Method for establishing database of urban building group basic information for providing earthquake service

Abstract

The present application relates to a kind of urban building group foundation information database establishment method for providing earthquake service, comprising: determining research area, for each single building of research area, establishing unique building identification;Obtain the building plan profile and multi-source spatial data of region, unify each data to the same coordinate system, and take the standardized building profile and its unique building identification as spatial correlation reference, after mapping each data to single building scale, it is associated to the corresponding unique building identification;Establish multi-dimensional decision rule base;Verify and correct multi-dimensional decision rule base according to verification result, output standardized urban building group foundation database.The present application takes single building profile and unique identification as data correlation core, in turn completes multi-source data acquisition, coordinate and format unification, building scale spatial mapping, missing attribute inference, sampling field verification and rule correction, and finally forms standardized foundation database that can be used for building group seismic analysis, digital twin modeling and post-earthquake function evaluation.

Description

Technical Field

[0001] This invention relates to the fields of earthquake services, geographic information processing, and intelligent modeling of urban building clusters. Specifically, it relates to a method for establishing a basic information database of urban building clusters that provides earthquake services, which is particularly suitable for seismic performance analysis, disaster risk assessment, and post-earthquake functional evaluation at the urban scale. Background Technology

[0002] Seismic performance analysis, disaster scenario simulation, and post-earthquake functional evaluation of urban building complexes require individual buildings as the basic object. Input must include a variety of fundamental information such as building plan outline, length, width, height, number of stories, latitude and longitude coordinates, structural type, function, construction year, seismic intensity, facade window openings, and building occupancy. The completeness and accuracy of this information directly affect the reliability of the results of reconstructing hidden building information, predicting structural seismic response, and evaluating the functional performance of the building complex.

[0003] Existing high-precision building data typically comes from architectural design drawings, housing surveys, or manual field investigations. However, cities have a large number of existing buildings spanning a long period, with many design drawings missing or difficult to obtain centrally. High-precision survey data is subject to data security and privacy management restrictions, making it difficult to use directly for scientific research and engineering applications. Comprehensive field investigations suffer from high costs, long cycles, a large demand for professional personnel, and low update efficiency, making it difficult to meet the need for rapid modeling of urban-scale building complexes.

[0004] With the rapid development of publicly available data such as open maps, remote sensing imagery, digital terrain models, digital elevation models, population rasters, points of interest (POIs), and areas of interest (AOIs), information such as building outlines, geographical locations, building heights, surrounding road networks, population distribution, functional cues, and construction timelines can be obtained from different platforms. However, existing technologies still have the following significant shortcomings when processing the aforementioned multi-source data:

[0005] First, different data sources differ significantly in spatial resolution, coordinate system, update frequency, data format, and coverage. A single data source cannot provide the complete building attributes required for seismic analysis. While existing methods attempt to overlay multiple data types onto the same digital map system, they essentially remain at the level of simple data layering and do not address the fundamental problem of uniformly mapping data from different spatial units to the scale of individual buildings. For example, population rasters typically use regular grids (e.g., 100 m × 100 m) as statistical units, DSM / DEMs use regular pixels as spatial units, POIs exist as discrete points, and AOIs use irregular polygons as spatial units. None of these data have ever been designed with "individual buildings" as the basic statistical unit. Directly overlaying multi-source data can lead to problems such as overlapping building outlines, spatial misalignment, attribute duplication, mismatch between points of interest and buildings, and difficulty in mapping population rasters to individual buildings.

[0006] Second, there is often a significant lack of publicly available data regarding key attributes essential for seismic analysis, such as building function, number of floors, structural type, and facade window openings. Existing methods often rely on manual, building-by-building analysis or establish isolated identification models for only one attribute. For example, some methods classify urban building clusters into functional units based on POI heat vectors (see "Research on Functional Classification of Urban Buildings Based on High-Resolution Remote Sensing Imagery Assisted by POI", Journal of Geoinformation Science, 2017, No. 6), or classify building functions using multi-source POI and AOI data (see "Research on Urban Built Environment Identification Based on Multi-Source Spatial Data Fusion", Shandong Transportation Science and Technology, 2024, No. 5). However, these methods only address the issue of single-function classification and do not involve systematic reasoning regarding attributes such as number of floors, structural type, and window openings; furthermore, their spatial scale is "functional unit" rather than individual building, which cannot meet the seismic analysis requirements based on individual buildings. Patent document CN202510571919 involves extracting three-dimensional information of buildings using multimodal data, but its technical means rely on pre-trained neural network models (multimodal semantic segmentation network, height inversion network), which belongs to data-driven deep learning methods rather than reasoning methods based on domain knowledge rules. Patent documents CN107229913A and CN110298253B involve using remote sensing data or population big data for population density analysis or building population prediction, but neither involves finely allocating population raster data to individual buildings and uniformly associating it with other building attributes to correspond to the structural type. No method has been found in the prior art to establish multidimensional decision rules for systematic inference using building age, number of floors, and function as joint inputs.

