Method, device and equipment for virtual restoration of cultural relics based on digital twinning technology and medium

Abstract

The invention relates to a cultural relic virtual restoration method and device based on a digital twinning technology, equipment and a medium. The method comprises the following steps: acquiring original data of overseas lost cultural relics; performing standardized preprocessing on the original data to obtain a cultural relic data packet; performing three-dimensional geometric reconstruction based on the cultural relic data packet to obtain a cultural relic three-dimensional digital model; identifying diseases appearing in the cultural relic three-dimensional digital model through an image segmentation model to obtain a cultural relic disease list; based on the cultural relic disease list, predicting the risk of the structural stability of the cultural relic three-dimensional digital model through mechanical simulation to obtain a cultural relic risk prediction map; and based on the cultural relic disease list and the cultural relic risk prediction map, parallel generation of restoration strategies with different emphasis points is carried out, and a virtual restoration scheme is obtained. By adopting the method, the initial state of the cultural relic can be reproduced in the virtual environment, the restoration effect can be reviewed, and the potential risk can be accurately predicted.

Description

Technical Field

[0001] This invention belongs to the field of 3D modeling, and in particular relates to methods, devices, equipment and media for virtual restoration of cultural relics based on digital twin technology. Background Technology

[0002] With breakthroughs in high-precision 3D acquisition and generative artificial intelligence technologies, significant progress has been made in virtual restoration technology for cultural relics. This technology can non-destructively deduce and reconstruct missing parts of cultural relics lost overseas, and highly restore their original materials and visual forms, providing a new path for the digital reconstruction of cultural relics. In the traditional process of physical restoration, due to reverence for the historical authenticity of cultural relics and the constraint of the principle of minimal intervention, restorers are usually extremely restrained. They often adopt conservative strategies for completing missing parts, or even leave them incomplete. While this approach preserves the historical information of cultural relics to the greatest extent, it also makes it difficult for the public to see their complete artistic design and technological achievements. Cultural relics lost overseas are characterized by their large quantity, variety, and incomplete data. Existing data is mainly text and images, which are insufficient to meet the needs of high-precision research. Furthermore, the fact that physical relics are scattered overseas has led to a gradual fading of cultural identity among the public due to limited knowledge of these relics. Traditional methods of cultural relic restoration have inherent limitations in fully leveraging the social and educational functions of cultural relics. They struggle to directly convey the complete aesthetic imagery, craftsmanship, and social functions of cultural relics in their respective eras to the public, thus hindering the public's comprehensive and in-depth understanding of the value of cultural relics. Furthermore, the traditional experience-based approach suffers from problems such as difficulty in quantifying evaluation standards, the inability to predict restoration effects, the difficulty in accurately predicting potential risks, and a lack of transparency in the decision-making process. This is especially true for cultural relics with complex structures or severe damage, where the intervention process itself may bring unpredictable risks. Summary of the Invention

[0003] Therefore, it is necessary to address the aforementioned technical issues by providing a method, device, equipment, and medium for virtual restoration of cultural relics based on digital twin technology. This method and medium can use artificial intelligence to reproduce the original state of cultural relics lost overseas in a virtual environment, simulate restoration effects, and accurately predict potential risks.

[0004] Firstly, this application provides a virtual restoration method for cultural relics based on digital twin technology, including:

[0005] The original data of cultural relics lost overseas was obtained; and the original data was standardized and preprocessed to obtain a cultural relic data package; the original data includes high-resolution images, 3D point clouds and historical documents;

[0006] Based on the cultural relic data package, three-dimensional geometric reconstruction is performed to obtain a three-dimensional digital model of the cultural relic.

[0007] Using an image segmentation model, we identify defects in the three-dimensional digital model of cultural relics and obtain a list of cultural relic defects. Defects include at least one of the following: cracks, missing parts, and corrosion.

[0008] Based on the list of cultural relics defects, the risk of structural stability of the three-dimensional digital model of the cultural relics is predicted by mechanical simulation, and a risk prediction map of the cultural relics is obtained.

[0009] Based on the list of cultural relic defects and the risk prediction map of cultural relic, restoration strategies with different focuses are generated in parallel to obtain virtual restoration schemes.

[0010] Furthermore, based on the list of cultural relic defects, the structural stability risk of the three-dimensional digital model of the cultural relic is predicted through mechanical simulation, resulting in a cultural relic risk prediction map, including:

[0011] The three-dimensional digital model of the cultural relic is re-divided into computational meshes to obtain a mesh model; and the mesh model is then geometrically cleaned to obtain a computational preparation model.

[0012] Based on a pre-defined knowledge base of cultural relics materials, mechanical parameters and constraints are assigned to the calculation preparation model to obtain a finite element model.

[0013] Stress distribution calculations are performed on the finite element model to obtain a stress distribution cloud map; the stress distribution cloud map is used to characterize the magnitude and distribution of stress within the entire cultural relic.

[0014] A quantitative analysis was performed by overlaying the list of cultural relic defects and stress distribution cloud maps to obtain a risk prediction map for cultural relics.

[0015] Furthermore, stress distribution calculations are performed on the finite element model to obtain stress distribution contour maps, including:

[0016] The element stiffness matrix is ​​calculated for each element in the finite element model to obtain a set of element stiffness matrices; and all the set of element stiffness matrices are superimposed to form the global stiffness matrix; the global stiffness matrix is ​​used to characterize the structural stiffness characteristics of the cultural relic;

[0017] Based on gravity simulation, the distributed forces on the finite element model are distributed to each node to obtain the overall load vector;

[0018] Based on the overall stiffness matrix and the overall load vector, the displacement of each node under the load is calculated to obtain the displacement vector of all nodes.

[0019] Based on the displacement vector, the stress components of each element in the element stiffness matrix are calculated; and the stress components are averaged across the nodes to obtain a stress distribution cloud map.

[0020] Furthermore, based on the list of cultural relic defects and the risk prediction map of cultural relics, restoration strategies with different focuses are generated in parallel to obtain virtual restoration plans, including:

[0021] Based on preset priority principles, a basic repair plan is generated; the priority principles include security priority, historical authenticity priority, and educational priority.

[0022] Based on the cultural relic risk prediction map, a quantitative analysis was conducted on each dimension of the basic restoration plan to obtain a quantitative score table; the dimensions include at least one of the following: structural safety improvement degree, minimum intervention compliance degree, historical information preservation degree, reversibility and aesthetic harmony.

[0023] Based on the repair objectives, weights are assigned to the dimensions to obtain an indicator weight table.

[0024] Based on the indicator weight table, the quantitative score table is weighted and summed using the following formula to obtain the score list; the score list includes the comprehensive score corresponding to each basic repair plan:

[0025]

[0026] Where S is the overall score, i is the dimension index, n is the total number of dimensions, and w i s is the weighting coefficient. i The score for the i-th dimension;

[0027] The basic repair scheme with the highest overall score is selected as the virtual repair scheme.

[0028] Furthermore, based on preset priority principles, a basic repair plan is generated, including:

[0029] Based on the principle of prioritizing educational value and a quantitative list of cultural relic defects, a complete morphological hypothesis for the three-dimensional digital model of the cultural relic is generated through a pre-trained generative AI model, resulting in a set of morphological prediction hypotheses. The generative AI model is trained using data contained in an art style database.

