursive subspace set to obtain a hierarchical boundary table.

[0026] Geometric feature detection is performed on the hierarchical boundary table to obtain a splitting criterion field. Subtrees are generated from the splitting criterion field using an octree segmentation method to obtain octree node groups.

[0027] A tree structure is constructed based on the octree node group to obtain a multi-level voxel hierarchical tree.

[0028] As a further technical solution, the step of performing orientation field calculations on the non-uniform voxel mesh structure to obtain the printing orientation tensor field includes:

[0029] Stress field analysis is performed on the non-uniform voxel mesh structure to obtain the stress distribution of voxel elements, and the stress distribution of voxel elements is decomposed to obtain the principal stress direction field.

[0030] The curvature mapping of the principal stress direction field is performed by differential geometry calculation to obtain the curvature tensor field, and an anisotropy measurement is performed based on the curvature tensor field to obtain the shape characteristic coefficient table.

[0031] Tensor interpolation is performed on the shape feature coefficient table to obtain a continuous direction field, and singular point detection is performed based on the continuous direction field to obtain a set of local singular regions;

[0032] The set of local singular regions is subjected to rotation field calculation by vector field decomposition to obtain a coordinated direction field, and a smoothness constraint is applied based on the coordinated direction field to obtain a smooth tensor field.

[0033] Based on the smooth tensor field, the orientation field of the non-uniform voxel mesh structure is optimized to obtain the printing orientation tensor field.

[0034] As a further technical solution, the step of generating a spiral trajectory for the non-uniform voxel mesh structure based on the printing direction tensor field to obtain a continuous spiral printing path includes:

[0035] Isosurface analysis is performed on the printing direction tensor field to obtain the interlayer equidistant surface sequence, and the interlayer equidistant surface sequence is mapped by a helical angle to obtain the reference helical direction field.

[0036] The reference spiral direction field is geometrically expanded to obtain a planar spiral layout diagram, and the path density is adjusted based on the planar spiral layout diagram to obtain a variable density spiral network.

[0037] A topology connection analysis is performed on the variable density spiral network to obtain a path connectivity graph. Based on the path connectivity graph, a transition bridging design is performed to obtain a path set for the transition region.

[0038] The curvature continuity of the path set in the transition region is processed by geometric optimization to obtain a smooth transition curve set, and the path is reconstructed based on the smooth transition curve set to obtain the initial spiral trajectory.

[0039] Based on the initial spiral trajectory, the path constraint of the non-uniform voxel mesh structure is checked to obtain a continuous spiral printing path.

[0040] As a further technical solution, the reference spiral direction field is geometrically expanded to obtain a planar spiral layout diagram, including:

[0041] Manifold degradation calculations are performed on the reference spiral direction field to obtain a degenerate surface sequence, and topological compression is performed on the degenerate surface sequence to obtain a simplified topological structure;

[0042] The simplified topology is geometrically flattened by discrete differentiation to obtain an initial plane mapping, and boundary constraint processing is performed based on the initial plane mapping to obtain a constrained plane domain.

[0043] The constrained plane domain is divided into equidistant grids to obtain regular grid cell groups, and spiral interpolation calculations are performed based on the regular grid cell groups to obtain a spiral path network;

[0044] Based on the spiral path network, a topology reconnection is performed to obtain a planar spiral layout diagram.

[0045] A second aspect of the present invention provides a 3D printing path planning system.

[0046] A 3D printing path planning system, comprising:

[0047] The feature extraction module is configured to perform multi-scale geometric feature extraction on the 3D model to be printed, and obtain a hierarchical feature mapping set.

[0048] The decomposition module is configured to: perform adaptive voxelization decomposition on the 3D model to be printed based on the hierarchical feature mapping set to obtain a non-uniform voxel mesh structure.

[0049] The orientation field calculation module is configured to perform orientation field calculations on the non-uniform voxel mesh structure to obtain the printing orientation tensor field.

[0050] The printing path generation module is configured to generate a spiral trajectory for the non-uniform voxel mesh structure based on the printing direction tensor field to obtain a continuous spiral printing path.

[0051] The printing module is configured to perform material deposition printing based on a preset printhead controlled by the continuous spiral printing path.

[0052] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of a 3D printing path planning method as described in the first aspect of the present invention.

[0053] A fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of a 3D printing path planning method as described in the first aspect of the present invention.

[0054] The above one or more technical solutions have the following beneficial effects:

[0055] (1) This invention effectively overcomes the defects of uneven filling, excessive support structure and decreased surface accuracy in existing methods by extracting multi-scale geometric features and adaptive voxelization decomposition, combined with continuous spiral path generation, and greatly improves the forming quality of printed parts.

[0056] (2) Multi-scale feature extraction can capture geometric information from macroscopic contours to microscopic details. Adaptive voxelization can dynamically adjust the voxel scale according to the feature density, ensuring the printing accuracy of complex structures (such as high curvature and thin areas) while reducing computational redundancy in non-critical areas, thus balancing accuracy and efficiency. By integrating stress field and curvature analysis through orientation field calculation, the generated printing orientation tensor field can guide the material to be deposited along the optimal direction, reducing internal stress accumulation, lowering the risk of deformation and cracking, and enhancing the structural strength and stability of the printed parts.

[0057] (3) The spiral trajectory is generated by non-uniform voxel mesh structure. The continuous spiral trajectory eliminates path interruption and corner accumulation through topological connection and smooth transition design, strengthens interlayer bonding force, improves material utilization, and reduces the support structure requirements. It is suitable for the efficient manufacturing of complex high-precision components in aerospace, biomedicine and other fields.

