Ancient building culture information extraction method and system based on clustering algorithm
Through three-dimensional laser scanning and clustering algorithm, high-precision point cloud data of ancient buildings are extracted and vectorized, which solves the problems of insufficient information and high cost in the extraction and protection of ancient buildings' cultural information, and achieves efficient and accurate cultural information Protect and inherit.
Patent Information
- Application Number
- CN202411133379.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-12-06
AI Technical Summary
The existing technology has problems such as insufficient information integrity and accuracy, complex data management and high cost in the extraction and protection of ancient building cultural information, and it is especially difficult to adapt to areas with limited resources.
Three-dimensional laser scanning is used to obtain high-precision point cloud data, and vectorized text and patterns are extracted through K-Means and IsoData clustering algorithms to generate a special map for cultural information extraction.
It improves the protection efficiency and accuracy of ancient building cultural information, reduces the risk of physical damage, saves storage space, and promotes cultural heritage.
Smart Images

Figure CN119091042A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of historical architectural cultural information extraction, and more specifically to a method and system for extracting ancient architectural cultural information based on a clustering algorithm. Background Art
[0002] Chinese ancient buildings are regarded as "historical books standing on the land of China" and have a long history and cultural value. They are excellent cultural heritages that are "rare to come by and can never be recovered" and have been preserved by all mankind. However, the ancient buildings have experienced hundreds and thousands of years of changes and vicissitudes, and are in an extremely fragile situation. The protection and repair of ancient buildings cannot be delayed.
[0003] The existing strategies for extracting and preserving cultural information of historical buildings include text records, image capture, and video recording. These methods have shown their unique value in the initial stage. For example, professional photographers or drone technology can capture detailed images of text inscriptions, complex patterns, and even the overall style of historical buildings, supplemented by detailed text descriptions, in order to record these precious cultural heritages as comprehensively as possible. However, with the in-depth application of technology and the growing demand, the limitations of these traditional methods are becoming increasingly prominent.
[0004] First, although text, image and video records have advantages in terms of intuitiveness and timeliness, they are insufficient in terms of information integrity and accuracy. In particular, for the fine text and complex patterns in historical buildings, traditional recording methods often have difficulty in accurately capturing their vector information, that is, they cannot perfectly reproduce their shape, size, proportion and color details in digital form, resulting in the loss of some key information during the conversion process.
[0005] Secondly, the mixed recording mode of various types of data has brought great challenges to the subsequent information management and archiving work. Data from different sources and formats face compatibility issues when integrating. In addition, the huge amount of data and the wide variety of data make information retrieval, analysis and utilization complicated and inefficient, making it difficult to form a systematic cultural heritage database, thus limiting the effective dissemination and in-depth research of cultural information.
[0006] Furthermore, from a cost-effectiveness perspective, relying on manual or high-tech equipment to record and protect large-scale historical architectural cultural information undoubtedly requires huge human and material resources. This includes not only the training costs of professionals, the purchase and maintenance costs of equipment, but also long-term data storage, updating and maintenance expenses. Therefore, this method is not enough to be extended to all types of historical architectural cultural information protection, especially for areas with limited resources and poor economic conditions, it is even more unbearable.
[0007] Therefore, how to provide a fast, accurate and efficient method and system for extracting cultural information for ancient building protection is an urgent problem that technicians in this field need to solve. Summary of the invention
[0008] In view of this, the present invention provides a method and system for extracting cultural information of ancient buildings based on clustering algorithm. High-precision point cloud data of ancient buildings are obtained through three-dimensional laser scanning, and K-Means clustering technology is used to accurately cluster text and patterns to realize their vectorized extraction, which greatly improves the efficiency and accuracy of protecting cultural information of ancient buildings.
[0009] In order to achieve the above object, the present invention adopts the following technical solution:
[0010] A method for extracting ancient architectural cultural information based on a clustering algorithm comprises the following steps:
[0011] Conduct 3D laser scanning on ancient buildings to obtain complete 3D point cloud model data of ancient buildings;
[0012] Generate a cultural information point cloud orthophoto according to the three-dimensional point cloud model data of the ancient building;
[0013] Clustering the cultural information point cloud orthophoto using K-Means and IsoData clustering algorithms to obtain the classification results of text and patterns and the corresponding raster data;
[0014] Converting the raster data into vector editable data;
[0015] A cultural information extraction thematic map is generated based on the vector editable data.
