A modeling method and system for ancient building restoration

Abstract

The application provides a modeling method and system for ancient building repair, and belongs to the field of building repair modeling. The method comprises the following steps: S1, collecting three-dimensional point cloud data of an ancient building to be repaired, and performing preprocessing to record damage state and deformation characteristics; S2, matching standardized data with an ancient building parameterized component library and performing attitude correction; S3, dividing an initial repair model into a plurality of sub-regions, calculating a shape difference measurement value of each sub-region, and dividing the initial repair model; S4, using the ancient building parameterized component library to register a completely reconstructed region and a partially repaired region to obtain a general part model, and using a fitting algorithm to match a region to be kept as original to obtain a detail part model; and S5, fusing the general part model and the detail part model to obtain a three-dimensional model of the ancient building. The application realizes accurate digital modeling of the ancient building, guarantees the overall structural standardization of the ancient building, and can retain the characteristics and details of the original building.

Description

Technical Field

[0001] This invention relates to the field of architectural restoration modeling technology, and in particular to a modeling method and system for the restoration of ancient buildings. Background Technology

[0002] Ancient architecture is an important material carrier of human civilization. Its systems are not only abundant in quantity but also possess unique characteristics in timber frame construction and mortise and tenon joinery, forming intangible cultural heritage. For a long time, traditional ancient building restoration relied primarily on experienced craftsmen who determined restoration plans through visual inspection, manual surveying, and experience. While this approach ensured the smooth progress of restoration work to some extent, it also faced many challenges, such as the limited accuracy of traditional surveying methods, making it difficult to accurately capture minute deformations and damage to ancient buildings; the lack of scientific basis for restoration plans; secondary damage to original components during restoration; and discrepancies between the restored building and its original appearance.

[0003] In recent years, the application of modeling technology in the field of ancient building restoration has become an important development trend. By establishing digital models of ancient buildings, comprehensive analysis, research, and planning can be carried out in a virtual environment, providing scientific basis and technical support for restoration work. However, existing modeling technologies struggle to balance overall structure and local details. They either oversimplify the overall structure while neglecting local details, or overemphasize local precision, resulting in an overly large and complex overall model. Furthermore, existing technologies cannot effectively handle the irregular deformations and defects commonly found in ancient buildings. This leads to either forcibly mapping the actual structure onto an ideal model, causing distortion, or abandoning parametric structures, making the model difficult to modify and analyze.

[0004] Therefore, finding a method that can balance the overall structure and detailed expression of ancient architecture while improving data processing efficiency is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] This invention provides a modeling method and system for the restoration of ancient buildings, which solves the defects of existing technologies in the restoration of ancient buildings, such as poor results and easy neglect of details. It realizes accurate digital modeling of ancient buildings, and can preserve the original features and details of the buildings while ensuring the overall structural standardization of the ancient buildings.

[0006] This invention provides a modeling method for the restoration of ancient buildings, comprising the following steps:

[0007] S1. Collect three-dimensional point cloud data of the ancient building to be restored, preprocess the three-dimensional point cloud data to obtain standardized data, and record the damage state and deformation characteristics of the ancient building to be restored.

[0008] S2. Obtain historical data of ancient buildings to construct a parametric component library of ancient buildings. Match and correct the orientation of the standardized data with the parametric component library of ancient buildings to obtain an initial restoration model of the ideal state of the ancient buildings before damage.

[0009] S3. Divide the initial restoration model into several sub-regions according to the restoration requirements. Calculate the morphological difference metric value of each sub-region based on the damage state and deformation characteristics of the ancient building to be restored. Based on the morphological difference metric value, divide the initial restoration model into a fully reconstructed region, a partially restored region, and a region that retains its original state.

[0010] S4. The conventional part model is obtained by registering the fully reconstructed area and the partially restored area using the parametric component library of ancient buildings, and the detailed part model is obtained by matching the original preserved area using a fitting algorithm.

[0011] S5. The conventional part model and the detailed part model are fused in a unified coordinate system to obtain a three-dimensional model of the ancient building, and a digital restoration archive is established based on the three-dimensional model of the ancient building and the damage status and deformation characteristics of the ancient building to be restored.

[0012] According to the present invention, a modeling method for the restoration of ancient buildings is provided, wherein the parametric component library of ancient buildings includes a basic construction model set, a mortise and tenon connection relationship library, a historical style database, a material and process information database, and a dynamic update mechanism.

[0013] According to the modeling method for ancient building restoration provided by the present invention, the step of matching and correcting the standardized data with a parametric component library of ancient buildings specifically includes:

[0014] The key geometric features of the standardized data are extracted and aligned with the geometric reference points of the ancient building parametric component library to generate preliminary matching results; the key geometric feature points include structural feature points, curvature feature points, edge feature points and planar feature points;

[0015] Based on the preliminary matching results, relative relationship constraints between components are established by combining the constraints of ancient building construction principles, including flatness, verticality and symmetry.

[0016] Based on the relative relationship constraints between components and the preliminary matching results, the iterative nearest point algorithm is used to process the standardized data.

[0017] According to the modeling method for ancient building restoration provided by the present invention, the iterative nearest point algorithm specifically includes:

[0018] Weights are assigned to feature points of different categories for key geometric features. Based on the weights of feature points of different categories, a registration objective function for fusion structural constraints is constructed. The registration objective function for fusion structural constraints includes a point-to-point distance term, constraints based on the principles of ancient building construction, and regularization constraints. The point-to-point distance term is the sum of the distances between points in the standardized data and the parametric construction models in the parametric component library of ancient buildings.

[0019] The preliminary matching results are coarsely registered with the main components of the ancient building parametric component library to determine the basic framework of the standardized data.