[0030] Based on the set of morphological inference hypotheses, the three-dimensional digital model of the cultural relic is virtually completed to obtain the completed three-dimensional model.

[0031] Based on an art style database, color rendering is performed on the completed 3D model to obtain the initial 3D model for manufacturing.

[0032] By integrating the initial 3D model of manufacturing and the virtual repair operation sequence, a basic repair plan is obtained.

[0033] Furthermore, after determining the basic repair scheme with the highest overall score as the virtual repair scheme, it also includes:

[0034] Based on a real-time rendering engine, the 3D model in the virtual restoration scheme is rendered to obtain the restored cultural relic model.

[0035] The virtual repair operation sequence is animated to obtain a virtual repair animation;

[0036] By integrating the restored artifact model, the artifact's 3D digital model, and the virtual restoration animation, a visualized result of the artifact restoration is obtained.

[0037] Furthermore, based on the artifact data package, three-dimensional geometric reconstruction is performed to obtain a three-dimensional digital model of the artifact, including:

[0038] Feature enhancement processing is performed on the 3D point cloud in the cultural relic data package to obtain the baseline 3D point cloud;

[0039] Based on the point cloud registration algorithm, the reference 3D point cloud and the high-resolution image are spatially aligned to obtain the aligned 3D point cloud;

[0040] A 3D network model is generated based on aligned 3D point clouds using a surface reconstruction algorithm.

[0041] Based on the texture mapping algorithm, the color and texture information corresponding to the high-definition image is projected onto the three-dimensional network model to obtain a three-dimensional digital model of the current state of the cultural relics.

[0042] The detailed information from the high-resolution images is added as an additional layer to the existing 3D digital model of the cultural relic, resulting in a 3D digital model of the cultural relic.

[0043] Secondly, this application also provides a virtual restoration device for cultural relics based on digital twin technology, comprising:

[0044] The preprocessing module is used to acquire the raw data of cultural relics lost overseas; and to perform standardized preprocessing on the raw data to obtain a cultural relic data package; the raw data includes high-resolution images, 3D point clouds and historical documents;

[0045] The reconstruction module is used to perform three-dimensional geometric reconstruction based on the cultural relic data package to obtain a three-dimensional digital model of the cultural relic.

[0046] The identification module is used to identify defects appearing in the three-dimensional digital model of cultural relics through image segmentation model, and obtain a list of cultural relics defects; the defects include at least one of cracks, missing parts and corrosion;

[0047] The risk module is used to predict the structural stability risk of a three-dimensional digital model of a cultural relic based on a list of cultural relic defects and through mechanical simulation, thereby obtaining a risk prediction map of the cultural relic.

[0048] The solution module is used to generate virtual restoration solutions by generating restoration strategies with different focuses based on the list of cultural relic defects and the risk prediction map of cultural relics.

[0049] Thirdly, this application also provides a computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement any step of the method provided in the first aspect of this application.

[0050] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any step of the method provided in the first aspect of this application.

[0051] The aforementioned method, apparatus, equipment, and media for virtual restoration of cultural relics based on digital twin technology acquire raw data of cultural relics lost overseas; standardize and preprocess the raw data to obtain a cultural relic data package; the raw data includes high-resolution images, 3D point clouds, and historical documents; based on the cultural relic data package, perform 3D geometric reconstruction to obtain a 3D digital model of the cultural relic; identify defects appearing in the 3D digital model of the cultural relic through an image segmentation model to obtain a list of cultural relic defects; defects include at least one of cracks, missing parts, and corrosion; based on the list of cultural relic defects, predict the structural stability risk of the 3D digital model of the cultural relic through mechanical simulation to obtain a cultural relic risk prediction map; based on the list of cultural relic defects and the cultural relic risk prediction map, generate restoration strategies with different focuses in parallel to obtain a virtual restoration plan. By quantifying the impact of cultural relic defects on the cultural relic and pre-simulating the restoration effect, a basic plan is provided for the cultural relic restoration process, avoiding the risks brought about by direct human intervention in the cultural relic. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 A schematic diagram illustrating the process of a virtual restoration method for cultural relics based on digital twin technology according to an embodiment of the present invention;

[0054] Figure 2 This is a schematic diagram of the structure of a virtual cultural relic restoration device based on digital twin technology provided in an embodiment of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0056] In one embodiment, such as Figure 1 As shown, a virtual restoration method for cultural relics based on digital twin technology is provided. This embodiment illustrates the method applied to a terminal, but it is understood that the method can also be applied to a server, or to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0057] Step 101: Obtain the original data of cultural relics lost overseas; and perform standardized preprocessing on the original data to obtain a cultural relic data package; the original data includes high-resolution images, 3D point clouds, and historical documents.

[0058] The raw data consists of various types of data directly collected through physical means and not yet standardized. High-resolution images are high-resolution photographs of cultural relics taken from multiple angles, used to record the surface color, texture, pattern, and microscopic details of relics lost overseas. 3D point clouds are collections of massive data points obtained through technologies such as 3D laser scanning; each point has three-dimensional spatial coordinates, and together they delineate the precise external geometry of the cultural relic. Historical documents are background information related to the cultural relic, including excavation reports, historical records, data on similar cultural relics, and material analysis reports, providing historical and scientific basis for subsequent restoration decisions. The cultural relic data package is a standardized, structured data set containing registered and aligned high-resolution images and 3D point clouds, as well as associated historical documents; it is the sole and reliable data source for all subsequent work. The terminal performs color correction and distortion correction on high-definition images to ensure true colors and no image distortion. It also performs noise reduction and encryption on 3D point clouds to optimize data quality. The high-definition images and 3D point clouds are precisely spatially aligned so that each pixel in the photo corresponds to the 3D coordinates of the point cloud. Historical documents are digitized and key information is extracted. Multiple heterogeneous data are integrated into a whole to obtain a cultural relic data package.

[0059] Step 102: Based on the cultural relic data package, perform three-dimensional geometric reconstruction to obtain a three-dimensional digital model of the cultural relic.

[0060] Specifically, a 3D digital model of a cultural relic is a virtual representation of the relic in a computer, also known as a digital twin. It includes a 3D mesh that defines the shape of the object and high-resolution texture maps attached to the mesh. The terminal performs feature enhancement on the point cloud to make key contours clearer. Through surface reconstruction algorithms, isolated point clouds are connected into a continuous, watertight triangular mesh surface. Texture mapping is then performed, and multiple high-resolution images are seamlessly stitched onto the 3D mesh surface according to their correspondence with the model, giving the model realistic color and texture. Extremely high-frequency details from the images are added as an additional layer to enhance the model, improving its precision and generating a high-fidelity digital copy of the cultural relic.

[0061] Step 103: Identify the defects appearing in the three-dimensional digital model of the cultural relic through the image segmentation model to obtain a list of cultural relic defects; the defects include at least one of cracks, missing parts and corrosion.