[0016] Preferably, the ancient building is subjected to three-dimensional laser scanning to obtain complete three-dimensional point cloud model data of the ancient building, which specifically includes:
[0017] The original point cloud data was obtained by scanning the text and pattern information inside the ancient building using the 3D laser scanning method;
[0018] The original point cloud data is processed by point cloud texture mapping, registration and splicing, coordinate conversion and denoising filtering before model construction to generate true color point cloud data, i.e., the three-dimensional point cloud model data of the ancient building.
[0019] Preferably, generating a cultural information point cloud orthophoto according to the ancient building three-dimensional point cloud model data specifically includes:
[0020] A number of feature points are selected in the text and pattern point cloud wall area to create an orthographic projection surface of the wall and determine the horizontal direction, and an orthographic image of the text and pattern point cloud wall area is generated according to the orthographic projection surface and the horizontal direction.
[0021] Preferably, the cultural information point cloud orthophoto is clustered by combining K-Means and IsoData clustering algorithms to classify the cultural information into raster data, and the specific process includes:
[0022] Clean and standardize cultural information point cloud orthophotos;
[0023] The IsoData algorithm was used to perform preliminary clustering on the cleaned and standardized cultural information point cloud orthophotos to obtain preliminary clustering results.
[0024] The preliminary clustering results are used as the initial center point of the K-Means algorithm to further cluster the cleaned and standardized cultural information point cloud orthophotos to obtain secondary clustering results;
[0025] Determine whether the secondary clustering result meets the preset conditions; if not, fine-tune the iterative IsoData algorithm of the secondary clustering result to optimize the cluster boundaries and centers; if the preset conditions are met, end the iteration and output the clustering result.
[0026] Preferably, generating a thematic map for cultural information extraction based on the vector editable data specifically includes:
[0027] Processing the vector editable data;
[0028] Creating elements of the processed vector editable data according to the classification results;
[0029] Adding attribute information to the element;
[0030] Lay out and adjust the elements to which attribute information has been added to generate a thematic map for cultural information extraction.
[0031] Preferably, the attribute information includes: text meaning and text color.
[0032] On the other hand, the present invention also provides an ancient architectural cultural information extraction system based on a clustering algorithm, comprising:
[0033] The acquisition module is used to perform 3D laser scanning on the ancient building to obtain complete 3D point cloud model data of the ancient building;
[0034] A projection module, used for generating a cultural information point cloud orthophoto according to the ancient building three-dimensional point cloud model data;
[0035] A classification module is used to cluster the cultural information point cloud orthophotos by combining K-Means and IsoData clustering algorithms to obtain the classification results of text and patterns and the corresponding raster data;
[0036] A format conversion module, used for converting the raster data into vector editable data;
[0037] The thematic map production module is used to generate a cultural information extraction thematic map based on the vector editable data.
[0038] Through the above technical solutions, it can be known that compared with the prior art, the present invention discloses a method and system for extracting cultural information of ancient buildings based on clustering algorithms. Aiming at the shortcomings of traditional methods for protecting cultural information of ancient buildings, such as time-consuming and labor-intensive methods, difficult to record vector information of text and pattern information, and difficult archiving and management of cultural information of ancient buildings, the present invention adopts three-dimensional laser scanning technology to collect high-precision three-dimensional point cloud model data of ancient buildings in a non-contact manner, avoiding direct contact with ancient buildings, thereby reducing the risk of physical damage or destruction to ancient buildings caused by the measurement process; further, the present invention adopts a clustering method combined with K-Means to cluster text information and pattern information in ancient buildings, and the text information and pattern in ancient buildings are often complex and diverse, including different shapes, sizes, directions and other characteristics. It can better handle the complex and diverse data of text information and pattern in ancient buildings, and gradually approach the real clustering structure through the process of iteration and optimization; finally, the clustered text and pattern are vectorized and extracted to produce the cultural information digital archiving thematic map, which can improve data quality and availability, save storage space and promote cultural inheritance. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0040] Figure 1 It is a schematic flow diagram of the extraction method of the present invention;
[0041] Figure 2 It is a flowchart combining K-Means and IsoData clustering algorithms;
[0042] Figure 3 This is a text classification effect diagram obtained by using the clustering algorithm disclosed in the present invention;
[0043] Figure 4 This is a classification effect diagram of carved patterns obtained by using the clustering algorithm disclosed in the present invention;
[0044] Figure 5 This is a pattern classification effect diagram obtained by using the clustering algorithm disclosed in the present invention;
[0045] Figure 6 This is a schematic diagram of the structure of the extraction system provided by the present invention. DETAILED DESCRIPTION
[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0047] The embodiment of the present invention discloses a method for extracting ancient architectural cultural information based on a clustering algorithm, such as Figure 1 As shown, the following steps are included:
[0048] Step 1: Perform 3D laser scanning on the ancient building to obtain complete 3D point cloud model data of the ancient building.
