Method and system for constructing building earthquake damage scenarios based on UAV remote sensing images

Abstract

The present application provides a method and system for constructing earthquake damage scenarios for buildings based on UAV remote sensing images, which relates to the field of urban earthquake disaster prevention technology, including: using an urban earthquake damage simulation system to construct a three-dimensional spatial model of a building and determine the nonlinear model parameters of the building; using an elastic-plastic time-history analysis method to construct earthquake damage scenarios with a single building as the granularity and determine the building damage index; determining the earthquake damage risk level of each building based on the mapping relationship between the preset building damage index and the building earthquake damage risk level; testing the earthquake resistance of buildings with a specified earthquake damage risk level based on a building list, and generating earthquake resistance measures corresponding to buildings with a specified earthquake damage risk level. The present application constructs a three-dimensional spatial model based on the geometric attribute information of the building in combination with the urban earthquake damage simulation system, accurately divides the earthquake damage risk level of a single building, and combines earthquake resistance measures to improve the refinement level of urban earthquake disaster prevention.

Description

Technical Field

[0001] The present application relates to the field of urban earthquake resistance and disaster prevention technology, and in particular to a method and system for constructing building earthquake damage scenarios based on drone remote sensing images. Background Art

[0002] With rapid economic and social development and the continued advancement of urbanization, urban systems are becoming increasingly complex, and the risks of economic and human losses caused by natural disasters such as earthquakes are becoming more severe. Buildings, as the primary vehicles for urban activities, have a seismic performance that is closely linked to the city's earthquake resilience. Previous earthquake disaster experiences have shown that building collapse is often the primary cause of casualties and property loss. Currently, the seismic resistance of buildings within cities varies widely. In the future, destructive earthquakes in cities and their surrounding areas will cause far greater casualties and economic losses than elsewhere, increasing the risk of cities suffering significant earthquake losses exponentially. Under new economic and technological conditions, understanding of urban disaster mechanisms and the underlying risk factors is limited. However, the development of technologies such as big data, simulation, and artificial intelligence has provided new tools for urban disaster prevention and mitigation. Urban disaster prevention and earthquake emergency response are evolving towards a more refined, scientific, intelligent, and real-time dynamic approach. Efficient and refined disaster risk management has become a crucial tool for promoting urban safety and sustainable economic development. Earthquake risk assessment analyzes and estimates the losses caused by potential future earthquakes, determining the extent of building damage, the scale of casualties, and other damage scenarios. This provides a foundation for effective disaster response and loss reduction. Therefore, constructing urban earthquake disaster loss scenarios and conducting detailed earthquake disaster risk assessments to gain a detailed understanding of the seismic resistance of urban buildings can provide a scientific basis for cities to prevent and mitigate future major earthquake disasters. It can also identify areas of weak seismic resistance in urban buildings and provide targeted recommendations and suggestions for urban building reinforcement projects and key hidden danger inspections, thereby minimizing casualties and economic losses caused by earthquakes. This is of great significance for improving urban earthquake disaster resilience and ensuring smooth social and economic development.

[0003] The existing scheme is based on the earthquake damage simulation technology of finite element analysis. By establishing a simplified linear model of the building, combining it with preset seismic motion parameters to calculate the overall response, and using empirical formulas to make probabilistic predictions on component damage.

[0004] Existing solutions rely on simplified linear assumptions, making it difficult to accurately characterize the nonlinear hysteretic behavior of buildings under strong earthquakes, resulting in large deviations between damage prediction results and actual failure modes. At the same time, empirical formulas are unable to dynamically reflect the interaction between components on different floors and lack the ability to provide a detailed analysis of the spatial distribution of local damage. Summary of the Invention

[0005] The present application provides a method and system for constructing building earthquake damage scenarios based on drone remote sensing images, which is used to solve the problems of low accuracy in constructing building earthquake damage scenarios in the existing technology, such as inaccurate characterization of nonlinear hysteretic behavior of buildings, insufficient analysis of dynamic interactions between components, and lack of refined analysis capabilities of the spatial distribution of local damage.

[0006] In a first aspect, the present application provides a method for constructing building earthquake damage scenarios based on drone remote sensing images, comprising:

[0007] Use drones to take multi-angle aerial photos of the target area, obtain drone remote sensing image data, and use image recognition algorithms to extract the geometric attribute information of the building;

[0008] According to the geometric attribute information of the building, a three-dimensional spatial model of the building is constructed using an urban earthquake damage simulation system, and nonlinear model parameters of the building are determined based on the three-dimensional spatial model;

[0009] Based on the nonlinear model parameters and the preset earthquake motion parameters, an elastic-plastic time history analysis method is used to construct earthquake damage scenarios with a single building as the granularity, to obtain the spatial distribution of damage to each building and determine the building damage index;

[0010] Determine the earthquake damage risk level of each building based on the mapping relationship between the preset building damage index and the building earthquake damage risk level;

[0011] According to the list of buildings with specified earthquake damage risk levels, the seismic performance of buildings with specified earthquake damage risk levels is tested, and based on the seismic performance test results, seismic measures corresponding to buildings with specified earthquake damage risk levels are generated, wherein the specified earthquake damage risk level is a level greater than a preset earthquake damage risk level.

[0012] Optionally, the nonlinear model parameters include: floor mass, floor stiffness, and a relationship curve between floor shear force and displacement; and the elastic-plastic time history analysis method includes the Newmark-β method;

[0013] Based on the nonlinear model parameters and in combination with preset earthquake motion parameters, an elastic-plastic time history analysis method is used to construct earthquake damage scenarios with a single building as the granularity, thereby obtaining the spatial distribution of damage to each building and determining the building damage index, including:

[0014] Normalizing the floor mass and the floor stiffness respectively, and discretizing the relationship curve between the floor shear force and displacement to obtain a standardized nonlinear parameter input file;

[0015] Based on the preset seismic motion parameters, the Newmark-β method is used to construct earthquake damage scenarios at the granularity of a single building, and the time-history response dataset of each building is solved.

[0016] Inputting the time-history response dataset into a pre-trained hybrid damage assessment model to output the damage spatial distribution of the corresponding building, wherein the damage spatial distribution includes the damage distribution of concrete components and steel structures on each floor;

[0017] The damage distribution of concrete components and steel structures on all floors are fused to obtain the building damage index.

[0018] Optionally, the step of inputting the time-history response dataset into a pre-trained hybrid damage assessment model to output a spatial distribution of damage corresponding to the building, wherein the spatial distribution of damage includes the damage distribution of concrete components and the damage distribution of steel structures on each floor, including:

[0019] For concrete components, the inter-story displacement angle, cumulative energy dissipation, and steel strain time history in the time history response dataset are extracted and input into the concrete damage assessment module in the hybrid damage assessment model. The concrete crack width, percentage of cover loss area, and damage depth of the compression zone of each floor are output;

[0020] The spatial distribution of concrete components is combined with the concrete crack width, the proportion of the cover shedding area, and the damage depth of the compression zone on each floor to obtain the concrete component damage distribution on each floor.

[0021] For steel structural components, the interlayer shear force time history, node plastic rotation angle, and local buckling strain are extracted and input into the steel structure damage assessment model in the hybrid damage assessment model. The fracture probability of the steel beam-column connection node, the web buckling index, and the overall instability risk level of the steel structure are output;

[0022] The spatial distribution of steel structure components is combined with the fracture probability of steel beam-column connection nodes, the web buckling index and the overall instability risk level of the steel structure to obtain the steel structure damage distribution on each floor.