[0007] Third, existing methods generally lack quantifiable data quality verification mechanisms and closed-loop correction processes. Most methods involve one-time data processing, making it impossible to know whether the inference results are correct or accurate, let alone iteratively optimize the inference rules based on the verification results. Although some technical solutions involve data quality monitoring or verification, they are all aimed at Building Information Modeling (BIM) or construction data streams, rather than at quality verification and rule iterative correction of inference results from multi-source data fusion of urban-scale building complexes.

[0008] Therefore, there is an urgent need to propose a low-cost, high-efficiency, batch-updable method for acquiring and improving basic information on urban building clusters with a quality verification mechanism. This method aims to construct a basic database of urban building clusters with relatively complete attributes, a unified format, and direct service to subsequent engineering analysis, without relying on complete design drawings and large-scale household surveys. Summary of the Invention

[0009] The purpose of this invention is to overcome the problems of scattered sources of basic information on urban building clusters, missing key attributes, inconsistent spatial scales, and difficulty in verifying data quality. It proposes a method for establishing a basic information database for urban building clusters based on multi-source publicly available data and domain knowledge reasoning. This method uses the outline and unique identifier of individual buildings as the core of data association, sequentially completing multi-source data acquisition, coordinate and format standardization, building scale spatial mapping, missing attribute reasoning, on-site sampling verification, and rule correction. Ultimately, it forms a standardized basic database that can be used for seismic analysis of building clusters, digital twin modeling, and post-earthquake functional evaluation.

[0010] To achieve the above objectives, the present invention provides a method for establishing a basic information database of urban building clusters that provides earthquake services, comprising:

[0011] Step S1: Determine the research area, target building objects, and basic information fields.

[0012] Based on the engineering analysis requirements, the research area boundary is determined, and a basic information field system for individual buildings is established. This field includes at least the following: building geometric information, geographic information, attribute information, and population information. Specifically, building geometric information includes planar outline, centroid, length, width, floor area, height, and number of stories; geographic information includes latitude and longitude, projected coordinates, surrounding road network, and functional area; attribute information includes building function, structural type, construction year, seismic intensity, and facade window opening information; and population information includes the resident population or service population within the individual building.

[0013] Furthermore, a unique building identifier, BuildingID, is established for each individual building within the study area. Subsequent data from different sources are all associated with the BuildingID through spatial inclusion, spatial intersection, nearest neighbor, or raster index relationships.

[0014] Step S2: Obtain and process the building plan outline.

[0015] The building plan outlines of the study area are obtained from open maps, online map services, remote sensing imagery, or existing vector data, and standardized building outline data are formed through raster segmentation, vectorization, geometric cleanup, coordinate transformation, and parameter calculation.

[0016] Further, step S2 includes:

[0017] Step 2.1, Building Outline Extraction. Obtain map images or vector data of the study area through an online map interface; when the input is a map image, perform raster binarization based on building color banding or image features, convert the binary raster into vector polygons, and smooth the polygon boundaries.

[0018] Step 2.2, Geometric Anomaly Cleanup. Delete geometrically empty elements and tiny contours with areas smaller than a preset threshold; perform anomaly detection on contours with significantly larger areas; identify overlapping building contours and rationally divide the overlapping parts using centerline segmentation or intersecting area reconstruction methods.

[0019] Step 2.3: Coordinate and Geometric Parameter Standardization. Calculate the centroid latitude and longitude coordinates of each building outline, transform the building outline to a suitable projected coordinate system, and calculate the planar area; determine the main direction, length, and width of the building using the minimum bounding rectangle, and store the processed outline corresponding to the BuildingID.

[0020] Step S3: Obtain and map the basic attributes of the multi-source building.

[0021] Data on building height, road network, population, POI, AOI, seismic intensity, and building age were acquired and mapped to the scale of individual buildings after coordinate unification and spatial registration.

[0022] It should be noted that the multi-source spatial data involved in step S3 originates from different data platforms, possessing different spatial resolutions, coordinate systems, and data formats. Specifically, the building height indicator data (DSM / DEM) is in a regular raster format, the population distribution data is in a regular grid format, the POI is in a discrete point format, the AOI is in an irregular polygon format, and the building age indicator data (time-series data of impermeable surfaces) is in a time-series raster format. These data organize information using different spatial units (pixels, grids, points, and polygons), and were not originally designed for the statistical unit of a "single building." Therefore, mapping these heterogeneous data uniformly to polygons of arbitrary shapes and scales of single buildings is a non-standard, unconventional cross-scale transformation problem, which cannot be simply summarized as "coordinate unification."