[0049] The 3D laser scanner uses the principle of laser ranging. By scanning the 3D coordinates, reflectivity, texture and other information of the surface points of the measured target with laser, it can quickly and non-contact reconstruct a high-precision and high-resolution 3D model of the measured target. Because its measurement data not only contains the x, y, z point information of the measured target, but also includes R, G, B color information, this technology is widely used in various architectural surveying and mapping, and has become a mainstream technology in the protection of ancient buildings. It overcomes the limitations of traditional measurement and modeling, and can realize fast and accurate 3D information extraction and high-precision and high-resolution 3D modeling of various large and complex entities, providing important data support and technical guarantee for the protection of ancient buildings and the extraction of architectural information.
[0050] The present invention uses the Trimble SX10 3D scanner to acquire the point cloud data of the ancient building. To ensure that the 3D point cloud model data of the ancient building has high accuracy and data integrity, a total of 98 site data were collected after field surveys of the building layout and structure, including 90 inside the building and 8 outside. Among them, the texture details inside the ancient building are complex, and the adjacent scanning sites are ensured to be no more than 5m apart and the overlap rate is not less than 50%. The overlap rate is not less than 30% during external scanning. The size of the final collected point cloud and image data is 25.6GB.
[0051] Step 2: Generate cultural information point cloud orthophotos based on the ancient building 3D point cloud model data.
[0052] Step 3: Combine K-Means and IsoData clustering algorithms to cluster the cultural information point cloud orthophotos to obtain the classification results of text and patterns and the corresponding raster data.
[0053] The present invention adopts a partitioning clustering method for classifying text and pattern information. The main algorithms used are K-Means and IsoData clustering algorithms. The partitioning clustering method divides the data set into non-overlapping subsets. Objects in the same subset are "similar", and objects in different subsets are "different".
[0054] Step 4: Convert raster data into vector editable data; optionally, use data processing software such as ENVI and ArcGIS to convert cultural information raster data into vector data in .shp format.
[0055] Step 5: Generate a cultural information extraction thematic map based on the vector editable data.
[0056] Furthermore, the ancient building is scanned by 3D laser to obtain complete 3D point cloud model data of the ancient building, including:
[0057] The original point cloud data was obtained by scanning the text and pattern information inside the ancient building using the 3D laser scanning method.
[0058] The original point cloud data is processed by point cloud texture mapping, registration and splicing, coordinate conversion and denoising filtering before model construction to generate true color point cloud data, i.e., the three-dimensional point cloud model data of the ancient building.
[0059] The amount of point cloud data obtained by 3D laser scanning is large and there are many sites. The data needs to be processed by a series of processes such as point cloud texture mapping, registration splicing, coordinate conversion and denoising filtering before model construction. In order to preserve the real 3D information of ancient buildings, in addition to the 3D spatial information of ancient buildings, the architectural texture, text, pattern and color are all rich in historical and cultural value. Therefore, the acquired point cloud information needs to be combined with image photos for texture mapping to generate true color point cloud data of ancient buildings. Point cloud data processing can be automated by Trimble Business Center software. After point cloud data texture mapping and site registration, a true color 3D point cloud model of the ancient building is generated, and then the complete point cloud model of the ancient building cultural information is retained after point cloud denoising filtering.