[0023] Optionally, the fusion processing of the concrete component damage distribution and the steel structure damage distribution of all floors to obtain the building damage index includes:

[0024] Determine the corresponding concrete weight factor and steel structure weight factor according to the structural type of the building;

[0025] performing spatial integration of the first damage value of each floor in the concrete component damage distribution and the second damage value in the steel structure damage distribution according to the concrete weight factor and the steel structure weight factor to obtain a total damage value of each floor;

[0026] The aging coefficient is introduced according to the construction year to correct the total damage value of each floor, and the corrected total damage value of each floor is obtained;

[0027] The entropy weight method is used to fuse the corrected total damage values of each floor to obtain the building damage index.

[0028] Optionally, the geometric attribute information includes construction year, number of floors, height, structure type and usage type;

[0029] The extraction of geometric attribute information of buildings by combining an image recognition algorithm includes:

[0030] Extracting the building's facade reflectance spectrum data from drone remote sensing image data, identifying the building's facade material based on a preset library of matching between reflectance spectra and materials, and using a random forest classifier to identify the building's construction year based on the facade material's spectral fingerprint;

[0031] Determine the structural type based on the beam-column node morphological features from the UAV remote sensing image data, and identify the usage type in combination with the red-hot external data;

[0032] Based on the three-dimensional point cloud data, the number of point cloud density mutation layers is counted along the elevation direction. Combined with the window opening distribution verification, the number of floors is determined, and the height of the building is inverted and calculated to obtain the height.

[0033] Optionally, the UAV remote sensing image data includes visible light image data and infrared light image data;

[0034] The method of determining the structural type based on the beam-column node morphological features from the UAV remote sensing image data and identifying the use type in combination with the red-hot external data includes:

[0035] Input the visible light image data into the improved ResNet-50 network and output pixel-level beam-column node segmentation mask;

[0036] Based on the geometric shape and size ratio of the segmentation mask, the preset structure type template library is matched to determine the structure type of the building;

[0037] The temperature distribution features are extracted from the infrared image data, and combined with the day-night temperature difference pattern to construct a use type thermal feature map; the texture features in the visible light image are integrated with the use type thermal feature map, and the use type is output through a support vector machine classifier.

[0038] Optionally, the method of counting the number of layers of point cloud density mutations along the elevation direction based on the three-dimensional point cloud data, combining the window opening distribution verification to determine the number of layers, and performing height inversion calculation on the building to obtain the height includes:

[0039] De-noising and ground filtering are performed on the three-dimensional point cloud data to obtain the main point cloud of the building;

[0040] Based on the main point cloud of the building, vertical slices are divided at intervals of preset units along the elevation direction, the point cloud density in each vertical slice is counted, and the peak point of density mutation is detected by Gaussian difference operator to determine the candidate floor interface;

[0041] The number of floors is determined based on the number of candidate floor interfaces. The difference between the highest interface elevation and the base elevation is calculated based on the building base elevation. Combined with the scale verification in the drone image, the building height value is output.

[0042] In a second aspect, the present application provides a system for constructing building earthquake damage scenarios based on drone remote sensing images, comprising:

[0043] The acquisition and extraction module is used to use drones to take multi-angle aerial photos of the target area, obtain drone remote sensing image data, and extract the geometric attribute information of the building in combination with image recognition algorithms;

[0044] A construction and determination module is used to construct a three-dimensional spatial model of the building using the urban earthquake damage simulation system according to the geometric attribute information of the building, and determine the nonlinear model parameters of the building based on the three-dimensional spatial model;

[0045] A simulation module is used to construct earthquake damage scenarios based on the nonlinear model parameters and preset earthquake motion parameters using an elastic-plastic time history analysis method with a single building as the granularity, obtain the spatial distribution of damage to each building, and determine the building damage index;

[0046] A determination module is used to determine the earthquake damage risk level of each building based on a preset mapping relationship between the building damage index and the building earthquake damage risk level;

[0047] The detection generation module is used to detect the seismic performance of buildings with a specified earthquake damage risk level based on a list of buildings with a specified earthquake damage risk level, and generate seismic measures corresponding to buildings with a specified earthquake damage risk level based on the seismic performance test results. The specified earthquake damage risk level is a level greater than a preset earthquake damage risk level.

[0048] In a third aspect, the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for constructing a building earthquake damage scenario based on drone remote sensing images as described in any one of the first aspects.

[0049] In a fourth aspect, the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements a method for constructing a building earthquake damage scenario based on drone remote sensing images as described in any one of the first aspects.

[0050] In the present application, a method for constructing a building earthquake damage scenario based on drone remote sensing images is provided, the method comprising: taking multi-angle aerial photos of a target area by a drone to obtain drone remote sensing image data, and extracting geometric attribute information of the building in combination with an image recognition algorithm; constructing a three-dimensional spatial model of the building using an urban earthquake damage simulation system based on the geometric attribute information of the building, and determining the nonlinear model parameters of the building based on the three-dimensional spatial model; based on the nonlinear model parameters and in combination with preset seismic motion parameters, constructing an earthquake damage scenario with a single building as the granularity using an elastic-plastic time-history analysis method, obtaining the spatial distribution of damage and destruction of each building, and determining the building damage index; determining the earthquake damage risk level of each building based on the mapping relationship between the preset building damage index and the building earthquake damage risk level; testing the seismic performance of buildings with the specified earthquake damage risk level based on a list of buildings with the specified earthquake damage risk level, and generating earthquake resistance measures corresponding to the buildings with the specified earthquake damage risk level based on the seismic performance test results, wherein the specified earthquake damage risk level is a level greater than the preset earthquake damage risk level.

[0051] This application uses drones to take multi-angle aerial photos of the target area, and combines image recognition algorithms to efficiently obtain geometric attribute information of buildings, providing a high-precision data foundation for model construction; establishes a refined physical model of the building based on three-dimensional spatial models and nonlinear parameters to accurately characterize the structural dynamic characteristics; utilizes elastic-plastic time-history analysis methods to simulate earthquake damage responses at the granularity of individual buildings, revealing the spatial distribution of damage and quantifying the damage index; implements risk classification management through preset grade mapping rules, accurately locks in high-risk targets, and improves the accuracy of building earthquake damage scenario construction; combines on-site detection data to generate targeted earthquake resistance measures, forming a closed-loop process from assessment, diagnosis to governance, and improving the refinement and scientific nature of urban earthquake disaster prevention.

[0052] Furthermore, based on the nonlinear model parameters and preset seismic motion parameters, a standardized input file is generated by normalizing the floor mass, stiffness, and discretized floor shear force-displacement curves. The Newmark-β method is used for elastic-plastic time-history analysis to solve the time-history response data of individual buildings. Using a hybrid damage assessment model, key parameters such as inter-story displacement angle, cumulative energy dissipation, and shear force time history are extracted, dynamically analyzing damage characteristics such as concrete crack width, percentage of protective layer detachment area, and fracture probability of steel beam-column connection nodes. Finally, the multiple damage distributions of all floors are integrated to generate an overall damage index. Through refined nonlinear modeling and hybrid damage assessment, nonlinear hysteretic behaviors such as concrete cracking, steel yielding, and plastic deformation of steel structure nodes are accurately quantified, and interactions between components are dynamically characterized. Combined with the floor damage spatial distribution fusion algorithm, a refined analysis of local damage patterns is achieved, improving the consistency between earthquake damage prediction and actual failure patterns, and providing high-precision data support for seismic reinforcement of high-risk buildings.

