Historical building-oriented real-scene three-dimensional GIS-based health monitoring data prediction method

Abstract

The invention relates to a historical building-oriented real-scene three-dimensional GIS-based health monitoring data prediction method, and belongs to the technical field of building monitoring, and the method comprises the steps: obtaining modeling data and monitoring data, and constructing a building geometric model based on the modeling data; matching the monitoring data with the building geometric model so as to map the monitoring data to a corresponding position of the building geometric model; based on a preset prediction and early warning model, learning the monitoring data, and predicting and outputting the change trend of the monitoring data in a specified time period in the future; and analyzing monitoring data corresponding to the building geometric model based on a pre-constructed health analysis model, and outputting a structural damage assessment result of the building geometric model. The method has the advantages that the historical building is efficiently monitored, potential health problems of the historical building are found in time, and the maintenance efficiency of the historical building is improved.

Description

Technical Field

[0001] This application relates to the technical field of building monitoring, and in particular to a prediction method for health monitoring data of historical buildings based on real-scene three-dimensional GIS. Background Art

[0002] With the acceleration of the urbanization process, the protection and restoration of historical buildings have become increasingly important. Traditional building health monitoring methods mainly rely on regular manual inspections and simple sensor monitoring. When the historical building structures to be monitored are relatively large and the monitoring scope is wide, a large amount of manpower and material resources are required to complete one round of monitoring. Moreover, the relatively long monitoring cycle corresponding to this monitoring process will also result in a long interval between adjacent two monitoring processes, thus making it impossible to timely discover potential structural problems of historical buildings and even causing delays in the maintenance of historical buildings. Therefore, there is a need for improvement. Summary of the Invention

[0003] In order to achieve efficient monitoring of historical buildings, timely discover potential health problems of historical buildings, and improve the maintenance efficiency of historical buildings, this application provides a prediction method for health monitoring data of historical buildings based on real-scene three-dimensional GIS.

[0004] In a first aspect, this application provides a prediction method for health monitoring data of historical buildings based on real-scene three-dimensional GIS, including: Obtain modeling data and monitoring data, and construct a building geometric model based on the modeling data; match the monitoring data with the building geometric model to map the monitoring data to the corresponding positions of the building geometric model; Based on a preset prediction and early warning model, learn the monitoring data and predict and output the change trend of the monitoring data within a specified future time period; Based on a pre-constructed health analysis model, analyze the monitoring data corresponding to the building geometric model, and output the structural damage assessment result of the building geometric model.

[0005] By adopting the above technical solutions, a corresponding three-dimensional building geometric model is constructed for a historical building based on the on-site collected modeling data to achieve a refined reproduction of the historical building. Then, the detection data is mapped to the corresponding positions of the building geometric model, so that the monitoring personnel can more intuitively know the monitoring data of each part of the historical building in combination with this building geometric model. In addition, this application also pre-constructs a prediction and early warning model and a health analysis model. Through the health analysis model, the monitoring data can be automatically analyzed and the determination and assessment of whether the historical building structure is damaged can be realized. And through the prediction and early warning model, the prediction of future monitoring data can be realized based on the current monitoring data, so as to predict and early warn of possible potential safety hazards based on this change trend.

[0006] Optionally, the modeling data includes multi-view image data; Constructing the building geometric model based on the modeling data includes: Obtain multi-view image data of the historical building, extract key feature points in the multi-view image data through a preset feature extraction algorithm, generate multi-view point cloud data, and use a point cloud fusion technology based on geometric constraints to fuse and process the point cloud data to construct an initial point cloud model of the historical building; Based on the grid reconstruction technology of semantic segmentation, perform grid processing on the initial point cloud model to obtain a building geometric model.

[0007] By adopting the above technical solution, multi-view image data of the historical building can be obtained through the drone oblique photography and the ground close-range photography technology, and then key feature points are extracted from the multi-view image data. These key feature points can be regarded as the specific structure of the historical building, thereby realizing the preliminary matching of the image data and the historical building structure. Then, point cloud data is generated based on the multi-view image data, and an initial point cloud model is constructed. Next, the grid reconstruction technology of semantic segmentation is used to perform grid processing on the initial point cloud model, and the grid model is optimized according to the semantic information of the historical building. Finally, a high-precision real-scene three-dimensional fine model of the historical building is obtained, improving the refinement degree of the constructed building geometric model.

[0008] Optionally, the monitoring data includes environmental data and structural data; The method further includes: Based on the real-time acquired monitoring data, by analyzing the change trends of the corresponding structural data when different environmental data are obtained, infer the material properties and / or structural features at the mapping positions of the structural data on the building geometric model, generate and output optimization suggestions based on the inference results, and perform optimization processing based on the feedback results of the monitoring personnel on the optimization suggestions; the optimization suggestions at least include correcting and optimizing the monitoring data at the mapping positions of the building geometric model, and / or obtaining on-site survey data at the corresponding mapping positions of the building geometric model.