[0020] Based on the relative relationship constraints between components, medium-precision and fine-precision registration were performed with the component categories and component structures of the ancient building parametric component library.

[0021] According to the modeling method for ancient building restoration provided by the present invention, the step of calculating the morphological difference metric value of each sub-region based on the damage state and deformation characteristics of the ancient building to be restored specifically includes:

[0022] To establish characteristic description models and assessment benchmarks for different types of damage to ancient buildings to be restored, the damage status includes surface weathering, component fracture and structural deformation, and the deformation characteristics include component bending, warping and traces of historical repair.

[0023] Adaptive downsampling is performed on the standardized data for each sub-region to generate a low-resolution point cloud representation;

[0024] For each point in each sub-region, a neighborhood set is determined, where the neighborhood size is dynamically adjusted based on the component type and local normalized data density.

[0025] Calculate the position difference vector between each point in the sub-region and all points in the neighborhood point set, and calculate the mean of all position difference vectors in the neighborhood;

[0026] The morphological difference metric for each point in the sub-region is calculated based on the position difference vector and the mean of the position difference vectors. The calculation formula is as follows:

[0027] ;

[0028] in, Point The morphological difference metric at the location, Represents any point in standardized data. Point The neighborhood point set, Describing the L2 norm, Representing the neighborhood any point in, This represents the mean of the vectors between points in the neighborhood. Point The number of neighborhood point clusters, This represents the i-th point in the sub-region. Represents points other than i in the subregion.

[0029] According to the present invention, a modeling method for the restoration of ancient buildings, wherein the sub-region is divided into a fully reconstructed region, a partially restored region, and a region preserving the original state based on morphological difference metrics, specifically includes:

[0030] Calculate the global distribution of morphological difference metrics for all sub-regions, and determine the division thresholds T1 and T2, where T1 is the standard for minor damage to ancient buildings and T2 is the standard for severe damage to ancient buildings.

[0031] Sub-regions whose average morphological difference metric value is lower than T1 are classified as areas to be preserved in their original state and only undergo protection and reinforcement treatment.

[0032] Sub-regions whose average morphological difference metric value is higher than T2 are classified as fully reconstructed regions, which need to be reconstructed in accordance with traditional ancient building styles and local architectural styles.

[0033] The spatial distribution characteristics of the morphological difference metric values ​​in the sub-regions whose average morphological difference metric values ​​are between T1 and T2 are analyzed. The spatial distribution characteristics analysis includes calculating the local clustering coefficient, geometric feature vector, and positional change rate of the morphological difference metric values ​​in the sub-regions.

[0034] According to the modeling method for ancient building restoration provided by the present invention, the method of registering the fully reconstructed area and the partially restored area using a parametric component library of ancient buildings to obtain a conventional partial model specifically includes:

[0035] Extract the principal axis direction and key geometric parameters of the standardized data of the fully reconstructed area and the partially restored area, and match them to the corresponding models in the ancient building parametric component library to obtain the initial conventional model;

[0036] Adjust the core parameters of the initial conventional model, and optimize the position and orientation of the components in the initial conventional model by combining the spatial relationships and connection constraints between the ancient architectural components; the core parameters include the basic shape, size and structural characteristics of the components.

[0037] The registration results are iterated step by step until the position and orientation of the components in the initial conventional model reach the preset standard, thus obtaining the conventional partial model.

[0038] According to the modeling method for the restoration of ancient buildings provided by the present invention, the step of matching the original preserved area with a fitting algorithm to obtain the detailed model specifically includes:

[0039] Extract basic template models of the same type as the original preserved area from the parametric component library of ancient buildings, and perform Gaussian regression on the standardized data of the original preserved area to generate surface weathering characteristics of components of different materials;

[0040] The skeleton of the basic template model is extracted, the mortise and tenon joints of the components, the center line of the load-bearing components and the key points of mechanical transmission are identified, and a deformation influence weighting system from the structural skeleton to the surface of the components is established.

[0041] Based on the aforementioned deformation influence weighting system and the surface weathering characteristics of components made of different materials, the optimal repair parameters for the traditional component skeleton are calculated. During the calculation of the repair parameters, the structural rules of ancient buildings are applied to constrain the process, resulting in a repair model. The repair parameters include the beam frame curvature, column tilt angle, and the amount of deformation of the bracket set.

[0042] Based on the deformed component skeleton and deformation influence weight system, control elements of key nodes are set, and radial basis functions that conform to the deformation characteristics of traditional materials are used for surface repair interpolation. Historical and cultural information of ancient buildings is extracted from standardized data of the original preservation area and applied to the surface of the repair model through normal replacement and texture mapping.

[0043] According to the modeling method for ancient building restoration provided by the present invention, the conventional partial model and the detailed partial model are fused in a unified coordinate system, specifically including:

[0044] Analyze the boundary features between the conventional part model and the detailed part model, establish the spatial correspondence between different model regions, and set a transition zone in the boundary region between the conventional part model and the detailed part model;

[0045] Within the transition zone, a distance-weighted fusion algorithm is used for smooth transition processing. By calculating the spatial distance between each point in the transition zone and the regular part model and the detail part model, a weight function that changes smoothly with distance is generated, and the geometric features and surface properties of the regular part model and the detail part model are weighted and fused according to the weight function.

[0046] This invention also provides a modeling system for the restoration of ancient buildings, implementing the modeling method described above, including:

[0047] The data acquisition and processing module is used to acquire three-dimensional point cloud data of the ancient building to be restored, preprocess the three-dimensional point cloud data to obtain standardized data, and record the damage state and deformation characteristics of the ancient building to be restored.

