A method and system for original appearance inference reconstruction of a decorative component of a damaged historic building

Abstract

The application belongs to the technical field of historical building repair, and discloses a damaged historical building decorative component original appearance reasoning reconstruction method and system. The method comprises the following steps: S1, constructing a database of texture symbols and connection state attributes of the connection part of the historical building decorative component; S2, fitting the shape contour of the damaged part, and constructing the connection state attributes of the damaged part; S3, comparing the obtained connection state attributes with the database, obtaining the texture symbol matched with the connection state attributes, constructing the point cloud model of the matched texture symbol, registering the point cloud model with the point cloud model of the historical building decorative component to be repaired, and obtaining the complete point cloud model after repair; and S4, using the complete point cloud model to perform surface reconstruction on the historical building decorative component to be repaired, so as to realize the original appearance reasoning reconstruction of the historical building decorative component to be repaired. Through the application, the problem that the original appearance of the damaged component is difficult to reason and repair in the prior art is solved.

Description

Technical Field

[0001] This invention belongs to the technical field of historical building restoration, and more specifically, relates to a method and system for reconstructing the original appearance of decorative components of damaged historical buildings. Background Technology

[0002] Decorative components of historical buildings often suffer from long-term wear and tear due to wind, sun, and rain, resulting in surface defects such as peeling, cracking, and damage. Restoring these damaged components to their original appearance is crucial for the preservation and inheritance of historical buildings. Furthermore, the restoration process should adhere to the principle of "restoring the old as it was, maintaining its original appearance," preserving as much of the original style and cultural connotations as possible.

[0003] Existing case-based methods for reasoning and reconstructing historical building components focus primarily on the repair of load-bearing structures (such as walls and roofs). The repair process largely involves reinforcing the original structure. The few existing studies on reasoning about the original appearance of historical building components still rely on existing case studies and the experience of manual restorers, and have not yet developed into intelligent and effective technical solutions. Therefore, how to achieve the original appearance reasoning and reconstruction of damaged parts of decorative components in historical buildings, while ensuring stylistic consistency after restoration, is a key challenge in the restoration process of decorative components in historical buildings.

[0004] To this end, the present invention proposes a method for reconstructing the original appearance of decorative components of damaged historical buildings by reasoning, which, while preserving the original cultural connotation of the components to be repaired, realizes the reasoning and reconstruction of the original appearance of the damaged parts of the components to be repaired. Summary of the Invention

[0005] In response to the above-mentioned defects or improvement needs of existing technologies, this invention provides a method and system for reconstructing the original appearance of decorative components of damaged historical buildings by reasoning, thereby solving the problem that it is difficult to reason about and repair the original appearance of damaged components in existing technologies.

[0006] To achieve the above objectives, according to one aspect of the present invention, a method for reconstructing the original appearance of decorative components of damaged historical buildings is provided, the method comprising the following steps:

[0007] S1 collects complete images of decorative symbols in the decorative components of historical buildings and constructs a database of connection status attributes at the connection points between the decorative symbols and the decorative components of historical buildings;

[0008] S2 acquires images of the damaged parts of the decorative components of the historical building to be repaired, and uses these images to construct point cloud models of the decorative components of the historical building to be repaired. It then fits the shape and contour of the damaged parts to construct the connection state attributes of the damaged parts.

[0009] S3 compares the connection state attribute obtained in step S2 with the database in step S1 to obtain the decorative symbol that matches the connection state attribute, constructs the point cloud model of the matching decorative symbol, and registers the point cloud model with the point cloud model of the decorative component of the historical building to be repaired to obtain the complete point cloud model after repair.

[0010] S4 uses the complete point cloud model to reconstruct the surface of the decorative components of the historical building to be restored, thereby realizing the original appearance reasoning and reconstruction of the decorative components of the historical building to be restored.

[0011] More preferably, in step S1, the decorative symbols include mushroom patterns, ram's horn patterns, cloud patterns, heart patterns, nail patterns, and rope patterns, and the connection status attributes include the number of angles with the main connecting line, angle values, pattern intersection types, and repeating pattern symbol array angles.