[0009] By adopting the above technical solution, since the building geometric model corresponding to the historical building is constructed based on multi-view image data; therefore, the building geometric model constructed above mainly replicates the historical building in terms of its external structure, and it is not easy to comprehensively and accurately determine the building materials or structural characteristics of the building geometric model during construction. For this reason, the present application proposes that the monitoring data also includes environmental data and structural data, and the material properties and building techniques of the structure at the mapping position are inferred by analyzing the changes in the structural data under different environmental conditions (using the environmental data to characterize the corresponding environment, and the environmental data may specifically include temperature data, humidity data, etc.), such as key parameters such as the thermal expansion coefficient and elastic modulus of the material can be inferred, and the optimization of the building geometric model is achieved based on the inferred results, such as using more refined sensors to conduct on-site surveys of the historical building to verify the aforementioned inferred results, or performing optimization processing on the building geometric model, such as correcting the material properties and structural characteristics of the corresponding mapping position on the building geometric model based on the aforementioned inferred results, thereby ultimately improving the accurate construction of the building geometric model of the historical building.

[0010] Optionally, the optimizing process based on the feedback result of the monitoring personnel on the optimization suggestion includes: Obtain the on-site survey data of the corresponding mapping position of the building geometric model, and determine whether the corresponding mapping position of the building geometric model meets the layering condition according to the survey data obtained from the on-site survey; If it meets the condition, perform layering processing on the corresponding mapping position to obtain a multi-level structure model; wherein, the multi-level structure model sequentially includes a minimum component layer, a combined module layer, and an overall structure layer from the inside out, and there is a progressive inclusion and being-included relationship between the minimum component layer, the combined module layer, and the overall structure layer; Describe the association relationship between different levels included in the multi-level structure model based on a preset representation form, so that the monitoring personnel can know the association relationship.

[0011] By adopting the above technical solution, by deploying refined sensors (such as micro-strain sensors, 3D laser scanners, six-axis force sensors, etc.) to further refine the survey of specific mapping positions of the building geometric model, if a certain mapping position of the building geometric model is constructed by complex structural characteristics or is formed by a composite of multiple materials, then it is considered to meet the layering condition. At this time, the present application proposes to perform layering processing on this mapping position to form a multi-level structure model, and there is an inclusion and being-included relationship between adjacent levels, so as to split and refine the building geometric model in a multi-level progressive manner, enabling the monitoring personnel to more intuitively understand and analyze the specific structure of the historical building.

[0012] Optionally, the method further includes: For the target mapping positions with a multi-level structure model, a data conduction scheme is executed through a preset data conduction engine; wherein, the data conduction scheme is to convert the surveyed data obtained corresponding to the direct level through the data conduction engine according to predefined data transfer rules and map the converted data to the indirect level; the direct level and the indirect level both belong to the levels included in the multi-level structure model, and are in an inclusion or being included relationship; the data transfer rules at least include sequentially transferring according to the hierarchical order from the inside out, and / or sequentially transferring according to the hierarchical order from the outside in.

[0013] By adopting the above technical solution, based on the correlation relationship between each level in the multi-level structure model, a data conduction mechanism is developed to achieve progressive data conduction between different levels. For example, taking the overall structure layer as the direct level and the minimum component layer and the combined module layer as the indirect levels, the boundary conditions and load information are sequentially transferred from the overall structure layer to the combined module layer and the minimum component layer, so that the stress conditions of the overall structure layer can be transferred to the combined module layer and the minimum component layer, and then the stress conditions of the combined module layer and the minimum component layer can be known. Similarly, the damage information and deformation data of the minimum component layer can also be sequentially transferred to the combined module layer and the overall structure layer, so that the influence of the cracks generated in the minimum component layer on the stability of the combined module layer and the overall structure layer can be known. In summary, through spatial indexing, data mapping, data conversion, and the conduction engine, the efficient and accurate transfer of data between different levels is finally achieved, providing a multi-scale analysis basis for the health diagnosis of large-span and high-density wooden building complexes.

[0014] Optionally, after converting the surveyed data obtained corresponding to the direct level and mapping the converted data to the indirect level, it further includes: Analyze whether the surveyed data belongs to preset damaged data, wherein the preset damaged data at least includes crack data and deformation data; If so, based on the surveyed data obtained at the direct level, calculate the influence range of the surveyed result corresponding to the surveyed data on the direct level; map the influence range on the direct level to the indirect level and calculate the influence range on the indirect level; wherein, the influence range at least includes the influence area and the influence region; Formulate a monitoring strategy for the corresponding mapping position based on the influence range, and the monitoring strategy at least includes the monitoring frequency.

[0015] By adopting the above technical solution, when the survey data is data used to describe damage conditions such as cracks and deformations, then such survey data belongs to the preset damage data. This application proposes to analyze the influence range of such survey data on the corresponding mapped position based on such survey data, that is, to analyze and obtain the influence area and influence area. And when the corresponding mapped position is a multi-level structure model, the influence range of each level will be analyzed layer by layer, so as to improve the accuracy of the health diagnosis of historical buildings, and also provide a scientific basis for structural optimization, real-time monitoring and early warning, which helps to realize the progressive protection of large-span and high-density wooden building complexes.

[0016] Optionally, the method further includes: Receiving a virtual display instruction, displaying the building geometric model in a pre-constructed virtual environment, and decomposing the building geometric model with the level as the smallest splitting unit based on the display granularity included in the virtual display instruction; Displaying the building geometric model corresponding to the health level based on the health level in the virtual display instruction.