[0058] Advantages of additional aspects 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

[0059] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0060] Figure 1 This is a flowchart of the method in the first embodiment.

[0061] Figure 2 This is a system structure diagram of the second embodiment. Detailed Implementation

[0062] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0063] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0064] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0065] Example 1

[0066] This embodiment discloses a 3D printing path planning method;

[0067] like Figure 1 As shown, a 3D printing path planning method includes:

[0068] Step S1: Perform multi-scale geometric feature extraction on the 3D model to be printed to obtain a hierarchical feature mapping set;

[0069] Step S2: Based on the hierarchical feature mapping set, perform adaptive voxelization decomposition on the 3D model to be printed to obtain a non-uniform voxel mesh structure.

[0070] Step S3: Perform orientation field calculation on the non-uniform voxel mesh structure to obtain the printing orientation tensor field;

[0071] Step S4: Generate a spiral trajectory for the non-uniform voxel mesh structure based on the printing direction tensor field to obtain a continuous spiral printing path;

[0072] Step S5: Based on the continuous spiral printing path, control the preset printing nozzle to perform material deposition printing.

[0073] Specifically, the above steps also include the following:

[0074] Step S1: Perform multi-scale geometric feature extraction on the 3D model to be printed to obtain a hierarchical feature mapping set.

[0075] Through multi-level geometric analysis, the geometric information of the 3D model at different scales is systematically captured and expressed, providing a structured feature foundation for subsequent path planning. First, the original 3D model is preprocessed in the form of digital geometry, including surface normal vector calculation, local curvature estimation, and boundary identification. Then, multi-scale analysis algorithms (such as wavelet transform, Gaussian filtering, or Laplacian smoothing) are used to extract features from the model surface at different resolutions, ensuring effective modeling from the overall macroscopic shape to microscopic local details. Based on this, the extracted geometric features are further organized according to scale hierarchy, forming a hierarchical feature mapping set, i.e., a layered feature mapping set.

[0076] Specifically, including:

[0077] Step S11: Perform shape tensor decomposition on the three-dimensional model to be printed to obtain the feature component field, and perform spectral analysis on the feature component field to obtain the multi-scale characterization spectrum.

[0078] When performing shape tensor decomposition on the 3D model to be printed, the model surface or volume data is represented as a higher-order tensor form and subjected to spectral decomposition to extract multiple physically meaningful feature components, forming a feature component field. These components typically correspond to the local geometric properties of the model in different directions, such as the principal direction of curvature and boundary strength.

[0079] Specifically, the outer surface of the model to be printed is abstracted as a smooth parametric surface. The first and second fundamental forms of the surface are calculated using differential geometry tools. Based on the first and second fundamental forms, the shape operator of the surface shape tensor is calculated, as shown in the following equation:

[0080]

[0081] in, For shape operators; It is the first basic form; This is the second basic form.

[0082] By performing spectral decomposition on the shape operator, the principal curvature and principal direction are obtained;

[0083]

[0084] in, It is an orthogonal matrix, and each column is an eigenvector, i.e., the direction of principal curvature; These are the two eigenvalues ​​of the shape operator, namely the principal curvatures.

[0085] The spectral decomposition results are backfilled into three dimensions to obtain direction-intensity pairs. The surface feature component field is defined by the direction and intensity, as shown below:

[0086]

[0087] in, For surface feature components; For directional intensity pairs; Let p be the unit normal vector of the surface.

[0088] For the volume model, a shape tensor is constructed using voxel gradient information. The three-dimensional gradient field is calculated using the voxel data of the volume model. The structure tensor is used to describe the local directional distribution characteristics of the voxels. The volume feature component field is obtained by spectral decomposition of the volume shape tensor, as shown below:

[0089]

[0090] in, For structure tensors; For Gaussian kernel; This is a convolution operation; For volume field The gradient; This is the outer product of the gradient vectors. By... Eigenvalue decomposition can robustly estimate the dominant anisotropic direction and define the volume characteristic component field accordingly.

[0091] Step S12: The multi-scale representation spectrum is hierarchically mapped by manifold embedding to obtain a feature layer map, and connectivity is calculated based on the feature layer map to obtain the inter-layer connection matrix.

[0092] By performing spectral analysis on the characteristic component field, including continuous-scale spatial solution, wavelet or bandpass energy decomposition and tensor spectral calculation, the energy distribution at different frequencies or scales is extracted to form a multi-scale characterization spectrum, so that the geometric information of the model can be uniformly described at both macroscopic and microscopic levels.

[0093] Next, the multi-scale representation spectrum is hierarchically mapped using the manifold embedding method, embedding it into a low-dimensional manifold space. Based on the mapping results, a feature layer map is constructed, thereby achieving a hierarchical abstract representation of the original geometric information. Furthermore, connectivity calculations are performed on the feature layer map to identify the topological relationships between layers, generating an inter-layer connectivity matrix to characterize the transmission paths and dependencies between features at different scales.

[0094] Specifically, the manifold embedding method includes: taking the sampling points on the model as nodes, and extracting the local multi-scale spectral fragment of the descriptor for each point:

[0095]

[0096] in, This is the multi-scale feature vector of the m-th sampling point, containing the feature vector of that point at different scales. The characteristic spectrum values ​​below; In order to scale The spectral operator or feature extraction function is used to obtain the local geometric features of a point; Let m be the spatial coordinates of the m-th sampling point (node) on the model; The number of scales represents the total number of different scales used to represent the point.

[0097] The weight matrix and degree matrix are obtained by constructing an affinity graph:

[0098]

[0099]

[0100] in, , where are elements of the weight matrix, representing the similarity between node m and node n; The scale parameter (kernel width) controls the decay rate of the similarity function; Represents the norm of a matrix; K-nearest neighbors means that connections are established only between the K nearest neighbors; It is a degree matrix; Let m be the degree of node m, which is the sum of its similarity to other nodes.