[0012] More preferably, in step S2, when fitting the shape contour of the damaged area, the following steps are performed:

[0013] S21 performs preprocessing on the image of the damaged area, including filtering and binarization;

[0014] S22 uses a line detection algorithm to fit lines, thereby obtaining a set of multiple fitted lines;

[0015] S23 filters the fitted lines, removing those that do not meet the preset requirements, and obtains the required set of fitted lines.

[0016] S24 For the lines in the set of lines in step S23, extract the two endpoints of each line, perform adaptive clustering on the two endpoints of the lines to obtain the clustered line endpoints, and the line formed by connecting the clustered line endpoints is the required fitted line.

[0017] More preferably, in step S23, the filtering is performed by calculating the degree of overlap between the straight line and the edge contour of the damaged part, and removing straight lines with an overlap degree lower than a preset threshold.

[0018] More preferably, the connection status attributes of the damaged part include: the number of angles between the decorative symbols at the damaged part and the main connecting line, the angle value, the type of decorative intersection, and the angle of the repeating decorative symbol array.

[0019] More preferably, in step S3, the point cloud model of the matched ornamental symbols is constructed according to the following steps:

[0020] Obtain the two-dimensional coordinates of the matching decorative symbols in the image coordinate system from the database, including points on the edge of the decorative symbol outline and points inside the outline;

[0021] For the points inside the contour, including the points inside the upper and lower surface contour edges of the ornamentation symbols, the points inside the upper surface contour edge are all assigned the same depth value Z1, and the points inside the lower surface contour edge are all assigned the same depth value Z2, where Z1≠Z2, so as to obtain the three-dimensional coordinates of the points on the upper and lower surfaces of the ornamentation symbols; for the points on the contour edge, different depth values Z3 are assigned, where Z1<Z3<Z2, so as to obtain the points on the side surface of the ornamentation symbol; thus, the three-dimensional point cloud model of the entire ornamentation symbol is obtained.

[0022] Further preferably, in step S3, the matching adopts the method of fuzzy matching.

[0023] Further preferably, in step S4, the Poisson reconstruction method is adopted for surface reconstruction of the damaged historical building decorative component using the complete point cloud model.

[0024] According to another aspect of the present invention, there is provided a system for reasoning and reconstructing the original appearance of a damaged historical building decorative component, including a processor, and the processor is used to execute the method for reasoning and reconstructing the original appearance of the damaged historical building decorative component described above.

[0025] According to yet another aspect of the present invention, there is provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for reasoning and reconstructing the original appearance of the damaged historical building decorative component described above is implemented.

[0026] Generally speaking, compared with the prior art through the above technical solutions conceived by the present invention, the following beneficial effects are achieved:

[0027] 1. The present invention proposes a method for reasoning and reconstructing the original appearance of a damaged historical building decorative component. The reasoning and reconstruction method divides the ornamentation symbols on the basis of the existing complete historical building decorative components and establishes a database of ornamentation symbols of historical building decorative components. For the damaged decorative component to be repaired, the connection state attributes are calculated and matched to obtain similar ornamentation symbols to repair the damaged part, providing a feasible idea and method for reasoning and repairing the original appearance of the damaged historical building decorative component, and retaining the original style and cultural connotation of the component while repairing;

[0028] 2. When fitting the shape contour of the damaged part in the present invention, the endpoints of the straight lines are clustered, and then the connection state attributes of the damaged part are calculated, and similar ornamentation symbols with similar features and styles are obtained by matching in the database of ornamentation symbols of historical building decorative components. By establishing the connection state attributes of the ornamentation symbols, the geometric and style features of different ornamentation symbols in historical building decorative components can be accurately described, providing a quantitative index for style description and feature matching of ornamentation symbols, and ensuring the consistency of the style and cultural connotation of the repaired component;

[0029] 3. In this invention, when constructing the point cloud model of the decorative symbol, the point sets of the upper and lower surfaces and the point sets of the side surfaces are obtained according to its image data, so as to realize the reasoning and reconstruction of the decorative symbol in the damaged part. This method can quickly and effectively obtain the point cloud model of the decorative symbol based only on the image data of the decorative symbol in the database. It does not rely on the real point cloud model of the decorative symbol during the repair process, thus reducing the cost of acquiring the three-dimensional data of the decorative symbol before repair.

