Historical building digital modeling and outer wall diagnosis method based on multi-source data fusion

Abstract

The invention provides a historical building digital modeling and outer wall diagnosis method based on multi-source data fusion, and relates to the technical field of historical building protection and digitization, the method aims at comprehensively obtaining historical building data and analyzing the condition of an outer wall body, firstly, a building main body image is captured through aerial photography of an unmanned aerial vehicle at multiple angles and multiple levels, and the flight height is 30-80 m; sites are arranged on the ground, millimeter-level scanning is conducted on the building facade through a three-dimensional laser scanner, the distance between every two adjacent sites is 10-30 m, and the sites cover the area 10-15 m away from the ground; the method comprises the following steps: firstly, carrying out multi-source data fusion on a wall body, then carrying out close-range shooting on upward view and typical local parts by using a micro lens of a single lens reflex, then preprocessing various data, and finally, measuring and analyzing the concave-convex condition of the wall body according to fused point cloud data, thereby providing accurate data support for historical building protection and research. By means of unmanned aerial vehicle aerial photography and three-dimensional laser scanning station arrangement, it is ensured that high-quality data is obtained.

Description

Technical Field

[0001] The present invention relates to the field of historical building protection and digitization technology, and in particular to a historical building digital modeling and exterior wall diagnosis method based on multi-source data fusion. Background Art

[0002] Historical buildings carry rich cultural, artistic, and historical value. However, with the passage of time, natural erosion, and human influence, their preservation faces many challenges. Traditional methods for surveying and diagnosing historical buildings often rely on a single data source or manual surveys, which suffer from insufficient accuracy, low efficiency, and difficulty in fully reflecting the overall and detailed condition of the building. For example, simple manual measurement cannot obtain high-precision geometric information of complex building facades, and two-dimensional drawings cannot intuitively present the three-dimensional spatial characteristics and artistic decorative details of the building. Single measurement technologies, such as drone aerial photography, cannot accurately capture the characteristics of buildings near the ground and hidden parts, and three-dimensional laser scanning from a ground perspective may miss fine structures in high-altitude areas. Therefore, a comprehensive multi-source data acquisition and processing method is urgently needed to achieve comprehensive and accurate digital modeling and exterior wall condition diagnosis of historical buildings.

[0003] The present invention aims to provide a method for digital modeling and exterior wall diagnosis of historical buildings using multi-source data fusion, overcoming the limitations of existing technologies in the surveying and diagnosis of historical buildings. By comprehensively utilizing multiple data acquisition methods to obtain all-round information, the present invention achieves high-precision three-dimensional digital modeling of historical buildings through efficient data fusion and processing, and accurately analyzes and diagnoses the structural stability and damage status of building exterior walls, providing strong support for the protection, restoration and research of historical buildings. Summary of the Invention

[0004] To achieve the above objectives, the present invention proposes a method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion, comprising:

[0005] Step S1: Drone aerial photography

[0006] The drones follow the predetermined routes to take multi-angle and multi-level aerial photos of the historical buildings, with the flight altitude covering a range of 30-80 meters above the main building.

[0007] Step S2: 3D laser scanning

[0008] Scanning stations are arranged on the ground around the building, and a 3D laser scanner is used to perform a millimeter-level full-scale scan of the building from the ground. Depending on the complexity of the building, the distance between adjacent scanning stations is 10-30 meters, covering the area from the ground to 10-15 meters on the building facade;

[0009] Step S3: Close-up photography

[0010] For upward-looking parts of buildings and local parts with typical features, use a professional SLR camera with a macro lens, set the aperture, shutter speed and sensitivity according to the lighting conditions and the texture characteristics of the target, and shoot multi-angle, high-definition image sequences at a distance of 0.5-5 meters.

[0011] Step S4: Data fusion and model reconstruction

[0012] Preprocessing of drone aerial photography data, 3D laser scanning data, and close-range photography data: distortion correction of drone images, geographic coordinate matching based on flight logs, denoising and stitching of 3D laser scanning data, and feature extraction and matching of homonymous points on close-range photography images.