[0101] The Laplacian is normalized based on the weight and degree matrices, and the minimum nontrivial feature pair after normalization is selected to embed the coordinates:

[0102]

[0103]

[0104] in, For normalized symmetric Laplace matrices; The weight matrix; for The component of the d-th eigenvector at node m; Let be the embedding coordinates of node m in the low-dimensional manifold space, which are composed of the components of several eigenvectors.

[0105] Ultimately through the Hierarchical clustering yields clusters with consistent scale and similar geometry, forming a feature hierarchical map. By statistically analyzing cross-cluster adjacencies between layers, the inter-layer connectivity matrix is ​​obtained, as shown below:

[0106]

[0107] in, The elements of the inter-layer connectivity matrix represent the layers. With layers The average connection strength between them; For the first The set of nodes in a layer (or cluster); For the first The set of nodes in a layer (or cluster); For node pairs, it indicates whether there is a connection between the m-th node and the n-th node.

[0108] Step S13: Perform structural feature analysis on the inter-layer connection matrix to obtain a hierarchical feature mapping set.

[0109] Based on this inter-layer connectivity matrix, structural feature analysis is carried out to extract key structural features and their interaction patterns in each layer, and finally integrate them into a hierarchical feature mapping set with clear structure and explicit semantics.

[0110] For example, when printing a biological scaffold model containing complex curved surfaces and fine branch structures, this method can extract the overall contour direction at a large scale, the direction of the main support structure at a medium scale, and the edge features of the micropores at a small scale. This ensures that subsequent path planning can maintain the overall structural stability while taking into account local accuracy requirements, ultimately achieving a high-quality, continuous, and adaptive 3D printing process.

[0111] Step S2: Based on the hierarchical feature mapping set, perform adaptive voxelization decomposition on the 3D model to be printed to obtain a non-uniform voxel mesh structure.

[0112] Specifically, by utilizing the multi-scale geometric information extracted in step S1 to guide the density and distribution of voxels, non-uniform discretization modeling of the internal space of the 3D model is achieved. In the obtained hierarchical feature map set, the geometric features at different levels correspond to the structural importance of the model at different scales, such as curvature changes, boundary continuity, and local complexity. This information is used to dynamically adjust the size and distribution density of voxels. For the obtained non-uniform voxel mesh structure, in regions with rich geometric features and complex details (such as high curvature regions or small branches), voxels are divided into smaller, denser units to ensure the fidelity of the local structure; while in regions with gentle geometric features and stable structures, larger and sparser voxel units are used to reduce computational redundancy and improve processing efficiency. The resulting non-uniform voxel mesh structure not only more accurately reflects the spatial topological relationships of the original model but also provides a more adaptive data foundation for subsequent orientation field calculations and trajectory generation.

[0113] Specifically, including:

[0114] Step S21: Perform geometric density calculation on the hierarchical feature mapping set to obtain the feature density distribution field, and perform gradient tensor analysis on the feature density distribution field to obtain the voxel scale mapping matrix, as shown below.

[0115]

[0116] in, It is the gradient tensor; Let x be the characteristic density field at position x in three-dimensional space; This is the gradient vector of the density field, used to characterize the variation of the density distribution in various directions; To form a local directional distribution tensor; To use the scale parameter as The Gaussian kernel is used to convolve and smooth the tensor field to enhance robustness.

[0117] By utilizing the existing geometric information (such as curvature, boundary strength, structural continuity, etc.) in the hierarchical feature mapping set, the geometric density at each location is weighted and calculated to form a feature density distribution field. This field reflects the distribution of geometric importance in different regions. For example, there are usually higher density values ​​at the edges of pores or thin-walled structures. Based on this, gradient tensor analysis is performed on the feature density distribution field. That is, by calculating its gradient direction and magnitude, the direction of the most drastic geometric change is identified, thus providing a basis for the dynamic adjustment of the voxel scale. The resulting voxel scale mapping matrix can be used as the basic data for subsequent spatial partitioning.

[0118] Step S22 involves spatially partitioning the voxel scale mapping matrix using an octree partitioning method to obtain a multi-level voxel hierarchy tree, and then performing boundary refinement based on this multi-level voxel hierarchy tree to obtain an adaptive mesh topology. Specifically, this includes the following steps:

[0119] Step S221: Perform eigenvalue decomposition on the voxel scale mapping matrix to obtain an eigenvector group, and align the eigenvector group with coordinate axes to obtain the principal direction component set, as shown below.

[0120]

[0121] in, This is the voxel-scale mapping matrix; The eigenvalues ​​of this matrix represent the local intensity in the three principal directions, respectively. For a set of feature vectors, each There is a corresponding main direction; by aligning Q with the global coordinate system, the main direction component set can be obtained, which provides a basis for guiding the direction of the subsequent printing path.

[0122] By performing eigenvalue decomposition on the voxel-scale mapping matrix, the spatial orientation information contained therein can be extracted. The voxel-scale mapping matrix is ​​essentially a field function describing the required voxel scale at each location in the model space. By performing eigenvalue decomposition on this matrix, a set of eigenvectors representing the main directions of the local space can be obtained. These eigenvectors reflect the main directions of the geometric changes in the current region, such as the direction of the surface normal or the direction of the boundary. Subsequently, the obtained eigenvector set is aligned with the coordinate axes to make it consistent with the standard Cartesian coordinate system, thereby obtaining the set of main direction components, which provides directional reference for subsequent spatial partitioning.