[0023] Further, step S3 includes:

[0024] Step 3.1, Building Height Estimation. Obtain the Digital Surface Model (DSM) and Digital Elevation Model (DEM) for the same area. Project the two types of rasters onto a unified coordinate system and resample and align them with the grid. Extract the corresponding raster elevation using the centroid of a single building or the coverage area of ​​its outline as an index. Estimate the building height using the elevation difference between the DSM and DEM.

[0025] Step 3.2, Road Network Information Processing. Obtain road vector data from the open road network platform, and filter the road network according to road level based on the purpose of engineering analysis; when used for functional sub-building group division, retain backbone roads such as expressways, arterial roads and secondary arterial roads, and remove branch roads, internal roads of residential areas and service passages that will cause excessive division of spatial units.

[0026] Step 3.3, Population Spatial Allocation. Acquire population raster data and resample it to a unified spatial grid; traverse each population raster and allocate the raster population to individual buildings based on the proportion of building outline area within the raster. For building i within population raster g, its allocated population P i Calculate P using the following formula: i =P g ×A i,g / ΣA j,g , where P g For the raster population, A i,g Let ΣA be the area of ​​the outline of building i within grid g. j,g It is the sum of the outline areas of all buildings within the grid.

[0027] Step 3.4, POI and AOI Processing. Acquire Points of Interest (POIs) and Areas of Interest (AOIs) data for the study area; verify and correct overlapping, contained, or anomalously sized AOI boundaries; determine the AOI to which a building belongs based on its centroid, prioritizing the smaller AOI with a clearer spatial orientation when a building is located within multiple AOIs. For POIs, only points within a preset distance of the building outline are retained; when a POI is adjacent to multiple buildings, it is assigned to the nearest building outline.

[0028] Step 3.5: Seismic fortification intensity matching. Spatial matching is performed based on the latitude and longitude coordinates of the centroid of the individual building and the seismic ground motion parameter zoning data to obtain the seismic fortification intensity or peak ground acceleration parameters of the building's location.

[0029] Step 3.6, Building Age Determination. Obtain the time-series raster of impermeable surfaces updated annually and project it onto the building outline. Using the building centroid or the area covered by the building outline as an index, search for the corresponding raster status by year. When the raster changes from impermeable to impermeable, determine that year as the building's construction time or construction period.

[0030] Step S4: Establish a rule base for missing attribute reasoning based on domain knowledge.

[0031] Based on building design codes, structural stress characteristics, regional construction practices, and expert experience, the information obtained in steps S2 and S3 is used to logically deduce and complete the information on building function, number of floors, structural type, and exterior window openings.

[0032] It is important to emphasize that the reasoning rule base established in this step is not a simple "data completion" or "adding search terms", but a domain knowledge-driven reasoning system specifically designed for the specific engineering scenario of urban seismic analysis. Specifically: the reasoning behind the number of building stories needs to be based on the typical floor height ranges of different functional buildings (residential 2.8–3.0 m, office 3.8–4.5 m, industrial and warehousing 5.0–6.0 m, etc.) and structural stress characteristics (whether a large space is set on the ground floor); the reasoning behind the structural type needs to be based on the development characteristics of structural systems in different eras (e.g., masonry was the main type before 1980, while frames and shear walls became popular after 2000), the stress requirements of different number of stories (masonry or frames can be used for 6 stories or less, shear walls or frame-shear walls are required for 7 stories or more), and the spatial requirements of different functions to make multi-dimensional decisions; the reasoning behind the window openings on the facade needs to consider the constraints of seismic fortification intensity on the minimum width of the wall between windows (not less than 1.0 m in 6-degree and 7-degree areas, not less than 1.2 m in 8-degree areas, and not less than 1.5 m in 9-degree areas), the lighting requirements of different functional buildings (1.5 m for residential windows, 1.8 m for office windows, and 0.6 m for industrial and warehousing windows), and the requirements of structural stress on the location of the openings. The aforementioned reasoning rules possess distinct knowledge attributes specific to the civil engineering field and cannot be replaced by general data processing methods.

[0033] Further, step S4 includes:

[0034] Step 4.1, Building Function Identification. First, determine the main function type of the area based on the AOI category to which the building belongs, and then refine the function of the individual building using the associated POIs. When the POI contains key function terms such as emergency, fire protection, disease control, government, hospital, transportation facilities, television or radio station, it is prioritized as the corresponding government, medical, emergency, or public service function; when a building is associated with multiple POIs of different function categories, it is labeled as a multi-functional building; for buildings not associated with AOIs or POIs, a default function is set based on the main land use type of the study area.