[0013] Step S5: External wall diagnosis and analysis

[0014] Based on the fused point cloud data, a comprehensive measurement and analysis of the concave and convex conditions of the walls of historical buildings was carried out. Points in the continuous wall area were selected in the point cloud data, the plane equation was fitted, and the distance deviation from each point to the fitted plane was calculated.

[0015] In one example, the drone aerial photography system is equipped with a high-resolution camera, and the flight altitude is set according to the scale of the building and the surrounding environment to ensure that an aerial bird's-eye view image of the entire building is obtained.

[0016] In one example, the drone aerial photography frequency is determined based on the flight speed and the required image overlap, ensuring that the overlap rate of two adjacent images is no less than 70%, meeting the data requirements for subsequent three-dimensional reconstruction. The collected data includes high-resolution RGB images and flight log data that records the drone's position and attitude information.

[0017] In one example, the three-dimensional laser scanner records the three-dimensional coordinates and reflection intensity information of each scanning point during the scanning process, obtains the fine geometric shape data of the ground part of the building, and reflects the actual size and spatial position relationship of the walls, door and window openings, and column base structures.

[0018] In one example, during the close-up photography process, the camera is calibrated using a calibration plate to record the shooting position and posture information of each photo, ensuring the precise correspondence between images from different perspectives, and providing detailed data for subsequent feature extraction and three-dimensional modeling, so as to restore the three-dimensional form of the building's fine decoration and complex structure.

[0019] In one example, the data fusion and model reconstruction adopt an adaptive 3D registration algorithm. The algorithm automatically selects appropriate registration primitives and registration strategies based on the data features of different data sources. The registration primitives can be based on feature points, edge lines or surface features. The registration strategy can be a combination of rigid registration and non-rigid registration. Based on the initial point cloud framework constructed by 3D laser scanning data, the aerial point cloud calculated from the drone aerial image and the feature part point cloud generated by close-range photography are gradually registered and fused to form a complete and unified multi-source point cloud data.

[0020] In one example, the data fusion and model reconstruction utilizes a point cloud data reconstruction algorithm, based on Poisson surface reconstruction and triangulated meshing methods, to construct a spatial digital model of a historical building. For the architectural art pattern area, the fine texture information obtained by close-up photography is used to accurately map the pattern to the corresponding three-dimensional model surface through texture mapping technology, thereby achieving high-precision modeling of the historical architectural art patterns.

[0021] In one example, the exterior wall diagnostic analysis determines the degree of deformation of the wall surface and judges whether the wall surface is beyond the normal range. The exterior wall diagnostic analysis selects characteristic lines along the longitudinal and transverse directions of the wall in the point cloud data, compares the position changes of the corresponding characteristic lines under different periods or design standards, and uses trigonometric function relationships to calculate the wall inclination angle and the size of the in-plane rotation angle to evaluate the stability of the wall structure.

[0022] In one example, the exterior wall diagnostic analysis visualizes wall damage, visually marking damaged areas in point cloud data on a three-dimensional model through color coding and highlighting, generating a wall damage distribution map while annotating detailed information on damage type, size, and depth. The depth information can be combined with texture features from close-range photographic images and actual measurement estimates to provide an accurate basis for formulating subsequent repair plans.

[0023] The digital modeling and exterior wall diagnosis method of historical buildings with multi-source data fusion proposed by the present invention can bring the following benefits:

[0024] Beneficial effects:

[0025] The benefits of this patent and prior art are mainly reflected in the following aspects:

[0026] 1. This invention can obtain more comprehensive and accurate historical building information through multi-source data fusion. It uses drone aerial photography, 3D laser scanning, and close-range photography technology to effectively capture the characteristics of complex facades and hidden parts of buildings. The site layout of 3D laser scanning ensures the acquisition of high-quality data. At the same time, the use of close-range photography can also be used to capture specific parts of the building in detail to meet data collection needs in different situations.

[0027] 2. This invention overcomes the limitations of existing technologies in the surveying and diagnosis of historical buildings. By comprehensively utilizing multiple data collection methods, it achieves comprehensive and accurate digital modeling and exterior wall condition diagnosis of historical buildings. Through the all-round measurement and visualization of point cloud data, it can intuitively identify damaged areas of the wall and generate detailed damage distribution maps, providing an accurate basis for the formulation of subsequent repair plans.