[0062] Specifically, image segmentation models are trained artificial intelligence algorithms that can identify different objects or regions in an image, much like the human eye. They can automatically classify each pixel or region on an image or 3D model into different categories. "Disease" refers to various types of damage to cultural relics, mainly including cracks (structural fissures), missing parts (partial structural detachment or loss), and corrosion (weathering and peeling of surface materials). A list of cultural relic defects is a structured diagnostic report containing quantitative information such as the type of defect, its precise location on the 3D model, its geometric dimensions, and its severity. The terminal inputs the 3D digital model of the cultural relic, including its geometric and texture information, into the pre-trained image segmentation model. The model scans and analyzes pixel by pixel, automatically marking all regions belonging to the defect category. The terminal calculates various parameters for these marked regions, ultimately generating a structured list of defects.

[0063] Step 104: Based on the list of cultural relic defects, predict the risk of structural stability of the three-dimensional digital model of the cultural relic through mechanical simulation, and obtain the cultural relic risk prediction map.

[0064] In this embodiment, mechanical simulation specifically refers to finite element analysis, a computer simulation technique that simulates and calculates the mechanical response of a complex structure under stress by discretizing it into a large number of small, simple elements. Structural stability risk refers to the possibility that a cultural relic may deform, crack, or even collapse due to its own inherent defects and external loads such as gravity. The cultural relic risk prediction map is a visualized analysis result map, overlaid on the 3D model in the form of a color cloud map. Different colors in the map represent different risk levels, intuitively showing which defective areas are the weak points in the structure and most prone to problems. The terminal converts the 3D model into a mesh model suitable for calculation and performs geometric cleanup. Based on the material of the cultural relic, it assigns realistic physical properties to the model, sets constraints, and applies loads. By solving complex mechanical equations, it calculates the stress magnitude at each point inside the model. The calculated stress distribution cloud map is overlaid with the list of cultural relic defects for analysis. If an area already has defects and the calculated stress is high, then that area will be marked as high-risk.

[0065] Step 105: Based on the list of cultural relic defects and the risk prediction map of cultural relic, generate restoration strategies with different focuses in parallel to obtain virtual restoration plans.

[0066] The restoration strategy is a plan outlining the specific methods, materials, and steps for restoring cultural relics. Different focuses refer to strategies prioritizing objectives such as structural safety, minimal intervention, historical authenticity, or public education and display. The virtual restoration plan is the optimal plan selected after comprehensive evaluation, containing a detailed sequence of restoration steps, information on the virtual materials used, and a model of the expected 3D effect after restoration. Based on different priority principles—for example, a safety-first principle focuses on reinforcing high-risk areas, while an education-first principle uses AI to predict and fill in missing parts—the terminal generates multiple basic restoration plans, establishing a multi-dimensional evaluation system. Weights are assigned to each dimension according to the restoration goals, and each plan is scored across all dimensions. A weighted summation is then used to calculate a comprehensive score, and the plan with the highest score is determined as the final virtual restoration plan.

[0067] This embodiment provides a virtual restoration method for cultural relics based on digital twin technology. It acquires the original data of cultural relics lost overseas and performs standardized preprocessing to obtain a cultural relic data package. The original data includes high-resolution images, 3D point clouds, and historical documents. Based on the cultural relic data package, 3D geometric reconstruction is performed to obtain a 3D digital model of the cultural relic. Through image segmentation, defects appearing in the 3D digital model are identified, resulting in a list of defects. Defects include at least one of cracks, missing parts, and corrosion. Based on the list of defects, mechanical simulation is used to predict the structural stability risk of the 3D digital model, resulting in a risk prediction map. Based on the list of defects and the risk prediction map, restoration strategies with different focuses are generated in parallel to obtain a virtual restoration plan. Through these methods, the impact of cultural relic defects on the relics is identified and quantified, the restoration effect is simulated, a basic plan is provided for the restoration process, and the risks associated with direct human intervention in the restoration of cultural relics are avoided.

[0068] In one embodiment, based on a list of cultural relic defects, the structural stability risk of the three-dimensional digital model of the cultural relic is predicted through mechanical simulation, resulting in a cultural relic risk prediction map, including:

[0069] Step 201: Re-divide the three-dimensional digital model of the cultural relic into a computational mesh to obtain a mesh model; and perform geometric cleanup on the mesh model to obtain a computational preparation model.

[0070] The computational mesh discretizes a complex continuous geometry into a large set of simple, standard-shaped elements. These elements are interconnected by nodes and form the mathematical foundation for finite element method (FEM) calculations. Complex mechanical problems are decomposed into problems that need to be solved on these simple elements. The mesh model is a model entirely composed of computational meshes, obtained after the 3D digital model of the artifact has been re-divided. It defines the computational domain but does not yet include physical properties. Geometric cleanup refers to eliminating meaningless geometric features in the original 3D model that negatively impact the calculation. Examples include removing tiny cracks, burrs, and non-manifold geometry caused by scanning noise, and merging surfaces that are too close together. The computational preparation model is a model with a high-quality computational mesh after geometric cleanup. The high mesh quality ensures the accuracy and stability of subsequent mechanical calculations. The terminal uses a specialized meshing algorithm to automatically mesh the 3D digital model of the artifact. During meshing, a balance between computational accuracy and efficiency must be considered. A denser mesh is used in areas where stress may concentrate. The generated mesh model undergoes geometric cleanup, checking and repairing problematic mesh elements to ensure there are no overlapping, distorted, or excessively deformed elements.

[0071] Step 202: Based on the preset knowledge base of cultural relics materials, mechanical parameters and constraints are assigned to the calculation preparation model to obtain the finite element model.

[0072] Specifically, the cultural relic materials knowledge base is a pre-defined database containing the mechanical parameters of various common materials used in cultural relics. Key parameters include elastic modulus, Poisson's ratio, and density. Mechanical parameters refer to the specific values ​​assigned to the computational model from the materials knowledge base; these parameters determine how the model responds under stress. Constraints are boundary conditions used in mechanical simulations to limit the displacement of certain parts of the model. For example, simulating an object placed on a table requires setting the displacement of its bottom nodes to zero to simulate the support of the tabletop, thus simulating the support state of the cultural relic in a real environment. The finite element model is a complete physical model that can be solved, containing geometric information, material properties, and boundary conditions. Based on the material of the cultural relic, the terminal queries and selects the most matching set of material parameters from the cultural relic materials knowledge base, assigns the parameters to each element of the mesh model, and applies constraints to the model according to the actual placement of the cultural relic, limiting the displacement of the nodes, thus obtaining the finite element model.

[0073] Step 203: Calculate the stress distribution of the finite element model to obtain a stress distribution cloud map; the stress distribution cloud map is used to characterize the magnitude and distribution of stress inside the entire cultural relic.

[0074] Specifically, stress distribution calculation refers to solving the governing equations of mechanics using the finite element method (FEM) to calculate the stress values ​​at each element and node in the finite element model. Stress is the internal force per unit area within an object; it is a vector that can be decomposed into components in different directions and is a key indicator of the severity of stress on a structure. A stress distribution cloud map is a visual chart, typically overlaid on a 3D model in color, with different colors representing the magnitude and distribution of stress. The final step involves assembling the overall stiffness matrix and overall load vector of the entire model, solving a large system of linear equations to obtain the displacement of each node, calculating the stress of each element based on the displacement results, and mapping the results back to the nodes to generate a continuous stress distribution cloud map.