[0017] By adopting the above technical solution, the virtual environment can specifically be a VR / AR environment. Monitoring personnel can enter the virtual environment based on a preset virtual interaction device to achieve an intuitive view of the building geometric model. In addition, this application also proposes optional functions for display granularity and health level, so that monitoring personnel can view historical buildings under different health levels with different display granularities.

[0018] In a second aspect, this application provides a prediction system for health monitoring data of historical buildings based on real-scene three-dimensional GIS, adopting the following technical solution: A geometric model construction module, configured to obtain modeling data and monitoring data, construct a building geometric model based on the modeling data; match the monitoring data with the building geometric model to map the monitoring data to the corresponding position of the building geometric model; A data change prediction module, configured to learn the monitoring data based on a preset prediction and early warning model and predict and output the change trend of the monitoring data within a specified future period; A building health assessment module, configured to analyze the monitoring data corresponding to the building geometric model based on a pre-constructed health analysis model, and output a structural damage assessment result for the building geometric model.

[0019] In a third aspect, this application provides a prediction device for health monitoring data of historical buildings based on real-scene three-dimensional GIS, including a memory and a processor, and a computer program capable of being loaded and executed by the processor as described in any one of the first aspects is stored on the memory.

[0020] In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program that can be loaded and executed by a processor to perform the method described in any one of the first aspects.

[0021] In summary, the present application includes the following beneficial technical effects: Based on the modeling data collected on-site, a corresponding three-dimensional building geometric model is constructed for the historical building to achieve a refined reproduction of the historical building. Then, the detection data is mapped to the corresponding positions of the building geometric model, so that the monitoring personnel can more intuitively obtain the monitoring data of each part of the historical building in combination with the building geometric model. In addition, the present application also pre-constructs a prediction and early warning model and a health analysis model. Through the health analysis model, the monitoring data can be automatically analyzed to determine and evaluate whether there is structural damage to the historical building, and through the prediction and early warning model, the future monitoring data can be predicted based on the current monitoring data, so as to predict and early warn of possible potential safety hazards in advance based on the change trend. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings without creative efforts based on these drawings.

[0023] Figure 1 is a flowchart of a method for predicting health monitoring data of historical buildings based on real-scene three-dimensional GIS according to an embodiment of the present application.

[0024] Figure 2 is a structural block diagram of a method for predicting health monitoring data of historical buildings based on real-scene three-dimensional GIS according to an embodiment of the present application.

[0025] Description of the reference numerals: 201, geometric model construction module; 202, data change prediction module; 203, building health assessment module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] The following will further describe the present application in detail Figure 1-2 with reference to the attached

[0027] An embodiment of the present application discloses a prediction method for health monitoring data based on real - scene three - dimensional GIS for historical buildings (hereinafter referred to as the health monitoring method), which is mainly used to scan historical buildings by borrowing intelligent detection instruments, construct a three - dimensional geometric model of historical buildings, and obtain health monitoring data of historical buildings in real time (including structural data, building environment data, etc.). Based on the health monitoring data, the damage degree of historical buildings is evaluated and future damage prediction of historical buildings is carried out to improve safety prevention. At the same time, it can also effectively combine the three - dimensional geometric model, real - time dynamic monitoring data and future prediction data, so that the monitoring personnel can more intuitively know the structural damage changes of historical buildings at present and in the future. The execution entity of the health monitoring method is a prediction system for health monitoring data based on real - scene three - dimensional GIS for historical buildings (hereinafter referred to as the health monitoring system). The following will be combined with the attached Figure 1 Specifically elaborate on the specific execution steps of the health monitoring system for the health monitoring method.

[0028] S101. Obtain modeling data and monitoring data, and construct a building geometric model based on the modeling data; match the monitoring data with the building geometric model to map the monitoring data to the corresponding positions of the building geometric model; Among them, "construct a building geometric model based on the modeling data" in S101 specifically includes the following sub - steps: Obtain multi - perspective image data of historical buildings, extract key feature points in the multi - perspective image data through a preset feature extraction algorithm to generate multi - perspective point cloud data, and use point cloud fusion technology based on geometric constraints to fuse and process the point cloud data to construct an initial point cloud model of historical buildings; Based on the mesh reconstruction technology of semantic segmentation, perform meshing processing on the initial point cloud model to obtain a building geometric model.

[0029] In implementation, the modeling data is specifically multi - perspective image data of historical buildings obtained by drone oblique photography and ground close - range photography. The health monitoring system is used to obtain the foregoing modeling data, extract key feature points in the multi - perspective image data based on a preset feature matching algorithm for preliminary matching; then use point cloud fusion technology based on geometric constraints to fuse and process the point cloud data to construct an initial point cloud model of historical buildings; finally, use the mesh reconstruction technology of predicted semantic segmentation to perform meshing processing on the initial point cloud model and optimize the mesh model according to the semantic information of historical buildings to finally obtain a high - precision real - scene three - dimensional fine model of historical buildings (i.e., the building geometric model).