[0035] Step 4.2, Building Count Estimation. Match typical floor height ranges to the building's function and use the total building height to calculate the number of floors. For residential, school, and industrial / warehouse buildings where the ground floor height is close to the standard floor height, divide the total building height by the average standard floor height and round down. For office and medical buildings where the ground floor may have an entrance lobby or waiting area, take the upper limit of the corresponding floor height range for the ground floor height and calculate the number of floors using the formula: "(Total Building Height - Ground Floor Height) / Average Standard Floor Height + 1".

[0036] Step 4.3, Structural Type Inference. Using the building's age, number of stories, and function as inputs, a multi-dimensional decision rule is established to infer the structural type of a single building: masonry structure, reinforced concrete frame structure, reinforced concrete frame-shear wall structure, or reinforced concrete shear wall structure. This rule is determined based on the structural system development characteristics of different eras, the stress requirements of different number of stories, and the spatial requirements of different building functions.

[0037] Step 4.4, Window opening estimation on the facade. Determine the minimum width of the inter-window wall, typical window width, and window spacing based on the seismic intensity, building function, structural type, and building length and width dimensions. Under the premise of meeting the seismic structural requirements along the long side of the building, arrange the window openings evenly in the center, and arrange one to two window openings symmetrically in the center along the short side according to functional requirements. Output the width, spacing, number, and location of the window openings.

[0038] Step S5: Conduct on-site sampling surveys and quantitative verification of data quality.

[0039] Stratified sampling was conducted based on building function, age, number of floors, and structural type. Typical areas were selected to verify building function, number of floors, structural type, and facade window opening information. The recognition accuracy of different attributes was quantified based on the verification results.

[0040] Furthermore, building function and structural type are evaluated using classification accuracy, while the number of building floors and window openings are evaluated using relative error: Building Function Accuracy func =N correct / N total Relative error of building number of floors (RE) floor =|N pred -N true | / N true Structure type accuracy (Acc) stru =N correct / N total ; Relative error of window opening RE win =|L pred -L true | / L true Among them, N correct To correctly identify the number of samples, N total To verify the total number of samples, N pred and N true These represent the predicted number of layers and the actual number of layers, respectively. pred and L true These represent the predicted and measured total lengths of window openings on the building facade, respectively.

[0041] Furthermore, based on the confusion matrix, error distribution, and typical error samples, the sources of deviation are identified, and the functional priorities, typical floor height parameters, structural type decision boundaries, or window opening layout modules are corrected. Attribute reasoning is then re-executed on the buildings in the study area, thus forming a data quality closed loop of "acquisition → reasoning → verification → correction".

[0042] It should be noted that this closed-loop correction mechanism is one of the key innovative features of this invention. In existing technologies, most methods are one-way, one-time data processing flows, making it impossible to know whether the inference result is correct or how accurate it is, let alone iteratively optimize the inference rules based on the verification results. This invention obtains objective truth values ​​through stratified sampling field surveys, uses quantitative accuracy indicators (accuracy rate, relative error) as feedback signals, systematically corrects rule parameters, and re-executes the inference, making data quality have traceable quantitative evidence and continuously improving inference accuracy as verification samples accumulate.

[0043] Step S6: Output a standardized urban building cluster basic database.

[0044] Using BuildingID as the primary key, the building outline, geometric parameters, geographic coordinates, road network relationship, population, POI / AOI, seismic intensity, construction year, and inferred function, number of floors, structural type, and window opening information are uniformly written into a vector file, relational database, or geographic database. The data source, acquisition or inference method, and verification status are recorded for subsequent digital twin modeling of building complexes, seismic response prediction, damage assessment, and post-earthquake functional evaluation.

[0045] Compared with the prior art, the present invention has the following advantages:

[0046] 1. Open information sources and low data requirements. This invention uses publicly available data such as open maps, remote sensing elevation data, impermeable surfaces, population rasters, and POI / AOI as the main inputs. It does not require complete building design drawings or door-to-door surveys covering the entire area, making it suitable for the rapid construction of basic data for large-scale urban building clusters.

[0047] 2. Unifying multi-source heterogeneous data to the scale of a single building. This invention uses building outline and BuildingID as the core association, and through methods such as projection transformation, grid alignment, spatial inclusion, nearest neighbor, and area weight allocation, it unifies heterogeneous data from different data platforms, with different spatial resolutions and data formats, to the scale of a single building, thus solving the problem of the difficulty in directly associating multi-source data.

[0048] 3. It can systematically supplement the key attributes required for seismic analysis. This invention establishes a domain knowledge rule base based on building function, age, number of stories, seismic intensity, and geometric characteristics, and performs unified reasoning on building function, number of stories, structural type, and window openings, avoiding the limitations of existing methods that can only handle single attributes or rely on manual interpretation of each building.

