A digital mapping method for cultural relics based on multi-scale point cloud fusion

Abstract

The application provides a cultural relic digital surveying and mapping method based on multi-scale point cloud fusion, applied to the technical field of surveying and mapping, in the process of surveying and mapping of a target cultural relic, macro point cloud data, meso point cloud data and micro point cloud data are collected by using different surveying and mapping equipment respectively, and corresponding surveying and mapping accuracy attributes are generated based on the quantitative performance parameters of each device, the method selects and fuses the redundant point clouds at the same space position by the surveying and mapping accuracy attributes, avoids the influence of low-precision data on high-precision micro details, and combines local and global registration to ensure the spatial consistency of the overall structure and local details of the cultural relic, thereby improving the digital surveying and mapping accuracy of the cultural relic, and providing reliable data support for cultural relic recording, protection, repair analysis and digital display.

Description

Technical Field

[0001] This application relates to the field of surveying and mapping technology, and in particular to a method for digital surveying and mapping of cultural relics based on multi-scale point cloud fusion. Background Technology

[0002] With the development of cultural relic protection and digital technology, digital mapping of cultural relics has become an important technical means for recording, restoring, analyzing and displaying cultural relics. By conducting three-dimensional mapping of cultural relics and constructing digital models, it is possible to achieve long-term preservation and accurate recording of the morphological characteristics and spatial structure of cultural relics without touching the relics themselves.

[0003] In existing methods for digital mapping of cultural relics, in order to comprehensively obtain the structural information of the target cultural relic, it is usually necessary to use a variety of mapping equipment to collect three-dimensional data. Through these devices, the spatial geometric information of the cultural relic can be obtained at different scales, thereby realizing the recording of the overall structure and local details.

[0004] However, in the actual surveying and mapping process of cultural relics, there are significant differences in the surveying and mapping performance of different surveying and mapping equipment. Existing methods usually do not quantify and standardize the surveying and mapping performance parameters of each equipment, which makes it difficult to have a unified evaluation basis for the accuracy, reliability and spatial consistency of point cloud data from different sources. Therefore, the data selection and processing in the point cloud data fusion process often rely on human experience, which results in deviations in the overall structure and local details of the fused digital surveying and mapping model of cultural relics, leading to lower accuracy of digital surveying and mapping of cultural relics. Summary of the Invention

[0005] This application provides a digital mapping method for cultural relics based on multi-scale point cloud fusion. Its core lies in the following: during the mapping of the target cultural relic, different mapping devices are used to collect macroscopic point cloud, mid-scale point cloud, and microscopic point cloud data. Based on the mapping performance parameters of each device, corresponding mapping accuracy attributes are quantified and generated. During data processing, this method selects and fuses redundant point clouds at the same spatial location through the mapping accuracy attributes, avoiding the influence of low-precision data on high-precision microscopic details. Simultaneously, by combining local and global registration, the spatial consistency between the overall structure and local details of the cultural relic is ensured, thereby improving the accuracy of digital mapping of cultural relics and providing reliable data support for the recording, protection, restoration analysis, and digital display of cultural relics.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] This application provides a method for digital mapping of cultural relics based on multi-scale point cloud fusion, the method may include:

[0008] During the surveying and mapping of the target cultural relic, different surveying and mapping equipment are controlled to collect macroscopic point cloud data, mid-level point cloud data and microscopic point cloud data of the target cultural relic respectively;

[0009] Based on the surveying performance data of the surveying equipment, the corresponding surveying accuracy attributes are associated and quantified for the macro point cloud data, mid-point point cloud data, and micro point cloud data, respectively. The surveying accuracy attributes are used as the basis for point cloud data selection when fusing point cloud data.

[0010] Based on the aforementioned Zhongguan Cloud data, semantic segmentation of the cultural relic components is performed to determine multiple component regions corresponding to the target cultural relic:

[0011] In each of the component regions, the mid-view cloud data is registered to the macro point cloud data to establish a first mapping fusion region. Based on the first mapping fusion region, the micro point cloud data is locally registered with the corresponding mid-view cloud data to obtain a second mapping fusion region.

[0012] In the second mapping fusion area, based on the mapping accuracy attribute, the macro point cloud data, the mid-point cloud data and the micro point cloud data at the same spatial location are fused accordingly to obtain the mapping data of the target cultural relic;

[0013] Based on the surveying data, a digital surveying model of the target cultural relic is generated.

[0014] In some possible implementations, the surveying equipment includes a first device, a second device, and a third device. The surveying performance data based on the surveying equipment is used to quantify and associate corresponding surveying accuracy attributes with the macroscopic point cloud data, mid-level point cloud data, and microscopic point cloud data, respectively, including:

[0015] Based on the mapping performance data of the first device, the first mapping accuracy attribute corresponding to the macro point cloud data is quantized and associated.

[0016] Based on the mapping performance data of the second device, the second mapping accuracy attribute corresponding to the cloud data of the central viewpoint is quantized.

[0017] Based on the mapping performance data of the third device, the corresponding third mapping accuracy attribute is quantized and associated with the micro-viewpoint cloud data.

[0018] In some possible implementations, the mapping performance data based on the first device is used as the first mapping accuracy attribute corresponding to the macroscopic point cloud data, including:

[0019] The first mapping performance parameters of the first device are obtained, including resolution parameters, imaging geometry parameters, and positioning and attitude determination accuracy parameters.

[0020] Based on the resolution parameters, determine the spatial resolution accuracy of the macro point cloud data at the overall scale of the target cultural relic;

[0021] Based on the imaging geometric parameters, the error accuracy of the macroscopic point cloud data in the spatial coordinate calculation process is determined;

[0022] Based on the positioning and attitude determination accuracy parameters, the spatial positioning accuracy of the macro point cloud data in the mapping coordinate system is determined;

[0023] The spatial resolution accuracy, the error accuracy, and the spatial positioning accuracy are normalized to generate the first mapping accuracy attribute corresponding to the macroscopic point cloud data.

[0024] In some possible implementations, the mapping performance data based on the second device is a second mapping accuracy attribute corresponding to the central viewpoint cloud data association quantization, including:

[0025] The second surveying performance parameters of the second device are obtained, including distance measurement accuracy parameters, angle measurement accuracy parameters, and station registration accuracy parameters.

[0026] Based on the ranging accuracy parameters, determine the spatial ranging accuracy of each point in the central point cloud data along the ranging direction;

[0027] Based on the aforementioned angular measurement accuracy parameters, the angular unfolding error accuracy of the midpoint cloud data at the component scale is determined;

[0028] Based on the site registration accuracy parameters, the registration accuracy of the central point cloud data in the mapping coordinate system is determined;

[0029] The spatial ranging accuracy, the angular unfolding error accuracy, and the registration accuracy are normalized to generate the second mapping accuracy attribute corresponding to the central point cloud data.