[0075] Step 204: Perform a quantitative analysis by overlaying the list of cultural relic defects and the stress distribution cloud map to obtain a cultural relic risk prediction map.

[0076] The overlay quantitative analysis refers to the spatial matching and numerical comprehensive calculation of the list of cultural relic defects and stress distribution cloud maps. The cultural relic risk prediction map is a risk level distribution map based on comprehensive assessment, containing risk level information for different areas. The judgment is based on whether high stress and severe defects spatially overlap. The terminal traverses each defect area in the list of cultural relic defects. For each defect area, it finds the stress value at the corresponding location on the stress distribution cloud map and judges it according to a set of risk assessment rules. For example, a long and deep crack located in a high-stress area will be marked as extremely high in comprehensive risk level; while the same crack located in a low-stress area may have a medium risk level. Through overlay and quantification, the final risk prediction map is generated.

[0077] This embodiment combines disease and stress through a risk prediction diagram, thereby improving the diagnosis from physical phenomena to structural safety prognosis and providing the most direct and scientific basis for the formulation of repair plans.

[0078] In one embodiment, stress distribution calculation is performed on the finite element model to obtain a stress distribution cloud map, including:

[0079] Step 301: Calculate the element stiffness matrix for each element in the finite element model to obtain a set of element stiffness matrices; and superimpose all the set of element stiffness matrices into an overall stiffness matrix; the overall stiffness matrix is ​​used to characterize the structural stiffness characteristics of the cultural relic.

[0080] The element stiffness matrix is ​​a mathematical expression describing the deformation resistance of a single mesh element. It is a square matrix whose element values ​​are determined by the shape, size, and material properties of the element, defining the relationship between nodal forces and nodal displacements within the element. The set of element stiffness matrices refers to the sum of the stiffness matrices of all elements in the finite element model. The global stiffness matrix is ​​a large, globally sparse matrix assembled by superimposing the stiffness matrices of all elements in the model according to their corresponding node numbers. It is used to characterize the overall stiffness characteristics of the entire cultural relic structure, that is, the total ability of the structure to resist deformation under any load. For each element in the finite element model, an element stiffness matrix is ​​generated using standard mathematical formulas based on its geometric parameters and mechanical parameters obtained from the material knowledge base. Based on the node connection information of each element in the global model, the values ​​in each element stiffness matrix are accumulated and added to the corresponding position in the global stiffness matrix. This process is repeated for all elements to form a complete global stiffness matrix. For example, for a triangular element that connects nodes 1, 2, and 3, the element in its stiffness matrix that describes the relationship between nodes 1 and 2 will be added to the position of the first row and second column of the overall stiffness matrix.

[0081] Step 302: Based on gravity simulation, distribute the distributed forces on the finite element model to each node to obtain the overall load vector.

[0082] Specifically, gravity simulation refers to applying loads equivalent to real-world gravity in the simulation. Distributed force in this embodiment specifically refers to gravity, which is the force acting on the entire volume of the artifact, not a force concentrated at a single point. The overall load vector is a vector with a length equal to the total degrees of freedom of the model. The value of each element in the vector represents the equivalent concentrated force acting on the corresponding node, used to characterize the external loads experienced by the entire artifact structure. The terminal uses the principle of virtual work to equivalently distribute gravity to each node, calculates the total force generated by gravity in each element, and then distributes this force to each node of that element according to rules related to the element's shape function. The forces distributed to the nodes of all elements are then assembled into the corresponding positions of the overall load vector according to the node numbers.

[0083] Step 303: Based on the overall stiffness matrix and the overall load vector, calculate the displacement of each node under the load to obtain the displacement vector of all nodes.

[0084] Specifically, displacement refers to the amount of movement of each node relative to its original position under load, including linear and angular displacement. The displacement vector is a vector whose elements are an ordered set of all nodal displacements, used to characterize the deformation state of the entire cultural relic structure after being subjected to force. The final step involves solving a large system of linear equations, with the basic form: global stiffness matrix × displacement vector = global load vector. After applying constraints, this equation is well-posed and solvable. Due to the enormous and sparse nature of the global stiffness matrix, the Choliski decomposition method or the conjugate gradient method is used to solve for this displacement vector.

[0085] Step 304: Based on the displacement vector, calculate the stress components of each element in the element stiffness matrix; and average the stress components to the nodes to obtain the stress distribution cloud map.

[0086] The stress components indicate that stress is a tensor that can be decomposed into components in different directions. A stress distribution cloud map is a visual image that maps calculated stress values ​​back to the geometric model. The map typically uses continuously changing colors to represent the magnitude and distribution of stress; red usually represents high-stress areas, and blue represents low-stress areas. The terminal calculates the strain of each element based on the nodal displacement vector and the strain-displacement relationship. The stress components of that element are then calculated from the strain using the material's constitutive relations. Since the same node may belong to multiple adjacent elements, the stress values ​​calculated at that node by different elements may differ. Therefore, averaging is required. The stress values ​​of all elements sharing the same node are weighted and averaged to obtain the stress value of that node. The stress values ​​of all nodes are then rendered on the model using color, forming a continuous stress distribution cloud map.

[0087] This embodiment transforms abstract numerical results into intuitive visual images through stress distribution cloud maps, enabling easy identification of weak points and high-risk areas in the structure during the restoration of cultural relics. It transforms physical problems into mathematical problems, allowing for the quantitative solution of structural risks of cultural relics through mathematical methods, thereby improving the accuracy and safety of virtual restoration of cultural relics.

[0088] In one embodiment, based on the list of cultural relic defects and the cultural relic risk prediction map, restoration strategies with different focuses are generated in parallel to obtain a virtual restoration plan, including:

[0089] Step 401: Generate a basic repair plan based on preset priority principles; the priority principles include security priority principle, historical authenticity priority principle, and educational priority principle.

[0090] Among these principles, the priority principle refers to the fundamental value orientation and highest standard followed when formulating a restoration plan, defining the focus and starting point of the plan's generation. The safety priority principle prioritizes eliminating structural safety hazards, and under this principle, the plan tends to adopt the most effective reinforcement measures, potentially sacrificing other aspects to some extent. The historical authenticity priority principle, also known as the minimum intervention principle, focuses on preserving the historical information of the artifact to the greatest extent possible. The plan tends to perform only necessary stabilization treatments, avoiding the addition of any new or non-original materials or structures, emphasizing restoring the old as it was. The educational priority principle focuses on showcasing the complete historical appearance and educational value of the artifact to the public. The plan allows for reasonable morphological speculation and virtual reconstruction based on academic research to facilitate understanding. The basic restoration plan is a preliminary restoration plan generated based on a specific priority principle, including the main restoration measures to be taken to achieve the core objectives of that principle, the virtual materials to be used, and a general sequence of restoration steps. For example, based on the selected priority principle, the terminal makes a solution decision by combining the list of cultural relic defects and the risk prediction map. If the safety priority principle is followed, the focus is on the high-risk areas in the cultural relic risk prediction map, and a solution based on structural reinforcement is generated, such as injecting high-strength adhesive into the cracks and adding invisible support frames in key parts. If the historical authenticity priority principle is followed, the solution will try its best to avoid changing the current state of the cultural relic, only suggesting surface cleaning and weathering prevention treatment for the cracks, and opposing any form of restoration or reconstruction. If the educational priority principle is followed, the AI ​​prediction process is initiated to generate a restoration model showing the possible original appearance of the cultural relic and to formulate a corresponding restoration operation sequence.