[0028] 3. The present invention adopts an adaptive three-dimensional registration algorithm and a point cloud data reconstruction algorithm, which can quickly process and fuse data from different sources to form complete multi-source point cloud data, thereby accelerating the three-dimensional modeling of historical buildings, overcoming the limitations of existing technologies in the surveying and diagnosis of historical buildings, and realizing comprehensive and accurate digital modeling of historical buildings through the comprehensive use of multiple data collection methods.

[0029] 4. The present invention uses a method of fitting point cloud planes to diagnose the unevenness and damage of walls. Through point cloud data processing and 3D model reconstruction, wall damage is visually identified on the 3D model by color coding and highlighting, generating a wall damage distribution map that clearly displays the damage type, size, and depth. Trigonometric functions are used to calculate the wall inclination angle and the size of the in-plane rotation angle to assess the structural stability of the wall. Combining the texture features of close-range photographic images with actual measurement estimates, depth information is provided to ensure the accuracy of the diagnostic results. The detailed damage information and structural stability assessment provide a reliable basis for the formulation of subsequent repair plans, helping to formulate more scientific and reasonable repair plans and improve the efficiency and quality of repair work. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0031] Figure 1 The figure is a flow chart of a method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to the present invention.

[0032] Figure 2 This is an example diagram of a wall point cloud of a multi-source data fusion method for digital modeling and exterior wall diagnosis of historical buildings in the present invention.

[0033] Figure 3 This is a wall point cloud rendering of a multi-source data fusion method for digital modeling and exterior wall diagnosis of historical buildings in the present invention.

[0034] Figure 4 This is a point cloud data flow chart of a multi-source data fusion method for digital modeling and exterior wall diagnosis of historical buildings in the present invention. DETAILED DESCRIPTION

[0035] In order to more clearly and completely illustrate the technical solution of the present invention, the present invention will be further described below with reference to the accompanying drawings.

[0036] Please refer to Figure 1 The present invention proposes a method for digital modeling and exterior wall diagnosis of historical buildings with multi-source data fusion. By comprehensively using drone aerial photography, 3D laser scanning and close-range photography techniques, all-round and high-precision data collection and processing of historical buildings are carried out. Drone aerial photography can cover a range of 30-80 meters above the main building, obtaining an aerial bird's-eye view image of the entire building. At the same time, it has a high-resolution camera to ensure that the image quality meets the data requirements of subsequent 3D reconstruction. 3D laser scanning uses scanning stations arranged on the ground to perform millimeter-level full-scale scanning of the building from a ground perspective, recording the 3D coordinates and reflection intensity information of each scanning point, reflecting the actual size and spatial position relationship of the wall, door and window openings, and column base structure. Close-range photography is aimed at the upward-looking parts of the building. The local parts with typical features are photographed using a professional SLR camera with a macro lens to ensure the accurate correspondence between images from different perspectives, providing detailed data for subsequent feature extraction and 3D modeling. The data fusion and model reconstruction use an adaptive 3D registration algorithm and a point cloud data reconstruction algorithm to efficiently pre-process and fuse data from different sources to form a complete and unified multi-source point cloud data, and construct a spatial digital model of the historical building. The external wall diagnosis and analysis is based on the fused point cloud data to conduct a full-scale measurement and analysis of the convexity and concavity of the historical building wall, to determine whether the concavity and convexity of the wall exceeds the normal range and to evaluate the stability of the wall structure. At the same time, the wall damage is presented in a visual way to provide an accurate basis for the formulation of subsequent repair plans.

[0037] Example 1: Small ancient temple building protection project

[0038] Drone aerial photography: Small ancient temples are usually located in relatively quiet environments, and may be surrounded by trees, small hillsides, etc. When choosing a DJI Phantom 4 RTK drone for aerial photography, it is important to survey the surrounding environment of the temple in advance. Fly at an altitude of 80 meters above the temple. Local wind conditions need to be considered. If the wind is strong, the flight speed and route need to be adjusted appropriately to ensure stable flight. Fly in a zigzag pattern at a speed of 5 meters per second, with the camera set to 48 million pixels, a shooting interval of 2 seconds, and an image overlap rate of 70%. This will allow you to obtain aerial images of the temple roof and courtyard layout, as well as flight data, providing a basis for subsequent 3D reconstruction.