[0048] The initial model building module is used to acquire historical data of ancient buildings to build a parametric component library of ancient buildings. The standardized data is matched with the parametric component library of ancient buildings and the posture is corrected to obtain an initial restoration model of the ideal state of the ancient buildings before damage.

[0049] The region division module is used to divide the initial restoration model into several sub-regions according to the restoration requirements. It calculates the morphological difference metric value of each sub-region based on the damage state and deformation characteristics of the ancient building to be restored. Based on the morphological difference metric value, the initial restoration model is divided into a fully reconstructed region, a partially restored region, and a region that retains its original state.

[0050] The registration module is used to register the fully reconstructed area and the partially restored area using the parametric component library of ancient buildings to obtain the general part model, and to use the fitting algorithm to match the original preserved area to obtain the detailed part model;

[0051] The model fusion module is used to fuse the regular part model and the detailed part model in a unified coordinate system to obtain the three-dimensional model of the ancient building, and to establish a digital restoration archive based on the three-dimensional model of the ancient building and the damage status and deformation characteristics of the ancient building to be restored.

[0052] This invention provides a modeling method and system for the restoration of ancient buildings. By processing the three-dimensional point cloud data of the component to be restored and dividing the area based on the morphological difference metric, it achieves accurate digital modeling of ancient buildings. While ensuring the overall structural standardization of the ancient buildings, it can preserve the original features and details of the buildings, significantly improving the accuracy, scientific nature and efficiency of the restoration of ancient buildings. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0054] Figure 1 This is a flowchart of a modeling method for the restoration of ancient buildings provided by the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0056] like Figure 1 As shown, this invention provides a modeling method for the restoration of ancient buildings, comprising the following steps:

[0057] S1. Collect three-dimensional point cloud data of the ancient building to be restored, preprocess the three-dimensional point cloud data to obtain standardized data, and record the damage state and deformation characteristics of the ancient building to be restored.

[0058] Understandably, 3D point cloud data of ancient buildings can be detected using instruments such as 3D laser scanners, handheld mobile laser scanner cameras, and thermal infrared imagers. In the actual data collection process for ancient buildings, multiple devices can be combined, and a detailed collection plan can be formulated based on the specific conditions of the ancient building. For example, for large palace complexes, drones may be used first for overall aerial photography and photogrammetry to obtain the overall layout and external form of the complex; ground-based laser scanners may be used to scan the main buildings with medium precision to obtain their basic geometric shapes; and high-precision structured light scanners may be used for detailed local scanning of exquisite carvings, paintings, and other areas. This invention does not impose specific limitations on the method of acquiring 3D point cloud data.

[0059] In one embodiment of the present invention, preprocessing includes data cleaning and noise removal, multi-site data registration, coordinate normalization, and orientation correction.

[0060] Understandably, since ancient buildings often have intricate carvings and weathering marks on their surfaces, it is necessary to use data cleaning and noise removal to distinguish between real surface details and noise, avoid over-smoothing and losing important historical information, and register the collected 3D point cloud data to a unified coordinate system, then transform it to a standard coordinate system to establish a local coordinate system with the main building as a reference, correct the main axis direction of the building to align it with the coordinate axis, and perform orientation correction according to the orientation rules of ancient buildings (such as facing south).

[0061] S2. Obtain historical data of ancient buildings to construct a parametric component library of ancient buildings. Match and correct the standardized data with the parametric component library of ancient buildings to obtain an initial restoration model of the ideal state of the ancient buildings before damage. The parametric component library of ancient buildings includes a basic construction model set, a mortise and tenon connection relationship library, a historical style database, a material and craftsmanship information database, and a dynamic update mechanism.

[0062] Understandably, a parametric component library for ancient architecture should not only include common component types found in traditional ancient buildings, but also consider regional differences, dynastic changes, and the diversity of special techniques. Therefore, constructing a parametric component library from historical data requires extracting the basic parameter relationships and geometric features of various architectural components. It necessitates collecting component examples from different regions, eras, and types of buildings, and considering the relationships between ancient architectural components. For example, the dimensions and forms of components such as column heads and brackets, beams and purlins are interconnected. Furthermore, given the rich variations in Chinese ancient architecture across different regions and historical periods—such as the differences between northern and southern architecture, and the stylistic shifts between the Tang and Song dynasties and the Ming and Qing dynasties—the parametric component library should provide parametric models with regional and chronological labels.

[0063] In one embodiment of the present invention, the basic component model set includes various types of basic component parametric models. Each basic component model is controlled by core parameters that define its geometry. These core parameters include the basic shape, size, and structural features of the component, such as length, width, height, cross-sectional shape, and curvature.

[0064] In one embodiment of the present invention, the mortise and tenon connection relationship library is used to record the connection methods and positional relationships between various components, including parameters such as mortise and tenon type, connection position, and connection angle.

[0065] In one embodiment of the present invention, the historical style database contains stylistic feature parameters of ancient buildings from different dynasties and regions, such as roof curve equations, bracket projection ratios, and decorative patterns.

[0066] In one embodiment of the present invention, the material and process information database is used to record the material properties, manufacturing processes and common deformation characteristics of various components.

[0067] In one embodiment of the present invention, the dynamic update mechanism can automatically expand the parametric component library of ancient buildings based on newly discovered component features, including processes such as feature extraction, parametric modeling, expert verification, and establishment of associations within the library. The dynamic update mechanism specifically includes:

[0068] When the matching degree between the standardized data and the existing models in the parametric component library of ancient buildings reaches a preset threshold (usually 70%), the new component identification process is triggered.