[0060] Furthermore, the cultural information point cloud orthophotos are generated based on the 3D point cloud model data of the ancient buildings, including:
[0061] A number of feature points are selected in the text and pattern point cloud wall area to create an orthographic projection surface of the wall and determine the horizontal direction, and an orthographic image of the text and pattern point cloud wall area is generated according to the orthographic projection surface and the horizontal direction.
[0062] Furthermore, if Figure 2-5As shown in the figure, the K-Means and IsoData clustering algorithms are combined to cluster the cultural information point cloud orthophotos and classify the cultural information into raster data. The specific process includes:
[0063] The cultural information point cloud orthophotos were cleaned and standardized to eliminate noise and unify the data scale.
[0064] The IsoData algorithm was used to perform preliminary clustering on the cleaned and standardized cultural information point cloud orthophotos to obtain preliminary clustering results in order to obtain a more detailed classification.
[0065] The preliminary clustering results were used as the initial center points of the K-Means algorithm, and further clustering analysis was performed on the cleaned and standardized cultural information point cloud orthophotos to obtain secondary clustering results.
[0066] Determine whether the secondary clustering results meet the preset conditions; if not, fine-tune the IsoData algorithm for the secondary clustering results to optimize the cluster boundaries and centers; if the preset conditions are met, the iteration ends and the clustering results are output. Furthermore, the clustering results can also be post-processed, such as smoothing and denoising, to improve data quality.
[0067] Among them, the steps of K-Means algorithm are as follows:
[0068] 1) Randomly select k sample points in the sample data set D. Assign the feature values of the k sample points to the initial cluster centers
[0069] At the jth iteration, the midpoint x of sample point D t (t=1,...,n), calculate to each cluster center in turn The Euclidean distance d(t,i):
[0070]
[0071] 3) Find x t about The minimum distance of x t Classified as about In the cluster with the smallest distance;
[0072] 4) Update the cluster center of each cluster:
[0073]
[0074] Calculate the objective function J of the updated cluster center i And with the previous objective function J i-1 For comparison:
[0075]
[0076] If |J i -J i-1 |<δ, the algorithm ends, otherwise it enters the second step and continues to iterate until the target condition is met.
[0077] The detailed steps of the IsoData algorithm are as follows:
[0078] For a set of N pattern samples, determine c initial cluster centers z1, z2, ..., z c , where c is not necessarily equal to K. Set the control parameters of cluster analysis, which can also be manually modified during the iteration process to distribute N pattern samples to each cluster center according to the index;
[0079] Calculate the distance index function of each sample in each category, and classify each sample according to the principle of minimum distance from the cluster center;
[0080] Split and merge the cluster sets obtained last time, and obtain new cluster centers through the above steps;
[0081] Re-iterate the operation, calculate various indicators, and judge whether the clustering results meet the requirements. After multiple iterations, if the results converge, the operation ends.
[0082] If it is satisfied, the algorithm ends, otherwise it continues to iterate until the target condition is met.
[0083] Furthermore, a thematic map of cultural information extraction is generated based on the vector editable data, including:
[0084] Process the vector editable data; the wall text and patterns are basically classified successfully through the clustering algorithm, but there are still some errors in the extraction, so it is necessary to convert the extracted edge results into vector data, and then delete the wrong part through vector data processing. Use ENVI, ArcGIS and other data processing software to convert the cultural information raster data into .shp format vector data, and delete the noise (wrong) part in the vector data.
[0085] Create features from the processed vector editable data according to the classification results;
[0086] Add attribute information to features;
[0087] The elements with added attribute information are laid out and adjusted to generate the cultural information extraction thematic map. Before generating the thematic map, the following steps are also included: inserting the map name, text legend and scale, etc.
[0088] Preferably, the attribute information includes: text meaning and text color.
[0089] On the other hand, the present invention also provides an ancient architectural cultural information extraction system based on clustering algorithm, such as Figure 6 As shown, including:
[0090] The acquisition module is used to perform 3D laser scanning on the ancient building to obtain complete 3D point cloud model data of the ancient building;
[0091] The projection module is used to generate cultural information point cloud orthophotos based on the ancient building three-dimensional point cloud model data;
[0092] The classification module is used to cluster the cultural information point cloud orthophotos by combining K-Means and IsoData clustering algorithms to obtain the classification results of text and patterns and the corresponding raster data;
[0093] Format conversion module, used to convert raster data into vector editable data;
[0094] Thematic map production module is used to generate cultural information extraction thematic maps based on vector editable data.