[0039] 3D laser scanning: When arranging scanning stations on the ground around the temple, attention should be paid to the flatness of the ground. If the ground has large undulations or debris, the ground needs to be properly processed to ensure that the scanner can be placed stably. Six scanning stations are arranged with a spacing of 15 meters. The FARO Focus S350 scanner is used with a scanning accuracy of 3 mm. The building facade is scanned from the ground to a height of 12 meters. During the scanning process, the branches and leaves of the trees around the temple blocked part of the wall. At this time, it is necessary to adjust the scanning angle or add scanning stations to ensure that the complete wall, doors and windows are obtained with fine geometric data.

[0040] Close-up photography: When taking close-up photos of typical local features such as temple eaves carvings and door hairpin carvings, lighting conditions are crucial. If you shoot under direct sunlight, reflections may occur, affecting image quality; if the light is too dark, the image will be blurred. You should choose the appropriate shooting time according to the actual lighting conditions. Use a Canon 5D Mark IV with a 100mm macro lens to shoot from multiple angles within a distance of 1-3 meters. Use a calibration plate to calibrate the camera and record the shooting position and posture information of each photo. When shooting door hairpin carvings, due to the rich details of the carvings, multi-angle shooting and precise calibration can clearly record every detail.

[0041] Data fusion and modeling, preprocessing: When preprocessing the collected drone aerial photography data, 3D laser scanning data, and close-range photography data, drone image distortion correction is performed using professional image processing software combined with relevant information in the flight log. When denoising the 3D laser scanning data, an appropriate denoising algorithm is selected based on the scanning environment and data characteristics to remove noise caused by environmental interference while retaining valid data. When performing feature extraction and same-name point matching on close-range images, inaccurate matching may occur due to the complex details of architectural decoration, requiring manual intervention and optimization to ensure data reliability.

[0042] Adaptive 3D registration: Based on the ground scanning point cloud, aerial photography is integrated with the near-field point cloud, and a strategy based on feature points and rigid registration is adopted. In actual operation, since the temple architectural structure is relatively simple, but there are some irregular decorative parts, when selecting feature points, the characteristics of these decorative parts must be fully considered to ensure the accuracy of registration. After generating the multi-source point cloud, the Poisson surface reconstruction algorithm is used to construct a digital model of the temple. During the model construction process, the parameters must be continuously adjusted to make the model more consistent with the shape and details of the actual building. When using near-field texture mapping to restore the carved artistic patterns, attention must be paid to the clarity and fit of the texture to ensure that the artistic patterns can be truly presented on the 3D model.

[0043] Exterior wall diagnosis: When measuring the concave and convex parts of the wall, points are selected in the wall point cloud to fit the plane and calculate the deviation. A 2 cm concave and convex deformation is found on the back wall. The abnormality is judged by comparing with historical data. In the process of comparing historical data, incomplete or inaccurate data may be encountered. At this time, it is necessary to combine other relevant information, such as local historical records and old photos, to make a comprehensive judgment. Feature lines are selected along the longitudinal and transverse directions of the wall, and the east wall is tilted by 1.5°. The corner changes are recorded to evaluate stability. When measuring the tilt angle and corner changes, multiple measurements are taken to take the average value to reduce errors. The wall damage is presented in a visual way. Cracks are marked in red on the model, and the size and depth are marked. A damage distribution map is generated to guide repair. When marking the depth information, the texture characteristics of the close-range photographic image are combined with the actual measurement estimate to ensure the accuracy of the marking, providing an accurate basis for the subsequent formulation of the repair plan.