[0030] The health monitoring data specifically includes environmental data and structural data. Environmental data refers to the data used to describe the physical environment where historical buildings are located, such as temperature data, humidity data, wind speed and direction, etc. Structural data refers to the data used to describe the physical structure of historical buildings, such as dimension data, stress data, displacement data, etc. The foregoing structural data and environmental data can be monitored by preset intelligent instruments (such as sensors). Correspondingly, the structural data can specifically include image data in the form of images, and specific numerical values used to describe the dimensions of structures (such as unhealthy structures like cracks, deformations, decay, etc.) in the corresponding image data. The health monitoring system is used to obtain the foregoing health monitoring data, perform standardization processing on it to make it match the spatial coordinate system of the building geometric model, and then based on a preset spatio-temporal data interpolation algorithm, map the discrete monitoring data to the corresponding positions of the building geometric model. This corresponding position (hereinafter simply referred to as the mapping position) forms a corresponding association relationship with the health monitoring data, and the health monitoring data is the data used to describe the structure and the environment at the corresponding mapping position.

[0031] S102, Based on a preset prediction and early warning model, learn the monitoring data and predict and output the change trend of the monitoring data within a specified future period.

[0032] S103, Based on a pre-constructed health analysis model, analyze the monitoring data corresponding to the building geometric model and output the structural damage assessment result of the building geometric model.

[0033] In implementation, the health analysis model can be a pre-constructed and trained CNN model. The health monitoring system is used to collect image data of the structural damage of historical buildings (including crack structure images, deformation structure images, decay structure images, etc.) in the initial stage, and perform annotation and classification. Then, use the foregoing image data to train the CNN model, thereby adjusting the network structure and parameters of the CNN model to enable it to accurately identify different types of structural damage. Finally, after completing the training, a health analysis model is formed. Therefore, the input of the health analysis model is the monitoring data, and the output is the damage assessment result. Among them, the damage assessment result includes the determination of whether there is damage, and the type of damage (such as crack type, deformation type, decay type, etc.) corresponding to the determination result of having damage, and the corresponding damage position (such as the mapping position in the building geometric model).

[0034] In addition, the prediction and early warning model is a pre-constructed and trained RNN model, which can predict the safety trend of historical buildings. The health monitoring system is used to collect the time-varying sequence of monitoring data (such as displacement data, strain data, vibration data, settlement data, tilt data, etc.) over time at the initial stage, and use this time-varying sequence as the input of the RNN model to train the RNN model to learn the time series characteristics and laws in the aforementioned monitoring data. Then, after the training is finally completed, a prediction and early warning model is formed. Then, the current monitoring data is input into the prediction and early warning model, so that the prediction and early warning model outputs the time series of monitoring data within a specified future period (such as the period corresponding to the preset duration after the current moment). In addition, in other embodiments, the health monitoring system can first predict the time series of monitoring data corresponding to a specified future period based on the current monitoring data through the prediction and early warning model, and then use the health analysis model to perform damage assessment on each of the current monitoring data and all the predicted monitoring data corresponding to the specified future period, so as to predict the structural damage at present and within the specified future period.

[0035] Optionally, the health monitoring method further includes the following steps: S104. Based on the real-time obtained monitoring data, by analyzing the change trends of the corresponding structural data when different environmental data are analyzed, infer the material properties and / or structural characteristics at the mapping position of the structural data on the building geometric model, generate and output optimization suggestions based on the inference results, and perform optimization processing based on the feedback results of the monitoring personnel on the optimization suggestions; the optimization suggestions at least include correcting and optimizing the monitoring data at the mapping position of the building geometric model, and / or obtaining the on-site survey data at the corresponding mapping position of the building geometric model; Among them, "performing optimization processing based on the feedback results of the monitoring personnel on the optimization suggestions" in S104 includes the following steps: Obtain the on-site survey data at the corresponding mapping position of the building geometric model, and determine whether the corresponding mapping position of the building geometric model meets the layering conditions according to the survey data obtained from the on-site survey; If it meets the conditions, perform layering processing on the corresponding mapping position to obtain a multi-level structure model; among them, the multi-level structure model includes a minimum component layer, a combined module layer, and an overall structure layer from the inside out in sequence, and there is a progressive inclusion and being included relationship between the minimum component layer, the combined module layer, and the overall structure layer; Describe the association relationship between different levels included in the multi-level structure model based on a preset representation form, so that the monitoring personnel can know the association relationship.

[0036] In implementation, as can be seen from the above, the monitoring data includes environmental data and structural data. Therefore, this application proposes to classify the structural data based on the environmental data, and the environmental data is specifically used to describe the environmental scenario where the historical building is located. So the aforementioned classification realizes the classification and summary of the structural data under different environmental scenarios, and arranges the structural data in sequence according to the acquisition time of the structural data. It should be noted here that there is a database preset in the health monitoring system, which prestores different materials, the material properties of each material under different environmental scenarios (such as thermal expansion coefficient, elastic modulus), the correlation between the material properties and the monitoring data (such as the thermal expansion coefficient of the material can reflect structural deformation, stress data, cracked structures, etc.), and different historical buildings constructed based on historical periods, and the monitoring data corresponding to different structural features (such as mortise and tenon structures, raised beam wooden frames, and penetrated column wooden frames) in the historical buildings collected.