[0030] In some possible implementations, the mapping performance data based on the third device is a third mapping accuracy attribute corresponding to the micro-viewpoint cloud data association quantization, including:

[0031] Acquire the third mapping performance parameters of the third device, which include local resolution parameters, reference working distance parameters, and device calibration parameters;

[0032] Based on the local resolution parameters, the local spatial resolution accuracy of the micro-view cloud data within the local detail area of ​​the target cultural relic is determined;

[0033] Based on the reference working distance parameter, determine the local spatial positioning accuracy of the micro-viewpoint cloud data within the reference working distance range;

[0034] Based on the device calibration parameters, the mapping stability accuracy of the micro-viewpoint cloud data under different viewing angles and working distances is determined.

[0035] The local spatial resolution accuracy, the local spatial positioning accuracy, and the mapping stability accuracy are normalized to generate a third mapping accuracy attribute corresponding to the micro-viewpoint cloud data.

[0036] In some possible implementations, within the second mapping fusion area, based on the mapping accuracy attribute, the macroscopic point cloud data, the mesoscopic point cloud data, and the microscopic point cloud data at the same spatial location are fused accordingly to obtain the mapping data of the target cultural relic, including:

[0037] The first surveying accuracy attribute, the second surveying accuracy attribute, and the third surveying accuracy attribute are respectively converted into accuracy indicators under the second surveying fusion area;

[0038] Within the second mapping fusion area, a three-dimensional spatial voxel mesh covering the target cultural relic is established;

[0039] In the three-dimensional spatial voxel mesh, the macroscopic point cloud data, the mesoscopic point cloud data, and the microscopic point cloud data at the same spatial location are aggregated to form multiple local point sets;

[0040] For each local point set, a first fusion is performed based on the aforementioned accuracy index to obtain initial fused data;

[0041] Based on the initial fusion data, a second fusion is performed based on the accuracy index to obtain fused point cloud data;

[0042] The fused point cloud data is integrated to obtain the mapping data of the target cultural relic.

[0043] In some possible implementations, the step of uniformly converting the first mapping accuracy attribute, the second mapping accuracy attribute, and the third mapping accuracy attribute into an accuracy index under the second mapping fusion region includes:

[0044] The first mapping accuracy attribute is converted into a first accuracy index under the second mapping fusion area;

[0045] The second mapping accuracy attribute is converted into a second accuracy index under the second mapping fusion area;

[0046] The third mapping accuracy attribute is converted into a third accuracy index under the second mapping fusion area.

[0047] In some possible implementations, the initial fusion based on the accuracy metric is performed for each local point set to obtain initial structural data, including:

[0048] If the first precision index of the macro point cloud data is greater than the second precision index of the mid-level point cloud data, then the macro point cloud data is selected as the structural reference point, and the mid-level point cloud data is used as the structural correction point for the first fusion to obtain the initial structural data.

[0049] If the first precision index of the macro point cloud data is less than the second precision index of the mid-level point cloud data, then the mid-level point cloud data is selected as the structural reference point, and the macro point cloud data is used as the structural correction point for the first fusion to obtain the initial structural data.

[0050] In some possible implementations, the second fusion based on the initial structural data and the accuracy metric to obtain fused point cloud data includes:

[0051] If the third precision index of the micro-viewpoint cloud data is greater than the second precision index of the mid-viewpoint cloud data, then the micro-viewpoint cloud data is selected as a detail correction point and fused with the initial structure data for a second time to obtain fused point cloud data.

[0052] If the third precision index of the micro-viewpoint cloud data is less than the second precision index of the mid-viewpoint cloud data, then the mid-viewpoint cloud data is selected as the detail correction point and fused with the initial structure data for the second time to obtain fused point cloud data.

[0053] In some possible implementations, the step of performing semantic segmentation of the target cultural relic based on the mid-view cloud data to determine multiple component regions corresponding to the target cultural relic includes:

[0054] Identify and segment the main structural units of the target cultural relic from the aforementioned central cloud data;

[0055] Based on the main structural unit, and combined with the preset topological rules for cultural relic components, the target cultural relic is divided into multiple component regions.

[0056] As can be seen from the above technical solution, this application has the following beneficial effects:

[0057] 1. This application collects macroscopic point cloud data, mid-scale point cloud data, and microscopic point cloud data of the target cultural relics respectively, and generates surveying accuracy attributes based on the performance data of each surveying and mapping equipment. This achieves accuracy control and correlation fusion of multi-scale point clouds, thereby obtaining high-precision cultural relic surveying and mapping data and improving the overall accuracy and completeness of digital surveying and mapping of cultural relics.

[0058] 2. This application achieves synchronous optimization of local structure and micro details by adopting a two-level point cloud registration and fusion strategy within the component area and making decisions based on the accuracy indicators of macro, meso, and micro point cloud data, thereby improving the spatial reconstruction accuracy of cultural relic components and effectively reducing surveying errors caused by insufficient point cloud at a single scale.

[0059] 3. This application uses semantic segmentation of cultural relic components based on mid-point cloud data, combines multi-scale fusion point cloud to generate digital mapping models, and can target the accuracy of different component areas, thereby providing a reliable data foundation and technical support for cultural relic protection, restoration and research. Attached Figure Description

[0060] The present application will be further described below with reference to the accompanying drawings.

[0061] Figure 1 A flowchart of the first digital mapping method for cultural relics based on multi-scale point cloud fusion provided in this application;

[0062] Figure 2 A flowchart of the second method for digital mapping of cultural relics based on multi-scale point cloud fusion provided in this application;

[0063] Figure 3 A flowchart of the third method for digital mapping of cultural relics based on multi-scale point cloud fusion provided in this application;

[0064] Figure 4 The flowchart of the fourth method for digital mapping of cultural relics based on multi-scale point cloud fusion provided in this application. Detailed Implementation

[0065] The terms "first," "second," and "third," etc., used in this application specification, claims, and drawings are used to distinguish different objects, not to limit a specific order.

[0066] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0067] To ensure clarity and conciseness in the description of the following embodiments, a brief introduction to the related technologies is given first:

[0068] Digital mapping of cultural relics is a technical means of digitally recording the morphological characteristics and spatial structure of cultural relics using 3D mapping technology. It is mainly applied to cultural relic protection, restoration analysis, damage assessment, and digital display. In actual mapping processes, to comprehensively acquire information on the overall structure and local details of cultural relics, various mapping devices are typically used for data acquisition. These include devices for acquiring the macroscopic outline of the relic, devices for acquiring component structural information, and high-precision scanning devices for acquiring microscopic details. The point cloud data acquired through these devices can be used to generate digital mapping models of cultural relics, achieving accurate description and digital preservation of the relic's morphological characteristics and spatial structure.