[0069] Deep feature analysis is performed on standardized data to extract geometric, topological, and semantic features. Geometric features include main size proportions, contour shape, and surface curvature distribution. Topological features include internal structural relationships of components and hole distribution. Semantic features are inferred based on the functional attributes and positional relationships of components.

[0070] Geometric features, topological features, and semantic features are compared with existing component types in the parametric component library of ancient buildings to calculate distance similarity. For example, for a special type of dougong, it may be identified as a variant of the "dougong type" component, and the most similar basic dougong models in the parametric component library of ancient buildings can be identified as references.

[0071] Standardized data is segmented to identify the main components and key feature lines of new components, and the relationships between geometric elements are analyzed to determine key parameters and constraints between them. For example, for dougong (bracket sets), key parameters such as the number of dou, the number of layers, the distance of the jump, and the curvature are identified, as well as the proportional relationships between these parameters.

[0072] The identified key parameters and their constraints are mapped to a predefined sequence of construction operations, automatically generating a parameter-controlled 3D model, which is then added to the ancient building parametric component library.

[0073] Furthermore, experts will supplement the generated 3D model with information on the cultural background, manufacturing process characteristics, or historical evolution of the corresponding components. They will also confirm the classification position of the new components and decide whether they are variants of a certain type of component or a completely new component category.

[0074] In one embodiment of the present invention, the matching and posture correction of standardized data with the parametric component library of ancient buildings specifically includes:

[0075] The key geometric features of the standardized data are extracted and aligned with the geometric reference points of the ancient building parametric component library to generate preliminary matching results; the key geometric feature points include structural feature points, curvature feature points, edge feature points and planar feature points;

[0076] Based on the preliminary matching results, relative relationship constraints between components are established by combining the constraints of ancient building construction principles, including flatness, verticality and symmetry.

[0077] Based on the relative relationship constraints between components and the preliminary matching results, the iterative nearest point algorithm is used to process the standardized data.

[0078] Specifically, the iterative nearest point algorithm includes:

[0079] Weights are assigned to feature points of different categories for key geometric features. A registration objective function based on these weights, incorporating structural constraints, is constructed. This objective function includes a point-to-point distance term, constraints based on the principles of ancient architectural construction, and regularization constraints. The point-to-point distance term is the sum of distances between points in the standardized data and the parametric construction models in the ancient architectural parametric component library. The formula for the registration objective function is:

[0080] Etotal=Edistance+λ1×Estructure+λ2×Eregularity

[0081] Where Etotal represents the overall registration objective function, Edistance represents the point-to-point term, λ1 represents the structural constraint weight coefficient, Estructure represents the constraint of the ancient building construction principle, λ2 represents the regularization weight coefficient, and Eregularity represents the regularization constraint term.

[0082] The preliminary matching results are coarsely registered with the main components of the ancient building parametric component library to determine the basic framework of the standardized data.

[0083] Based on the relative relationship constraints between components, medium-precision and fine-precision registration were performed with the component categories and component structures of the ancient building parametric component library.

[0084] Understandably, the constraints of the principles of ancient architectural construction include constraints on the positional relationships of components (such as columns must rest on the platform, beams must rest on the column heads, and brackets must be located in specific positions between the column heads and the eaves beams), structural geometric constraints (including the verticality of columns, the horizontality of transverse components (such as beams and purlins), the specific shape of the roof curve, and the rules for the stacking of brackets), proportional constraints (the dimensions of various parts of ancient buildings usually follow specific proportional relationships, such as the modular system of "cai" and "shan", as well as the different scale regulations for different grades of buildings), and the logic of component combination (such as the matching relationship between specific types of brackets and specific types of roofs, and the traditional combination methods between different components). The constraints of the principles of ancient architectural construction are the key rules of ancient architecture, such as columns should be perpendicular to the ground, beams should be placed horizontally, and bracket layers should have a regular stacking relationship.

[0085] S3. Divide the initial restoration model into several sub-regions according to the restoration requirements. Calculate the morphological difference metric value of each sub-region based on the damage state and deformation characteristics of the ancient building to be restored. Based on the morphological difference metric value, divide the initial restoration model into a fully reconstructed region, a partially restored region, and a region that retains its original state.

[0086] This invention calculates morphological difference metrics and divides areas into fully reconstructed, partially restored, and original-preserved regions. This allows for the adoption of the most suitable treatment strategies for ancient building areas in different states, ensuring the structural correctness of severely damaged parts while preserving the original features of higher-value parts to the greatest extent possible, thus improving the scientificity and rationality of the restoration plan.

[0087] Specifically, the calculation of the morphological difference metric value for each sub-region based on the damage state and deformation characteristics of the ancient building to be restored includes:

[0088] To establish characteristic description models and assessment benchmarks for different types of damage to ancient buildings to be restored, the damage status includes surface weathering, component fracture and structural deformation, and the deformation characteristics include component bending, warping and traces of historical repair.

[0089] Adaptive downsampling is performed on the standardized data for each sub-region to generate a low-resolution point cloud representation;

[0090] For each point in each sub-region, a neighborhood set is determined, where the neighborhood size is dynamically adjusted based on the component type and local normalized data density.

[0091] Calculate the position difference vector between each point in the sub-region and all points in the neighborhood point set, and calculate the mean of all position difference vectors in the neighborhood;

[0092] The morphological difference metric for each point in the sub-region is calculated based on the position difference vector and the mean of the position difference vectors. The calculation formula is as follows:

[0093] ;

[0094] in, Point The morphological difference metric at the location, Represents any point in standardized data. Point The neighborhood point set, Describing the L2 norm, Representing the neighborhood any point in, This represents the mean of the vectors between points in the neighborhood. Point The number of neighborhood point clusters, This represents the i-th point in the sub-region. This represents the points in the subregion other than i.