[0037] Based on the environmental data in the currently acquired monitoring data, the health monitoring system determines the material properties and structural features consistent with the structural data in the currently acquired monitoring data from the database, so as to realize the speculation of the material properties and structural features at each mapping position in the building geometric model, outputs an optimization suggestion with the speculation result, and pushes the optimization suggestion to the intelligent terminal of the monitoring personnel for them to know. Moreover, the health monitoring system is also used to obtain the feedback result of the monitoring personnel on the corresponding optimization suggestion, and the feedback result can include whether to agree to the optimization content, and the material properties and structural features of the corresponding mapping position entered by the monitoring personnel.

[0038] The health monitoring system is used to realize optimization based on the feedback result of the monitoring personnel. The specific optimization method is as follows: If the feedback result includes the material properties and structural features entered by the monitoring personnel, map the material properties and structural features into the building geometric model to use the material properties and structural features to mark and describe the corresponding mapping position. If the feedback result does not include the material properties and structural features entered by the monitoring personnel, and the monitoring personnel agree to the optimization, map the material properties and structural features included in the optimization suggestion to the corresponding mapping position of the building geometric model to realize the description and marking of the corresponding mapping position.

[0039] If the feedback results of the monitoring personnel include on-site survey location information (the on-site survey location is a specific part of the historical building and corresponds to the mapped location in the building geometric model), and the survey instruments deployed at the corresponding on-site survey locations (such as microstrain sensors, 3D laser scanners, six-axis force sensors, etc.), then the optimization plan of the health monitoring system is to obtain the survey data monitored by the corresponding survey instruments and determine whether the mapped location described by the survey data in the building geometric model meets the layering conditions. The determination method of whether it meets the layering conditions can be: pushing the survey data to the intelligent terminal of the monitoring personnel for the monitoring personnel to determine whether layering is required, such as when the corresponding mapped location is formed by multiple materials or has a complex structure (such as multiple components being joined to form an overall structure).

[0040] If the layering conditions are met, the health monitoring system performs layering processing on the corresponding mapped location. The specific layering processing method can be: using technologies such as 3D laser scanning and photogrammetry to obtain high-precision point cloud data of building components, establishing a three-dimensional model with millimeter-level accuracy (i.e., the three-dimensional model corresponding to the smallest component described above). On the basis of the millimeter-level model, simplify the geometric model (such as using mesh simplification algorithms like QEM and edge collapse to reduce the number of meshes in the millimeter-level model while retaining the main geometric features); then extract features (such as edges, faces, corner points, etc.), identify the connection relationships between components (such as mortise and tenon, bolts, etc.), material properties (such as wood species, density, etc.) and boundary conditions (such as support points, loads, etc.). Finally, generate a combined model with centimeter-level accuracy (i.e., the three-dimensional model corresponding to the combined module described above), and then perform a higher degree of geometric simplification on the basis of the centimeter-level accuracy combined model. Use abstract geometric shapes to represent the overall structure of the building and generate an overall model with meter-level accuracy (i.e., the three-dimensional model corresponding to the overall structure described above). Perform topological optimization on the overall model to remove parts with less impact on the overall performance. This overall model is the overall structure of the corresponding mapped location.

[0041] It should be noted that the splitting method of the overall structure of the mapping position can determine the minimum structural unit according to the size (such as millimeter level, centimeter level, meter level) described above, then combine the minimum structural units to form a combined model with centimeter-level accuracy, and finally form the overall model by combining the combined models with centimeter-level accuracy. In other embodiments, the common structures included in each structural feature of the historical building can also be used as the minimum structural units. For example, the short columns, beam heads, purlins, etc. included in the beam-lifting frame can all be used as independent minimum components. According to the installation sequence of the short columns, beam heads, and purlins, the structure formed by combining more than one minimum component after installation can form a combined module, and finally the structure formed by combining and installing all the minimum components is the overall structure, thus finally realizing layering. Among them, the minimum component forms the innermost structural layer, that is, the minimum component layer. The outer layer of the minimum component layer is successively provided with a combined module layer and an overall structure layer from the inside out, finally realizing a multi-level structural model with a progressive layer-by-layer structure.

[0042] Since the combined module is composed of minimum components and the overall structure is composed of combined modules, there is an inclusion and being-included relationship between different levels. Specifically, the health monitoring system is used to establish the spatial correspondence relationship between different level models (that is, the models corresponding to the minimum component layer, the combined module layer, and the overall structure layer) by using spatial indexing techniques (such as R-tree, Octree, etc.), construct a hierarchical tree structure, organize the different level models into a tree-like hierarchical relationship, use the overall structure layer as the root node, the combined module layer as the intermediate node, and the minimum component layer as the leaf node to clearly describe the association relationship, and corresponding interfaces can also be set for the monitoring personnel to view the association relationship. In addition, in other embodiments, a unique identifier (ID) can also be assigned to each minimum component, combined module, and overall structure, and the mapping relationship between levels can be established through the ID.