[0030] 4. Compared with existing restoration methods, this invention does not rely on past restoration cases or the experience guidance of manual restorers. It can realize the original appearance reasoning and reconstruction of the damaged parts of the decorative components of historical buildings. At the same time, by applying the decorative symbols in the decorative components of historical buildings to the restoration of similar components, it also ensures that the restored decorative components of historical buildings can be consistent with the original components in terms of style and cultural connotation, thus achieving "restoring the old as before and maintaining the original appearance". Attached Figure Description

[0031] Figure 1 This is a flowchart of the method for reconstructing the original appearance of decorative components of damaged historical buildings according to an embodiment of the present invention;

[0032] Figure 2 This is a schematic diagram of the complete method for classifying decorative symbols of historical building decorative components and the connection status attributes of decorative joints provided in the embodiments of the present invention;

[0033] Figure 3 This is a schematic diagram of the fitting and linear clustering results of the main connecting lines at the damaged part of the component to be repaired, provided in an embodiment of the present invention; wherein: (a) is a schematic diagram of the linear fitting results at the damaged part of the component to be repaired; (b) is a schematic diagram of the linear endpoint extraction results at the damaged part of the component to be repaired; (c) is a schematic diagram of the linear clustering results at the damaged part of the component to be repaired; and (d) is a schematic diagram of the linear endpoint clustering results at the damaged part of the component to be repaired.

[0034] Figure 4 This is a schematic diagram of similar pattern symbol matching results provided in an embodiment of the present invention, wherein: (a) is a similar pattern symbol obtained by matching based on the connection status attribute at the damaged part; (b) is the connection status attribute of the matched similar pattern symbol;

[0035] Figure 5 This is a schematic diagram of the result after registering the damaged part of the component to be repaired with the point cloud model of the matching similar decorative symbols provided in the embodiment of the present invention; wherein: (a) is the complete point cloud model of the component after repair based on similar decorative symbols 1; (b) is the complete point cloud model of the component after repair based on similar decorative symbols 2;

[0036] Figure 6This is a schematic diagram of a complete component model after the decorative components of a damaged historical building have been repaired according to the original appearance reasoning reconstruction method of the decorative components of a damaged historical building provided in this embodiment of the invention; wherein: (a) is a schematic diagram of a complete component model after repair based on similar decorative symbol 1; (b) is a schematic diagram of a complete component model after repair based on similar decorative symbol 2; Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0038] This invention provides a method for reconstructing the original appearance of decorative components of damaged historical buildings, such as... Figure 1 As shown, the reasoning and reconstruction method includes the following steps:

[0039] S1 identifies decorative symbols from existing images of complete historical building decorative components and obtains the connection status attributes of each symbol to the component. A database of decorative symbols for historical building decorative components is then established based on the image data of the symbols and the connection status attributes at the connection points.

[0040] S2 acquires an image of the damaged part of the decorative component to be repaired, fits the shape contour of the damaged part through a line detection algorithm, removes misfitted lines, and clusters the fitted line clusters through an adaptive clustering algorithm to obtain the main fitted lines and calculates the connection state attributes of the damaged part.

[0041] S3 performs fuzzy matching in the established pattern symbol database based on the connection status attributes of the damaged part of the decorative component to be repaired, and obtains pattern symbols with similar connection status attributes.

[0042] The decorative component to be repaired is reconstructed in three dimensions to obtain its point cloud model. Based on the edge contours of the similar decorative symbols obtained by matching, its point cloud model is established. The component to be repaired is registered with the point cloud model of the similar decorative symbols to obtain the point cloud model of the repaired decorative component.