[0053] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0055] Figure 1 A flowchart of a method for constructing building earthquake damage scenarios based on drone remote sensing images provided in an embodiment of the present application;

[0056] Figure 2 A schematic diagram of a building damage level indicator provided in an embodiment of the present application;

[0057] Figure 3 A schematic diagram of the structure of a building earthquake damage scenario construction system based on drone remote sensing images provided in an embodiment of the present application;

[0058] Figure 4 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0060] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 11, 12, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0061] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0062] In order to solve the problem of low accuracy in constructing earthquake damage scenarios for buildings caused by inaccurate characterization of nonlinear hysteretic behavior of buildings, insufficient analysis of dynamic interactions between components, and lack of ability to fine-tune the spatial distribution of local damage in the existing technology, an embodiment of the present application provides a method for constructing earthquake damage scenarios for buildings based on drone remote sensing images. The method adopts the following ideas: quickly acquire building geometric data through drone aerial photography and image recognition technology, establish a refined three-dimensional model and derive nonlinear parameters; use elastic-plastic time-history analysis to simulate the seismic response of individual buildings, quantify the damage index and divide the risk level; implement on-site inspection and seismic performance evaluation in combination with a list of high-risk buildings, and finally generate targeted reinforcement strategies, forming a full-chain technical system of "data collection-model construction-simulation analysis-risk classification-precise governance", realizing refined management of urban seismic capacity assessment and governance, and improving the accuracy of building earthquake damage scenario construction.

[0063] Figure 1 A flowchart of a method for constructing a building earthquake damage scenario based on drone remote sensing images is provided in an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0064] S11. Use drones to take multi-angle aerial photos of the target area, obtain drone remote sensing image data, and use image recognition algorithms to extract the geometric attribute information of the building.

[0065] UAV remote sensing image data refers to high-resolution image data acquired by drone-mounted cameras that includes the location, shape, and texture of buildings. Building geometric attribute information refers to spatial characteristic parameters such as the building's length, width, height, number of floors, and floor plan.

[0066] In an embodiment of the present application, a certain type of drone is used to perform multi-angle aerial photography of the target area to obtain high-resolution drone remote sensing image data; the YOLOv5 target detection algorithm and edge recognition technology can be used to pre-process the above image data, segment the building outline and extract geometric attribute information, and eliminate image distortion through a three-dimensional point cloud registration algorithm, and finally generate a structured attribute database containing position coordinates and geometric dimensions.

[0067] In practical applications, the embodiment of the present application can analyze the regional division, geological characteristics, and historical earthquakes of a province, and carry out a refined assessment of the earthquake disaster risk in the urban area with County-level City A as the assessment area. Historical earthquakes can be obtained through seismic data, and then regional destructive earthquakes greater than magnitude 4.7 can be compiled. Because the seismic data of the area can be stored in the form of a table, it includes: the year, month, day, longitude and latitude of the earthquake, magnitude, focal depth, reference location, epicenter intensity, and accuracy.

[0068] For example, this assessment used building data obtained from the first national comprehensive natural disaster risk survey in a certain province. This building data includes spatial layers and attribute data. The spatial layers include urban residential buildings, urban non-residential buildings, rural single-family homes, rural residential complexes, and rural non-residential buildings. The data is formatted as a shapefile, with relevant survey attributes organized within the shapefile's attribute fields. These include building area, number of floors, construction year, and structural type. Geographic Information System (GIS) software was used to process the building data, creating vector SHP data, and extracting building units in City A.

[0069] For example, there are 205,715 buildings in City A, including 9,668 urban residential buildings, 18,634 urban non-residential buildings, 125,951 rural independent residential buildings, 2,960 rural collective residential buildings, and 48,502 rural non-residential buildings. The statistics of urban residential structure types are shown in Table 1:

[0070] Table 1 Statistics of urban residential structure types

[0071]

[0072] S12. According to the geometric attribute information of the building, a three-dimensional spatial model of the building is constructed using the urban earthquake damage simulation system, and nonlinear model parameters of the building are determined based on the three-dimensional spatial model.

[0073] The urban earthquake damage simulation system is a software platform based on the finite element or discrete element method, used to simulate the dynamic response of buildings under earthquakes. A three-dimensional spatial model of a building is a digital model of the building's geometry and structural topology described in a three-dimensional coordinate system. Nonlinear model parameters characterize the nonlinear behavior of materials, such as concrete cracking strength, steel yield strain, and hysteretic energy dissipation coefficient.

[0074] In an embodiment of the present application, the geometric attribute information of the building is imported into the urban earthquake damage simulation system, and a three-dimensional space model is constructed based on the number of building floors, plane layout and height parameters; the floor mass and floor stiffness are generated through finite element meshing, and the nonlinear model parameters are derived in combination with the material constitutive model to form a standardized parameter input file.

[0075] S13. Based on the nonlinear model parameters and the preset seismic motion parameters, the elastic-plastic time-history analysis method is used to construct earthquake damage scenarios with a single building as the granularity, obtain the spatial distribution of damage to each building, and determine the building damage index.

[0076] Among them, the preset seismic motion parameters refer to the input seismic wave characteristic parameters, including peak acceleration, spectral characteristics, duration, etc. The elastoplastic time-history analysis method refers to a numerical method for solving the nonlinear dynamic response of a structure under earthquake action through step-by-step integration. The earthquake damage scenario refers to the simulation of the possible damage mode and degree of a building under a specific earthquake motion. The spatial distribution of damage to a building refers to the spatialized data describing the location and severity of damage to components on each floor. The building damage index refers to a quantitative indicator of overall damage calculated by comprehensively considering the degree of damage on each floor.

[0077] In an embodiment of the present application, based on the nonlinear model parameters and preset seismic motion parameters, an elastic-plastic time-history analysis method is adopted to solve the dynamic equations under the action of an earthquake with a single building as the granularity, and obtain a displacement, velocity, and acceleration time-history response data set of each floor; the time-history data is input into a hybrid damage assessment model, and the damage indicators of each floor are weighted and fused to generate a spatial distribution map of building damage and destruction, and the overall damage index is calculated.

[0078] S14. Determine the earthquake damage risk level of each building based on the mapping relationship between the preset building damage index and the building earthquake damage risk level.

[0079] The building damage risk level refers to the seismic resistance rating of a building based on its damage index. The building inventory contains a table or database of building numbers, locations, and risk levels. The mapping between building damage indices and building damage risk levels can be determined using Tables 2 and 3. Table 2 provides the correspondence between building damage levels and risk types, while Table 3 provides the correspondence between risk types and earthquake damage risk levels.

[0080] Table 2 Building damage level table

[0081]

[0082] Table 3 Building earthquake damage risk level table

[0083]

[0084] In an embodiment of the present application, based on the preset mapping relationship between building damage index and earthquake risk level, the damage index of each building is mapped to five risk levels through interval division rules, and a list containing building number, location, and risk level is generated and output in a map visualization form.