[0095] Understandably, the restoration of ancient buildings requires precise identification of which areas are well-preserved, which are partially damaged, and which are completely destroyed. Traditional methods that directly compare the geometric distance between point clouds and parametric models are too simplistic and cannot reflect the structural changes of a region. Furthermore, the damage to ancient buildings often manifests as changes in local structural relationships, rather than just point position shifts, such as the warping and deformation of wooden components. These changes are characterized by an overall change in shape, but the destruction of local structural relationships. Therefore, this invention calculates the morphological difference metric value of the sub-regions of the ancient building to be restored and divides the region accordingly. This can capture changes in structural relationships, more accurately reflect the damage status and restoration needs of the ancient building, and thus guide more scientific and reasonable restoration decisions.

[0096] Furthermore, the division of the sub-region into a fully reconstructed region, a partially repaired region, and a region preserving the original state based on the morphological difference metric specifically includes:

[0097] Calculate the global distribution of morphological difference metrics for all sub-regions, and determine the division thresholds T1 and T2, where T1 is the standard for minor damage to ancient buildings and T2 is the standard for severe damage to ancient buildings.

[0098] Sub-regions whose average morphological difference metric value is lower than T1 are classified as areas to be preserved in their original state and only undergo protection and reinforcement treatment.

[0099] Sub-regions whose average morphological difference metric value is higher than T2 are classified as fully reconstructed regions, which need to be reconstructed in accordance with traditional ancient building styles and local architectural styles.

[0100] The spatial distribution characteristics of the morphological difference metric values ​​in sub-regions where the average value of the morphological difference metric values ​​is between T1 and T2 are analyzed. This spatial distribution characteristic analysis includes calculating the local clustering coefficient, geometric feature vector, and positional change rate of the morphological difference metric values ​​in the sub-regions. The local clustering coefficient reflects the spatial distribution characteristics of the damage or deformation of the ancient building to be restored; the geometric feature vector reflects whether the deformation of the ancient building to be restored originates from the design features of the components themselves or from later damage; and the positional change rate reflects the gradual nature of the deformation or damage of the components in the ancient building to be restored.

[0101] If the local clustering coefficient of the sub-region morphological difference metric is greater than the clustering threshold, then the geometric feature vector of the sub-region morphological difference metric is further calculated:

[0102] If the geometric feature vector of the morphological difference metric of a sub-region has a high degree of matching with the components in the parametric component library of ancient buildings, then the sub-region is divided into an area to be preserved in its original state.

[0103] If the geometric feature vector of the morphological difference metric of a sub-region does not match the component in the parametric component library of ancient buildings, the sub-region will be divided into a partially restored area.

[0104] If the local clustering coefficient of the morphological difference metric in the sub-region is less than the clustering threshold, then the rate of change of the morphological difference metric with spatial location is calculated:

[0105] If the rate of change of the morphological difference metric with spatial location is greater than the rate of change threshold, then the sub-region is a partially repaired region.

[0106] If the rate of change of the morphological difference metric with spatial location is less than the rate of change threshold, then the sub-region is classified as a fully reconstructed region.

[0107] It is understandable that T1 is the average value of the global morphological difference metric plus one standard deviation, and T2 is the average value of the global morphological difference metric plus two standard deviations. The thresholds T1, T2, aggregation threshold, and rate of change threshold can all be set according to the actual situation of the ancient building to be restored and the restoration requirements. This invention does not impose any specific restrictions on this.

[0108] Specifically, the calculation of the local clustering coefficient, geometric feature vector, and rate of change of position of the sub-region morphological difference metric includes:

[0109] Determine the set of points HR in the subregion whose morphological difference metric values ​​are higher than the local threshold, and calculate the spatial proximity between these points. Then the local clustering coefficient = actual number of adjacent high difference point pairs / maximum possible number of adjacent high difference point pairs, where the local threshold is the mean of the morphological difference metric values ​​of the subregion plus half a standard deviation.

[0110] Principal component analysis is performed on the point set HR to extract the first three principal direction vectors and their corresponding eigenvalue ratios. The principal curvature and changes of the surface formed by the point set HR in different directions are calculated. Shape descriptors (including geometric indices such as compactness, linearity, and flatness) are calculated. The morphological feature distribution at different scales is extracted and the morphological features at different scales are integrated into a multi-dimensional vector, denoted as the geometric feature vector.

[0111] Calculate the spatial gradient vector of the difference metric for each point within the sub-region, and calculate the average magnitude of the gradient vectors of all points within the sub-region, denoted as the rate of change of location.

[0112] S4. The conventional part model is obtained by registering the fully reconstructed area and the partially restored area using the parametric component library of ancient buildings, and the detailed part model is obtained by matching the original preserved area using a fitting algorithm.

[0113] Specifically, the registration of the fully reconstructed area and the partially restored area using a parametric component library of ancient buildings to obtain the conventional partial model includes:

[0114] Extract the principal axis direction and key geometric parameters of the standardized data of the fully reconstructed area and the partially restored area, and match them to the corresponding models in the ancient building parametric component library to obtain the initial conventional model;

[0115] Adjust the core parameters of the initial conventional model, and optimize the position and orientation of the components in the initial conventional model by combining the spatial relationships and connection constraints between the ancient architectural components; the core parameters include the basic shape, size and structural characteristics of the components.

[0116] The registration results are iterated step by step until the position and orientation of the components in the initial conventional model reach the preset standard, thus obtaining the conventional partial model.