[0123] Step S222: The main direction component set is hierarchically partitioned by spatial bisection to obtain a recursive subspace set, and boundary nodes are extracted based on the recursive subspace set to obtain a hierarchical boundary table, as shown below:

[0124]

[0125] in, This is the set of space partitions obtained after the binary search operation; For three-dimensional space Along direction The space binary search operation was performed. An eigenvector in the principal direction component set; the formula represents the eigenvector along the direction. A binary search operation is performed on the three-dimensional space to divide the space into two subspaces. This binary search operation is applied recursively multiple times to obtain a set of recursive subspaces. Nodes are extracted on the boundaries of the divided subspaces to obtain a hierarchical boundary table, which is used for subsequent path generation and structural constraints.

[0126] Spatial bisection is performed based on the principal direction component set. The entire space is gradually divided into several subspaces in a recursive manner to form a recursive subspace set. This partitioning method is similar to the traditional KD-Tree structure, but it further considers the influence of the dominant local geometric direction, making the partitioning result more consistent with the actual geometric distribution characteristics of the model. At the same time, the boundary nodes of each subspace are extracted in each partitioning process to generate a hierarchical boundary table, which is used to record the set of boundary points formed after each partitioning layer, so as to be used for subsequent boundary refinement and splitting judgment.

[0127] Step S223: Perform geometric feature detection on the hierarchical boundary table to obtain the splitting criterion field. Then, generate subtrees from the splitting criterion field using the octree segmentation method to obtain octree node groups, as shown below:

[0128]

[0129] in, The geometric characteristic function at the boundary node x can be curvature, rate of change of normal, or density gradient, etc. is the splitting threshold; 1(•) is an indicator function that takes the value 1 when the condition is met, and 0 otherwise; This constitutes a splitting criterion field, marking which boundary regions need further subdivision.

[0130] After completing spatial partitioning and boundary information extraction, geometric feature detection is further performed on the hierarchical boundary table to identify geometric abrupt changes, high curvature regions, or weak structural connections in each boundary region, and a splitting criterion field is constructed accordingly. This splitting criterion field is essentially a spatial weight field used to guide whether a certain region needs to be further subdivided during the octree partitioning process. Gradient magnitude, curvature extrema, or local density change rate are typically used as splitting criteria to ensure higher precision partitioning details are preserved in geometrically complex regions, while the partitioning hierarchy is appropriately simplified in smooth regions, thus achieving adaptive control. Subsequently, based on this splitting criterion field, the octree segmentation method is applied to perform subtree generation, that is, each spatial unit that does not meet the termination condition is further subdivided into eight sub-cubes, and a corresponding octree node is established for each sub-cube. These nodes not only record their respective spatial extents but also carry related geometric feature information (such as splitting depth, principal direction, boundary attributes, etc.), ultimately forming an octree node group composed of multiple octree nodes.

[0131] Step S224: Based on the octree node group, construct a tree structure to obtain a multi-level voxel hierarchy tree, as shown below:

[0132]

[0133] in, This is for the final construction of a multi-level voxel hierarchy tree; To represent a node in a tree, it contains: For the octet unit (spatial center and scale information) corresponding to this node and It is the set of its child nodes (generated by octree partitioning); by recursively organizing all octree node groups, a tree hierarchy from coarse to fine is obtained; this hierarchical tree provides multi-level voxel indexes and local geometric constraints for subsequent path planning.

[0134] A tree structure is constructed based on octree node groups, organizing all nodes into a complete multi-level voxel hierarchy tree according to parent-child relationships. This tree structure not only clearly expresses the scale variation law of the model from the whole to the local, but also provides efficient data access interface and spatial index support for subsequent boundary refinement, mesh deformation, and path planning. Simultaneously, through principal direction alignment and splitting criteria guidance, the voxel partitioning direction is kept consistent with the principal stress direction of the support structure, avoiding structural distortion caused by misalignment of partitioning directions. This significantly improves the accuracy and stability of the printing path while ensuring modeling efficiency.

[0135] Step S23: Perform anisotropic analysis on the adaptive mesh topology to obtain the deformation feature tensor field, and perform mesh deformation calculation based on the deformation feature tensor field to obtain the deformed mesh element set.

[0136] Anisotropic analysis was performed on the adaptive mesh topology to identify the deformation trends of each voxel element in different directions, and a deformation characteristic tensor field was constructed accordingly. This tensor field describes the tensor field of each voxel element that may be stretched or compressed during stress or material deposition, which helps to reasonably adjust the voxel shape during subsequent mesh deformation calculations so that it better fits the actual geometry. Subsequently, mesh deformation calculations were carried out based on the deformation characteristic tensor field. An elasticity model or finite element analysis method was used to apply corresponding deformation constraints to each voxel element, ultimately generating a set of optimized deformable mesh elements, so that each voxel not only adapts to the local geometry in size, but also matches the original model more closely in shape.

[0137] Step S24: Based on the deformable mesh unit set, the three-dimensional model to be printed is reconstructed using voxels to obtain a non-uniform voxel mesh structure.

[0138] Based on the deformed mesh elements, the original 3D model is reconstructed using voxels. This involves projecting the model's surface or volume information onto these deformed meshes to form a complete non-uniform voxel mesh structure. This structure not only preserves the key geometric features of the model but also provides a high-quality spatial discretization foundation for subsequent orientation field calculations and spiral trajectory generation.

[0139] By refining the boundary and calculating the mesh deformation, geometric distortion caused by voxel misalignment or shape mismatch can be effectively avoided, ensuring that the printing path is accurately aligned in the microstructure and continuous and smooth in the macrostructure, thereby significantly improving the quality and functionality of the final molded part.