[0085] S15. Based on the list of buildings with a specified earthquake damage risk level, the seismic performance of the buildings with the specified earthquake damage risk level is tested, and based on the seismic performance test results, seismic measures corresponding to the buildings with the specified earthquake damage risk level are generated, where the specified earthquake damage risk level is a level greater than a preset earthquake damage risk level.

[0086] The designated earthquake damage risk level can be Level I, corresponding to high risk, and the preset earthquake damage risk level can be Level II, corresponding to medium-to-high risk. A building's seismic performance refers to its ability to resist earthquake damage, including indicators such as strength, stiffness, and ductility. Seismic measures refer to reinforcement plans for weak links in a building, such as adding supports and strengthening joints.

[0087] In an embodiment of the present application, based on a list of high-risk buildings, a rebound tester, a steel bar scanner, and a floor thickness gauge are used to conduct on-site seismic performance testing; combining the test data with the simulation results of the structural analysis software, targeted seismic measures are generated to form a reinforcement plan report.

[0088] In a specific example, a three-dimensional spatial model of a building is constructed by combining basic information such as the building structure type, construction year, and site type, and its nonlinear model parameters are calculated, including floor area, floor mass, floor stiffness, and floor shear force-displacement relationship curve. The elastic-plastic time-history analysis method is used to simulate the three-dimensional earthquake damage of individual buildings in City A, obtain the spatial distribution of damage to each building, and determine the building damage index. The damage index of a building is as follows: Figure 2The area between the nominal yield point and the nominal collapse point is classified as slightly damaged, moderately damaged, and severely damaged, while the area after the nominal collapse point is classified as destroyed or collapsed, as shown in Figure 2. Risk management and response recommendations are proposed for buildings of different damage levels, aiming to provide reference and inspiration for the government to adopt reinforcement measures, secondary disaster avoidance measures, monitoring and maintenance, and emergency response plans.

[0089] Here's a specific example: First, a certain type of drone was used to take aerial photos of the urban area of City A at a resolution of 0.1 meters. The YOLOv5 algorithm was used to extract the geometric attributes of 205,715 buildings, and the floor identification error was controlled within ±1 floor. Secondly, a three-dimensional model was constructed in the urban earthquake damage simulation system, nonlinear parameters were calculated, and seismic motion records with a peak acceleration of 0.15g were input. The simulation results showed that 23.02% of urban buildings were high-risk. Subsequently, on-site inspections were conducted on 652 of the 6,516 high-risk buildings, and it was found that the average strength of the concrete in the masonry structures had decreased. The generated reinforcement plan reduced the estimated repair cost. Finally, the risk level distribution was superimposed on the emergency resource deployment map to improve the efficiency of post-earthquake rescue route planning.

[0090] Optionally, the detailed assessment of urban earthquake disaster risks will be carried out mainly from six levels, including basic data collection, risk assessment calculation, risk hazard sorting, on-site field review, establishment of risk files, and proposal of countermeasures and suggestions.

[0091] By executing S11 to S15, the embodiment of the present application quickly obtains building geometric data through drone aerial photography and intelligent identification, and combines refined nonlinear modeling and elastic-plastic time-history analysis to achieve accurate prediction of the spatial distribution of damage in individual buildings; based on damage index classification and on-site detection data, targeted reinforcement strategies are generated to form a closed loop of "data collection-simulation evaluation-risk classification-precise governance", thereby improving the refinement level of urban earthquake resistance and disaster prevention and the efficiency of emergency response.

[0092] In one possible embodiment, the nonlinear model parameters include: floor mass, floor stiffness, and a relationship curve between floor shear force and displacement, and the elastic-plastic time history analysis method includes the Newmark-β method. S13. Based on the nonlinear model parameters and in combination with preset earthquake motion parameters, an elastic-plastic time history analysis method is used to construct earthquake damage scenarios at the granularity of a single building, obtain the spatial distribution of damage to each building, and determine the building damage index, including:

[0093] Step 131: Normalize the floor mass and floor stiffness respectively, discretize the relationship curve between floor shear force and displacement, and obtain a standardized nonlinear parameter input file.

[0094] Among them, the standardized nonlinear parameter input file refers to a parameter file that has been normalized and discretized, containing floor mass, stiffness, and shear force-displacement relationship data, which is used to unify the input format and improve calculation efficiency.

[0095] In an embodiment of the present application, the floor mass and floor stiffness are normalized respectively, and the mass and stiffness values are scaled as percentages of the maximum values; the floor shear force and displacement relationship curve is discretized, and the continuous curve is divided into multiple linear segments to generate a standardized nonlinear parameter input file.

[0096] Step 132: Based on the preset earthquake motion parameters, the Newmark-β method is used to construct earthquake damage scenarios with a single building as the granularity, and the time-history response dataset of each building is solved.

[0097] Among them, the time-history response data set refers to the dynamic response data set of the structure under earthquake action, including the displacement, velocity, acceleration and energy dissipation values at each time point.

[0098] In the embodiment of the present application, based on preset seismic parameters, the Newmark-β method is used to solve the dynamic equation with a single building as the granularity; the displacement, velocity, and acceleration of each time step are iteratively calculated to generate a time history response data set.

[0099] Step 133: Input the time-history response dataset into a pre-trained hybrid damage assessment model, and output the spatial distribution of damage corresponding to the building. The spatial distribution of damage includes the damage distribution of concrete components and steel structures on each floor.

[0100] Hybrid damage assessment models combine machine learning and physical principles, including independent modules for damage prediction for concrete and steel structures. Spatial damage distribution refers to spatialized data describing the location and extent of damage to components on each floor of a building, typically presented as a heat map or grid.

[0101] In an embodiment of the present application, the time-history response dataset is input into a pre-trained hybrid damage assessment model, in which the concrete damage module predicts the width of concrete cracks on each floor through a convolutional neural network, and the steel structure damage module calculates the fracture probability of steel beam-column nodes through a random forest algorithm; finally, the concrete component damage distribution and steel structure damage distribution of each floor are output.

[0102] Step 134: Fuse the concrete component damage distribution and the steel structure damage distribution of all floors to obtain a building damage index.

[0103] The concrete component damage distribution of each floor refers to the spatial distribution of the damage characteristics of the concrete components within the floor. The steel structure damage distribution refers to the spatial distribution of the damage characteristics of the steel components within the floor.

[0104] In an embodiment of the present application, the damage distribution of concrete components and steel structures on all floors are weighted and fused to calculate the overall building damage index, where the floor damage weight is determined by the floor height and the functional importance coefficient.

[0105] The following is a specific example: First, the floor mass of a masonry structure building in City A was normalized to reduce the mass proportion of the bottom floor and increase the proportion of the top floor. Secondly, the shear force-displacement curve was processed using a discretization step size of 0.1 mm to generate an input file containing 1,200 data points. Subsequently, the Newmark-β method was applied to simulate the seismic motion with a peak acceleration of 0.25g, reducing the time-consuming time-history analysis. Next, the hybrid damage model identified that the crack density of the 3-story concrete wall reached 15 cracks / square meter, and the buckling strain of the 2-story steel beam exceeded the limit by 40%. Finally, the fusion damage index was 0.72, which was judged to be a high-risk building. A reinforcement plan of adding ring beams and carbon fiber cloth was generated, which reduced the estimated repair cost.