[0117] By continuously parametrically registering components to ensure they conform to the standard form and proportions of ancient buildings, errors caused by incomplete point clouds or noise are reduced. The unique construction rules and connection constraints of ancient buildings are also taken into account, making the reconstruction model of the ancient building to be rebuilt closer to the original building in form. This ensures the standardization of ancient building components (conforming to traditional construction rules and process requirements) while retaining the unique characteristics of specific buildings through parameter adjustments.

[0118] Specifically, a concrete example will be used for illustration:

[0119] The core parameter adjustments for the initial conventional model include adjustments to the corner bracket parameters (such as adjusting the overall height of the initial conventional model, adjusting the proportions of each layer of the bracket, and adjusting the angle of the purlin) and adjustments to the eaves beam parameters (such as adjusting the cross-section of the eaves beam model, setting the center sag, and adjusting the mortise and tenon structural dimensions at the connection between the beam end and the bracket). The spatial relationships and connection constraints between ancient architectural components specifically apply the connection rules from the Qing Dynasty's *Engineering Practices and Regulations*: for example, the connection between the bracket and the eaves beam uses the "mounted bracket" method, and the mortise and tenon connections between the layers of the bracket use the "interlocking and overlapping" method. Connection constraint points include the connection points between the beam end and the bracket (3 points), the connection points between the bracket and the wall / column (4 points), and key internal nodes of the bracket (12 points), etc.

[0120] Furthermore, the process of using a fitting algorithm to match the original preserved region to obtain the detailed model specifically includes:

[0121] Extract basic template models of the same type as the original preserved area from the parametric component library of ancient buildings, and perform Gaussian regression on the standardized data of the original preserved area to generate surface weathering characteristics of components of different materials;

[0122] The skeleton of the basic template model is extracted, the mortise and tenon joints of the components, the center line of the load-bearing components and the key points of mechanical transmission are identified, and a deformation influence weight system from the structural skeleton to the component surface is established, wherein the weight value decreases smoothly as the distance from the surface point to the skeleton increases.

[0123] Based on the aforementioned deformation influence weighting system and the surface weathering characteristics of components made of different materials, the optimal repair parameters for the traditional component skeleton are calculated. During the calculation of the repair parameters, the structural rules of ancient buildings are applied to constrain the process, resulting in a repair model. The repair parameters include the beam frame curvature, column tilt angle, and the amount of deformation of the bracket set.

[0124] Based on the deformed component skeleton and deformation influence weight system, control elements corresponding to key nodes of traditional woodwork are set, and radial basis functions that conform to the deformation characteristics of traditional materials are used for surface repair interpolation. Historical and cultural information of ancient buildings is extracted from standardized data of the original preservation area and applied to the surface of the repair model through normal replacement and texture mapping.

[0125] The process of applying normal displacement and texture mapping to the surface of the restored model involves extracting cultural information such as textures, scratches, and paintings from standardized data. This extracted cultural information is then applied to the reconstructed surface through normal displacement and texture mapping, resulting in a detailed model. In the deformation influence weighting system, the weight value represents the degree to which skeletal deformation affects a point on the surface: a weight of 1 indicates that the point completely follows skeletal deformation; a weight of 0 indicates that the point is unaffected by skeletal deformation; and a weight between 0 and 1 indicates that the point is partially affected by skeletal deformation.

[0126] This invention achieves more accurate reconstruction by adopting corresponding weathering models and deformation characteristics based on the characteristics of different materials (such as different types of wood, stone, bricks, etc.). It also preserves historical and cultural information that is difficult to express directly with geometric shapes (such as paintings, engravings, and traces of use) through normal displacement and texture mapping, thus preserving the cultural value of ancient buildings while meeting the structural safety requirements of the restoration parameters.

[0127] S5. The conventional part model and the detailed part model are fused in a unified coordinate system to obtain a three-dimensional model of the ancient building. Based on the three-dimensional model of the ancient building and the damage status and deformation characteristics of the ancient building to be restored, a digital restoration archive is established. The digital restoration archive includes component number, restoration process, material selection and cultural value assessment.

[0128] Specifically, the fusion of the conventional partial model and the detailed partial model in a unified coordinate system includes:

[0129] The boundary features of the conventional part model and the detailed part model are analyzed to establish the spatial correspondence between different model regions, and a transition zone is set in the boundary region between the conventional part model and the detailed part model; the width of the transition zone is adaptively adjusted according to the dimensional characteristics and structural properties of the ancient building components.

[0130] Within the transition zone, a distance-weighted fusion algorithm is used for smooth transition processing. By calculating the spatial distance between each point in the transition zone and the regular part model and the detail part model, a weight function that changes smoothly with distance is generated, and the geometric features and surface properties of the regular part model and the detail part model are weighted and fused according to the weight function.

[0131] By setting adaptive transition zones in the boundary areas of the regular and detailed models and using a distance-weighted fusion algorithm for smooth transition processing, the connection problem between regions with different processing strategies is effectively solved, avoiding unnatural breaks or jumps at the boundaries of the model. This improves the overall integrity and visual effect of the final 3D model. At the same time, the digital restoration archives have accumulated valuable digital resources for subsequent research and conservation work, promoting the inheritance and development of ancient building conservation technology.

[0132] This invention processes the three-dimensional point cloud data of the firmware to be repaired and divides the area based on the morphological difference metric, thereby achieving precise digital modeling of ancient buildings. While ensuring the overall structural standardization of the ancient buildings, it can preserve the original features and details of the buildings, significantly improving the accuracy, scientific nature and efficiency of ancient building restoration.

[0133] This invention also provides a modeling system for the restoration of ancient buildings, implementing the modeling method described above, including:

[0134] The data acquisition and processing module is used to acquire three-dimensional point cloud data of the ancient building to be restored, preprocess the three-dimensional point cloud data to obtain standardized data, and record the damage state and deformation characteristics of the ancient building to be restored.