[0069] Research has revealed that in actual cultural relic surveying, different surveying equipment exhibits significant differences in spatial resolution, coordinate reference, and error distribution when acquiring macroscopic point cloud data, mesoscopic point cloud data, and microscopic point cloud data due to variations in working principles, imaging methods, working distances, and measurement accuracy. Macroscopic surveying equipment can quickly acquire the overall outline information of cultural relics, but its spatial positioning accuracy is relatively low; mesoscopic surveying equipment can provide high-precision structural information at the component scale of cultural relics; and microscopic high-precision scanning equipment can capture minute features on the surface of cultural relics, but its acquisition range is limited and it is sensitive to surveying environmental conditions. These multi-source, multi-scale data exhibit fundamental differences in spatial scale, accuracy level, and error characteristics. This makes it difficult to determine the reliability and priority of different point cloud data during the data fusion stage if the quantitative attributes of the mapping performance of each device are not fully considered. This can easily lead to low-precision data covering high-precision details, geometric misalignment between data of different scales, or the inability to accurately map local microscopic information into the overall model. Furthermore, in the actual application of data processing and fusion by surveyors, when facing the mapping tasks of large cultural relics or groups of cultural relics, manual judgment and operation are not only inefficient but also easily affected by experience level. This may result in the overall spatial structure and local details not being taken into account in the digital model, affecting the accuracy and reliability of the model.

[0070] As can be seen from the above analysis, existing methods for digital mapping of cultural relics still have shortcomings in the processing and fusion of multi-source, multi-scale point cloud data. This makes it difficult for the fused digital mapping model of cultural relics to simultaneously take into account the overall spatial structure and local micro-details, thereby reducing the accuracy of digital mapping of cultural relics and affecting its application in the protection, restoration analysis and digital display of cultural relics.

[0071] To address the aforementioned issues, this application provides a method for digital mapping of cultural relics based on multi-scale point cloud fusion. Please refer to [link to relevant documentation]. Figure 1 .

[0072] S101, during the surveying and mapping of the target cultural relic, control different surveying and mapping equipment to collect macroscopic point cloud data, mid-level point cloud data and microscopic point cloud data of the target cultural relic respectively.

[0073] To ensure clarity and conciseness in the description of the following embodiments, a brief introduction to the relevant terms is given first.

[0074] Target cultural relics refer to individual cultural relics or groups of cultural relics that require digital mapping and 3D modeling, including sculptures, architectural components, artifacts, and components of large-scale archaeological sites.

[0075] Macroscopic point cloud data refers to point cloud data acquired by the first device, used to record the overall outline and spatial layout of cultural relics. It has a wide coverage but relatively low resolution.

[0076] Point cloud data refers to point cloud data acquired by a second device, used to record structural information at the level of cultural relic components, with high resolution and spatial accuracy.

[0077] Micro-point cloud data refers to point cloud data acquired by a third-party device, used to record minute features on the surface of cultural relics, such as carving textures, cracks, or microscopic damage. It has extremely high resolution but a limited acquisition range.

[0078] In the process of surveying and mapping target cultural relics, multiple surveying and mapping devices are required to collect macroscopic point cloud data, mesoscopic point cloud data, and microscopic point cloud data. To comprehensively record the overall shape and local details of the cultural relic, three types of surveying and mapping devices are typically used: the first device is used to acquire the overall outline and spatial layout of the cultural relic; the second device is used to acquire structural information at the component scale; and the third device is used to acquire surface fine textures and decorative details. Macroscopic point cloud data is generated by the first device, characterized by its wide coverage and fast acquisition speed, enabling rapid capture of the overall spatial shape of the cultural relic, suitable for overall outline recording and spatial layout analysis. Mesoscopic point cloud data is generated by the second device, possessing high resolution and spatial accuracy, capable of representing the local component shapes of the cultural relic, providing a foundation for subsequent component recognition, semantic segmentation, and local registration. Microscopic point cloud data is generated by the third device, with extremely high resolution, accurately depicting features such as carvings, patterns, micro-cracks, or damage on the surface of the cultural relic, providing precise data for cultural relic restoration analysis and high-fidelity digital display.

[0079] In some possible implementation methods, macroscopic point cloud data (macroscopic point cloud data) is typically acquired by a primary device suitable for rapidly scanning the overall outline and spatial layout of cultural relics, capable of covering large areas or large artifacts. Commonly used primary devices include terrestrial laser scanners, UAV aerial survey systems equipped with high-resolution cameras, and total stations. These devices can acquire the overall three-dimensional point cloud of the cultural relic and its surrounding environment in a relatively short time, providing a spatial reference framework for subsequent mesoscopic and microscopic mapping.

[0080] For meso-level point cloud data (meso-level point cloud data), a second device is used to acquire it. This device is mainly used to record fine structural information at the scale of cultural relic components, and has higher spatial resolution and accuracy than macro-level data. Commonly used second devices include structured light scanners, portable handheld laser scanners, and close-range photogrammetry equipment. These devices can accurately capture the geometric shape and structural features of the surface components of cultural relics, such as the joints of sculptures and the concave and convex shapes of architectural components, providing basic data for semantic segmentation and detailed local modeling of cultural relic components.

[0081] For point cloud data at the microscopic level (micro-point cloud data), a third-party device is used to capture minute details and fine patterns on the surface of the artifact. This third-party device is typically a high-precision microscopic scanning device, such as a miniature structured light scanner, a microscopic 3D scanner, or a high-precision optical scanner, capable of recording features such as microtextures, cracks, wear marks, and carved patterns on the artifact's surface. Because such devices have limited acquisition range, they usually perform detailed scans of specific local areas to ensure that microscopic information is accurately mapped into the overall digital model. In practice, the three types of devices are rationally deployed according to the needs of the target artifact mapping task, and their acquisition parameters and positions are controlled to ensure that macroscopic, mesoscopic, and microscopic point cloud data cover the overall structure and local detailed areas of the target artifact.

[0082] For example, when digitally mapping an ancient building, a layered, multi-device approach can be used. LiDAR (ground-based laser scanner) or a drone equipped with a high-resolution camera can scan the overall appearance of the building, generating macroscopic point cloud data. This data is used to quickly establish the building's overall outline, spatial layout, and reference coordinate system. This stage of data coverage is broad and accurately reflects the building's overall form. For local components such as columns, window frames, eaves, and porches, structured light scanners, portable handheld laser scanners, or multi-view photogrammetry equipment are used to scan and obtain mid-level point cloud data. Mid-level point cloud data has high spatial accuracy and can precisely record the geometric shape and structural features of components. For minute details such as reliefs, carvings, patterns, and local damage on the building surface, miniature high-precision structured light scanners or microscopic 3D scanners are used to scan and generate microscopic point cloud data. Microscopic point cloud data has extremely high resolution and can accurately capture subtle textures and surface micro-damage. Through this layered, multi-device point cloud acquisition, a basic mapping dataset covering both the overall structure and local details can be established.