[0106] By executing steps 131 to 134, the embodiment of the present application accurately simulates the nonlinear hysteretic behavior of a building under a strong earthquake through standardized parameter processing and elastic-plastic time-history analysis; utilizes a hybrid damage assessment model to dynamically analyze the local damage characteristics of concrete and steel structures, and combines a floor fusion algorithm to generate a high-precision damage index, thereby improving the consistency between earthquake damage prediction and actual damage mode, and providing a reliable basis for the reinforcement of high-risk buildings.

[0107] In one possible embodiment, step 133 inputs the time-history response dataset into a pre-trained hybrid damage assessment model, and outputs a spatial distribution of damage corresponding to the building. The spatial distribution of damage includes the damage distribution of concrete components and the damage distribution of steel structures on each floor, including:

[0108] Step a1: For concrete components, extract the inter-story displacement angle, cumulative energy dissipation, and steel strain time history from the time-history response dataset and input them into the concrete damage assessment module in the hybrid damage assessment model. Output is the concrete crack width, percentage of cover loss area, and damage depth of the compression zone for each floor.

[0109] The inter-story drift angle, which measures the ratio of the horizontal displacement difference between floors to the floor height, reflects the degree of structural deformation. Cumulative energy dissipation, the sum of the energy dissipated by each structural component during an earthquake, is used to assess the cumulative effects of damage. The steel bar strain history is a curve showing the strain of steel bars over time under earthquake action.

[0110] In an embodiment of the present application, for concrete components, the inter-story displacement angle, cumulative energy consumption and steel strain time history are extracted from the time-history response data set and input into the concrete damage assessment module of the hybrid damage assessment model; the module predicts the width of concrete cracks on each floor through regression analysis, and calculates the proportion of protective layer detachment area through image segmentation algorithm, combines finite element inversion to obtain the damage depth of the compressive zone, and outputs quantitative damage parameters.

[0111] Step a2: Combine the spatial distribution of concrete components with the concrete crack width, the proportion of the protective layer falling off area, and the damage depth of the compression zone on each floor to obtain the concrete component damage distribution on each floor.

[0112] The concrete crack width refers to the maximum opening width of a crack on the concrete surface. The cover loss area ratio refers to the percentage of the concrete cover loss area to the component surface area. The compression zone damage depth refers to the depth of material degradation caused by damage in the compression zone of the concrete.

[0113] In an embodiment of the present application, the spatial distribution of concrete components is spatially superimposed with the output concrete crack width, the proportion of protective layer detachment area and the damage depth of the compression zone, and a rasterization algorithm is used to generate a concrete component damage distribution map for each floor. Through weighted fusion, the overall concrete damage score of the floor is obtained.

[0114] Step a3: For steel structural components, extract the interlayer shear force time history, node plastic rotation angle, and local buckling strain, input them into the steel structure damage assessment model in the hybrid damage assessment model, and output the fracture probability of the steel beam-column connection node, the web buckling index, and the overall instability risk level of the steel structure.

[0115] The interlayer shear force time history refers to the shear force curve of each floor under earthquake action over time. The node plastic rotation angle refers to the cumulative rotation angle of the plastic hinge area of the steel beam-column node, reflecting the degree of plastic deformation of the node. Steel structural components refer to the load-bearing components such as steel beams, steel columns, and nodes used in buildings. Local buckling strain is the critical strain value when buckling occurs in a local area of a steel component. The fracture probability of a steel beam-column connection node refers to the possibility of fracture of the node due to plastic deformation or fatigue.

[0116] In an embodiment of the present application, for steel structure components, the interlayer shear force time history, node plastic rotation angle and local buckling strain are extracted from the time history response data set and input into the steel structure damage assessment module of the hybrid damage assessment model; the module calculates the fracture probability of the steel beam-column connection node through a probabilistic model, evaluates the web buckling risk through the buckling stability formula, and determines the overall instability risk level of the steel structure in combination with the overall stiffness degradation coefficient.

[0117] Step a4: Combine the spatial distribution of steel structure components with the fracture probability of steel beam-column connection nodes, the web buckling index, and the overall instability risk level of the steel structure to obtain the steel structure damage distribution on each floor.

[0118] The web buckling index, the ratio of the actual height-to-thickness ratio of the web to the critical height-to-thickness ratio, is used to assess buckling risk. The overall instability risk level of a steel structure refers to the risk of overall structural instability loss due to local damage.

[0119] In an embodiment of the present application, the spatial distribution of steel structure components is spatially correlated with the output node fracture probability, web buckling index and overall instability risk level, high-risk areas are divided through cluster analysis, and a steel structure damage distribution map of each floor is generated. The degree of steel structure damage on each floor is quantified through the risk superposition formula.

[0120] By executing steps a1 to a4, the embodiment of the present application accurately quantifies the nonlinear damage characteristics of key components by separating the damage assessment modules of concrete and steel structures; combined with the spatial distribution fusion algorithm, it dynamically analyzes the impact of local damage on the overall structure, and realizes multi-scale earthquake damage assessment from micro components to macro floors, thereby improving the accuracy and operability of earthquake damage prediction for complex structures.

[0121] In a possible embodiment, step 134, fusing the concrete component damage distribution and the steel structure damage distribution of all floors to obtain a building damage index, includes:

[0122] Step c1: Determine the corresponding concrete weight factor and steel structure weight factor according to the structural type of the building.

[0123] The building's structural type refers to the type of load-bearing system, such as frame structure, shear wall structure, or frame-core tube structure. The concrete weight factor refers to the weight coefficient of the concrete component's impact on overall damage. The steel structure weight factor refers to the weight coefficient of the steel component's impact on overall damage.

[0124] Step c2: Based on the concrete weight factor and the steel structure weight factor, spatially integrate the first damage value of each floor in the concrete component damage distribution and the second damage value in the steel structure damage distribution to obtain the total damage value of each floor.

[0125] The first damage value refers to the damage score of concrete components. The second damage value refers to the damage score of steel components. The total damage value is the weighted comprehensive score of the concrete and steel damage of a single floor.

[0126] Step c3: introducing an aging coefficient according to the construction year, correcting the total damage value of each floor, and obtaining a corrected total damage value of each floor.

[0127] The aging factor is a correction factor for material degradation based on the building's service life. The corrected total damage value is the floor damage score that takes aging effects into account.

[0128] Step c4: Use the entropy weight method to fuse the corrected total damage values of each floor to obtain the building damage index.

[0129] Among them, the entropy weight method refers to a method of calculating indicator weights based on information entropy theory, which reflects the impact of data dispersion on weights.

[0130] By executing steps c1 to c4, the embodiment of the present application accurately quantifies the synergistic damage effects of concrete and steel structures through differentiated weight allocation and spatial integration of structural types; dynamically corrects the impact of material degradation in combination with the aging coefficient, and objectively integrates multi-floor damage data using the entropy weight method, thereby improving the reliability and scientific nature of the overall damage assessment of the building and providing a quantitative basis for seismic reinforcement priority decision-making.

[0131] In a possible embodiment, the geometric attribute information includes the construction year, number of floors, height, structure type, and usage type. S11, extracting the geometric attribute information of the building in combination with the image recognition algorithm includes:

[0132] Step 111: Extract the building's facade reflectance spectrum data from the drone remote sensing image data, identify the building's facade material based on a preset matching library between the reflectance spectrum and the material, and use a random forest classifier to identify the building age based on the spectral fingerprint of the facade material.