[0140] Step S3: Perform orientation field calculation on the non-uniform voxel mesh structure to obtain the printing orientation tensor field.

[0141] Specifically, based on the constructed non-uniform voxel mesh, and combining the local geometric features of the model with printing process constraints, a tensor with orientation information is calculated on each voxel cell to describe the optimal deposition orientation of the printing material in that region. Specifically, this includes:

[0142] Step S31: Perform stress field analysis on the non-uniform voxel mesh structure to obtain the stress distribution of voxel elements, and perform characteristic decomposition on the stress distribution of voxel elements to obtain the principal stress direction field.

[0143] Stress field analysis based on non-uniform voxel mesh structures typically employs finite element simulation or elasticity models to map the internal stress distribution that may occur during material deposition onto each voxel element, thereby obtaining the stress distribution of the voxel element. Subsequently, the stress distribution is decomposed using features such as principal component analysis or tensor diagonalization to extract the principal stress directions of each voxel element, i.e., the principal stress direction field. This directional information reflects the natural flow trend and dominant force direction of the material during deposition at that location, providing a physical basis for subsequent directional field construction.

[0144] Step S32: Curvature mapping is performed on the principal stress direction field through differential geometry calculation to obtain a curvature tensor field, and an anisotropy measurement is performed based on the curvature tensor field to obtain a shape characteristic coefficient table.

[0145] By using the surface differential operator in differential geometry to perform curvature mapping on the principal stress direction field, the curvature tensor field of the region where each voxel element is located is calculated. This tensor field not only contains information about the degree of local surface curvature, but also reflects the degree of drastic change in direction.

[0146] Based on this, an anisotropy measurement method is further combined to quantitatively evaluate the orientation consistency of each voxel unit and generate a shape feature coefficient table. This table records the comprehensive characteristics of each voxel in terms of orientation stability, curvature intensity, etc., which helps in the weight allocation in the subsequent orientation field interpolation and optimization process.

[0147] Step S33: Perform tensor interpolation on the shape feature coefficient table to obtain a continuous direction field, and perform singular point detection based on the continuous direction field to obtain a set of local singular regions, as shown below:

[0148]

[0149] in, This represents the continuous directional field obtained at spatial location x; This refers to the local tensor (orientation information) in the shape characteristic coefficient table; The interpolation weighting function is often a Gaussian kernel or a distance-based weighting function, used to ensure the smoothness and continuity of the orientation field. Through this tensor interpolation, a continuous orientation distribution can be constructed throughout the voxel space. In this orientation field, if... ×F(x) or • The location of abnormal mutation in F(x) is identified as a singularity; these singularity regions are clustered to obtain a set of local singular regions, which are areas that need to be avoided or specially treated during path optimization.

[0150] Tensor interpolation is performed on the shape characteristic coefficient table to expand the discrete directional information distributed on each voxel into a continuous spatial directional field, so that the directional field maintains a smooth transition in space.

[0151] To ensure the topological rationality of the orientation field, singularity detection is also required for the continuous orientation field to identify local singular regions where the orientation changes abruptly or the main orientation cannot be defined, such as key structural parts like the intersection of pores or the turning point of thin walls, thereby forming a set of local singular regions.

[0152] Step S34: Perform rotation field calculation on the set of local singular regions through vector field decomposition to obtain a harmonized direction field, and apply smoothness constraints based on the harmonized direction field to obtain a smooth tensor field, as shown below:

[0153]

[0154] in, This represents the original orientation field within a set of local singular regions; The gradient component of the direction field is used to characterize the irrotational component. The rotational component (curl field) of the orientation field is used to characterize the rotational properties around singular points. Through this vector field decomposition, the rotational component can be extracted separately to construct a harmonized orientation field. On this basis, a smoothness constraint is applied to the orientation field to obtain a globally consistent and continuous smooth tensor field. This smooth tensor field provides smooth and singular-free directional guidance for subsequent path generation.

[0155] To eliminate directional conflicts caused by these singular regions, a vector field decomposition method is used to perform rotation field calculations on the local singular region set. That is, through local coordinate transformation and rotation matrix adjustment, the directional fields of adjacent regions tend to be consistent, thereby constructing a coordinated directional field. On this basis, a smoothness constraint mechanism is introduced, usually using harmonic field optimization or minimum energy functional methods to perform overall smoothing processing on the coordinated directional field, ensuring that the directional field is consistent not only in local regions, but also maintains continuity and integrability in the global scope, ultimately obtaining a smooth tensor field.

[0156] Step S35: Optimize the orientation field of the non-uniform voxel mesh structure based on the smooth tensor field to obtain the printing orientation tensor field.

[0157] Based on this smooth tensor field, the orientation field of the original non-uniform voxel mesh structure is optimized. This involves remapping the optimized orientation information onto each voxel unit, resulting in a more reasonable orientation distribution in space. This generates a printing orientation tensor field for subsequent spiral trajectory generation. This tensor field not only accurately reflects the orientation of the model's surface and internal structure but also effectively avoids problems such as abrupt orientation changes and path breaks in traditional path planning, thus significantly improving the continuity and fit of the printing path.

[0158] Step S4: Generate a spiral trajectory for the non-uniform voxel mesh structure based on the printing direction tensor field to obtain a continuous spiral printing path.

[0159] By utilizing the constructed printing direction tensor field as a directional guide and combining it with the spatial distribution characteristics of non-uniform voxel meshes, an efficient and high-quality spiral printing trajectory suitable for 3D printing is generated while ensuring path continuity and directional consistency.