[0043] S4 performs surface reconstruction and texture mapping on the point cloud model of the repaired decorative components to obtain the solid model of the repaired decorative components, thus completing the repair work of the decorative components of the damaged historical building.

[0044] Furthermore, in S1, the decorative symbols of the historical building's decorative components include patterns such as mushroom patterns, ram's horn patterns, cloud patterns, heart patterns, nail patterns, and rope patterns. Based on the above patterns, a series of decorative symbols can be obtained by dividing the existing complete images of the historical building's decorative components. The connection status attributes of the decorative symbols and the component outline edges include the number of angles with the main connecting line, angle values, intersection types, and the angle of the repeating decorative symbol array. The connection status attributes of each decorative symbol can be manually recorded based on its image information. The connection status attributes and image data of each decorative symbol are used as the main fields in a data table to establish a decorative symbol database.

[0045] Furthermore, in S2, the main fitted straight line is the connection line between the damaged part and the central pattern; the connection status attributes include the number of angles between the pattern at the damaged part and the main connecting line, the angle value, the intersection type at the pattern connection, and the angle of the repeating pattern symbol array; when fitting the main straight line, the image data of the damaged part of the component to be repaired is first extracted and preprocessed by filtering, binarization, etc., and the set of straight lines L can be obtained by using the straight line detection algorithm.

[0046] For the detected set of straight lines L, misfitted lines are filtered by combining the binary image of the damaged area. Specifically, the straight line L is calculated. i The degree of overlap P with the edge contour of the damaged area i P i m is the number of overlapping pixels i With line L i The number of pixels occupied by itself n i The ratio of the two lines. Set a threshold t, and discard lines with an overlap ratio lower than t as misfitted lines.

[0047] Furthermore, calculate the endpoint distribution set e1={(x1,y1),...,(x i ,y i ),...,(x m ,y m )},e2={(x1,y1),...,(x i ,y i ),...,(x m ,y m )}, e 1i and e 2i They represent lines L respectively i The two endpoints. Using an adaptive clustering algorithm to cluster sets e1 and e2, we can obtain the set... e 1i ′ and e 2i'' represents the two endpoints of line i after clustering. Based on the lines obtained from clustering, the number of angles and their values ​​between each line can be calculated.

[0048] Furthermore, in S3, the fuzzy matching condition is the connection status attribute of the damaged part of the component, and the pattern symbol with the closest connection status attribute is matched in the database.

[0049] Furthermore, in S3, images of the decorative component to be repaired are first captured from different camera perspectives. The point cloud model of the component is then obtained using the SFM (Structure From Motion) algorithm. The point cloud model of the decorative symbol is established based on the edge contour of the symbol; specifically, the two-dimensional coordinates (x, y, x) of the symbol in the image coordinate system can be obtained from the symbol image. i ,y i For points inside the edge contour in the image, given depths z1 = h1 and z2 = h2, the point set p1 = {(x1, y1, h1), ..., (x...} on the upper and lower surfaces of the decorative symbol can be obtained. i ,y i ,h1),...,(x n ,y n p2 = {(x1,y1,h2),..., ... i ,y i ,h2),...,(x n ,y n For points on the image edge contour, sampling between depths z1 and z2 yields the point set p3 = {(x1, y1, z3), ..., (x, h2)}, which represents the side of the decorative symbol. i ,y i ,z3),...,(x n ,y n Given the point sets p1, p2, and p3, we can construct a point cloud model of the decorative symbols.

[0050] The component to be repaired is registered with a point cloud model of a similar decorative symbol at the damaged area of ​​the component to obtain a complete point cloud model of the repaired component.

[0051] Furthermore, in S4, the surface of the point cloud model of the repaired decorative component is reconstructed using the Poisson reconstruction method, and the texture of the repaired decorative component model is consistent with that before the repair.