[0095] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0096] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for extracting ancient architectural cultural information based on clustering algorithm, characterized in that: The following steps are involved: Conduct 3D laser scanning on ancient buildings to obtain complete 3D point cloud model data of ancient buildings; Generate a cultural information point cloud orthophoto according to the three-dimensional point cloud model data of the ancient building; Clustering the cultural information point cloud orthophoto using K-Means and IsoData clustering algorithms to obtain the classification results of text and patterns and the corresponding raster data; Converting the raster data into vector editable data; A cultural information extraction thematic map is generated based on the vector editable data.
2. The ancient architectural cultural information extraction method based on clustering algorithm according to claim 1 is characterized in that: Perform 3D laser scanning on the ancient building to obtain complete 3D point cloud model data of the ancient building, including: The three-dimensional laser scanning method is used to scan the text and pattern information inside the ancient building to obtain the original point cloud data; the original point cloud data is subjected to point cloud texture mapping, registration splicing, coordinate conversion and denoising filtering, and then the model is constructed to generate true color point cloud data, that is, the three-dimensional point cloud model data of the ancient building.
3. The ancient architectural cultural information extraction method based on clustering algorithm according to claim 1 is characterized in that: Generating a cultural information point cloud orthophoto according to the ancient building three-dimensional point cloud model data specifically includes: A number of feature points are selected in the text and pattern point cloud wall area to create an orthographic projection surface of the wall and determine the horizontal direction, and an orthographic image of the text and pattern point cloud wall area is generated according to the orthographic projection surface and the horizontal direction.
4. The ancient architectural cultural information extraction method based on clustering algorithm according to claim 1 is characterized in that: The cultural information point cloud orthophotos are clustered by combining K-Means and IsoData clustering algorithms to classify the cultural information into raster data. The specific process includes: Clean and standardize cultural information point cloud orthophotos; The IsoData algorithm was used to perform preliminary clustering on the cleaned and standardized cultural information point cloud orthophotos to obtain preliminary clustering results. The preliminary clustering results are used as the initial center point of the K-Means algorithm to further perform clustering analysis on the cleaned and standardized cultural information point cloud orthophotos to obtain secondary clustering results; Determine whether the secondary clustering result meets the preset conditions; if not, fine-tune the iterative IsoData algorithm of the secondary clustering result to optimize the cluster boundaries and centers; if the preset conditions are met, end the iteration and output the clustering result.
5. The ancient architectural cultural information extraction method based on clustering algorithm according to claim 1 is characterized in that: Generating a cultural information extraction thematic map according to the vector editable data specifically includes: Processing the vector editable data; Creating elements of the processed vector editable data according to the classification results; Adding attribute information to the element; Lay out and adjust the elements to which attribute information has been added to generate a thematic map for cultural information extraction.
6. The ancient architectural cultural information extraction method based on clustering algorithm according to claim 5 is characterized in that: The attribute information includes: text meaning and text color.
7. An ancient architectural cultural information extraction system based on clustering algorithm, characterized in that: include: The acquisition module is used to perform 3D laser scanning on the ancient building to obtain complete 3D point cloud model data of the ancient building; A projection module, used for generating a cultural information point cloud orthophoto according to the ancient building three-dimensional point cloud model data; A classification module is used to cluster the cultural information point cloud orthophotos by combining K-Means and IsoData clustering algorithms to obtain the classification results of text and patterns and the corresponding raster data; A format conversion module, used for converting the raster data into vector editable data; The thematic map production module is used to generate a cultural information extraction thematic map based on the vector editable data.
Citation Information
Cited By
Clustering analysis-based Laue microcell diffraction grid scanning segmentation method and device, electronic equipment and medium
CN119295757A
Laue micro-area diffraction grating scanning segmentation method, device, electronic device and medium based on cluster analysis
CN119295757B