[0160] In the specific implementation process, each voxel in the non-uniform voxel mesh is first used as the basic operation unit. The material deposition direction of the region is determined based on the principal direction information of its corresponding position in the printing direction tensor field. The path is then gradually extended along this direction using an integral curve or streamline tracing algorithm. Based on this, a spiral topology is introduced, allowing the path to exhibit a smooth transition and natural rotation during layer-by-layer stacking, thereby avoiding common problems in traditional scanning paths such as path interruption, corner stacking, and interlayer misalignment. Specifically, this includes:

[0161] Step S41: Perform isosurface analysis on the printing direction tensor field to obtain an interlayer equidistant surface sequence, and perform helical angle mapping on the interlayer equidistant surface sequence to obtain a reference helical direction field.

[0162] First, isosurface analysis is performed on the tensor field of the printing direction to extract multiple interlayer equidistant surfaces that maintain a certain distance from the model surface, forming an interlayer equidistant surface sequence. The spacing between these equidistant surfaces can be set between 0.1mm and 0.3mm according to the printing layer thickness to ensure that each layer can be accurately covered.

[0163] A spiral angle mapping operation is performed on the interlayer equidistant surface sequence, that is, the direction vector on each equidistant surface is transformed by angle along a preset rotation axis (usually the Z-axis of the model) to make it exhibit a certain spiral trend, and finally a reference spiral direction field is generated. This direction field not only inherits the spatial direction characteristics of the original direction tensor field, but also introduces the geometric features of spiral ascent, providing a directional basis for subsequent path unfolding.

[0164] Step S42: Geometrically expand the reference spiral direction field to obtain a planar spiral layout diagram, and adjust the path density based on the planar spiral layout diagram to obtain a variable density spiral network.

[0165] The reference spiral direction field is geometrically unfolded and projected onto a two-dimensional plane to form a planar spiral layout diagram. This process is similar to "flattening" a three-dimensional spiral structure into a two-dimensional curve, allowing path planning to be carried out in a more intuitive space.

[0166] Based on this, the path density of the planar spiral layout is adjusted according to the geometric complexity and material requirements of each region to generate a variable density spiral network. For example, the path density is automatically increased to 2-3 lines per millimeter at the edge of pores or in thin-walled regions, while it is reduced to 1 line per millimeter in structurally stable regions, thereby achieving a balance between local enhancement of material deposition and overall efficiency.

[0167] Step S421: Perform manifold degradation calculation on the reference spiral direction field to obtain a degenerate surface sequence, and perform topological compression on the degenerate surface sequence to obtain a simplified topological structure.

[0168] The reference spiral direction field is subjected to manifold degradation calculation, which involves gradually simplifying the spiral direction information originally embedded in the three-dimensional curved space and extracting a set of degenerate surface sequences that can represent the main features of the original direction field. These degenerate surfaces are usually obtained by projecting the original model surface through the principal direction and constraining the local curvature. Their number can be set to 5-15 layers according to the model complexity to ensure that geometric details are preserved without increasing redundant computational burden.

[0169] The degenerate surface sequence is subjected to topological compression processing to identify and merge repeating or highly similar topological units, thereby obtaining a simplified topological structure that is simpler in structure but can reflect the overall geometric orientation.

[0170] Step S422: The simplified topology is geometrically flattened by discrete differentiation to obtain an initial plane mapping, and boundary constraint processing is performed based on the initial plane mapping to obtain a constrained plane domain.

[0171] The simplified topology is geometrically flattened using discrete differential methods, "flattening" it from three-dimensional space to two-dimensional planar space to form an initial planar mapping. This mapping process employs a conformal transformation strategy, meaning that mapping is performed while preserving the original angular relationships as much as possible, ensuring that the spiral path maintains good directional consistency in two-dimensional space. Based on this, the initial planar mapping is further constrained by boundary conditions. According to the contour boundary information of the original model under two-dimensional projection, the effective printing area is defined, forming a constrained planar domain. For example, in the projection of a biological scaffold model, this constrained planar domain might be an elliptical region with a diameter of approximately 30mm-50mm, ensuring that all paths are within the effective printing range and avoiding exceeding the nozzle movement limits.

[0172] Step S423: Divide the constrained plane domain into equidistant grids to obtain regular grid cell groups, and perform spiral interpolation calculations based on the regular grid cell groups to obtain a spiral path network.

[0173] The constrained planar domain is divided into equidistant grids to construct regular grid cell groups. These grid cells are typically set as square grids with a side length of 0.5mm-1mm to balance path accuracy and computational efficiency. The spiral direction vector at the corresponding position is recorded in each grid cell for subsequent interpolation calculations.

[0174] By performing spiral interpolation calculations, that is, based on the directional information of adjacent grid cells, a spiral path network that runs through the entire planar domain is generated using bilinear interpolation or spline interpolation methods. This network exhibits a shape that gradually rotates and unfolds from the center outwards, similar to an Archimedean spiral curve. Its pitch (i.e., the radial distance between adjacent turns) can be controlled between 1mm and 3mm to match the printing requirements of different scales.

[0175] Step S424: Perform topology reconnection based on the spiral path network to obtain a planar spiral layout diagram.

[0176] Local adjustments are made to path intersections, breakpoints, or areas of abrupt directional changes to ensure that the entire path has good connectivity and no self-intersection characteristics in a two-dimensional plane. For example, when two paths are too close or even intersect in a certain area, the system will automatically insert a transition segment or readjust the local path direction so that the final planar spiral layout not only has visual continuity but also meets the physical feasibility requirements of the actual printing path.

[0177] Step S43: Perform topology connection analysis on the variable density spiral network to obtain a path connectivity graph, and design a transition bridging based on the path connectivity graph to obtain a path set for the transition region.