[0043] Optionally, the health monitoring method further includes the following steps: For the target mapping position with a multi-level structural model, execute the data conduction scheme through a preset data conduction engine; among them, the data conduction scheme is to convert the surveyed data obtained corresponding to the direct level through the data conduction engine according to the predefined data transfer rules and map the converted data to the indirect level; both the direct level and the indirect level belong to the levels included in the multi-level structural model and are in an inclusion or being-included relationship; the data transfer rules at least include sequentially transferring according to the hierarchical order from the inside out and / or sequentially transferring according to the hierarchical order from the outside in; Analyze whether the surveyed data belongs to the preset damaged data, where the preset damaged data at least includes crack data and deformation data; If so, based on the survey data obtained from the direct level, calculate the influence range of the survey results corresponding to the survey data on the direct level; map the influence range on the direct level to the indirect level and calculate the influence range on the indirect level; where the influence range includes at least the influence area and the influence region. Based on the influence range, formulate a monitoring strategy for the corresponding mapped position, and the monitoring strategy includes at least the monitoring frequency.

[0044] In implementation, the target mapped position refers to the mapped position with a multi-level structure model. The health monitoring system presets a data mapping table to record the data correspondence between different levels, such as recording the position and attributes of each minimum component model in the combined module and the overall structure; the health monitoring system also presets a data conversion algorithm to convert data from one precision to another, such as converting the point cloud data of the millimeter-level minimum component model into the simplified geometric data of the centimeter-level combined module model. The data conduction engine is a pre-developed engine for supporting data query, mapping, conversion, and transfer functions. The data transfer rules are specifically divided into two types. One is to sequentially perform data conversion based on the hierarchical order from the inside to the outside of the multi-level structure model, and the other is to sequentially perform data conversion based on the hierarchical order from the outside to the inside of the multi-level structure. Exemplarily, when the stress data in a certain area of the overall structure model is large, at this time, the health monitoring system takes the overall structure layer as the direct level and uses data mapping and conversion technologies to sequentially map and convert the aforementioned stress data to the combined module layer (i.e., the indirect level) and the minimum component layer (i.e., the indirect level) to know the stress condition of the minimum component model when the stress on the overall structure model is refined to the minimum component layer. In addition, when the monitoring data of the crack type is detected in the minimum component model, the health monitoring system takes the minimum component layer as the direct level and sequentially maps and converts the corresponding crack data to the combined model layer (i.e., the indirect level) and the overall structure layer (i.e., the indirect level), so as to obtain the corresponding crack position in the overall model, so that the monitoring personnel can generate maintenance suggestions.

[0045] Furthermore, for the survey data belonging to the preset damage data, the health monitoring data is also used to determine the influence range of the corresponding damage type (such as cracks, decay, etc.) on the historical building structure (corresponding to the mapped position in the building geometric model), and if the mapped position is a multi-level structure model, the health monitoring model will respectively determine the influence range of the damage type on each level model. The specific method for determining the influence range is as follows: Use buffer analysis or spatial interpolation method to calculate the influence area at the position corresponding to the direct action layer (such as the minimum component layer) of the survey data belonging to the preset damage data. For example, when the damage type is a crack, calculate the influence area around the crack, and then simulate the crack propagation through finite element analysis to determine its influence region, so as to obtain the influence area and influence region of the minimum component layer. Map the influence range of the minimum component layer onto the combined module model. Based on this mapped position, analyze its influence on the minimum components adjacent to this mapped position in the combined module model (i.e., further determine the influence area around this mapped position using buffer analysis or spatial interpolation), and then use a mechanical model to analyze the propagation of stress on the combined module model to determine the influence area, thereby obtaining the influence area and influence region of the combined module model; Similarly, map the influence range of the combined module onto the overall structure model again. Based on this mapped position, analyze its influence on the combined modules adjacent to this mapped position in the overall structure model (i.e., further determine the influence area around this mapped position using buffer analysis or spatial interpolation), and then use a damage mechanics model to predict the expansion trend of damage to determine the influence area on the overall model, and finally obtain the influence area and influence region of the overall structure model.

[0046] In summary, by using the above refined calculation of the influence range, the damaged area can be accurately located. The maintenance warning priority can be determined according to the size of the influence range. By comparing the dynamic changes of the aforementioned influence range in real time, the warning timeliness can be improved.

[0047] Optionally, the health monitoring method further includes the following steps: Receive a virtual display instruction, display the building geometric model in a pre-constructed virtual environment, and decompose the building geometric model with the hierarchy as the minimum splitting unit based on the display granularity included in the virtual display instruction; Display the building geometric model corresponding to the health degree based on the health degree in the virtual display instruction.

[0048] In implementation, the health monitoring system can use VR / AR technology to construct a virtual environment and display the corresponding building geometric model in the virtual environment, so that the monitoring personnel can use a preset interaction tool (such as an AR / VR helmet) to view the virtual environment, thereby realizing the visualized and immersive access to the building geometric model. In addition, this application also proposes a function for the monitoring personnel to select the display granularity and health degree. Combining with the multi-level model mentioned above, the display granularity here is the display method of the building geometric model displayed in the virtual environment, that is, the specific hierarchical model shown. Exemplarily, the display granularity can be divided into three categories: coarse granularity, fine granularity, and fine-grained granularity. Among them, the coarse granularity corresponds to the overall structure model, that is, the smallest component structure of the building geometric model displayed in the virtual environment is the overall structure model; the fine granularity corresponds to the combined module model, that is, the smallest component structure of the building geometric model displayed in the virtual environment is the combined module model; similarly, the fine-grained granularity corresponds to the minimum component model, that is, the smallest component structure of the building geometric model displayed in the virtual environment is the minimum component model.