[0049] 4. It has a quantifiable data quality verification and correction mechanism. This invention verifies the inference results through stratified sampling field surveys, confusion matrices, and relative error indicators, and corrects the rules and parameters according to the error distribution, so that the database accuracy has a traceable quantitative basis, forming a complete technical closed loop of "acquisition → inference → verification → correction".

[0050] 5. High degree of automation, easy to update and expand. Data sources can be acquired in batches according to the study area, and coordinate processing, attribute association and rule reasoning can be executed automatically by computer programs; when data sources are updated or the study area changes, the same process can be reused to quickly update the database.

[0051] 6. Excellent engineering application results. In the embodiment within the Harbin Ring Expressway area, this method constructed a basic database containing the outlines of 39,200 individual buildings. Sampling field verification results show that the accuracy rate of building function identification reached 92.3%, the average relative error of building number of floors was 3.2%, the accuracy rate of structural type identification reached 96.7%, and the average relative error of the total length of window openings on building facades was 8.9%, which can meet the basic input requirements for urban-scale seismic performance assessment and post-earthquake functional evaluation. Attached Figure Description

[0052] Figure 1 This is a flowchart illustrating the overall process of acquiring and improving basic information about urban building complexes according to the present invention.

[0053] Figure 2 This is a flowchart of the building plan outline acquisition and geometric standardization process of the present invention.

[0054] Figure 3 This is a schematic diagram illustrating the spatial mapping relationship between multi-source data and individual building scale in this invention.

[0055] Figure 4 This is a flowchart of the missing attribute reasoning based on domain knowledge in this invention.

[0056] Figure 5 This is a flowchart of the data quality verification and iterative correction process for this invention.

[0057] Figure 6 This is a schematic diagram illustrating the organization of fields in the standardized urban building cluster basic database of this invention.

[0058] Figure 7This is a schematic diagram illustrating the results of a field survey conducted according to an embodiment of the present invention. Detailed Implementation

[0059] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are for illustrative purposes only and should not be considered as limiting the scope of the invention.

[0060] This implementation method focuses on the urban building complex within the Harbin Ring Expressway area. The study area is approximately 28.34 km long from east to west and 25.02 km wide from north to south, with a total area of ​​approximately 591.58 km². The area covers both the central old city and the peripheral new city, and includes various building types such as masonry, reinforced concrete frames, frame-shear wall structures, and shear wall structures. The steps are as follows... Figure 1 As shown.

[0061] (a) Setting the research area and target fields

[0062] Based on the input requirements for building cluster seismic response prediction and post-earthquake functional evaluation, the basic data is divided into four categories: The first category is building geometric information, including building plan outline, centroid, length, width, floor area, height, and number of stories; the second category is building geographic information, including latitude and longitude coordinates, projected coordinates, surrounding road network, and associated AOI; the third category is building attribute information, including structural type, building function, construction year, seismic intensity, and facade window openings; and the fourth category is resident distribution information, including the number of residents or service personnel within the building.

[0063] A unique BuildingID is established for each building outline within the study area, and the original source, processing steps, and final attributes of various data are all associated with the BuildingID. The main data sources and processing results in this embodiment are shown in Table 1.

[0064] Table 1. Main data sources and individual building output information for the embodiments

[0065] Building outline Online map services / map imagery Thresholding, vectorization, smoothing, geometric cleaning Outline, centroid, area, length and width Building height DSM and DEM Projection unification, resampling, grid alignment, and difference calculation Building height road network Open road vector data Road grade filtering and topology cleaning Backbone road network, adjacency relationships population Approximately 100 m population grid Nearest neighbor resampling, building area weighting Population of a single building POI / AOI Open map points of interest / areas of interest Boundary correction, spatial inclusion, 5 m neighbor assignment Functional clues, AOI attribution fortification intensity Seismic ground motion parameter zoning data centroid coordinate space matching Fortification Intensity / Peak Acceleration Construction Era Annual data on impermeable surfaces The first change of year in time-series retrieval Year or period of construction

[0066] (II) Acquisition and Processing of Building Plan Outline

[0067] like Figure 2 As shown, this embodiment acquires high-resolution map images of the study area through an online map interface. First, raster binarization is performed based on building color banding values ​​and RGB thresholds; then, the binary raster is converted into shapefile vector polygons, and the boundaries are smoothed to reduce jagged edges and abnormal nodes caused by pixelation.