[0178] A topology connection analysis was performed on the variable density spiral network to identify the connection relationships between each spiral segment and to construct a path connectivity graph. This graph records the adjacency information between all path segments and is used to guide the subsequent transition and bridging design.

[0179] Step S44: Perform curvature continuity processing on the path set of the transition region through geometric optimization to obtain a smooth transition curve set, and reconstruct the path based on the smooth transition curve set to obtain the initial spiral trajectory.

[0180] Transition path sets are designed for path breakpoints or abrupt directional changes. Smooth bridging curves are inserted to achieve seamless connections between adjacent path segments. The length of these transition paths is typically controlled between 0.5mm and 2mm to avoid material buildup due to excessive length, while ensuring path continuity and stability. To improve the overall smoothness of the path, the transition path sets undergo geometric optimization, focusing on curvature continuity adjustment. This involves local fitting and interpolation of key points on the path to eliminate sharp changes at path corners, achieving a smoother transition. The continuity standard ultimately forms a set of smooth transition curves. The continuity standard refers to the fact that two curves not only coincide in position at the splicing point ( ), tangent direction is consistent ( More importantly, its curvature also remains continuously changing (the second derivative is continuous). This ensures that the curvature of the printing path at the inflection point is smoothly transitioned, fundamentally eliminating abrupt changes in curvature. As a result, the print head does not need to experience a sharp change in acceleration when passing through the point, ultimately achieving the technical effects of effectively suppressing equipment vibration, avoiding material accumulation, and significantly improving the surface finish of the printed parts.

[0181] Subsequently, the path was reconstructed based on these smooth curve groups, and all segments were reintegrated into a complete initial spiral trajectory. This trajectory presents a natural spiral ascent in space and is geometrically highly consistent with the model surface.

[0182] Step S45: Based on the initial spiral trajectory, perform path constraint verification on the non-uniform voxel mesh structure to obtain a continuous spiral printing path.

[0183] Based on the initial spiral trajectory, the non-uniform voxel mesh structure is subjected to path constraint verification to determine whether the path conflicts with or is missing from the model boundary, and whether it meets the minimum radius of curvature requirement (usually greater than or equal to 1.5 mm) for nozzle movement. If non-compliant areas are found, they are locally corrected until all path segments meet the process constraints, and finally a continuous spiral printing path is output.

[0184] Step S5: Based on the continuous spiral printing path, control the preset printing nozzle to perform material deposition printing.

[0185] The generated continuous spiral printing path is converted into actual executable motion commands, which drive the preset printing nozzle to perform high-precision, continuous material deposition operations in three-dimensional space according to the path.

[0186] In practice, the continuous spiral printing path is first input into the 3D printing control system in the form of CNC code (such as G-code). The system calculates the nozzle's motion trajectory, speed curve, and material extrusion rate in real time based on the path data, and coordinates with the lifting and lowering movements of the printing platform to ensure that each layer of material is precisely laid along the set spiral direction. Based on this, and considering the rheological properties and curing behavior of the printing material, process parameters such as nozzle temperature and extrusion pressure are dynamically adjusted to ensure the stability of material deposition and the quality of interlayer bonding. For example, when printing a biological scaffold model with complex pore structures and thin-walled features, this step ensures that the printing nozzle runs smoothly along the spiral path, avoiding material accumulation or breakage problems caused by sudden path changes or frequent start-stop cycles, thereby achieving high-precision, high-strength, and high-surface-quality integrated printing of the biological scaffold.

[0187] Example 2

[0188] This embodiment discloses a 3D printing path planning system;

[0189] like Figure 2 As shown, a 3D printing path planning system includes:

[0190] A 3D printing path planning system, comprising:

[0191] The feature extraction module is configured to perform multi-scale geometric feature extraction on the 3D model to be printed, and obtain a hierarchical feature mapping set.

[0192] The decomposition module is configured to: perform adaptive voxelization decomposition on the 3D model to be printed based on the hierarchical feature mapping set to obtain a non-uniform voxel mesh structure.

[0193] The orientation field calculation module is configured to perform orientation field calculations on the non-uniform voxel mesh structure to obtain the printing orientation tensor field.

[0194] The printing path generation module is configured to generate a spiral trajectory for the non-uniform voxel mesh structure based on the printing direction tensor field to obtain a continuous spiral printing path.

[0195] The printing module is configured to perform material deposition printing based on a preset printhead controlled by the continuous spiral printing path.

[0196] Example 3

[0197] The purpose of this embodiment is to provide a computer-readable storage medium.

[0198] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a 3D printing path planning method as described in Example 1.

[0199] Example 4

[0200] The purpose of this embodiment is to provide an electronic device.

[0201] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in a 3D printing path planning method as described in Embodiment 1.