[0083] S102, based on the surveying performance data of each surveying equipment, associate and quantify the corresponding surveying accuracy attributes for macroscopic point cloud data, mid-level point cloud data, and microscopic point cloud data respectively. Please refer to [link / reference]. Figure 2 .

[0084] To ensure clarity and conciseness in the description of the following embodiments, a brief introduction to the relevant terms is given first.

[0085] Surveying performance data refers to a set of parameters provided by different surveying equipment that reflect its measurement capabilities, accuracy level, and working characteristics, including but not limited to resolution parameters, imaging geometric parameters, positioning and attitude accuracy, distance measurement accuracy, angle measurement accuracy, station registration accuracy, reference working distance, and equipment calibration accuracy.

[0086] Among them, the resolution parameters, imaging geometric parameters, and positioning and attitude accuracy parameters of the first device are used to describe the mapping accuracy of the macroscopic point cloud data. The resolution parameter refers to the smallest spatial spacing that the first device can resolve at the overall scale, and is used to evaluate the spatial detail capture capability of the macroscopic point cloud data; the imaging geometric parameters refer to the spatial coordinate calculation errors caused by optical structure, scanning angle, projection model, etc. during the imaging or scanning process of the first device, and are used to measure the geometric accuracy of the point cloud data; the positioning and attitude accuracy parameters refer to the spatial positioning and attitude measurement accuracy of the first device in the mapping coordinate system, including GPS positioning accuracy, inertial measurement unit attitude accuracy, and device installation stability, and are used to ensure the accurate positioning of the macroscopic point cloud data in the overall spatial coordinate system.

[0087] The ranging accuracy parameters, angle measurement accuracy parameters, and station registration accuracy parameters of the second device are used to quantify the mapping accuracy of the mid-view cloud data. The ranging accuracy parameter refers to the precision of the distance from the second device's measurement point to the surface of the artifact, ensuring the reliability of the spatial scale information of local components. The angle measurement accuracy parameter refers to the error range of the point cloud angle unfolding at the component scale, ensuring the unfolding accuracy of the point cloud in local geometry. The station registration accuracy parameter refers to the accuracy of the alignment of multi-station scan data from the second device in the mapping coordinate system, ensuring that mid-view clouds acquired from different perspectives can be accurately registered to form an overall local model.

[0088] The local resolution parameters, reference working distance parameters, and equipment calibration parameters of the third device are used to quantify the mapping accuracy of the micro-viewpoint cloud data. The local resolution parameter refers to the smallest spatial distance that the third device can resolve within a local detail area, used to capture subtle features such as patterns, cracks, or micro-damage. The reference working distance parameter refers to the spatial positioning accuracy of the third device within a set scanning distance range, used to ensure that the micro-viewpoint cloud data can be accurately mapped to the overall model. The equipment calibration parameters refer to the mapping stability of the third device under different viewing angles and working distances, including optical system calibration, projection accuracy, and scanning repeatability, used to ensure the consistency and reliability of the micro-viewpoint cloud data under multiple angles and conditions.

[0089] Mapping accuracy attributes refer to indicators generated through normalization, weighting, or comprehensive calculation based on mapping performance data. They are used to reflect the comprehensive level of spatial resolution accuracy, error accuracy, and positioning accuracy of different point cloud data.

[0090] In some possible implementation methods, the correlation quantification between macro point cloud data, meso point cloud data, and micro point cloud data and their respective mapping accuracy attributes can be achieved through a multi-step approach.

[0091] For the first device, its main performance parameters include resolution parameters, imaging geometric parameters, and positioning and attitude accuracy parameters. Each parameter undergoes unit unification and normalization; for example, the resolution parameter is mapped to the 0-1 range, the imaging geometric error is mapped using a reciprocal mapping or an exponential decay function, and the positioning and attitude accuracy is also normalized to ensure that all parameters can be compared on the same quantization scale. The three normalized dimensions are combined to form the first mapping accuracy attribute, with each dimension remaining independent. This first mapping accuracy attribute represents the comprehensive characteristics of the macroscopic point cloud data in terms of spatial resolution accuracy, geometric accuracy, and spatial positioning accuracy, and is stored in the metadata of the macroscopic point cloud data.

[0092] For the second device, after acquiring its ranging accuracy parameters, angle measurement accuracy parameters, and station registration accuracy parameters, each parameter is standardized. The ranging accuracy is mapped to a spatial ranging accuracy dimension, the angle measurement accuracy to an angular unfolding accuracy dimension, and the station registration accuracy to a registration accuracy dimension. Through linear or nonlinear function mapping, a second mapping accuracy attribute is formed. Each dimension parameter independently reflects the reliability and accuracy of the central viewpoint cloud data at the local component scale. This accuracy attribute is also appended to the metadata of the central viewpoint cloud data.

[0093] For the third device, after acquiring its local resolution parameters, reference working distance parameters, and device calibration parameters, normalization mapping is performed on each. The local resolution is mapped to a local spatial resolution accuracy dimension, the reference working distance is mapped to a local spatial positioning accuracy dimension, and the device calibration parameters are mapped to a mapping stability accuracy dimension. These three dimensions are combined to form a third mapping accuracy attribute, which fully reflects the comprehensive capabilities of Microview Cloud data in detail capture, positioning accuracy, and multi-angle stability.

[0094] For example, taking the first device as an example, after acquiring the resolution parameters, imaging geometric parameters, and positioning and attitude accuracy parameters, the original parameters are processed to unify the units. For instance, the resolution parameters are converted from millimeters or centimeters to standardized scale values, the imaging geometric error is converted into a spatial error percentage based on the scanning range and angle, and the positioning and attitude accuracy is achieved by mapping GPS accuracy and attitude error to a unified scale. Each parameter is mapped to the interval between 0 and 1 using a normalization function (such as linear normalization or exponential decay function), generating three independent dimensions of values. These three dimensions of values ​​are then combined using a matrix or vector structure to form the first mapping accuracy attribute.