[0091] Step 402 involves a quantitative analysis of each dimension of the basic restoration plan based on the cultural relic risk prediction map, resulting in a quantitative score table. The dimensions include at least one of the following: structural safety improvement, minimum intervention compliance, historical information retention, reversibility, and aesthetic harmony.

[0092] The dimensions are a series of independent evaluation criteria or perspectives used to assess the quality of a restoration plan, forming an evaluation index system. Structural safety improvement refers to the extent to which the assessment plan eliminates the structural risks identified in the cultural relic risk prediction map. Minimum intervention compliance is the degree of intervention the assessment plan makes on the cultural relic itself; less intervention results in a higher score. Historical information preservation is the degree to which the assessment plan protects the existing historical traces, materials, and techniques of the cultural relic. Reversibility assesses whether the restoration measures can be safely removed and the original state restored when future technological conditions improve. Aesthetic harmony assesses the visual harmony between the restored portion and the overall cultural relic. The quantitative score table is a table or data structure that clearly lists the specific scores obtained by each basic restoration plan on each evaluation dimension. The terminal evaluates the performance of each basic restoration plan across all dimensions based on the cultural relic risk prediction map, plan details, and professional knowledge. For example, when assessing the improvement in structural safety, the score is based on how much the area of ​​high-risk areas in the cultural relic risk prediction map has decreased after the implementation of the plan; when assessing the compliance of minimum intervention, the score is based on the volume and area involved in the restoration measures and whether new materials have been added.

[0093] Step 403: Based on the repair objective, assign weights to the dimensions to obtain the indicator weight table.

[0094] The restoration objective refers to the core purpose to be achieved in this specific restoration project. The indicator weight table is a list defining the importance of each evaluation dimension, with weight coefficients ranging from 0 to 1 (decimals), and the sum of the weights for all dimensions is 1. The terminal assigns weights to each evaluation dimension based on the specific objectives of this restoration. For example, if the objective is emergency reinforcement, the weight for structural safety improvement will be assigned the highest value, while the weight for aesthetic harmony will be lower; if the objective is exhibition and display, the weight for aesthetic harmony will be increased; if the objective is archaeological research, the weights for historical information preservation and reversibility will be increased.

[0095] Step 404: Based on the indicator weight table, the quantitative score table is weighted and summed using the following formula to obtain a score list; the score list includes the comprehensive score corresponding to each basic repair plan:

[0096]

[0097] Where S is the overall score, i is the dimension index, n is the total number of dimensions, and w i s is the weighting coefficient. i Let be the score for the i-th dimension.

[0098] Specifically, the score list is a clear list that shows the final comprehensive score corresponding to each basic repair plan. The score reflects the overall performance of the plan after weighting. For each basic repair plan, the terminal uses a formula to calculate the comprehensive score. It takes the score of the plan in each dimension from the quantitative score table, multiplies it by the corresponding weight in the indicator weight table, and adds the products. This process is repeated for each plan to obtain the comprehensive score for each plan.

[0099] Step 405: The basic repair scheme with the highest comprehensive score is determined as the virtual repair scheme.

[0100] Specifically, the virtual repair scheme is the basic repair scheme with the highest overall score selected, which serves as the direct basis for subsequent virtual simulation and actual repair work.

[0101] This embodiment ensures that the final selected virtual repair solution is the optimal solution that best balances the needs of all parties and best meets the core repair objectives under the current understanding conditions through a complete, transparent, and traceable scientific decision-making process, from generating options to quantitative evaluation and then to objective selection, thereby improving the accuracy of virtual repair.

[0102] In one embodiment, a basic repair plan is generated based on a preset priority principle, including:

[0103] Step 501: Based on the principle of prioritizing education and the quantitative list of cultural relic defects, a complete morphological hypothesis of the three-dimensional digital model of the cultural relic is generated through a pre-trained generative AI model, resulting in a set of morphological prediction hypotheses; the generative AI model is trained using data contained in the art style database.

[0104] The principle of prioritizing educational value is a restoration philosophy whose core objective is to maximize the public's understanding of the historical, artistic, and scientific value of cultural relics. It allows for reasonable and identifiable restoration, provided sufficient evidence is available, to help viewers understand the original function and appearance of the relics. A quantitative inventory of cultural relic defects precisely describes the location, size, and shape of missing parts. A pre-trained generative AI (Artificial Intelligence) model is an AI algorithm trained on a large amount of data, learning the artistic style, modeling rules, structural proportions, and decorative features of cultural relics from specific periods and of specific types. When given information about a damaged cultural relic, it can generate a complete form that conforms to its stylistic logic. An art style database, containing a large amount of digitized data on similar complete cultural relics, is used to train the generative AI model, enabling it to acquire the necessary artistic style knowledge. A set of morphological inference hypotheses is used because there may be multiple reasonable inferences about the original form of the missing parts of a cultural relic. The AI ​​model typically generates multiple possible complete form schemes, forming a set of hypotheses. This reflects academic rigor and provides multiple options for subsequent decision-making. The terminal inputs the current three-dimensional digital model of the cultural relic and its quantitative list of defects into the pre-trained generative AI model. Based on the knowledge it has learned from the art style database, the model analyzes the stylistic characteristics of the existing parts of the cultural relic and generates one or more virtual complete parts that can be naturally connected with the existing parts and have a unified style. The generated results together constitute a set of morphological inference hypotheses.

[0105] Step 502: Based on the set of morphological inference hypotheses, the three-dimensional digital model of the cultural relic is virtually completed to obtain the completed three-dimensional model.

[0106] Specifically, virtual completion refers to the seamless stitching and fusion of inferred morphology with the original, incomplete 3D digital model of a cultural relic in a computer environment using digital geometric processing technology. The completed 3D model is a geometrically complete, watertight 3D mesh model that includes both the existing real parts of the cultural relic and the virtually generated parts based on inference, together forming a complete geometric body. The terminal selects the optimal hypothesis from a set of morphological inference assumptions and uses 3D modeling software to precisely align the inferred virtual components with the incomplete edges of the existing model of the cultural relic. Through surface reconstruction, mesh Boolean operations, and smooth edge transitions, it ensures a smooth geometric connection between the virtual and real parts, forming a seamless overall model.

[0107] Step 503: Based on the art style database, perform color rendering on the completed 3D model to obtain the initial 3D model for manufacturing.

[0108] Specifically, in this embodiment, color rendering refers to assigning appropriate colors, patterns, and material textures to the virtual parts of the completed 3D model based on historical research and an art style database. The initial 3D model is a highly realistic model, not only with complete geometric shapes but also with surface textures and colors that reproduce the artifact's pristine state as closely as possible, containing high-precision material and texture information. The terminal, based on the art style database and relevant historical documents, determines the appropriate materials and color patterns for the completed parts, and precisely projects this information onto the corresponding areas of the completed 3D model using digital rendering and material mapping technologies.