[0052] like Figure 2As shown, the images of existing complete historical building decorative components are divided into decorative symbols. The components in the image can be divided into two different styles of ram's horn patterns. Based on their image information, the following information can be obtained in Table 1: the angle between both ram's horn patterns and the main connecting line is 1, with an angle value of approximately 90 degrees. The intersection type is rigid connection. Both ram's horn patterns are circular array elements with an array angle of 90 degrees. The connection status attributes of the two ram's horn patterns and the corresponding image data correspond to the data table fields in the established decorative symbol database, and can be added to the database as decorative symbols. Figure 2 As shown, the main connecting line refers to the outline line that is directly connected to the decorative symbol.

[0053] Table 1 Connection Status Attributes

[0054]

[0055] like Figure 3 As shown in (a), the figure shows the fitting result of the connecting line between the damaged part and the component to be repaired. The endpoints of the fitted straight line cluster are extracted as cluster objects, and the main contour line of the damaged part of the component to be repaired is fitted with a straight line. The enlarged area in the figure is the approximate straight line cluster obtained by fitting. Figure 3 In step (b), the endpoints of each line in the fitted line cluster are extracted. The endpoints of each line are divided into two endpoint clusters. The figure shows the relative positional relationship of each line and its endpoints in the two-dimensional coordinate system. The endpoints marked by blue and red circles belong to two different endpoint clusters. Figure 3 Figure (c) shows the clustering results of the fitted straight lines of the damaged area. The magnified area in the figure is the main contour line obtained after clustering. Figure 3 In step (d), adaptive clustering is performed on the two endpoint clusters respectively. Connecting the endpoints in the two endpoint clusters one by one after clustering yields the straight lines after clustering in the figure. Based on the clustering, it can be calculated that there are two angles between the connecting lines between the damaged part and the component to be repaired, one of which is less than 90° and the other is greater than 90°. At the same time, the pattern at the damaged part is an array element, and the array angle is 90° as can be seen from the image.

[0056] like Figure 4 As shown in (a), based on the connection status attributes of the damaged part of the component to be repaired, similar pattern symbols 1 and 2 with similar connection status attributes can be obtained by querying the established pattern database. Figure 4 In (b), the connection status attributes of the similar decorative symbols obtained by matching are similar to the features of the damaged part of the component to be repaired in terms of geometric features such as the number of included angles, included angles, and array angles.

[0057] like Figure 5As shown, 3D reconstruction of the existing component to be repaired yields a point cloud model of the damaged component. Based on image data of similar decorative symbols, point sets of its upper and lower surfaces and sides are obtained by specifying a depth. These point sets are then combined to form a point cloud model of the similar decorative symbols. By stitching the point cloud model of the similar decorative symbols onto the damaged area of ​​the point cloud model of the damaged component, a complete point cloud model of the repaired component can be obtained. Figure 5 In (a), the point cloud at the selected area is the point cloud model of similar decorative symbol 1, and the rest is the point cloud model of the damaged component. The two are spliced ​​together at the damaged area to obtain the complete point cloud model. Figure 5 As shown in (b), similarly, the point cloud at the selected area is the point cloud model of the similar decorative symbol 2, and the rest is the point cloud model of the damaged component. The two are spliced ​​together at the damaged part to obtain the complete point cloud model.

[0058] like Figure 6 As shown, based on the registered point cloud models of the two complete components, the Poisson reconstruction method is used to reconstruct the surface of the point cloud models, thereby obtaining the solid model of the repaired complete component. Furthermore, texture mapping is applied to the surface-reconstructed complete component solid model to ensure that the texture of the repaired component is consistent with that before repair. After surface reconstruction and texture mapping, the final inference and reconstruction results of the component to be repaired can be obtained, and the repaired component maintains stylistic consistency with the original component. Figure 6 (a) is a schematic diagram of the complete component solid model after restoration based on the decorative symbol 1. Figure 6 (b) is a schematic diagram of the complete component solid model after repair based on similar decorative symbols 2.