[0133] Facade reflectance spectrum data refers to the reflectance distribution of a building's surface under different wavelengths of light. Preset reflectance spectra are databases that establish a pre-established relationship between materials and reflectance at specific wavelengths. Building facade material refers to the type of building surface material, such as concrete, glass, or tile. Spectral fingerprints are the characteristic combinations of a material's reflectance at specific wavelengths, used to uniquely identify it.

[0134] Step 112: Determine the structural type based on the beam-column node morphological features in the UAV remote sensing image data, and identify the usage type in combination with the red-hot external data.

[0135] The random forest classifier is an ensemble learning algorithm that performs classification through voting among multiple decision trees. The beam-column joint morphological characteristics refer to the geometric shape characteristics of the connection between the beam and the column.

[0136] Step 113: Based on the three-dimensional point cloud data, the number of layers with sudden changes in point cloud density is counted along the elevation direction. Combined with the window opening distribution verification, the number of layers is determined, and the height of the building is inverted and calculated to obtain the height.

[0137] The elevation direction refers to the vertical direction perpendicular to the ground. Point cloud density mutation layers refer to the number of layers where the point cloud data density changes in the vertical direction, which is used to determine floor divisions. Window opening distribution verification verifies the rationality of floor divisions by analyzing the regularity of the horizontal spacing of windows. Height inversion calculation calculates the total building height based on the vertical distance between the highest point in the point cloud data and the ground reference.

[0138] By executing steps 111 to 113, the embodiment of the present application realizes non-contact and accurate identification of the building facade material, age, structural type and purpose through multi-spectral and morphological feature fusion analysis; combined with point cloud density mutation and window opening distribution verification, the calculation reliability of the number of floors and height inversion is improved, providing efficient and automated data collection and processing technical support for the digitization of urban building information.

[0139] In one possible embodiment, the drone remote sensing image data includes visible light image data and infrared light image data. Step 112: Determine the structural type based on the beam-column node morphological features in the drone remote sensing image data, and identify the usage type in combination with the infrared external data, including:

[0140] Step b1: Input the visible light image data into the improved ResNet-50 network and output the pixel-level beam-column node segmentation mask.

[0141] Among them, the pixel-level beam-column node segmentation mask refers to the binary image generated by the image segmentation algorithm, which accurately marks the pixel position of the beam-column node.

[0142] Step b2: Based on the geometric shape and size ratio of the segmentation mask, match it with a preset structure type template library to determine the structure type of the building.

[0143] The preset structural template library refers to a database containing typical beam-column node morphological parameters of different structural types. The structural type of a building refers to the classification of the building's load-bearing system, such as frame structure, shear wall structure, etc.

[0144] Step b3: Extract temperature distribution features from the infrared image data and, combined with the diurnal temperature difference pattern, construct a thermal signature map of the usage type. Fusion the texture features from the visible light image with the thermal signature map of the usage type, and output the usage type using a support vector machine classifier.

[0145] Temperature distribution features refer to the temperature matrix data for different areas of a building's surface, reflecting its thermal radiation characteristics. Diurnal temperature difference patterns refer to the pattern of changes in building surface temperature over a 24-hour period. Thermal signature maps by usage type refer to the characteristic patterns in thermal radiation data for buildings of different uses. Texture features in visible light images refer to the spatial distribution of pixel grayscale values within an image. A support vector machine classifier is a supervised learning algorithm that achieves data classification by constructing a hyperplane.

[0146] By executing steps b1 to b3, the embodiment of the present application achieves high-precision identification of structural types through pixel-level beam-column node segmentation and template library matching; the integration of temperature distribution, texture features and machine learning classification improves the accuracy and robustness of use type discrimination, providing technical support for multi-source data collaborative analysis for the digitization of urban building information.

[0147] In one possible embodiment, S12, based on the three-dimensional point cloud data, counting the number of layers with sudden changes in point cloud density along the elevation direction, combining with window opening distribution verification to determine the number of layers, and performing height inversion calculation on the building to obtain the height, includes:

[0148] Step 121: De-noise and perform ground filtering on the three-dimensional point cloud data to obtain a point cloud of the building body.

[0149] Ground filtering involves layering the raw point cloud using a cloth simulation filtering algorithm. This process simulates the physical properties of cloth sagging under gravity to isolate point clouds that conform to the ground's geometric features. This process involves two core steps: slope continuity detection and surface fitting. The building's main point cloud is a collection of 3D spatial coordinates that has undergone denoising and ground filtering. This includes key structural features such as the building's exterior walls and roof, excluding interfering elements like vegetation and ground attachments, and resulting in clear spatial contours.

[0150] In this embodiment, a spatial filtering algorithm is first used to denoise the 3D point cloud data, removing discrete noise points caused by scanning errors. Cloth simulation filtering is then used to separate the ground point cloud. By iteratively simulating the process of flexible cloth covering the point cloud surface, continuous areas below a set slope threshold are identified as ground point clouds. The remaining non-ground point clouds represent the point cloud data containing the main building structure.

[0151] Step 122: Based on the main point cloud of the building, vertical slices are divided along the elevation direction at intervals of preset units. The point cloud density in each vertical slice is counted, and the peak point of density mutation is detected by the Gaussian difference operator to determine the candidate floor interface.

[0152] Among them, the point cloud density within the vertical slice refers to the statistical value of the number of valid points contained in the unit volume of each slice after the point cloud is layered along the elevation direction, reflecting the spatial distribution characteristics of the building structure in the vertical direction. The Gaussian difference operator refers to the process of constructing Gaussian kernel functions of different scales, performing convolution operations on the density sequences of adjacent slices, and then obtaining the difference, which is used to enhance the density mutation characteristics at the junction of floors. The density mutation peak point refers to the local extreme point where the density change gradient exceeds the set threshold after Gaussian difference processing, corresponding to the structural boundary position between building floors. The candidate floor interface refers to the set of spatial planes determined by detecting the density mutation peak. Each plane contains elevation coordinates and confidence parameters, which are used for subsequent floor number determination.

[0153] In this embodiment, the building's main point cloud is first layered vertically at fixed height intervals, forming multiple vertical slices of equal thickness. Next, the number of point clouds per unit volume within each slice is counted as a density value. The density values of adjacent slices are then differentiated along the elevation direction, and a sliding window is used to detect sudden peaks where the density change exceeds a set threshold. Finally, these sudden peaks are identified as candidate floor interface locations, forming a set of candidate interfaces containing spatial coordinates.

[0154] Step 123: Determine the number of floors based on the number of candidate floor interfaces. Calculate the difference between the highest interface elevation and the base elevation based on the building base elevation. Combined with the scale verification in the drone image, output the building height value.

[0155] The highest interface elevation is the elevation corresponding to the plane with the largest spatial coordinates in the candidate interface set, representing the spatial location of the building's top floor. The base elevation difference is the vertical difference between the building's base plane and the highest interface. Calculated through three-dimensional coordinate transformation, it is the core parameter for building height calculation.

[0156] Alternatively, the number of building floors can be directly determined based on the number of candidate interfaces. Next, the base elevation obtained from drone aerial surveys is selected as the reference surface, and the absolute height difference between the highest interface and the base surface is calculated. The rationality of the elevation difference is then verified through 3D coordinate transformation, using reference objects of known dimensions and scale parameters in the imagery. Finally, combining the results of point cloud analysis and image verification, the building height measurement is output, optimized using spatial geometric constraints.