[0109] Step 504: Integrate the initial 3D model and virtual repair operation sequence to obtain the basic repair plan.

[0110] The virtual restoration operation sequence refers to a series of ordered digital operation instructions required to transform the existing 3D digital model of the cultural relic into a 3D model for initial manufacturing. The basic restoration plan includes the final visualization goal and specific methods and steps. The terminal associates and packages the initial manufacturing 3D model with the operation sequence that details the type, parameters, and virtual materials used in each step, forming a structured plan document.

[0111] This embodiment generates a 3D model of the artifact in its original state within a virtual environment, giving the digital model a visually fresh look and greatly enhancing its aesthetic appeal and educational value. In real-world artifact restoration, the principle of minimal interference is a fundamental rule; however, in a virtual environment, the public can intuitively understand the artifact's historical splendor through this model, thus fulfilling the core objective of prioritizing education.

[0112] In one embodiment, after determining the basic repair scheme with the highest overall score as the virtual repair scheme, the method further includes:

[0113] Step 601: Render the 3D model in the virtual restoration scheme based on the real-time rendering engine to obtain the restored cultural relic model.

[0114] The real-time rendering engine is a powerful computer graphics software capable of rapidly calculating and generating highly realistic images, simulating complex lighting, shadows, material textures, and physical effects. The 3D model in the virtual restoration scheme is the ideal 3D model after restoration, containing complete geometric shape and surface material information. Rendering refers to the process by which the rendering engine, based on the lighting, environment, and material parameters set in the scene, uses a series of complex mathematical calculations to convert the lighting, color, and texture information of the 3D model into a series of two-dimensional, photorealistic images or dynamic scenes. The restored artifact model refers to a high-fidelity image or interactive 3D scene processed by the rendering engine, used to showcase the final expected effect of the restoration scheme. The terminal sets up a virtual scene in the real-time rendering engine, including arranging lighting and selecting the background environment, importing the 3D model determined by the virtual restoration scheme, and assigning it detailed material properties. The engine calculates how light interacts with the model surface in real-time or offline rendering mode, generating a realistic artifact model.

[0115] Step 602: Animate the virtual repair operation sequence to obtain a virtual repair animation.

[0116] Specifically, a virtual restoration operation sequence refers to a detailed list of restoration steps recorded when generating a basic restoration plan. Animated operation refers to using 3D animation technology to transform a static operation sequence into a continuous, visual animated film, defining the start and end states of each operation and simulating the intermediate changes. A virtual restoration animation is a dynamic video that encompasses the entire process of a cultural relic being gradually restored from a damaged state to a complete state, clearly demonstrating the restoration sequence, the virtual technologies used, and key technical details. The terminal sets keyframes in the 3D animation software based on the operation sequence. For example, in the step of filling cracks, an animation is set to show the crack narrowing until it disappears; when filling in missing parts, an animation is set to show the missing parts gradually appearing from nothing, and through tweening calculations, smooth transition frames are automatically generated, ultimately synthesizing a complete animation.

[0117] Step 603: Integrate the restored artifact model, the artifact's 3D digital model, and the virtual restoration animation to obtain a visualized artifact restoration result.

[0118] Specifically, the visualized cultural relic restoration result is a comprehensive deliverable package, including a pre-restoration model, a post-restoration model, and an animation of the restoration process. The terminal is based on a visualization platform, with the pre-restoration model on the left, the post-restoration model on the right, a slider in the middle to smoothly transition between the two states, and an embedded control at the bottom to play the restoration animation.

[0119] This embodiment produces the final deliverables of the entire virtual restoration project, providing a comprehensive perspective that allows cultural relic restoration workers or the public to compare the significant differences before and after restoration, while also understanding the ins and outs of the restoration process. It has irreplaceable value for project archiving, academic exchange, public exhibition, and assisting actual restoration work, thereby enhancing the feasibility of virtual restoration of cultural relics.

[0120] In one embodiment, based on the artifact data package, three-dimensional geometric reconstruction is performed to obtain a three-dimensional digital model of the artifact, including:

[0121] Step 701: Perform feature enhancement processing on the 3D point cloud in the cultural relic data package to obtain the baseline 3D point cloud.

[0122] Feature enhancement processing involves optimizing the original 3D point cloud data using a series of algorithms. The focus is on noise removal, hole filling, surface smoothing, and enhancing the clarity of key geometric features. The baseline 3D point cloud, obtained after feature enhancement, is of higher quality and serves as the foundation for all subsequent geometric operations. It features lower noise, a more complete structure, and clearer feature contours. The terminal application uses outlier removal algorithms to filter out isolated noise points that are significantly deviated from the object's surface due to scanner errors or environmental interference. A smoothing filter algorithm is used to slightly smooth the point cloud to reduce data fluctuations and make the surface smoother. For sparse or missing data areas, interpolation algorithms are used to fill in the gaps. Feature point extraction algorithms are employed to enhance the saliency of key features such as corners and edges, significantly improving the data quality of the original point cloud.

[0123] Step 702: Based on the point cloud registration algorithm, spatially align the reference 3D point cloud and the high-resolution image to obtain an aligned 3D point cloud.

[0124] Specifically, point cloud registration algorithms are computational methods used to find spatial transformation relationships between two or more sets of data, aligning them in the same coordinate system. Spatial alignment refers to determining the precise correspondence between the shooting perspective of a high-resolution photograph and a 3D point cloud through calculation, ensuring that every pixel in the photograph corresponds to a specific 3D coordinate in the point cloud. The aligned 3D point cloud is a baseline 3D point cloud that has already established a precise spatial correspondence with the high-resolution image; it not only contains the 3D coordinates of each point but also associates correct color information from one or more high-resolution images. The terminal identifies common feature points on the point cloud and the high-resolution images, and by matching these feature points, calculates the camera parameters required to project the photograph onto the point cloud. This process is performed one by one on multiple high-resolution photographs taken from different angles to ensure that the entire point cloud surface is covered.

[0125] Step 703: Based on the aligned 3D point cloud, a 3D network model is generated using a surface reconstruction algorithm.

[0126] The surface reconstruction algorithm is a computer graphics algorithm that aims to connect and fit discrete 3D point cloud data to generate a continuous surface mesh composed of numerous polygons. The 3D network model is a triangular mesh model, consisting of vertices and triangular faces, defining the geometry of the artifact; at this stage, it is a model without color or texture. The terminal analyzes the spatial distribution of the point cloud through Poisson reconstruction, infers the surface orientation of the object, and connects triangular faces between points to form a watertight mesh surface that encloses all points.

[0127] Step 704: Based on the texture mapping algorithm, the color and texture information corresponding to the high-definition image is projected onto the three-dimensional network model to obtain a three-dimensional digital model of the current state of the cultural relic.

[0128] Specifically, the current 3D digital model of the cultural relic is a 3D model that combines accurate geometry and realistic texture. It includes triangular mesh geometry and texture maps attached to it, recording the current surface condition of the cultural relic, including all defects, stains, and signs of aging. The terminal uses alignment relationships to project multiple high-resolution images onto the 3D mesh model. For each triangular facet on the mesh, it calculates which pixel region corresponds to which image, extracts the color information of that pixel, and assigns it to that facet. Through texture fusion technology, it eliminates color differences at the seams of multiple images, generating a complete and seamless texture map.