[0049] The selection of the health level is for the monitoring personnel to select and display the building geometric models at different damage levels. The monitoring data in the building geometric models can be adjusted by using the preset simulation models, so as to simulate and obtain the building geometric models at different damage levels.

[0050] The embodiment of the present application also discloses a health monitoring data prediction system for historical buildings based on real-scene three-dimensional GIS, referring to Figure 2 , including: A geometric model construction module 201, configured to obtain modeling data and monitoring data, construct a building geometric model based on the modeling data; match the monitoring data with the building geometric model to map the monitoring data to the corresponding positions of the building geometric model; A data change prediction module 202, configured to learn the monitoring data based on a preset prediction and early warning model and predict and output the change trend of the monitoring data within a specified future time period; A building health assessment module 203, configured to analyze the monitoring data corresponding to the building geometric model based on a pre-constructed health analysis model and output a structural damage assessment result for the building geometric model.

[0051] Optionally, the geometric model construction module 201 is further configured to obtain multi-view image data of the historical building, extract key feature points in the multi-view image data through a preset feature extraction algorithm to generate multi-view point cloud data, and perform fusion processing on the point cloud data by using a point cloud fusion technology based on geometric constraints to construct an initial point cloud model of the historical building; and is further configured to perform grid processing on the initial point cloud model by using a grid reconstruction technology based on semantic segmentation to obtain a building geometric model.

[0052] Optionally, it further includes a geometric model optimization module, configured to, based on the real-time obtained monitoring data, infer the material properties and / or structural features at the mapping positions of the structural data on the building geometric model by analyzing the change trends of the corresponding structural data when analyzing different environmental data, generate and output an optimization suggestion based on the inference result, and perform optimization processing based on the feedback result of the monitoring personnel on the optimization suggestion; the optimization suggestion at least includes correcting and optimizing the monitoring data at the mapping positions of the building geometric model, and / or obtaining on-site survey data at the corresponding mapping positions of the building geometric model.

[0053] Optionally, the geometric model optimization module is further configured to obtain the on-site survey data of the corresponding mapping position of the building geometric model, and determine whether the corresponding mapping position of the building geometric model meets the layering condition according to the survey data obtained from the on-site survey; it is also configured to, if it meets the condition, perform layering processing on the corresponding mapping position to obtain a multi-level structure model; wherein, the multi-level structure model includes, from the inside out, a minimum component layer, a combined module layer, and an overall structure layer, and there is a progressive inclusion and being included relationship between the minimum component layer, the combined module layer, and the overall structure layer; it is also configured to describe the association relationship between different levels included in the multi-level structure model based on a preset representation form, so that the monitoring personnel can know the association relationship.

[0054] Optionally, it further includes a multi-level data conduction module, which is configured to execute a data conduction scheme for the target mapping position with a multi-level structure model through a preset data conduction engine; wherein, the data conduction scheme is to perform data conversion on the survey data obtained corresponding to the direct level through the data conduction engine according to the predefined data transfer rules and map the converted data to the indirect level; the direct level and the indirect level both belong to the levels included in the multi-level structure model and are in an inclusion or being included relationship; the data transfer rules at least include sequentially transferring according to the hierarchical order from the inside out, and / or sequentially transferring according to the hierarchical order from the outside in.

[0055] Optionally, the multi-level data conduction module is further configured to analyze whether the survey data belongs to preset damaged data, wherein the preset damaged data at least includes crack data and deformation data; if so, based on the survey data obtained at the direct level, calculate the influence range of the survey result corresponding to the survey data on the direct level; map the influence range on the direct level to the indirect level and calculate the influence range on the indirect level; wherein, the influence range at least includes the influence area and the influence area; it is also configured to formulate a monitoring strategy for the corresponding mapping position based on the influence range, and the monitoring strategy at least includes the monitoring frequency.

[0056] Optionally, it further includes a virtual display interaction module, which is configured to receive a virtual display instruction, display the building geometric model in a pre-constructed virtual environment, and decompose the building geometric model with the level as the minimum splitting unit based on the display granularity included in the virtual display instruction; it is also configured to display the building geometric model corresponding to the health level based on the health level in the virtual display instruction.

[0057] An embodiment of the present application also discloses a prediction device for health monitoring data of historical buildings based on real-scene three-dimensional GIS. The prediction device for health monitoring data of historical buildings based on real-scene three-dimensional GIS includes a memory and a processor. A computer program capable of being loaded and executed by the processor, such as the method for predicting health monitoring data of historical buildings based on real-scene three-dimensional GIS described above, is stored on the memory.

[0058] The embodiments of the present application also disclose a computer-readable storage medium, which stores a computer program that can be loaded and executed by a processor and is used for the prediction method of health monitoring data based on real-scene three-dimensional GIS for historical buildings as described above. Such computer-readable storage medium may include, for example, various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs.

[0059] It should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0060] The above embodiments are only used to illustrate the technical solutions of the present application, rather than limiting the protection scope of the application. Obviously, the described embodiments are only partial embodiments of the present application, rather than all embodiments. Based on these embodiments, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope to be protected by the present application.