[0202] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0203] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0204] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A 3D printing path planning method, characterized in that, include: Multi-scale geometric feature extraction is performed on the 3D model to be printed to obtain a hierarchical feature mapping set; Based on the hierarchical feature mapping set, the 3D model to be printed is adaptively decomposed into voxels to obtain a non-uniform voxel mesh structure. The orientation field of the non-uniform voxel mesh structure is calculated to obtain the printing orientation tensor field; Based on the printing direction tensor field, a spiral trajectory is generated for the non-uniform voxel mesh structure to obtain a continuous spiral printing path; specifically, isosurface analysis is performed on the printing direction tensor field to obtain a sequence of interlayer equidistant surfaces, and spiral angle mapping is performed on the interlayer equidistant surface sequence to obtain a reference spiral direction field. The reference spiral direction field is geometrically expanded to obtain a planar spiral layout diagram, and path density is adjusted based on the planar spiral layout diagram to obtain a variable-density spiral network. Obtaining the planar spiral layout diagram includes: performing manifold degradation calculations on the reference spiral direction field to obtain a degenerate surface sequence, and performing topological compression on the degenerate surface sequence to obtain a simplified topology; geometrically flattening the simplified topology through discrete differentiation to obtain an initial plane mapping, and performing boundary constraint processing based on the initial plane mapping to obtain a constrained plane domain; dividing the constrained plane domain into equidistant meshes to obtain regular mesh cell groups, and performing spiral interpolation calculations based on the regular mesh cell groups to obtain a spiral path network; and performing topological reconnection based on the spiral path network to obtain the planar spiral layout diagram. A topological connection analysis is performed on the variable-density spiral network to obtain a path connectivity graph. Based on the path connectivity graph, a transition bridging design is performed to obtain a path set for the transition region. The curvature continuity of the path set for the transition region is processed through geometric optimization to obtain a set of smooth transition curves. Based on the set of smooth transition curves, the path is reconstructed to obtain an initial spiral trajectory. Based on the initial spiral trajectory, the path constraint of the non-uniform voxel mesh structure is checked to obtain a continuous spiral printing path. Material deposition printing is performed using a preset printhead controlled by the continuous spiral printing path.

2. The 3D printing path planning method as described in claim 1, characterized in that, The process of extracting multi-scale geometric features from the 3D model to be printed, resulting in a hierarchical feature mapping set, includes: The shape tensor decomposition of the three-dimensional model to be printed is performed to obtain the feature component field, and the spectral analysis of the feature component field is performed to obtain the multi-scale characterization spectrum. The multi-scale representation spectrum is hierarchically mapped by manifold embedding to obtain a feature layer map, and connectivity is calculated based on the feature layer map to obtain the inter-layer connection matrix. Structural feature analysis is performed on the interlayer connection matrix to obtain a hierarchical feature mapping set.

3. The 3D printing path planning method as described in claim 1, characterized in that, The adaptive voxelization decomposition of the 3D model to be printed based on the hierarchical feature mapping set to obtain a non-uniform voxel mesh structure includes: Geometric density calculation is performed on the hierarchical feature mapping set to obtain the feature density distribution field, and gradient tensor analysis is performed on the feature density distribution field to obtain the voxel scale mapping matrix. The voxel scale mapping matrix is ​​spatially partitioned using an octree partitioning method to obtain a multi-level voxel hierarchical tree, and boundary refinement is performed based on the multi-level voxel hierarchical tree to obtain an adaptive mesh topology. Anisotropy analysis is performed on the adaptive mesh topology to obtain the deformation feature tensor field, and mesh deformation is calculated based on the deformation feature tensor field to obtain the deformable mesh element set. Based on the deformable mesh unit set, the three-dimensional model to be printed is reconstructed using voxels to obtain a non-uniform voxel mesh structure.

4. The 3D printing path planning method as described in claim 3, characterized in that, The step of spatially partitioning the voxel scale mapping matrix using an octree partitioning method to obtain a multi-level voxel hierarchical tree includes: The voxel scale mapping matrix is ​​decomposed into eigenvalues ​​to obtain a set of eigenvectors, and the eigenvectors are aligned with coordinate axes to obtain a set of principal direction components. The main direction component set is hierarchically divided by spatial bisection to obtain a recursive subspace set, and boundary nodes are extracted based on the recursive subspace set to obtain a hierarchical boundary table. Geometric feature detection is performed on the hierarchical boundary table to obtain a splitting criterion field. Subtrees are generated from the splitting criterion field using an octree segmentation method to obtain octree node groups. A tree structure is constructed based on the octree node group to obtain a multi-level voxel hierarchical tree.

5. The 3D printing path planning method as described in claim 1, characterized in that, The step of performing orientation field calculations on the non-uniform voxel mesh structure to obtain the printing orientation tensor field includes: Stress field analysis is performed on the non-uniform voxel mesh structure to obtain the stress distribution of voxel elements, and the stress distribution of voxel elements is decomposed to obtain the principal stress direction field. The curvature mapping of the principal stress direction field is performed by differential geometry calculation to obtain the curvature tensor field, and an anisotropy measurement is performed based on the curvature tensor field to obtain the shape characteristic coefficient table. Tensor interpolation is performed on the shape feature coefficient table to obtain a continuous direction field, and singular point detection is performed based on the continuous direction field to obtain a set of local singular regions; The set of local singular regions is subjected to rotation field calculation by vector field decomposition to obtain a coordinated direction field, and a smoothness constraint is applied based on the coordinated direction field to obtain a smooth tensor field. Based on the smooth tensor field, the orientation field of the non-uniform voxel mesh structure is optimized to obtain the printing orientation tensor field.

6. A 3D printing path planning system, employing a 3D printing path planning method as described in any one of claims 1-5, characterized in that, include: The feature extraction module is configured to perform multi-scale geometric feature extraction on the 3D model to be printed, and obtain a hierarchical feature mapping set. The decomposition module is configured to: perform adaptive voxelization decomposition on the 3D model to be printed based on the hierarchical feature mapping set to obtain a non-uniform voxel mesh structure. The orientation field calculation module is configured to perform orientation field calculations on the non-uniform voxel mesh structure to obtain the printing orientation tensor field. The printing path generation module is configured to generate a spiral trajectory for the non-uniform voxel mesh structure based on the printing direction tensor field to obtain a continuous spiral printing path. The printing module is configured to perform material deposition printing based on a preset printhead controlled by the continuous spiral printing path.

7. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in a 3D printing path planning method as described in any one of claims 1-5.

8. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the 3D printing path planning method as described in any one of claims 1-5.

Images