[0129] Step 705: Add the detailed information from the high-resolution image as an additional layer to the current 3D digital model of the cultural relic to obtain the 3D digital model of the cultural relic.

[0130] Specifically, detail information refers to microscopic features in high-resolution images that exceed the resolution of conventional texture maps, such as subtle scratches, brushstrokes, fabric fibers, and the peeling patterns of paint. Add-on layer enhancement typically uses advanced rendering techniques such as normal mapping or displacement mapping to store detail information as independent texture layers in grayscale or vector form. These layers are then overlaid on the base model during rendering, simulating uneven details without increasing the model's geometric complexity. A cultural relic's 3D digital model is an ultra-high-precision model containing a basic geometric mesh, basic color textures, and one or more additional detail layers recording microscopic details. The terminal uses detail extraction algorithms to calculate the microscopic geometric changes on the surface from ultra-high-resolution photographs, generating a normal map or displacement map. During rendering, the engine reads the map and simulates uneven details by changing the lighting and shadows on the model's surface, even if the model's geometry itself is smooth.

[0131] This embodiment constructs a digital twin that faithfully reflects the current state of the cultural relic and has a photorealistic quality through 3D modeling. This model can already be used for high-definition display and preliminary visual inspection. Furthermore, by adding layers, the visual realism and geometric accuracy of the model are greatly improved without significantly increasing the computational burden on the computer. This allows the model to not only be used for macroscopic display, but also to support the fine observation and measurement of microscopic defects on the surface of the cultural relic, providing a richer source of information for subsequent defect identification.

[0132] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0133] Based on the same inventive concept, this application also provides a digital twin-based virtual restoration device for cultural relics, used to implement the aforementioned method for virtual restoration of cultural relics based on digital twin technology. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the digital twin-based virtual restoration device for cultural relics provided below can be found in the limitations of the digital twin-based virtual restoration method for cultural relics described above, and will not be repeated here.

[0134] In one exemplary embodiment, such as Figure 2 As shown, a virtual artifact restoration device 800 based on digital twin technology is provided, comprising:

[0135] The preprocessing module 801 is used to acquire the raw data of cultural relics lost overseas; and to perform standardized preprocessing on the raw data to obtain a cultural relic data package; the raw data includes high-resolution images, 3D point clouds and historical documents;

[0136] The reconstruction module 802 is used to perform three-dimensional geometric reconstruction based on the cultural relic data package to obtain a three-dimensional digital model of the cultural relic;

[0137] The identification module 803 is used to identify defects appearing in the three-dimensional digital model of the cultural relic through the image segmentation model, and obtain a list of cultural relic defects; the defects include at least one of cracks, missing parts and corrosion;

[0138] Risk module 804 is used to predict the risk of structural stability of the three-dimensional digital model of cultural relics based on the list of cultural relics defects and through mechanical simulation, so as to obtain a cultural relics risk prediction map.

[0139] The solution module 805 is used to generate virtual restoration solutions by generating restoration strategies with different focuses in parallel based on the list of cultural relic defects and the cultural relic risk prediction map.

[0140] Furthermore, risk module 804 is also used for:

[0141] The three-dimensional digital model of the cultural relic is re-divided into computational meshes to obtain a mesh model; and the mesh model is then geometrically cleaned to obtain a computational preparation model.

[0142] Based on a pre-defined knowledge base of cultural relics materials, mechanical parameters and constraints are assigned to the calculation preparation model to obtain a finite element model.

[0143] Stress distribution calculations are performed on the finite element model to obtain a stress distribution cloud map; the stress distribution cloud map is used to characterize the magnitude and distribution of stress within the entire cultural relic.

[0144] A quantitative analysis was performed by overlaying the list of cultural relic defects and stress distribution cloud maps to obtain a risk prediction map for cultural relics.

[0145] Furthermore, risk module 804 is also used for:

[0146] The element stiffness matrix is ​​calculated for each element in the finite element model to obtain a set of element stiffness matrices; and all the set of element stiffness matrices are superimposed to form the global stiffness matrix; the global stiffness matrix is ​​used to characterize the structural stiffness characteristics of the cultural relic;

[0147] Based on gravity simulation, the distributed forces on the finite element model are distributed to each node to obtain the overall load vector;

[0148] Based on the overall stiffness matrix and the overall load vector, the displacement of each node under the load is calculated to obtain the displacement vector of all nodes.

[0149] Based on the displacement vector, the stress components of each element in the element stiffness matrix are calculated; and the stress components are averaged across the nodes to obtain a stress distribution cloud map.

[0150] Furthermore, solution module 805 is also used for:

[0151] Based on preset priority principles, a basic repair plan is generated; the priority principles include security priority, historical authenticity priority, and educational priority.

[0152] Based on the cultural relic risk prediction map, a quantitative analysis was conducted on each dimension of the basic restoration plan to obtain a quantitative score table; the dimensions include at least one of the following: structural safety improvement degree, minimum intervention compliance degree, historical information preservation degree, reversibility and aesthetic harmony.

[0153] Based on the repair objectives, weights are assigned to the dimensions to obtain an indicator weight table.

[0154] Based on the indicator weight table, the quantitative score table is weighted and summed using the following formula to obtain the score list; the score list includes the comprehensive score corresponding to each basic repair plan:

[0155]

[0156] Where S is the overall score, i is the dimension index, n is the total number of dimensions, and w i s is the weighting coefficient. i The score for the i-th dimension;

[0157] The basic repair scheme with the highest overall score is selected as the virtual repair scheme.

[0158] Furthermore, solution module 805 is also used for:

[0159] Based on the principle of prioritizing educational value and a quantitative list of cultural relic defects, a complete morphological hypothesis for the three-dimensional digital model of the cultural relic is generated through a pre-trained generative AI model, resulting in a set of morphological prediction hypotheses. The generative AI model is trained using data contained in an art style database.

[0160] Based on the set of morphological inference hypotheses, the three-dimensional digital model of the cultural relic is virtually completed to obtain the completed three-dimensional model.

[0161] Based on an art style database, color rendering is performed on the completed 3D model to obtain the initial 3D model for manufacturing.

[0162] By integrating the initial 3D model of manufacturing and the virtual repair operation sequence, a basic repair plan is obtained.

[0163] Furthermore, the device also includes an animation module for:

[0164] Based on a real-time rendering engine, the 3D model in the virtual restoration scheme is rendered to obtain the restored cultural relic model.

[0165] The virtual repair operation sequence is animated to obtain a virtual repair animation;

[0166] By integrating the restored artifact model, the artifact's 3D digital model, and the virtual restoration animation, a visualized result of the artifact restoration is obtained.

[0167] Furthermore, the reconstruction module 802 is also used for:

[0168] Feature enhancement processing is performed on the 3D point cloud in the cultural relic data package to obtain the baseline 3D point cloud;

[0169] Based on the point cloud registration algorithm, the reference 3D point cloud and the high-resolution image are spatially aligned to obtain the aligned 3D point cloud;

[0170] A 3D network model is generated based on aligned 3D point clouds using a surface reconstruction algorithm.