[0059] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for reconstructing the original appearance of decorative components of damaged historical buildings, characterized in that, The method includes the following steps: S1 Collect complete images of the decorative patterns on historical building decorative components, and construct a database of the connection state attributes at the joints between the decorative patterns and the historical building decorative components; S2 Obtain images of the damaged parts of the historical building decorative components to be repaired, construct a point cloud model of the historical building decorative components to be repaired based on these images, fit the shape outline of the damaged parts, and construct the connection state attributes of the damaged parts; S3 Compare the connection state attributes obtained in step S2 with those in the database in step S1, obtain the decorative patterns that match the connection state attributes, construct a point cloud model of the matching decorative patterns, register the point cloud model with the point cloud model of the historical building decorative components to be repaired, and obtain the complete point cloud model after repair; S4 Use the complete point cloud model to perform surface reconstruction on the historical building decorative components to be repaired, so as to realize the inference and reconstruction of the original appearance of the historical building decorative components to be repaired; In step S2, when fitting the shape outline of the damaged part, the following steps are carried out: S21 Perform preprocessing of filtering and binarization on the image of the damaged part; S22 Use a line detection algorithm for fitting to obtain a set of multiple fitted lines; S23 Filter the fitted lines, remove the lines that do not meet the preset requirements, and obtain the required set of fitted lines; S24 For the lines in the set of lines in step S23, extract the two endpoints of each line respectively, add the two endpoints of each line to two different endpoint clusters respectively, perform adaptive clustering on these two endpoint clusters to obtain the line endpoints after clustering, and the line formed by connecting the line endpoints after clustering is the required fitted line; The connection state attributes of the damaged part include: the number of included angles between the decorative pattern at the damaged part and the main connecting line, the angle values, the decorative cross type, and the angle of the repeated decorative pattern array.

2. The method for reconstructing the original appearance of decorative components of damaged historical buildings as described in claim 1, characterized in that, In step S1, the decorative patterns include mushroom patterns, ram's horn patterns, cloud-scroll patterns, central patterns, nail-shaped patterns, and rope patterns, and the connection state attributes include the number of included angles with the main connecting line, the angle values, the decorative cross type, and the angle of the repeated decorative pattern array.

3. The method for reconstructing the original appearance of decorative components of damaged historical buildings as described in claim 2, characterized in that, In step S23, the filtering is performed by calculating the coincidence degree between the line and the edge contour of the damaged part, and the lines with a coincidence degree lower than the preset threshold are removed.

4. The method for reconstructing the original appearance of decorative components of damaged historical buildings as described in claim 1, characterized in that, In step S3, the point cloud model of the matching decorative pattern is constructed according to the following steps: Obtain the two-dimensional coordinates of the matching decorative pattern in the image coordinate system in the database, including the points on the edge of the decorative pattern contour and the points inside the contour; For the points inside the contour, including the points inside the upper and lower surface contours of the decorative pattern, the points inside the upper surface contour edge are all given the same depth value Z1, and the points inside the lower surface contour edge are all given the same depth value Z2, where Z1≠Z2, so as to obtain the three-dimensional coordinates of the points on the upper and lower surfaces of the decorative pattern; for the points on the contour edge, different depth values Z3 are given, where Z1<Z3<Z2, so as to obtain the points on the side of the decorative pattern; thus, the three-dimensional point cloud model of the entire decorative pattern is obtained.

5. The method for reconstructing the original appearance of decorative components of damaged historical buildings as described in claim 4, characterized in that, In step S3, the matching is performed using a fuzzy matching method.

6. The method for reconstructing the original appearance of decorative components of damaged historical buildings as described in claim 1, characterized in that, In step S4, the surface reconstruction of the decorative components of the historical building to be restored using the complete point cloud model adopts the Poisson reconstruction method.

7. A system for reconstructing the original appearance of decorative components of damaged historical buildings, characterized in that, Includes a processor for executing the original appearance reasoning reconstruction method for decorative components of damaged historical buildings as described in any one of claims 1-6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the original appearance reasoning and reconstruction method for decorative components of damaged historical buildings as described in any one of claims 1-6.

Images