[0095] For the second device, after acquiring the ranging accuracy parameters, angle accuracy parameters, and station registration accuracy parameters, the ranging accuracy parameters are divided by the local component scale to obtain the relative error of the point in the ranging direction; then the angle accuracy parameters are converted into local linear displacement error by multiplying them by the component radius or the distance from the point to the rotation center, and normalized to the interval of 0 to 1 to obtain the angular unfolding accuracy dimension; the registration deviation of the multi-station scan is divided by the local scan coverage size to obtain the registration accuracy dimension, and these three dimensions are combined to form the second surveying accuracy attribute.

[0096] For the third device, the surveying performance data includes local resolution parameters, reference working distance parameters, and device calibration parameters. Dividing the local resolution parameter by the size of the local detail area yields the local spatial resolution accuracy; dividing the positioning error at the reference working distance by the local area size yields the local spatial positioning accuracy; calculating the mean and standard deviation of the deviations from multi-view scans of the same area and normalizing them forms the surveying stability accuracy dimension. Combining these three dimensions forms the third surveying accuracy attribute.

[0097] In practical applications, such as when digitally surveying an ancient building, macroscopic point cloud data is generated by a ground-based laser scanner or an unmanned aerial vehicle (UAV) surveying system, and the first surveying accuracy attribute is formed through the above-mentioned normalization and mapping methods; mid-level point cloud data is generated by a structured light scanner or a portable handheld laser scanner, and the second surveying accuracy attribute is formed through distance measurement, angle measurement, and station registration accuracy mapping; microscopic point cloud data is generated by a miniature high-precision structured light scanner, and the third surveying accuracy attribute is formed through local resolution, reference working distance, and equipment calibration parameter processing.

[0098] S103, based on the Zhongguan Cloud data, perform semantic segmentation of the cultural relic components to determine multiple corresponding component regions.

[0099] To ensure clarity and conciseness in the description of the following embodiments, a detailed introduction of the relevant terms is given first.

[0100] Semantic segmentation of cultural relic components refers to the structured analysis of point cloud data, classifying each point or set of points according to its corresponding cultural relic component type to form semantically meaningful component units. Semantic segmentation considers not only the geometric features of points (such as curvature, normal vector, and surface smoothness) but also spatial location, point cloud density, local connectivity, and the topological relationships between cultural relic components. The result of segmentation is the division of the point cloud into several functionally or structurally independent units, such as column units, beam units, window frame units, and sculpture body units, each unit corresponding to a specific component type and spatial boundary.

[0101] A component region refers to a subset of point clouds corresponding to a specific component of a cultural relic, formed after semantic segmentation. Each component region is a three-dimensional spatial set containing the complete geometric shape and boundary information of the component. The definition of a component region includes not only its spatial location but also its topological relationships and local geometric features, making it suitable for subsequent local registration, fine modeling, local damage detection, or texture mapping. The data structure of a component region typically includes a set of point cloud coordinates, surveying accuracy attributes, component type identifiers, and spatial boundary descriptions.

[0102] The main structural unit refers to the primary structural unit within a cultural relic, either as a whole or a partial component, that serves as the basis for segmentation and is the initial criterion for semantic segmentation. Main structural units typically possess clear spatial connectivity and morphological characteristics, such as the columns, beams, and walls of ancient buildings, or the main torso and head of a sculpture. In the segmentation process, the main structural unit is first identified, and then further subdivided into multiple component regions based on topological rules and morphological characteristics.

[0103] Topological rules for cultural relic components refer to segmentation guidelines formulated based on the spatial connectivity, component function, and geometric shape of the cultural relic structure. These rules are used to constrain and optimize semantic segmentation results. For example, the top of a column must be connected to a beam, window frames must be embedded in a wall, and the head and torso of a sculpture must be connected but separated from the base. Topological rules can be combined with spatial connectivity analysis of point clouds, point-to-point distance relationships, and spatial proximity constraints between components to ensure that the segmented component regions conform to the actual structure and function, while reducing erroneous segmentation or over-segmentation.

[0104] In some possible implementation methods, the main spatial units of the target artifact are extracted from the midpoint point cloud data, laying the foundation for subsequent component segmentation. Local geometric features, such as curvature, normal vector, and point density, are calculated for each point. Curvature is used to determine whether a point lies on a plane, curved surface, or sharp edge; the normal vector is used to determine the consistency of local surface orientation; and the point density is used to assess whether the point cloud in that area is dense enough to support reliable region growing. Through spatial connectivity analysis, such as using an octree structure, the set of neighboring points for each point is determined, and whether these points belong to the same connected unit is determined based on whether the distance between points within the neighborhood is below a preset threshold. Based on this, density clustering or region growing methods can be used to aggregate dense points with consistent geometric features into preliminary main structural units. Each unit retains its point cloud set and corresponding mapping accuracy attributes for use in subsequent processing.

[0105] The initial main units are corrected based on the actual structural characteristics of the cultural relic. Through morphological examination, the spatial bounding box, voxel dimensions, and main orientations of each main unit are calculated, and the aspect ratio and curvature distribution of the unit are evaluated to ensure they conform to the typical scale and morphological characteristics of the cultural relic components. If units are found to be too large or too small, or have abnormal curvature distributions, merging or segmentation operations can be performed based on the geometric relationships of neighboring units to avoid missegmentation or over-merging. The spatial proximity and topological relationships between main units are also analyzed, such as whether adjacent units have logical connectivity, hierarchical subordination, or symmetry, and the unit boundaries are adjusted accordingly to make the segmentation results more consistent with the actual structural layout of the cultural relic. The topologically constrained main structural units are further refined into specific component regions, and local geometric feature analysis (such as curvature changes and boundary shapes) is used to accurately delineate the boundaries of each component, resulting in multiple component regions corresponding to the target cultural relic structure.

[0106] For example, taking the digital mapping of an ancient building as an example, the second device is controlled to collect point cloud data of local components such as columns, beams, window frames, and door frames, obtaining central point cloud data. In the preliminary analysis of the central point cloud data, the local geometric features of each point are calculated, including curvature, normal vector, and point density. Based on these features, points with similar geometric features and spatial connectivity are aggregated using region growing or density clustering methods to form preliminary main structural units. For example, the point cloud of the column area usually has a columnar spatial distribution, with relatively consistent curvature and normal vector directions, thus it can be identified as a main structural unit; planar components such as window frames or door frames are identified as independent main structural units through region growing based on the point-to-plane fitting error.

[0107] After obtaining the initial main structural units, corrections and optimizations are performed based on the topological rules of the cultural relic components. The spatial bounding box, voxel dimensions, and main orientations of each main unit are calculated, and the typical scale and morphological characteristics of the cultural relic components are compared. If column or beam units are found to be too large or too small, or if the curvature distribution of planar components is abnormal, they can be merged or segmented based on the geometric relationships of neighboring units. Then, the spatial proximity and topological relationships between the main units are analyzed. For example, columns must be connected to beams, window frames are embedded in walls, and the head of a sculpture is connected to the torso but separated from the base. Unit boundaries are adjusted according to these rules to ensure that the segmentation results conform to the actual layout of the cultural relic structure.