[0171] Based on the texture mapping algorithm, the color and texture information corresponding to the high-definition image is projected onto the three-dimensional network model to obtain a three-dimensional digital model of the current state of the cultural relics.

[0172] The detailed information from the high-resolution images is added as an additional layer to the existing 3D digital model of the cultural relic, resulting in a 3D digital model of the cultural relic.

[0173] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of a virtual restoration method for cultural relics based on digital twin technology as described above.

[0174] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0175] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0176] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A method for virtual restoration of cultural relics based on digital twin technology, characterized in that, The method includes: The original data of cultural relics lost overseas is obtained; and the original data is standardized and preprocessed to obtain a cultural relic data package; the original data includes high-resolution images, 3D point clouds and historical documents; Based on the cultural relic data package, three-dimensional geometric reconstruction is performed to obtain a three-dimensional digital model of the cultural relic; Using an image segmentation model, defects appearing in the three-dimensional digital model of the cultural relic are identified to obtain a list of cultural relic defects; the defects include at least one of cracks, missing parts, and corrosion. Based on the list of cultural relic defects, the risk of structural stability of the three-dimensional digital model of the cultural relic is predicted by mechanical simulation, and a risk prediction map of the cultural relic is obtained. Based on the list of cultural relic defects and the risk prediction map of cultural relic, restoration strategies with different focuses are generated in parallel to obtain virtual restoration schemes.

2. The method according to claim 1, characterized in that, Based on the list of cultural relic defects, the risk of structural stability of the three-dimensional digital model of the cultural relic is predicted through mechanical simulation, resulting in a cultural relic risk prediction map, including: The three-dimensional digital model of the cultural relic is re-divided into computational meshes to obtain a mesh model; and the mesh model is geometrically cleaned to obtain a computational preparation model. Based on a pre-set knowledge base of cultural relics materials, mechanical parameters and constraints are assigned to the calculation preparation model to obtain a finite element model. Stress distribution calculations are performed on the finite element model to obtain a stress distribution cloud map; the stress distribution cloud map is used to characterize the magnitude and distribution of stress within the entire cultural relic. By overlaying and quantitatively analyzing the list of cultural relic defects and the stress distribution cloud map, a risk prediction map of the cultural relic is obtained.

3. The method according to claim 2, characterized in that, The process of calculating the stress distribution of the finite element model to obtain a stress distribution cloud map includes: For each element in the finite element model, the element stiffness matrix is ​​calculated to obtain a set of element stiffness matrices; and all the sets of element stiffness matrices are superimposed to form an overall stiffness matrix; the overall stiffness matrix is ​​used to characterize the structural stiffness characteristics of the cultural relic; Based on gravity simulation, the distributed forces on the finite element model are allocated to each node to obtain the overall load vector; Based on the overall stiffness matrix and the overall load vector, the displacement of each node under the load is calculated to obtain the displacement vector of all nodes. Based on the displacement vector, the stress components of each element in the element stiffness matrix are calculated; and the stress components are averaged over the nodes to obtain the stress distribution cloud map.

4. The method according to claim 1, characterized in that, Based on the list of cultural relic defects and the risk prediction map of cultural relics, the virtual restoration plan is obtained by generating restoration strategies with different focuses in parallel, including: Based on preset priority principles, a basic repair plan is generated; the priority principles include security priority, historical authenticity priority, and educational priority. Based on the cultural relic risk prediction map, a quantitative analysis is performed on each dimension of the basic restoration plan to obtain a quantitative score table; the dimensions include at least one of structural safety improvement, minimum intervention compliance, historical information retention, reversibility, and aesthetic harmony. Based on the repair objectives, weights are assigned to the dimensions to obtain an indicator weight table. Based on the aforementioned indicator weight table, the quantitative score table is weighted and summed using the following formula to obtain a score list; the score list includes the comprehensive score corresponding to each of the aforementioned basic repair solutions: Where S is the overall score, i is the dimension index, n is the total number of dimensions, and w i s is the weighting coefficient. i The score for the i-th dimension; The basic repair scheme with the highest overall score is determined as the virtual repair scheme.

5. The method according to claim 4, characterized in that, The basic repair plan generated based on a preset priority principle includes: Based on the aforementioned educational priority principle and the aforementioned quantitative list of cultural relic defects, a complete morphological hypothesis of the three-dimensional digital model of the cultural relic is generated through a pre-trained generative AI model, resulting in a set of morphological prediction hypotheses; the generative AI model is trained using data contained in an art style database; Based on the set of morphological inference hypotheses, the three-dimensional digital model of the cultural relic is virtually completed to obtain a completed three-dimensional model. Based on the art style database, the completed 3D model is color-rendered to obtain the initial 3D model for manufacturing. By integrating the initial 3D model of manufacturing and the virtual repair operation sequence, the basic repair scheme is obtained.

6. The method according to claim 5, characterized in that, After determining the basic repair scheme with the highest comprehensive score as the virtual repair scheme, the process further includes: Based on a real-time rendering engine, the three-dimensional model in the virtual restoration scheme is rendered to obtain the restored cultural relic model. The virtual repair operation sequence is animated to obtain a virtual repair animation; By integrating the restored artifact model, the artifact's three-dimensional digital model, and the virtual restoration animation, a visualized artifact restoration result is obtained.

7. The method according to claim 1, characterized in that, The process of performing three-dimensional geometric reconstruction based on the cultural relic data package to obtain a three-dimensional digital model of the cultural relic includes: The 3D point cloud in the cultural relic data package is subjected to feature enhancement processing to obtain a baseline 3D point cloud. Based on the point cloud registration algorithm, the reference 3D point cloud and the high-definition image are spatially aligned to obtain an aligned 3D point cloud. Based on the aligned 3D point cloud, a 3D network model is generated using a surface reconstruction algorithm; Based on the texture mapping algorithm, the color and texture information corresponding to the high-definition image is projected onto the three-dimensional network model to obtain a three-dimensional digital model of the current status of the cultural relics. The detailed information in the high-definition image is added as an additional layer to the current three-dimensional digital model of the cultural relic to obtain the three-dimensional digital model of the cultural relic.

8. A virtual restoration device for cultural relics based on digital twin technology, characterized in that, The device includes: The preprocessing module is used to acquire the raw data of cultural relics lost overseas; and to perform standardized preprocessing on the raw data to obtain a cultural relic data package; the raw data includes high-resolution images, 3D point clouds and historical documents; The reconstruction module is used to perform three-dimensional geometric reconstruction based on the cultural relic data package to obtain a three-dimensional digital model of the cultural relic; The identification module is used to identify defects appearing in the three-dimensional digital model of the cultural relic through an image segmentation model, and obtain a list of cultural relic defects; the defects include at least one of cracks, missing parts and corrosion; The risk module is used to predict the structural stability risk of the three-dimensional digital model of the cultural relic based on the list of cultural relic defects through mechanical simulation, and to obtain a cultural relic risk prediction map. The solution module is used to generate virtual restoration solutions in parallel based on the list of cultural relic defects and the risk prediction map of cultural relic.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images