[0044] Example 2: European-style castle restoration project

[0045] Drone aerial photography: European castles are generally large in scale and have complex surrounding environments. They may have moats, gardens, etc. Professional surveying and mapping drones are dispatched to an altitude of 150 meters and fly around the castle according to a preset route. Before the flight, various parameters of the drone must be strictly checked to ensure flight safety and data collection quality. The camera resolution is 60 million pixels, the flight speed is 8 meters / second, and 65% image overlap is guaranteed to obtain panoramic views of the castle and images of the top of the tower. Due to the unique architectural style of the castle, there are many complex structures such as spires and towers. During the shooting process, attention should be paid to the choice of shooting angle to ensure that the characteristics of these structures can be fully captured. For example, when shooting the top of the tower, it is necessary to shoot from multiple angles so that the three-dimensional model of the tower can be accurately reconstructed later.

[0046] 3D laser scanning: Ten scanning stations were arranged around the castle, with a spacing of 20 meters. Leica ScanStation P50 scanners were used, with an accuracy of 2 mm and a scanning height of up to 15 meters, covering complex structural data of exterior walls and battlements. When arranging the scanning stations, the structural characteristics of the castle and the surrounding environment must be considered. For complex and high-rise structures such as battlements, appropriate scanning positions and angles need to be selected to ensure that complete data can be obtained. During the scanning process, it may be affected by the surrounding light, resulting in deviations in the scanned data. In this case, the scanned data needs to be calibrated and verified multiple times to ensure data accuracy.

[0047] Close-up photography: The entrance sculptures and fine details of the window lattice carvings of the ancient castle were photographed with a Nikon D850 and a 60mm macro lens at a distance of 2-4 meters. The camera was accurately calibrated to record posture information. When photographing the entrance sculptures, since the sculptures may have dust and dirt on their surfaces, they need to be properly cleaned before shooting to ensure that the captured images can clearly present the details of the sculptures. At the same time, the shooting angle and camera parameters should be adjusted according to the shape of the sculptures and the lighting conditions to obtain the best shooting effect. When photographing the window lattice carvings, attention should be paid to the complex texture of the carvings. Through multi-angle shooting and precise calibration, detailed data is provided for the subsequent restoration of the fine decoration of the window lattice.

[0048] Data fusion and modeling, preprocessing: Similar to the preprocessing process of Example 1, but due to the more complex structure of European castles and the larger amount of data, more challenges will be faced in the data processing process. UAV image distortion correction and geographic registration require higher precision to ensure that the image accurately corresponds to the actual geographic location. When denoising and stitching 3D laser scanning data, a large amount of point cloud data needs to be processed, and a point cloud data reconstruction algorithm needs to be adopted to improve processing efficiency. When extracting close-range image features and matching homonymous points, due to the numerous architectural decoration details, matching is more difficult, and it is necessary to combine multiple matching algorithms and manual intervention to ensure data accuracy.

[0049] Adaptive 3D registration: Due to the complex structure of the castle, a combination of feature lines and surfaces and a non-rigid registration auxiliary strategy are used to fuse multi-source point clouds. When selecting registration primitives and registration strategies, the structural characteristics and data characteristics of the castle must be fully considered. For the main structure of the castle, a rigid registration strategy based on surface features can be adopted; for some decorative parts, due to their irregular shapes, a non-rigid registration auxiliary strategy is required to ensure the accurate fusion of point cloud data, and a high-precision digital model is reconstructed through triangulation. During the reconstruction process, the model must be optimized and adjusted multiple times to make it more consistent with the actual structure and appearance of the castle. When finely restoring the texture of sculptures and carvings, the high-resolution texture information obtained by close-up photography must be used. The texture is accurately mapped to the 3D model surface through texture mapping technology to show the artistic charm of the castle.

[0050] Exterior wall diagnosis: When measuring the concavity and convexity of the wall, multiple deformations due to uneven foundation settlement were discovered, with the largest reaching 3 cm. Areas for repair were marked. A comprehensive analysis was conducted based on the castle's historical data and geological conditions to determine the cause of the deformation. Characteristic lines were selected along the longitudinal and transverse directions of the wall and their displacement was analyzed. A local tilt of 2° was determined for the south wall. Anomalies in the corners were investigated for structural hazards. When assessing the wall's structural stability, the castle's overall structure and stress conditions were considered. A reasonable survey plan was developed, visualizing the wall damage. Different colors were used to distinguish cracks, spalling, and erosion damage. A 3D damage map was generated, quantifying the extent of damage to assist in repair design. When generating the 3D damage map, the accuracy and completeness of the damage information must be ensured to provide a reliable basis for the development of the repair plan. The above two examples show that in practical applications, data acquisition and processing parameters must be flexibly adjusted based on the building's characteristics, fully considering various practical factors, effectively achieving high-precision digital modeling and accurate exterior wall diagnosis to meet the diverse needs of historical building protection.