[0108] The topologically constrained main structural units are further refined into specific component regions. Through local geometric feature analysis, such as curvature variations, boundary contours, and corner point positions, the spatial boundaries of each component unit are precisely delineated. For example, the column component region retains the point cloud set of the entire column, the beam component region contains the complete point cloud and boundary information of the beam, and the window frame component region contains the planar point cloud of the window frame and its boundary data embedded in the wall.

[0109] S104. In each component region, the mid-view cloud data is registered to the macro point cloud data to establish the first mapping fusion region, and the micro view cloud data and mid-view cloud data are locally registered in this region to obtain the second mapping fusion region.

[0110] To clearly describe the following embodiments, relevant terms will first be explained in detail and technically expanded.

[0111] Registration refers to the precise alignment of point cloud data sets from different surveying devices within the same spatial coordinate system, ensuring that the same spatial location corresponds and matches in different point clouds. The registration process typically consists of two stages: coarse registration and fine registration. Coarse registration is used to initially align the translation, rotation, and scale relationships of the point clouds. It is usually achieved based on the geometric center, principal direction, or feature point matching of the point clouds, such as centroid matching after voxel mesh downsampling or keypoint descriptor matching, to quickly obtain approximate alignment results. The fine registration stage, based on iterative optimization algorithms, finely adjusts the rotation matrix and translation vector between point clouds by minimizing the sum of squares of point-to-point or point-to-surface distances until the spatial deviation of the point clouds meets the preset accuracy requirements.

[0112] The first mapping fusion region refers to the spatial area formed by aligning the mid-level point cloud data with the macro-level point cloud data after coarse and fine registration within each component area. In this region, the mid-level point cloud data and the macro-level point cloud data have spatial overlap, forming a multi-level point cloud set with macro-scale and local-scale data.

[0113] The second mapping fusion region is a spatial region formed by locally registering micro-viewpoint cloud data and meso-viewpoint cloud data based on the first mapping fusion region. Micro-viewpoint cloud data usually contains high-precision textures or minute features of component surfaces. Through local feature matching (such as high curvature points, edge points, or texture feature points) and local optimization, the micro-viewpoint cloud data is accurately aligned with the meso-viewpoint cloud data, ultimately resulting in a multi-scale fused point cloud set.

[0114] In some possible implementations, coarse registration is performed between the mid-point cloud data and the macro-point cloud data for each component region. At this stage, the spatial position and orientation features of the component region in the mid-point cloud data are obtained by calculating the geometric center, bounding box, and principal direction vector. The macro-point cloud data already possesses mapping accuracy attributes in the global coordinate system and can be used as a reference benchmark. By translating the geometric center of the mid-point cloud to the corresponding center position in the macro-point cloud data and calculating a preliminary rotation matrix based on the principal direction vector, the mid-point cloud data is mapped to the spatial position in the macro-point cloud data. To increase stability, a set of point cloud center points can be generated using voxel mesh downsampling, and keypoint matching can be performed using nearest neighbors or feature descriptors to achieve coarse registration.

[0115] Based on the coarse registration, the fine registration stage begins. This stage primarily employs the iterative nearest point algorithm or its improved version. The fine registration process iteratively calculates the sum of squared distances between each point in the mid-view point cloud data and the nearest point in the macro point cloud data, and optimizes the rotation matrix and translation vector by minimizing this distance error. During the iteration process, a convergence threshold can be set, such as stopping the iteration when the average point distance change is less than a specified millimeter or the angle change is less than a specified radian, to ensure the stability of the registration results.

[0116] Within the first mapping fusion region of each component area, micro-viewpoint cloud data and meso-viewpoint cloud data are locally registered. Local registration extracts high-curvature areas, edge points, corner points, or texture-significant feature points from the micro-viewpoint cloud data as local features to ensure detailed structural capture. The meso-viewpoint cloud data also extracts similar feature points within the corresponding component area. A preliminary correspondence is established through feature point descriptor matching, and a coarse-aligned local rotation matrix and translation vector are calculated. The point correspondence between the micro-viewpoint cloud data and the meso-viewpoint cloud data is iteratively adjusted until the local registration error meets the actual mapping requirements, thus obtaining the second mapping fusion region. This region not only preserves the macroscopic and mesoscopic spatial morphology but also accurately reflects microscopic details, providing a high-precision foundation for subsequent multi-scale fusion.

[0117] S105, within the second surveying and mapping fusion area, based on the surveying accuracy attribute, macroscopic point cloud data, mid-level point cloud data, and microscopic point cloud data at the same spatial location are fused to obtain the surveying data of the target cultural relic. A digital surveying and mapping model of the cultural relic is then generated based on this data. Please refer to [link / reference]. Figure 3 and Figure 4 .

[0118] To ensure clarity and conciseness in the description of the following embodiments, a detailed introduction of the relevant terms is given first.

[0119] Accuracy metrics are numerical indicators that unify and quantify the mapping accuracy attributes of point clouds from different sources and at different scales. They are used to measure the credibility and reliability of point cloud data during point cloud fusion and processing. They not only reflect the spatial resolution capability of the mapping equipment itself, but also integrate factors such as the mapping environment, scanning distance, point cloud density, measurement noise, and data acquisition methods, thus providing a comparable reference for multi-scale point cloud fusion.

[0120] Accuracy metrics are typically obtained by mapping the original mapping accuracy attributes (such as the first mapping accuracy attribute for macroscopic point cloud data, the second mapping accuracy attribute for mesoscopic point cloud data, and the third mapping accuracy attribute for microscopic point cloud data) to the same quantization scale. Mapping methods can be based on error estimation models. For example, for laser-scanned point clouds, scanner calibration error, point-to-point repeatability bias, or system noise variance can be used for quantization; for photogrammetric point clouds, pixel resolution, camera intrinsic error, and spatial reprojection error can be considered; and for structured light or close-range scanned point clouds, projection accuracy and ranging noise can be combined for evaluation.

[0121] In some possible implementation methods, within each second mapping fusion region, macroscopic point cloud data, corresponding mid-level point cloud data, and microscopic point cloud data are acquired, and their mapping accuracy attributes within that region are extracted, including a first mapping accuracy attribute, a second mapping accuracy attribute, and a third mapping accuracy attribute. These three attributes are then converted into accuracy indices within the second mapping fusion region, namely, a first accuracy index, a second accuracy index, and a third accuracy index, to unify the quantification scale and facilitate subsequent fusion processing.