[0135] The initial model building module is used to acquire historical data of ancient buildings to build a parametric component library of ancient buildings. The standardized data is matched with the parametric component library of ancient buildings and the posture is corrected to obtain an initial restoration model of the ideal state of the ancient buildings before damage.

[0136] The region division module is used to divide the initial restoration model into several sub-regions according to the restoration requirements. It calculates the morphological difference metric value of each sub-region based on the damage state and deformation characteristics of the ancient building to be restored. Based on the morphological difference metric value, the initial restoration model is divided into a fully reconstructed region, a partially restored region, and a region that retains its original state.

[0137] The registration module is used to register the fully reconstructed area and the partially restored area using the parametric component library of ancient buildings to obtain the general part model, and to use the fitting algorithm to match the original preserved area to obtain the detailed part model;

[0138] The model fusion module is used to fuse the regular part model and the detailed part model in a unified coordinate system to obtain the three-dimensional model of the ancient building, and to establish a digital restoration archive based on the three-dimensional model of the ancient building and the damage status and deformation characteristics of the ancient building to be restored.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A modeling method for the restoration of ancient buildings, characterized in that, Includes the following steps: S1. Collect three-dimensional point cloud data of the ancient building to be restored, preprocess the three-dimensional point cloud data to obtain standardized data, and record the damage state and deformation characteristics of the ancient building to be restored. S2. Obtain historical data of ancient buildings to construct a parametric component library of ancient buildings. Match and correct the orientation of the standardized data with the parametric component library of ancient buildings to obtain an initial restoration model of the ideal state of the ancient buildings before damage. S3. Based on the restoration requirements, the initial restoration model is divided into several sub-regions. The morphological difference metric value of each sub-region is calculated based on the damage state and deformation characteristics of the ancient building to be restored. Based on the morphological difference metric value, the initial restoration model is divided into a fully reconstructed region, a partially restored region, and a region preserving the original state. The calculation of the morphological difference metric value for each sub-region based on the damage state and deformation characteristics of the ancient building to be restored specifically includes: To establish characteristic description models and assessment benchmarks for different types of damage to ancient buildings to be restored, the damage status includes surface weathering, component fracture and structural deformation, and the deformation characteristics include component bending, warping and traces of historical repair. Adaptive downsampling is performed on the standardized data for each sub-region to generate a low-resolution point cloud representation; For each point in each sub-region, a neighborhood set is determined, where the neighborhood size is dynamically adjusted based on the component type and local normalized data density. Calculate the position difference vector between each point in the sub-region and all points in the neighborhood point set, and calculate the mean of all position difference vectors in the neighborhood; The morphological difference metric for each point in the sub-region is calculated based on the position difference vector and the mean of the position difference vectors. The calculation formula is as follows: ； in, Point The morphological difference metric at the location, Represents any point in standardized data. Describing the L2 norm, Representing the neighborhood any point in, This represents the mean of the vectors between points in the neighborhood. Point The number of neighborhood point clusters, This represents the i-th point in the sub-region. This represents the points in the sub-region other than i; S4. The conventional part model is obtained by registering the fully reconstructed area and the partially restored area using the parametric component library of ancient buildings, and the detailed part model is obtained by matching the original preserved area using a fitting algorithm. S5. The conventional part model and the detailed part model are fused in a unified coordinate system to obtain a three-dimensional model of the ancient building, and a digital restoration archive is established based on the three-dimensional model of the ancient building and the damage status and deformation characteristics of the ancient building to be restored.

2. The modeling method for ancient building restoration according to claim 1, characterized in that, The parametric component library for ancient buildings includes a set of basic construction models, a library of mortise and tenon joint relationships, a database of historical styles, a database of material and craftsmanship information, and a dynamic update mechanism.

3. The modeling method for ancient building restoration according to claim 1, characterized in that, The process of matching and correcting the standardized data with the parametric component library of ancient buildings specifically includes: The key geometric features of the standardized data are extracted and aligned with the geometric reference points of the ancient building parametric component library to generate preliminary matching results; the key geometric feature points include structural feature points, curvature feature points, edge feature points and planar feature points; Based on the preliminary matching results, relative relationship constraints between components are established by combining the constraints of ancient building construction principles, including flatness, verticality and symmetry. Based on the relative relationship constraints between components and the preliminary matching results, the iterative nearest point algorithm is used to process the standardized data.

4. The modeling method for ancient building restoration according to claim 3, characterized in that, The iterative nearest point algorithm specifically includes: Weights are assigned to feature points of different categories for key geometric features. Based on the weights of feature points of different categories, a registration objective function for fusion structural constraints is constructed. The registration objective function for fusion structural constraints includes a point-to-point distance term, constraints based on the principles of ancient building construction, and regularization constraints. The point-to-point distance term is the sum of the distances between points in the standardized data and the parametric construction models in the parametric component library of ancient buildings. The preliminary matching results are coarsely registered with the main components of the ancient building parametric component library to determine the basic framework of the standardized data. Based on the relative relationship constraints between components, medium-precision and fine-precision registration were performed with the component categories and component structures of the ancient building parametric component library.