[0068] The initial outline was cleaned up as follows: 1) Delete elements with empty geometric values; 2) Eliminate small outlines with a planar area of ​​less than 10 m²; 3) Check each outline with a significantly larger planar area to identify whether multiple adjacent buildings have been incorrectly merged; 4) Detect overlapping areas of building outlines and use the centerline segmentation method to divide the overlapping parts into adjacent buildings; 5) Calculate the latitude and longitude of the outline centroid and retain 6 decimal places; 6) Transform the coordinates to the WGS 1984 UTM Zone 52N projected coordinate system, calculate the planar area, and obtain the building length, width, and main direction through the minimum bounding rectangle.

[0069] After the above processing, a total of 39,200 valid individual building outlines were obtained in the study area. Each processed building outline has a unique BuildingID, which can be used as a spatial reference for subsequent information association such as height, population, POI, AOI, age, and fortification intensity.

[0070] (III) Acquisition of building height, road network, population and spatial attributes

[0071] 1. Building Height. This embodiment selects a digital surface model and a digital elevation model with a spatial resolution of 30 m. The two types of raster data are reprojected to a unified coordinate system and mesh aligned. DSM and DEM raster values ​​are extracted using the building centroid coordinates, and then calculated according to H... i = DSM i - DEM i Estimate the total height of the i-th building; for samples with negative differences or obvious anomalies, mark them as data to be verified.

[0072] 2. Road Network Information. Road vector data for the study area was obtained from an open geographic information platform. To ensure the integrity of subsequent spatial division of functional building clusters, only the backbone road network at three levels—expressways, arterial roads, and secondary arterial roads—was retained, while branch roads, internal residential roads, and service passages were removed. A total of 5172 road network topology segments were ultimately obtained.

[0073] 3. Spatial Distribution of Population. Population raster data with a spatial resolution of approximately 100 m is used. The population raster is resampled using the nearest neighbor method. Each raster is traversed and the area ratio of each building outline within it is calculated. The population of each raster is allocated to individual buildings according to area weight. If a building spans multiple rasteres, the allocation results of different rasteres are summed to obtain the residential population or service population of that building.

[0074] 4. POI and AOI. POI and AOI data are obtained through polygon search via the map's open interface. This embodiment associates 4256 AOI functional faces and 295735 POI attribute records. For AOIs, overlapping, containment, and area anomaly boundaries are corrected first, then the building centroid is determined; when a building falls into multiple AOIs, the AOI with the smaller area is selected. For POIs, only points within 5 m of the building outline are retained; when a POI is adjacent to multiple buildings, it is assigned to the nearest building.

[0075] 5. Seismic Intensity. The seismic intensity is automatically matched based on the building ID, the latitude and longitude of the building's centroid, and the seismic motion parameter zoning data. In this embodiment, most buildings are located in a seismic intensity 7 (0.10 g) zone, and a small number are located in a seismic intensity 6 (0.05 g) zone.

[0076] 6. Construction Date. Impermeable surface time-series data with a spatial resolution of 30 m and updated annually were used. Raster data from 1985 to 2024 were projected onto a unified coordinate system, and the raster status was retrieved year by year using the building centroid; when the corresponding position first changed from 0 to 1, that year was determined as the building's construction date. Buildings existing before the data's starting year were uniformly assigned to the earliest date range.

[0077] After coordinate unification and spatial registration, the above data is mapped to the scale of individual buildings and then associated with the corresponding unique building identifier, such as... Figure 3 As shown.

[0078] (iv) Reasoning based on missing building attributes

[0079] like Figure 4 As shown.

[0080] 1. Building Function Identification. First, the functional sub-area of ​​the building is determined based on the AOI (Area of ​​Interest), such as a residential community, campus, industrial park, or commercial district. Then, the individual building's use is refined based on the POI (Point of Interest). If the POI contains keywords such as "emergency," "fire protection," "disease control," "people's government," "hospital," "transportation facilities services," "television station," or "radio station," it is prioritized as the corresponding key public service function. If the same building contains multiple POIs of different categories, it is labeled as a composite function such as "residential + commercial." Buildings without associated AOIs or POIs are defaulted to residential in this embodiment.

[0081] 2. Building Stories Estimation. This embodiment establishes typical story height parameters for buildings with different functions based on architectural design experience, as shown in Table 2. Residential, school, and industrial / warehouse buildings are calculated by dividing the total height by the average standard story height. Office and medical buildings consider the large open space on the ground floor; the ground floor height is taken as the upper limit of the corresponding range, and the remaining floors are calculated based on the average standard story height. The calculation results are rounded to positive integers, and results with fewer than one story or significantly inconsistent with the building height are marked as abnormal.