[0157] Alternatively, the raw point cloud acquired by the LiDAR is first subjected to outlier removal and noise smoothing, and a cloth simulation algorithm is used to separate the ground and non-ground point clouds to obtain point cloud data containing the main building structure. The main point cloud is then vertically layered at 0.5-meter intervals, and the number of point clouds per cubic meter per layer is counted to form a density sequence. A third-order Gaussian kernel difference operation is used to detect areas of density mutations and identify candidate floor boundary surfaces. The number of candidate surfaces is then counted to determine the number of building floors. A foundation survey point is selected as the base elevation benchmark, and the relative height difference of the highest boundary surface is calculated. Scale verification is performed using window components of known dimensions in the aerial imagery, and the final output is a spatially geometrically corrected building height measurement.

[0158] By executing steps 121 through 123, this embodiment of the present application achieves automated and accurate building height measurement through a multi-stage data processing process, effectively integrating point cloud analysis technology with image verification methods. Physically-based ground filtering ensures the integrity of the main structure, while vertical layered density analysis accurately captures floor distribution characteristics. Combined with a multi-source data verification mechanism, the reliability of measurement results is enhanced, providing a highly efficient technical solution for building surveying and mapping.

[0159] Figure 3 A schematic diagram of a system for constructing earthquake damage scenarios for buildings based on drone remote sensing images is provided in an embodiment of the present application. Figure 3 As shown, the system includes:

[0160] The acquisition and extraction module 31 is used to use a drone to take multi-angle aerial photos of the target area, obtain drone remote sensing image data, and extract geometric attribute information of the building in combination with an image recognition algorithm.

[0161] The construction and determination module 32 is used to construct a three-dimensional space model of the building according to the geometric attribute information of the building using the urban earthquake damage simulation system, and determine the nonlinear model parameters of the building based on the three-dimensional space model.

[0162] The simulation module 33 is used to construct earthquake damage scenarios based on nonlinear model parameters and preset earthquake motion parameters using an elastic-plastic time-history analysis method with a single building as the granularity, obtain the spatial distribution of damage to each building, and determine the building damage index.

[0163] The determination module 34 is configured to determine the earthquake damage risk level of each building based on a preset mapping relationship between the building damage index and the building earthquake damage risk level.

[0164] The detection generation module 35 is used to detect the seismic performance of buildings with a specified earthquake damage risk level based on a list of buildings with a specified earthquake damage risk level, and generate seismic measures corresponding to buildings with a specified earthquake damage risk level based on the seismic performance test results. The specified earthquake damage risk level is a level greater than a preset earthquake damage risk level.

[0165] Figure 3 The system for constructing earthquake damage scenarios based on UAV remote sensing images can be used to Figure 1 The implementation principles and technical effects of the method for constructing earthquake damage scenarios based on drone remote sensing imagery described in the illustrated embodiment are not further elaborated. The specific manner in which each module and unit performs operations in the aforementioned embodiment of the system for constructing earthquake damage scenarios based on drone remote sensing imagery has been described in detail in the related embodiments and will not be further elaborated here.

[0166] In one possible design, Figure 3 The system for constructing building earthquake damage scenarios based on drone remote sensing images in the embodiment shown can be implemented as a computing device, such as Figure 4 As shown, the computing device may include a storage component 41 and a processing component 42 .

[0167] The storage component 41 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 42 .

[0168] The processing component 42 is used to: use a drone to take multi-angle aerial photos of the target area, obtain drone remote sensing image data, and extract geometric attribute information of the building in combination with an image recognition algorithm; based on the geometric attribute information of the building, use the urban earthquake damage simulation system to construct a three-dimensional spatial model of the building, and determine the nonlinear model parameters of the building based on the three-dimensional spatial model; based on the nonlinear model parameters and in combination with preset seismic motion parameters, use an elastic-plastic time-history analysis method to construct an earthquake damage scenario with a single building as the granularity, obtain the spatial distribution of damage and destruction of each building, and determine the building damage index; determine the earthquake damage risk level of each building based on the mapping relationship between the preset building damage index and the building earthquake damage risk level; based on the list of buildings with a specified earthquake damage risk level, test the seismic performance of buildings with a specified earthquake damage risk level, and generate earthquake resistance measures corresponding to buildings with a specified earthquake damage risk level based on the seismic performance test results, where the specified earthquake damage risk level is a level greater than the preset earthquake damage risk level.

[0169] The processing component 42 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0170] The storage component 41 is configured to store various types of data to support operations on the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as random access memory (RAM), static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0171] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0172] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0173] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0174] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0175] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment provides a method for constructing building earthquake damage scenarios based on drone remote sensing images.

[0176] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0177] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0178] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0179] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for constructing building earthquake damage scenarios based on UAV remote sensing images, characterized in that: include: Use drones to take multi-angle aerial photos of the target area, obtain drone remote sensing image data, and use image recognition algorithms to extract the geometric attribute information of the building; According to the geometric attribute information of the building, a three-dimensional spatial model of the building is constructed using an urban earthquake damage simulation system, and nonlinear model parameters of the building are determined based on the three-dimensional spatial model; Based on the nonlinear model parameters and the preset earthquake motion parameters, an elastic-plastic time history analysis method is used to construct earthquake damage scenarios with a single building as the granularity, to obtain the spatial distribution of damage to each building and determine the building damage index; Determine the earthquake damage risk level of each building based on the mapping relationship between the preset building damage index and the building earthquake damage risk level; testing the seismic performance of buildings at the specified seismic risk level based on a list of buildings at the specified seismic risk level, and generating seismic measures corresponding to the buildings at the specified seismic risk level based on the seismic performance test results, wherein the specified seismic risk level is a level greater than a preset seismic risk level; The nonlinear model parameters include: floor mass, floor stiffness, and a relationship curve between floor shear force and displacement. The elastic-plastic time history analysis method includes the Newmark-β method. Based on the nonlinear model parameters and in combination with preset earthquake motion parameters, an elastic-plastic time history analysis method is used to construct earthquake damage scenarios with a single building as the granularity, thereby obtaining the spatial distribution of damage to each building and determining the building damage index, including: Normalizing the floor mass and the floor stiffness respectively, and discretizing the relationship curve between the floor shear force and displacement to obtain a standardized nonlinear parameter input file; Based on the preset seismic motion parameters, the Newmark-β method is used to construct earthquake damage scenarios at the granularity of a single building, and the time-history response dataset of each building is solved. Inputting the time-history response dataset into a pre-trained hybrid damage assessment model to output the damage spatial distribution of the corresponding building, wherein the damage spatial distribution includes the damage distribution of concrete components and steel structures on each floor; The damage distribution of concrete components and steel structures on all floors is integrated to obtain the building damage index; The time-history response dataset is input into a pre-trained hybrid damage assessment model to output the damage spatial distribution of the corresponding building, wherein the damage spatial distribution includes the damage distribution of concrete components and steel structures on each floor, including: For concrete components, the inter-story displacement angle, cumulative energy dissipation, and steel strain time history in the time history response dataset are extracted and input into the concrete damage assessment module in the hybrid damage assessment model. The concrete crack width, percentage of cover loss area, and damage depth of the compression zone of each floor are output; The spatial distribution of concrete components is combined with the concrete crack width, the proportion of the cover shedding area, and the damage depth of the compression zone on each floor to obtain the concrete component damage distribution on each floor. For steel structural components, the interlayer shear force time history, node plastic rotation angle, and local buckling strain are extracted and input into the steel structure damage assessment model in the hybrid damage assessment model. The fracture probability of the steel beam-column connection node, the web buckling index, and the overall instability risk level of the steel structure are output; The spatial distribution of steel structure components is combined with the fracture probability of steel beam-column connection nodes, the web buckling index, and the overall instability risk level of the steel structure to obtain the steel structure damage distribution on each floor. The damage distribution of concrete components and steel structures on all floors is integrated to obtain the building damage index, including: Determine the corresponding concrete weight factor and steel structure weight factor according to the structural type of the building; performing spatial integration of the first damage value of each floor in the concrete component damage distribution and the second damage value in the steel structure damage distribution according to the concrete weight factor and the steel structure weight factor to obtain a total damage value of each floor; The aging coefficient is introduced according to the construction year to correct the total damage value of each floor, and the corrected total damage value of each floor is obtained; The entropy weight method is used to fuse the corrected total damage values of each floor to obtain the building damage index.