5. The modeling method for ancient building restoration according to claim 1, characterized in that, The division of the sub-region into a fully reconstructed region, a partially repaired region, and a region preserving the original state based on morphological difference metrics specifically includes: Calculate the global distribution of morphological difference metrics for all sub-regions, and determine the division thresholds T1 and T2, where T1 is the standard for minor damage to ancient buildings and T2 is the standard for severe damage to ancient buildings. Sub-regions whose average morphological difference metric value is lower than T1 are classified as areas to be preserved in their original state and only undergo protection and reinforcement treatment. Sub-regions whose average morphological difference metric value is higher than T2 are classified as fully reconstructed regions, which need to be reconstructed in accordance with traditional ancient building styles and local architectural styles. The spatial distribution characteristics of the morphological difference metric values ​​in the sub-regions whose average morphological difference metric values ​​are between T1 and T2 are analyzed. The spatial distribution characteristics analysis includes calculating the local clustering coefficient, geometric feature vector, and positional change rate of the morphological difference metric values ​​in the sub-regions.

6. The modeling method for ancient building restoration according to claim 1, characterized in that, The process of registering the fully reconstructed area and the partially restored area using a parametric component library of ancient buildings to obtain the conventional partial model specifically includes: Extract the principal axis direction and key geometric parameters of the standardized data of the fully reconstructed area and the partially restored area, and match them to the corresponding models in the ancient building parametric component library to obtain the initial conventional model; Adjust the core parameters of the initial conventional model, and optimize the position and orientation of the components in the initial conventional model by combining the spatial relationships and connection constraints between the ancient architectural components; the core parameters include the basic shape, size and structural characteristics of the components. The registration results are iterated step by step until the position and orientation of the components in the initial conventional model reach the preset standard, thus obtaining the conventional partial model.

7. A modeling method for the restoration of ancient buildings according to claim 6, characterized in that, The process of matching the original preserved region using a fitting algorithm to obtain the detailed model specifically includes: Extract basic template models of the same type as the original preserved area from the parametric component library of ancient buildings, and perform Gaussian regression on the standardized data of the original preserved area to generate surface weathering characteristics of components of different materials; The skeleton of the basic template model is extracted, the mortise and tenon joints of the components, the center line of the load-bearing components and the key points of mechanical transmission are identified, and a deformation influence weighting system from the structural skeleton to the surface of the components is established. Based on the aforementioned deformation influence weighting system and the surface weathering characteristics of components made of different materials, the optimal repair parameters for the traditional component skeleton are calculated. During the calculation of the repair parameters, the structural rules of ancient buildings are applied to constrain the process, resulting in a repair model. The repair parameters include the beam frame curvature, column tilt angle, and the amount of deformation of the bracket set. Based on the deformed component skeleton and deformation influence weight system, control elements of key nodes are set, and radial basis functions that conform to the deformation characteristics of traditional materials are used for surface repair interpolation. Historical and cultural information of ancient buildings is extracted from standardized data of the original preservation area and applied to the surface of the repair model through normal replacement and texture mapping.

8. The modeling method for ancient building restoration according to claim 1, characterized in that, The conventional and detailed models are fused in a unified coordinate system, specifically including: Analyze the boundary features between the conventional part model and the detailed part model, establish the spatial correspondence between different model regions, and set a transition zone in the boundary region between the conventional part model and the detailed part model; Within the transition zone, a distance-weighted fusion algorithm is used for smooth transition processing. By calculating the spatial distance between each point in the transition zone and the regular part model and the detail part model, a weight function that changes smoothly with distance is generated, and the geometric features and surface properties of the regular part model and the detail part model are weighted and fused according to the weight function.

9. A modeling system for the restoration of ancient buildings, characterized in that, Implementing the modeling method as described in any one of claims 1-8, comprising: The data acquisition and processing module is used to acquire three-dimensional point cloud data of the ancient building to be restored, preprocess the three-dimensional point cloud data to obtain standardized data, and record the damage state and deformation characteristics of the ancient building to be restored. The initial model building module is used to acquire historical data of ancient buildings to build a parametric component library of ancient buildings. The standardized data is matched with the parametric component library of ancient buildings and the posture is corrected to obtain an initial restoration model of the ideal state of the ancient buildings before damage. The region division module is used to divide the initial restoration model into several sub-regions according to the restoration requirements. It calculates the morphological difference metric value for each sub-region based on the damage state and deformation characteristics of the ancient building to be restored. Based on the morphological difference metric value, the initial restoration model is divided into a fully reconstructed region, a partially restored region, and a region preserving the original state. The calculation of the morphological difference metric value for each sub-region based on the damage state and deformation characteristics of the ancient building to be restored specifically includes: To establish characteristic description models and assessment benchmarks for different types of damage to ancient buildings to be restored, the damage status includes surface weathering, component fracture and structural deformation, and the deformation characteristics include component bending, warping and traces of historical repair. Adaptive downsampling is performed on the standardized data for each sub-region to generate a low-resolution point cloud representation; For each point in each sub-region, a neighborhood set is determined, where the neighborhood size is dynamically adjusted based on the component type and local normalized data density. Calculate the position difference vector between each point in the sub-region and all points in the neighborhood point set, and calculate the mean of all position difference vectors in the neighborhood; The morphological difference metric for each point in the sub-region is calculated based on the position difference vector and the mean of the position difference vectors. The calculation formula is as follows: ； in, Point The morphological difference metric at the location, Represents any point in standardized data. Describing the L2 norm, Representing the neighborhood any point in, This represents the mean of the vectors between points in the neighborhood. Point The number of neighborhood point clusters, This represents the i-th point in the sub-region. This represents the points in the sub-region other than i; The registration module is used to register the fully reconstructed area and the partially restored area using the parametric component library of ancient buildings to obtain the general part model, and to use the fitting algorithm to match the original preserved area to obtain the detailed part model; The model fusion module is used to fuse the regular part model and the detailed part model in a unified coordinate system to obtain the three-dimensional model of the ancient building, and to establish a digital restoration archive based on the three-dimensional model of the ancient building and the damage status and deformation characteristics of the ancient building to be restored.

Images