[0082] Table 2. Estimation methods for typical floor height and number of floors of buildings with different functions

[0083] Residential buildings 2.8～3.0 N = round(H / hm) The bottom layer is close to the standard layer. School buildings 3.8～4.5 N = round(H / hm) The teaching space has a moderate ceiling height. office buildings 3.8～4.5 N = round[(H-h1) / hm] + 1 Ground floor lobby (take the upper limit of the interval) Medical buildings 3.8～4.5 N = round[(H-h1) / hm] + 1 The outpatient and waiting areas have a higher floor level. Commercial / Industrial Warehousing 5.0～6.0 N = round(H / hm) Estimated based on typical floor height of large spaces

[0084] Where H is the total building height, hm is the median of the corresponding standard floor height range, and h1 is the floor height of the ground floor.

[0085] 3. Structural Type Inference. A decision table is established based on the building's age, number of stories, and function, as detailed in Table 3. This rule reflects the main application range of masonry, reinforced concrete frame, frame-shear wall, and shear wall structures in urban construction during different periods.

[0086] Table 3 Rules for Inferring Building Structure Type

[0087] 1979 and earlier all all Masonry structure 1980–2001 6 floors and below Business, Medical RC framework 1980–2001 6 floors and below other Masonry structure 1980–2001 7 floors and above all RC framework 2002-2008 6 floors and below Residential Masonry structure 2002-2008 6 floors and below other RC framework 2002-2008 7 floors and above Residential RC shear wall 2002-2008 7 floors and above other RC Frame - Shear Wall After 2009 6 floors and below all RC framework After 2009 7 floors and above Residential RC shear wall After 2009 7 floors and above other RC Frame - Shear Wall

[0088] 4. Estimation of Window Openings on Exterior Facades. The minimum width of load-bearing window piers in masonry buildings is determined based on the seismic intensity: not less than 1.0 m in seismic intensity zones 6 and 7, not less than 1.2 m in seismic intensity zone 8, and not less than 1.5 m in seismic intensity zone 9. Along the long side, typical window widths and spacings are matched to the building function and evenly distributed centrally; for residential buildings, window widths are 1.5 m and spacings are 1.5 m; for office, school, and medical buildings, window widths are 1.8 m and spacings are 1.5 m; for industrial and warehouse buildings, window widths are 0.6 m and spacings are 2.1 m. Along the short side, 0 to 2 windows are symmetrically arranged centrally according to functional requirements. During the arrangement process, the width of remaining wall sections is checked to ensure that the window piers meet the minimum width requirements.

[0089] (v) Sampling field survey and accuracy verification

[0090] like Figure 5 As shown, this embodiment selects typical areas within the study region to conduct a sampling field survey. Specifically, the functions, number of floors, and window opening information of 402 buildings were investigated, and the structural type of 57 buildings that met the survey criteria was further verified.

[0091] The building's function was determined through on-site visits and street view images; the number of floors was determined by visually counting the number of floors above ground; the structural type was determined by observing the arrangement of the internal vertical load-bearing components upon entering the building; and the window openings were measured using an industrial laser rangefinder combined with the principle of triangulation to measure the window width and window spacing on typical floors.

[0092] The results were evaluated using the aforementioned accuracy and relative error formulas, as follows: Figure 7The results show that the accuracy rate for building function identification is 92.3%; the average relative error for the number of building floors is 3.2%, and the proportion of samples with an error of no more than 2 floors is 93.3%; the accuracy rate for structure type identification is 96.7%; and the average absolute value of the relative error for the total length of window openings on the building facade is 8.9%.

[0093] Based on the functional identification confusion matrix and the error distribution of floor numbers and window openings, easily confused samples are reviewed. For example, when an office building is defaulted to residential due to a lack of POI, the priority of the office park category in the AOI is increased; when there is a systematic deviation in the number of floors in a medical building, the floor height parameters of the ground floor and standard floors are corrected; when masonry and frame structures are confused during the transition period, a joint determination condition of building function and floor number is added. After correction, the full-area attribute reasoning is re-executed and the database is updated.

[0094] (vi) Standardized database output

[0095] The final database uses BuildingID as the primary key and includes at least: building outline geometry, centroid latitude and longitude, projected coordinates, area, length, width, height, number of floors, building function, structural type, construction year, seismic intensity, window opening parameters, individual building population, AOI identification, and road network adjacency relationships. For example... Figure 6 As shown.

[0096] The database constructed in this embodiment contains 39,200 individual building outlines, 4,256 AOI functional surfaces, 295,735 POI attribute records, and 5,172 backbone road network segments. The database can be exported as a shapefile, GeoJSON, CSV, or relational database table, and can be used as direct input for reconstructing hidden building information, rapidly predicting structural seismic response, simulating building complex damage scenarios, and evaluating post-earthquake functional performance.

[0097] As can be seen, this invention achieves the systematic acquisition and improvement of basic information on urban-scale building clusters at a relatively low survey cost through a complete process of "multi-source public data acquisition → spatial association of individual buildings → domain knowledge reasoning → sampling verification and correction → standardized database construction".