[0051] In this embodiment, a DJI drone was selected for drone aerial photography. The drone flew in a zigzag pattern at an altitude of 80 meters above the temple at a speed of 5 meters per second. The camera was set to 48 megapixels, with a shooting interval of 2 seconds to ensure an image overlap rate of 70%. This allowed the acquisition of aerial images of the temple roof and courtyard layout, as well as flight data. These data provided the basis for subsequent 3D reconstruction.

[0052] During the 3D laser scanning phase, six scanning stations were arranged on the ground around the temple, with a spacing of 15 meters. Using the FARO Focus S350 scanner, the scanning accuracy reached 3 mm, and the building facade was scanned from the ground to a height of 12 meters. This step obtained fine geometric data of the walls, doors and windows, providing important information for the detailed representation of the model.

[0053] For close-up photography, the Canon 5D Mark IV with a 100mm macro lens was used to capture typical features of the temple's eaves carvings and door pins, taking multi-angle shots at a distance of 1-3 meters. During the shooting process, the camera was calibrated using a calibration plate, and the shooting position and posture information of each photo was recorded to ensure accurate correspondence between images from different perspectives. These detail-rich image data provided key support for restoring the three-dimensional form of the building's fine decoration and complex structure.

[0054] Subsequently, the collected drone aerial photography data, 3D laser scanning data and close-range photography data were preprocessed, and the drone images were subjected to distortion correction and geo-registration; the scanning data were denoised and spliced; and the close-range images were subjected to feature extraction and homonymous point matching. These preprocessing steps laid a solid foundation for subsequent data fusion and model reconstruction.

[0055] During the data fusion and model reconstruction stage, an adaptive 3D registration algorithm is used to fuse aerial photography with near-field point clouds based on ground scanning point clouds. The algorithm automatically selects appropriate registration primitives and registration strategies based on the data characteristics of different data sources, such as strategies based on feature points and rigid registration. Ultimately, a multi-source point cloud is generated, and a Poisson surface reconstruction algorithm is used to construct a digital model of the temple. In addition, near-field texture mapping technology is used to accurately map the carved art patterns to the corresponding 3D model surface, achieving high-precision modeling of historical architectural art patterns.

[0056] Finally, in the external wall diagnosis and analysis phase, the concave and convex conditions of the historical building walls were measured and analyzed in all directions based on the fused point cloud data. Specifically, points were selected in the wall point cloud to fit the plane, and the deviation was calculated. Through this step, the concave and convex deformation of the rear wall was discovered, and the anomaly was judged by comparing with historical data. At the same time, characteristic lines were selected along the longitudinal and transverse directions of the wall, and the trigonometric function relationship was used to calculate the wall inclination angle and the size of the in-plane rotation angle, so as to evaluate the stability of the wall structure. Finally, the wall damage was presented in a visual way, and the damaged area was intuitively identified on the three-dimensional model through color coding and highlighting, and a wall damage distribution map was generated. At the same time, detailed information on the damage type, size, and depth was marked, providing an accurate basis for the formulation of subsequent repair plans.

[0057] Of course, the present invention may have many other implementations. Based on this implementation, other implementations obtained by ordinary technicians in this field without any creative work are all within the scope of protection of the present invention.