2. The method according to claim 1, characterized in that The geometric attribute information includes construction year, number of floors, height, structure type and use type; The method of extracting geometric attribute information of a building by combining an image recognition algorithm includes: Extracting the building's facade reflectance spectrum data from drone remote sensing image data, identifying the building's facade material based on a preset library of matching between reflectance spectra and materials, and using a random forest classifier to identify the building's construction year based on the facade material's spectral fingerprint; Determine the structural type based on the beam-column node morphological features from the UAV remote sensing image data, and identify the usage type in combination with the red-hot external data; Based on the three-dimensional point cloud data, the number of point cloud density mutation layers is counted along the elevation direction. Combined with the window opening distribution verification, the number of floors is determined, and the height of the building is inverted and calculated to obtain the height.

3. The method according to claim 2, characterized in that The UAV remote sensing image data includes visible light image data and infrared light image data; The method of determining the structural type based on the beam-column node morphological features from the UAV remote sensing image data and identifying the use type in combination with the red-hot external data includes: Input the visible light image data into the improved ResNet-50 network and output pixel-level beam-column node segmentation mask; Based on the geometric shape and size ratio of the segmentation mask, the preset structure type template library is matched to determine the structure type of the building; Temperature distribution features are extracted from the infrared image data, and combined with the day-night temperature difference pattern to construct a use type thermal feature map; texture features in the visible light image are integrated with the use type thermal feature map, and the use type is output through a support vector machine classifier.

4. The method according to claim 2, characterized in that Based on the three-dimensional point cloud data, the number of point cloud density mutation layers is counted along the elevation direction, and the number of layers is determined by combining the window opening distribution verification, and the height of the building is inverted and calculated to obtain the height, including: De-noising and ground filtering are performed on the three-dimensional point cloud data to obtain the main point cloud of the building; Based on the main point cloud of the building, vertical slices are divided at intervals of preset units along the elevation direction, the point cloud density in each vertical slice is counted, and the peak point of density mutation is detected by Gaussian difference operator to determine the candidate floor interface; The number of floors is determined based on the number of candidate floor interfaces. The difference between the highest interface elevation and the base elevation is calculated based on the building base elevation. Combined with the scale verification in the drone image, the building height value is output.

5. A building earthquake damage scenario construction system based on drone remote sensing images, characterized by: include: The acquisition and extraction module is used to use drones to take multi-angle aerial photos of the target area, obtain drone remote sensing image data, and extract the geometric attribute information of the building in combination with image recognition algorithms; A construction and determination module is used to construct a three-dimensional spatial model of the building using the urban earthquake damage simulation system according to the geometric attribute information of the building, and determine the nonlinear model parameters of the building based on the three-dimensional spatial model; A simulation module is used to construct earthquake damage scenarios based on the nonlinear model parameters and preset earthquake motion parameters using an elastic-plastic time history analysis method with a single building as the granularity, obtain the spatial distribution of damage to each building, and determine the building damage index; A determination module is used to determine the earthquake damage risk level of each building based on a preset mapping relationship between the building damage index and the building earthquake damage risk level; a detection and generation module for testing the seismic performance of buildings at a specified seismic damage risk level based on a list of buildings at a specified seismic damage risk level, and generating seismic measures corresponding to the buildings at the specified seismic damage risk level based on the seismic performance test results, wherein the specified seismic damage risk level is a level greater than a preset seismic damage risk level; The nonlinear model parameters include: floor mass, floor stiffness, and a relationship curve between floor shear force and displacement. The elastic-plastic time history analysis method includes the Newmark-β method. Based on the nonlinear model parameters and in combination with preset earthquake motion parameters, an elastic-plastic time history analysis method is used to construct earthquake damage scenarios with a single building as the granularity, thereby obtaining the spatial distribution of damage to each building and determining the building damage index, including: Normalizing the floor mass and the floor stiffness respectively, and discretizing the relationship curve between the floor shear force and displacement to obtain a standardized nonlinear parameter input file; Based on the preset seismic motion parameters, the Newmark-β method is used to construct earthquake damage scenarios at the granularity of a single building, and the time-history response dataset of each building is solved. Inputting the time-history response dataset into a pre-trained hybrid damage assessment model to output the damage spatial distribution of the corresponding building, wherein the damage spatial distribution includes the damage distribution of concrete components and steel structures on each floor; The damage distribution of concrete components and steel structures on all floors is integrated to obtain the building damage index; The time-history response dataset is input into a pre-trained hybrid damage assessment model to output the damage spatial distribution of the corresponding building, wherein the damage spatial distribution includes the damage distribution of concrete components and steel structures on each floor, including: For concrete components, the inter-story displacement angle, cumulative energy dissipation, and steel strain time history in the time history response dataset are extracted and input into the concrete damage assessment module in the hybrid damage assessment model. The concrete crack width, percentage of cover loss area, and damage depth of the compression zone of each floor are output; The spatial distribution of concrete components is combined with the concrete crack width, the proportion of the cover shedding area, and the damage depth of the compression zone on each floor to obtain the concrete component damage distribution on each floor. For steel structural components, the interlayer shear force time history, node plastic rotation angle, and local buckling strain are extracted and input into the steel structure damage assessment model in the hybrid damage assessment model. The fracture probability of the steel beam-column connection node, the web buckling index, and the overall instability risk level of the steel structure are output; The spatial distribution of steel structure components is combined with the fracture probability of steel beam-column connection nodes, the web buckling index, and the overall instability risk level of the steel structure to obtain the steel structure damage distribution on each floor. The damage distribution of concrete components and steel structures on all floors is integrated to obtain the building damage index, including: Determine the corresponding concrete weight factor and steel structure weight factor according to the structural type of the building; performing spatial integration of the first damage value of each floor in the concrete component damage distribution and the second damage value in the steel structure damage distribution according to the concrete weight factor and the steel structure weight factor to obtain a total damage value of each floor; The aging coefficient is introduced according to the construction year to correct the total damage value of each floor, and the corrected total damage value of each floor is obtained; The entropy weight method is used to fuse the corrected total damage values of each floor to obtain the building damage index.

6. A computing device, characterized in that It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for constructing building earthquake damage scenarios based on drone remote sensing images as described in any one of claims 1 to 4.

7. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, the method for constructing a building earthquake damage scenario based on drone remote sensing images as described in any one of claims 1 to 4 is implemented.

Images