[0098] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and not restrictive.

Claims

1. A method for establishing a basic information database of urban building complexes to provide earthquake services, characterized in that, Includes the following steps: S1. Determine the study area, establish a unique building identifier for each individual building within the study area, and determine the basic information fields of the individual building; S2. Obtain the building plan outline of the study area, and perform standardization processing on the building plan outline to obtain a standardized building outline; S3. Acquire multi-source spatial data of the study area. The multi-source spatial data includes at least building height indication data, population distribution data, point of interest data, interest surface data, and building age indication data from different data platforms, with different spatial resolutions and coordinate systems. Unify all data into the same coordinate system, and use the standardized building outline and its unique building identifier as the spatial association benchmark. Map each data from the original spatial unit to the scale of a single building and then associate it with the corresponding unique building identifier. S4. Based on the building height mapped to the scale of a single building in step S3, the functional clues indicated by the associated interest surface and interest point, and the building age information, a multi-dimensional decision rule base is established in combination with building design codes, structural stress characteristics and regional construction habits. The building function, number of floors, structural type and facade window opening information of the single building are inferred and completed. S5. Verify the reasoning results of step S4 through stratified sampling field survey, calculate the accuracy evaluation index, and correct the judgment threshold or decision boundary in the multidimensional decision rule base according to the verification results, and re-execute the attribute reasoning of step S4 with the updated rule base. S6. Using the unique building identifier as the primary key, output a standardized urban building cluster basic database containing the basic information of the individual building.

2. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, In step S3, the spatial association methods for mapping each data from the original spatial unit to the scale of a single building include: spatial inclusion or spatial intersection to determine the attribution of area data, nearest neighbor matching for point data, and area weight allocation for raster data.

3. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, The building height indication data mentioned in step S3 is the digital surface model and the digital elevation model; after resampling and meshing the digital surface model and the digital elevation model, the elevation value is extracted from the area covered by the building centroid or building outline, and the building height is equal to the difference between the corresponding digital surface elevation and the digital ground elevation.

4. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, The building age indication data mentioned in step S3 is the time series data of impermeable ground surface year by year; the state of impermeable ground surface is read year by year with the building centroid or the area covered by the building outline as the index, and the year in which the corresponding location first changes from impermeable surface to impermeable surface is determined as the building construction time or construction age range.

5. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, The population distribution data mentioned in step S3 is allocated to individual buildings from the population raster in the following manner: For building i within population raster cell g, population P is allocated. i =P g ×A i,g / ΣA j,g , where P g For grid population, A i,g Let ΣA be the area of ​​the outline of building i within grid g. j,g It is the sum of the outline areas of all buildings within the grid.

6. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, The reasoning method for building functions in step S4 is as follows: first, determine the main function type of the area based on the category of the interest surface to which the building belongs, and then refine the use of individual buildings by matching the interest point keywords associated with the building; when the same building is associated with multiple interest points of different function categories, it is marked as a composite function building.

7. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, The reasoning method for the number of building floors in step S4 is as follows: based on the typical floor height range corresponding to the building function, the number of floors is estimated by dividing the total building height by the average floor height of the standard floors and rounding it down; for buildings with tall spaces on the ground floor, the floor height of the ground floor is taken as the upper limit of the corresponding floor height range, and the number of floors is calculated by rounding it down to the average floor height of (total building height - ground floor height) / average floor height of the standard floors + 1.

8. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, The reasoning method for the structural type mentioned in step S4 is as follows: taking the building age, number of floors, and building function as inputs, a multi-dimensional decision table is established according to the development law of structural systems in different ages, the stress requirements of different number of floors, and the spatial requirements of different functions, and the building is determined to be a masonry structure, a reinforced concrete frame structure, a reinforced concrete frame-shear wall structure, or a reinforced concrete shear wall structure.

9. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, The reasoning method for the exterior window opening information in step S4 is as follows: determine the minimum window wall width based on the seismic fortification intensity, match the typical window width and window spacing according to the building function, arrange window openings evenly in the center along the long side of the building, and arrange zero to two window openings symmetrically in the center along the short side of the building, and the width of the window wall after arrangement is not less than the minimum window wall width.

10. The method for establishing a basic information database of urban building clusters providing earthquake services according to claim 1, characterized in that, The accuracy evaluation indicators mentioned in step S5 include: the classification accuracy evaluation of building function and structural type, and the relative error evaluation of building number and total length of window openings on facade; the correction includes: adjusting the typical floor height parameters of each functional building according to the error distribution, modifying the floor number threshold or function determination conditions in the structural type decision table, or adjusting the window opening layout module.

Images