[0122] A three-dimensional voxel grid covering the target cultural relic is established within the second mapping fusion area. Each voxel is used to aggregate point cloud data of different scales at the same spatial location. Based on the voxel structure, points within the same voxel from macroscopic, mesoscopic, and microscopic point cloud data are aggregated to form local point sets. Each local point set contains point cloud points at different scales and their corresponding accuracy indicators, providing basic data units for multi-scale fusion.

[0123] For each local point set, an initial fusion is performed to generate initial structural data. This involves comparing the first precision index of the macroscopic point cloud data with the second precision index of the mid-range point cloud data. The data with higher precision is selected as the structural baseline point, while the data with lower precision is used as structural correction points for fusion. The fusion can employ voxel-weighted averaging, Gaussian weighting, or least squares optimization methods to ensure that high-precision data plays a dominant role in the spatial structure, while simultaneously correcting the structure of low-precision data to guarantee the accuracy of the overall spatial morphology.

[0124] A second fusion is performed based on the initial structural data to incorporate detailed information from the micro-view cloud data. If the third precision index of the micro-view cloud data is higher than the second precision index of the mid-view cloud data, the micro-view cloud data is used as a detail correction point and locally weighted fused with the initial structural data to enhance minute features, such as sculpted patterns, cracks, or decorative details. If the precision of the micro-view cloud data is lower than that of the mid-view cloud data, the mid-view cloud data is retained as the primary source of detail, while the micro-view cloud data serves as a supplement. This hierarchical fusion strategy ensures high-precision reconstruction of both structural integrity and local details.

[0125] After completing the two-step fusion of all local point sets, the fused point clouds are integrated, and duplicate, outlier, or noisy points are removed to generate complete fused point cloud data. Based on this fused point cloud data, a digital mapping model can be further generated, including surface reconstruction, normal vector estimation, texture mapping, and geometric optimization, thereby obtaining a digital mapping model of the target cultural relic, which can be used for cultural relic protection, restoration, display, and scientific analysis.

[0126] In some possible implementation methods, within each second mapping fusion region, macroscopic point cloud data, mid-level point cloud data, and microscopic point cloud data are hierarchically sorted according to their corresponding accuracy indicators, with high-precision point cloud data placed at the top and low-precision data at the bottom, to ensure the dominant role of high-precision data in the spatial structure during the fusion process. For each local point set, a dynamic weighting function is introduced. This function not only considers the accuracy indicator of the point itself but also combines local geometric consistency, point density, and spatial deviation of neighboring points to assign weighting coefficients to each point, thereby dynamically adjusting the influence of each point on the final result during fusion.

[0127] The fusion process can be divided into three layers. The first layer is dominated by macro point cloud data to achieve coarse alignment and preliminary fusion of the overall spatial structure of the target cultural relic, while using low-precision point cloud data to supplement the structure to avoid information loss. The second layer is dominated by meso point cloud data to make local fine adjustments to the preliminary structure, optimize component boundaries, connection relationships and geometric continuity to make meso-scale features more accurate. The third layer uses micro point cloud data as detail correction points to enhance micro features, such as carving textures, cracks or decorative lines, through accuracy indicators and local geometric consistency weights, to ensure high-precision reconstruction of micro structures.

[0128] In some possible implementation methods, a series of digital modeling steps can be used to generate digital mapping models of cultural relics. The fused point cloud data is then reconstructed into surfaces using voxel-based Poisson reconstruction, transforming the discrete point cloud into a continuous 3D surface model while preserving local details. The normal vector and curvature information of each surface unit are calculated to ensure the geometric accuracy of the model during lighting, rendering, or subsequent analysis. For areas with patterns, cracks, or carvings, high-precision micro-point cloud data can be combined for texture mapping and detail enhancement, generating digital models with realistic surface features. In practical applications, the generated digital mapping models have multiple functions. For example, in cultural relic conservation and restoration, they can be used to accurately assess damaged areas and guide restoration techniques; in museum displays and virtual exhibitions, they enable 3D visualization and interactive operation, allowing the public to observe the relics in detail without physical contact; in scientific research, the model can serve as an important data foundation for analyzing the manufacturing processes, material distribution, stylistic characteristics, and historical evolution of cultural relics. This model can also serve as a long-term digital archive, supporting dynamic monitoring of the condition of cultural relics and comparison with historical data, enabling preventative protection. In this way, digital mapping models not only restore the geometric and detailed features of cultural relics, but also provide reliable, quantifiable, and reusable digital tools for the protection, display, and scientific research of cultural relics.

[0129] This application utilizes a layered, multi-scale point cloud acquisition and fusion mechanism, combining macroscopic, mesoscopic, and microscopic point cloud data with surveying accuracy attributes to achieve high-precision reconstruction of the overall structure and local details of cultural relics. Layered fusion based on surveying accuracy attributes ensures the dominant role of high-precision data in representing the spatial structure, while introducing local geometric consistency and point density constraints to optimize detail presentation. The fused point cloud undergoes surface reconstruction, normal vector calculation, and texture mapping to generate a digital surveying model of cultural relics suitable for protection, restoration, display, and scientific research. This improves the accuracy, completeness, and application value of digital surveying of cultural relics.

[0130] The foregoing has shown and described the basic principles, main features, and advantages of this application. Those skilled in the art should understand that this application is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this application. Various changes and modifications can be made to this application without departing from the spirit and scope thereof, and all such changes and modifications fall within the scope of this application as claimed. The scope of protection of this application is defined by the appended claims and their equivalents.

Claims

1. A digital mapping method for cultural relics based on multi-scale point cloud fusion, characterized in that, The method includes: During the surveying and mapping of the target cultural relic, different surveying and mapping equipment are controlled to collect macroscopic point cloud data, mesoscopic point cloud data, and microscopic point cloud data of the target cultural relic. Macroscopic point cloud data refers to the point cloud data obtained by the surveying and mapping equipment, which is used to record the overall outline and spatial layout of the cultural relic. Mesoscopic point cloud data refers to the point cloud data obtained by the surveying and mapping equipment, which is used to record the structural information of the cultural relic components. Microscopic point cloud data refers to the point cloud data obtained by the surveying and mapping equipment, which is used to record the minute features of the surface of the cultural relic. Based on the surveying performance data of the surveying equipment, the corresponding surveying accuracy attributes are associated and quantified for the macro point cloud data, mid-point point cloud data, and micro point cloud data, respectively. The surveying accuracy attributes are used as the basis for point cloud data selection when fusing point cloud data. Based on the aforementioned Zhongguan Cloud data, semantic segmentation of the cultural relic components is performed to determine multiple component regions corresponding to the target cultural relic: In each of the component regions, the mid-view cloud data is registered to the macro point cloud data to establish a first mapping fusion region. Based on the first mapping fusion region, the micro point cloud data is locally registered with the corresponding mid-view cloud data to obtain a second mapping fusion region. In the second mapping fusion area, based on the mapping accuracy attribute, the macro point cloud data, the mid-point cloud data and the micro point cloud data at the same spatial location are fused accordingly to obtain the mapping data of the target cultural relic; Based on the surveying data, a digital surveying model of the target cultural relic is generated.