Claims

1. A method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion, characterized in that: include: Step S1: Drone aerial photography The drones follow the predetermined routes to take multi-angle and multi-level aerial photos of the historical buildings, with the flight altitude covering a range of 30-80 meters above the main building. Step S2: 3D laser scanning Scanning stations are arranged on the ground around the building, and a 3D laser scanner is used to perform a millimeter-level full-scale scan of the building from the ground. Depending on the complexity of the building, the distance between adjacent scanning stations is 10-30 meters, covering the area from the ground to 10-15 meters on the building facade; Step S3: Close-up photography For upward-looking parts of buildings and local parts with typical features, use a professional SLR camera with a macro lens, set the aperture, shutter speed and sensitivity according to the lighting conditions and the texture characteristics of the target, and shoot multi-angle, high-definition image sequences at a distance of 0.5-5 meters. Step S4: Data fusion and model reconstruction Preprocessing of drone aerial photography data, 3D laser scanning data, and close-range photography data: distortion correction of drone images, geographic coordinate matching based on flight logs, denoising and stitching of 3D laser scanning data, and feature extraction and matching of homonymous points on close-range photography images. Step S5: External wall diagnosis and analysis Based on the fused point cloud data, a comprehensive measurement and analysis of the concave and convex conditions of the walls of historical buildings was carried out. Points in the continuous wall area were selected in the point cloud data, the plane equation was fitted, and the distance deviation from each point to the fitted plane was calculated.

2. The method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to claim 1 is characterized in that: The drone aerial photography is equipped with a high-resolution camera, and the flight altitude is set according to the scale of the building and the surrounding environment to ensure that an aerial bird's-eye view image of the entire building is obtained.

3. The method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to claim 1 is characterized in that: The drone's aerial photography frequency is determined based on the flight speed and the required image overlap, ensuring that the overlap rate of two adjacent images is no less than 70%, meeting the data requirements for subsequent 3D reconstruction. The collected data includes high-resolution RGB images and flight log data that records the drone's position and attitude information.

4. The method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to claim 1 is characterized in that: The three-dimensional laser scanner records the three-dimensional coordinates and reflection intensity information of each scanning point during the scanning process, obtains the fine geometric shape data of the ground part of the building, and reflects the actual size and spatial position relationship of the wall, door and window openings, and column base structure.

5. The method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to claim 1 is characterized in that: During the close-up photography process, the camera is calibrated using a calibration plate to record the shooting position and posture information of each photo, ensuring the precise correspondence between images from different perspectives. This provides detailed data for subsequent feature extraction and three-dimensional modeling, which is used to restore the three-dimensional form of the building's fine decoration and complex structure.

6. The method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to claim 1 is characterized in that: The data fusion and model reconstruction adopt an adaptive 3D registration algorithm. The algorithm automatically selects appropriate registration primitives and registration strategies based on the data features of different data sources. The registration primitives can be based on feature points, edge lines or surface features. The registration strategy can be a combination of rigid registration and non-rigid registration. Based on the initial point cloud framework constructed by 3D laser scanning data, the aerial point cloud calculated from the drone aerial image and the feature part point cloud generated by close-up photography are gradually registered and fused to form complete and unified multi-source point cloud data.

7. The method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to claim 1 is characterized in that: The data fusion and model reconstruction utilizes a point cloud data reconstruction algorithm, based on Poisson surface reconstruction and triangulation meshing methods, to construct a spatial digital model of the historical building. For the architectural art pattern area, the fine texture information obtained by close-range photography is used, and the pattern is accurately mapped to the corresponding three-dimensional model surface through texture mapping technology, thereby achieving high-precision modeling of the historical architectural art pattern.

8. The method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to claim 1 is characterized in that: The exterior wall diagnostic analysis determines the degree of concave-convex deformation on the wall surface and judges whether the wall concave-convexity exceeds the normal range. The exterior wall diagnostic analysis selects characteristic lines along the longitudinal and transverse directions of the wall in the point cloud data, compares the position changes of corresponding characteristic lines under different periods or design standards, and uses trigonometric functions to calculate the wall inclination angle and the size of the in-plane rotation angle to evaluate the stability of the wall structure.

9. The method for digital modeling and exterior wall diagnosis of historical buildings based on multi-source data fusion according to claim 1 is characterized in that: The exterior wall diagnostic analysis visualizes wall damage, visually identifying damaged areas in point cloud data on a three-dimensional model through color coding and highlighting, generating a wall damage distribution map while also noting detailed information on damage type, size, and depth. Depth information can be combined with texture features from close-range photographic images and actual measurement estimates to provide an accurate basis for formulating subsequent repair plans.

Images