Claims

1. A health monitoring data prediction method based on real-life 3D GIS for historical buildings, characterized in that: include: Acquire modeling data and monitoring data, and construct a building geometric model based on the modeling data; Matching the monitoring data with the building geometric model to map the monitoring data to corresponding positions of the building geometric model; Based on the preset prediction and early warning model, the monitoring data is learned and the change trend of the monitoring data within a specified period of time in the future is predicted and output; Based on the pre-built health analysis model, the monitoring data corresponding to the building geometric model is analyzed, and a structural damage assessment result of the building geometric model is output.

2. The method for predicting health monitoring data based on real-scene three-dimensional GIS for historical buildings according to claim 1 is characterized in that: The modeling data includes multi-view image data; The step of constructing a building geometric model based on the modeling data comprises: Acquire multi-view image data of historical buildings, extract key feature points in the multi-view image data by using a preset feature extraction algorithm, generate multi-view point cloud data, fuse the point cloud data using a point cloud fusion technology based on geometric constraints, and construct an initial point cloud model of the historical building; The initial point cloud model is meshed based on the semantic segmentation grid reconstruction technology to obtain a building geometric model.

3. The method for predicting health monitoring data based on real-scene three-dimensional GIS for historical buildings according to claim 1 is characterized in that: The monitoring data includes environmental data and structural data; The method further comprises: Based on the monitoring data acquired in real time, by analyzing the changing trends of the corresponding structural data under different environmental data, the material properties and / or structural features of the structural data at the mapping position on the building geometric model are inferred, and optimization suggestions are generated and output based on the inference results, and optimization processing is performed based on the feedback results of the monitoring personnel on the optimization suggestions; the optimization suggestions at least include correcting and optimizing the monitoring data at the mapping position of the building geometric model, and / or obtaining field survey data of the corresponding mapping position of the building geometric model.

4. The method for predicting health monitoring data based on real-scene three-dimensional GIS for historical buildings according to claim 3 is characterized in that: The optimizing process based on the feedback result of the monitoring personnel on the optimization suggestion includes: Acquire field survey data of a corresponding mapping position of the building geometric model, and determine whether the corresponding mapping position of the building geometric model satisfies a layering condition according to the survey data obtained through the field survey; If the conditions are met, the corresponding mapping positions are hierarchically processed to obtain a multi-level structural model; wherein the multi-level structural model includes, from the inside to the outside, a minimum component layer, a combination module layer, and an overall structure layer, wherein the minimum component layer, the combination module layer, and the overall structure layer are in a progressive inclusion and inclusion relationship; The association relationship between different levels included in the multi-level structure model is described based on a preset representation form, so that the monitoring personnel can know the association relationship.

5. The method for predicting health monitoring data based on real-scene three-dimensional GIS for historical buildings according to claim 4 is characterized in that: The method further comprises: For the target mapping position where there is a multi-level structural model, a data transmission scheme is executed through a preset data transmission engine; wherein, the data transmission scheme is to perform data conversion on the survey data corresponding to the direct level and map the converted data to the indirect level through the data transmission engine according to a predefined data transmission rule; the direct level and the indirect level both belong to the levels included in the multi-level structural model, and are in a containing or being included relationship; the data transmission rules at least include sequential transmission in a hierarchical order from the inside to the outside, and / or sequential transmission in a hierarchical order from the outside to the inside.

6. The method for predicting health monitoring data based on real-scene three-dimensional GIS for historical buildings according to claim 5 is characterized in that: The step of converting the survey data corresponding to the direct level and mapping the converted data to the indirect level further includes: Analyzing whether the survey data belongs to preset damage data, wherein the preset damage data at least includes crack data and deformation data; If yes, then based on the survey data obtained at the direct level, calculate the impact range of the survey result corresponding to the survey data on the direct level; map the impact range on the direct level to the indirect level, and calculate the impact range on the indirect level; wherein the impact range at least includes the impact region and the impact area; A monitoring strategy is formulated for the corresponding mapping position based on the impact range, and the monitoring strategy at least includes a monitoring frequency.

7. The method for predicting health monitoring data based on real-scene 3D GIS for historical buildings according to claim 4 is characterized in that: The method further comprises: receiving a virtual display instruction, displaying the building geometric model in a pre-built virtual environment, and decomposing the building geometric model based on a display granularity contained in the virtual display instruction and taking a hierarchy as a minimum splitting unit; Based on the health level in the virtual display instruction, a building geometric model corresponding to the health level is displayed.

8. A real-life 3D GIS-based health monitoring data prediction system for historical buildings, characterized in that: include: A geometric model building module (201) is used to obtain modeling data and monitoring data, and to build a building geometric model based on the modeling data; Matching the monitoring data with the building geometric model to map the monitoring data to corresponding positions of the building geometric model; A data change prediction module (202) is used to learn the monitoring data and predict and output the monitoring data change trend within a future specified period of time based on a preset prediction and early warning model; The building health assessment module (203) is used to analyze the monitoring data corresponding to the building geometric model based on a pre-built health analysis model, and output a structural damage assessment result for the building geometric model.

9. A health monitoring data prediction device based on real-life three-dimensional GIS for historical buildings, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the method according to any one of claims 1 to 7.

Images