2. The method of claim 1, wherein, The surveying equipment includes a first device, a second device, and a third device. The surveying performance data based on the surveying equipment is used to associate and quantify the corresponding surveying accuracy attributes of the macroscopic point cloud data, mid-level point cloud data, and microscopic point cloud data, including: Based on the mapping performance data of the first device, the first mapping accuracy attribute corresponding to the macro point cloud data is quantized and associated. Based on the mapping performance data of the second device, the second mapping accuracy attribute corresponding to the cloud data of the central viewpoint is quantized. Based on the mapping performance data of the third device, the corresponding third mapping accuracy attribute is quantized and associated with the micro-viewpoint cloud data.

3. The method of claim 2, wherein, The mapping performance data based on the first device is the first mapping accuracy attribute corresponding to the macroscopic point cloud data association quantization, including: The first mapping performance parameters of the first device are obtained, including resolution parameters, imaging geometry parameters, and positioning and attitude determination accuracy parameters. Based on the resolution parameters, determine the spatial resolution accuracy of the macro point cloud data at the overall scale of the target cultural relic; Based on the imaging geometric parameters, the error accuracy of the macroscopic point cloud data in the spatial coordinate calculation process is determined; Based on the positioning and attitude determination accuracy parameters, the spatial positioning accuracy of the macro point cloud data in the mapping coordinate system is determined; The spatial resolution accuracy, the error accuracy, and the spatial positioning accuracy are normalized to generate the first mapping accuracy attribute corresponding to the macroscopic point cloud data.

4. The method of claim 2, wherein, The mapping performance data based on the second device is the second mapping accuracy attribute corresponding to the central viewpoint cloud data association quantization, including: The second surveying performance parameters of the second device are obtained, including distance measurement accuracy parameters, angle measurement accuracy parameters, and station registration accuracy parameters. Based on the ranging accuracy parameters, determine the spatial ranging accuracy of each point in the central point cloud data along the ranging direction; Based on the aforementioned angular measurement accuracy parameters, the angular unfolding error accuracy of the midpoint cloud data at the component scale is determined; Based on the site registration accuracy parameters, the registration accuracy of the central point cloud data in the mapping coordinate system is determined; The spatial ranging accuracy, the angular unfolding error accuracy, and the registration accuracy are normalized to generate the second mapping accuracy attribute corresponding to the central point cloud data.

5. The method according to claim 2, characterized in that, The mapping performance data based on the third device is the third mapping accuracy attribute corresponding to the micro-viewpoint cloud data association quantization, including: Acquire the third mapping performance parameters of the third device, which include local resolution parameters, reference working distance parameters, and device calibration parameters; Based on the local resolution parameters, the local spatial resolution accuracy of the micro-view cloud data within the local detail area of ​​the target cultural relic is determined; Based on the reference working distance parameter, determine the local spatial positioning accuracy of the micro-viewpoint cloud data within the reference working distance range; Based on the device calibration parameters, the mapping stability accuracy of the micro-viewpoint cloud data under different viewing angles and working distances is determined. The local spatial resolution accuracy, the local spatial positioning accuracy, and the mapping stability accuracy are normalized to generate a third mapping accuracy attribute corresponding to the micro-viewpoint cloud data.

6. The method of claim 2, wherein, Within the second mapping fusion area, based on the mapping accuracy attribute, the macroscopic point cloud data, the mid-level point cloud data, and the micro-level point cloud data at the same spatial location are fused accordingly to obtain the mapping data of the target cultural relic, including: The first surveying accuracy attribute, the second surveying accuracy attribute, and the third surveying accuracy attribute are respectively converted into accuracy indicators under the second surveying fusion area; Within the second mapping fusion area, a three-dimensional spatial voxel mesh covering the target cultural relic is established; In the three-dimensional spatial voxel mesh, the macroscopic point cloud data, the mesoscopic point cloud data, and the microscopic point cloud data at the same spatial location are aggregated to form multiple local point sets; For each local point set, a first fusion is performed based on the aforementioned accuracy index to obtain initial fused data; Based on the initial fusion data, a second fusion is performed based on the accuracy index to obtain fused point cloud data; The fused point cloud data is integrated to obtain the mapping data of the target cultural relic.

7. The method of claim 6, wherein, The step of uniformly converting the first surveying accuracy attribute, the second surveying accuracy attribute, and the third surveying accuracy attribute into accuracy indicators under the second surveying fusion region includes: The first mapping accuracy attribute is converted into a first accuracy index under the second mapping fusion area; The second mapping accuracy attribute is converted into a second accuracy index under the second mapping fusion area; The third mapping accuracy attribute is converted into a third accuracy index under the second mapping fusion area.

8. The method of claim 7, wherein, For each local point set, a first fusion is performed based on the accuracy index to obtain initial structural data, including: If the first precision index of the macro point cloud data is greater than the second precision index of the mid-level point cloud data, then the macro point cloud data is selected as the structural reference point, and the mid-level point cloud data is selected as the structural correction point for the first fusion to obtain the initial structural data. If the first precision index of the macro point cloud data is less than the second precision index of the mid-level point cloud data, then the mid-level point cloud data is selected as the structural reference point, and the macro point cloud data is used as the structural correction point for the first fusion to obtain the initial structural data.

9. The method according to claim 7, characterized in that, The step of performing a second fusion based on the initial structural data and the accuracy index to obtain fused point cloud data includes: If the third precision index of the micro-viewpoint cloud data is greater than the second precision index of the mid-viewpoint cloud data, then the micro-viewpoint cloud data is selected as a detail correction point and fused with the initial structure data for a second time to obtain fused point cloud data. If the third precision index of the micro-viewpoint cloud data is less than the second precision index of the mid-viewpoint cloud data, then the mid-viewpoint cloud data is selected as the detail correction point and fused with the initial structure data for the second time to obtain fused point cloud data.

10. The method of claim 1, wherein, The step of performing semantic segmentation of the target cultural relic based on the central cloud data to determine multiple component regions corresponding to the target cultural relic includes: Identify and segment the main structural units of the target cultural relic from the aforementioned central cloud data; Based on the main structural unit, and combined with the preset topological rules for cultural relic components, the target cultural relic is divided into multiple component regions.

Images