A refined method for estimating the population buried by earthquakes and the medium

Abstract

This invention relates to a refined method and medium for estimating the population buried in earthquakes, belonging to the field of population data extrapolation technology. First, based on remote sensing data, the functional types of buildings in the target area are identified, thereby obtaining the population weight of each functional type of building. This allows rescue forces to accurately locate each high-risk building after an earthquake, rather than a vague area. Second, for each building, based on its functional type, building area, and population weight, the building-scale population is calculated, and the collapse rate and occupancy rate at the time of the earthquake are obtained, thus dynamically estimating the number of people buried. In this way, a more refined spatial distribution of the buried population can be obtained by integrating various complex nonlinear potential influencing factors, providing a more scientific and effective decision-making basis for earthquake emergency response.

Description

Technical Field

[0001] This invention relates to the field of population data estimation technology, specifically to a refined method and medium for estimating the population buried by earthquakes. Background Technology

[0002] In the field of earthquake disaster assessment, early assessments typically used population data based on administrative regions, obtained through censuses and sampling statistics, which were often disconnected from the spatial characteristics of the respective regions and could not accurately reflect the spatial distribution of the population. The main solution to these problems is the spatialization of statistical population data. This refers to using statistical data and qualitative or quantitative analytical models to reconstruct the distribution of population in a given time and geographic space, achieving the conversion and estimation of spatial units at different scales. With the rapid development of Earth observation, modern information networks and communications, and especially geographic big data analysis and mining, including remote sensing and social sensing, geographic big data provides rich and diverse data resources for the rapid and accurate acquisition of large-scale disaster-bearing body exposure and vulnerability parameters. "Spatialization of economic and social data" has gradually become a research hotspot in academia, with population data spatialization being a particularly important research direction. Allocating statistical population data to fine-grained spatial units is of great significance for refined earthquake disaster assessment.

[0003] In recent years, various methods have been developed to spatially decompose census data into grid cells, such as area-weighted, geographic-weighted regression, and zonal density mapping, resulting in many gridded population datasets covering large geographic areas. However, most of these datasets have relatively low spatial resolution, making it difficult to meet the needs of fine-scale research, especially in the area of ​​refined earthquake disaster risk assessment.

[0004] Population spatial distribution is influenced by a combination of factors, including environment, transportation, and resources. Combining this with relevant auxiliary data helps achieve accurate grid-based population allocation. Remote sensing satellite imagery, due to its high spatial resolution and short acquisition cycle, is widely used in population spatialization research. Currently, population spatialization modeling data is becoming increasingly diversified, refined, and dynamic. Some studies improve the accuracy of population estimation by integrating various data, such as land use and nighttime light data. However, these methods also have some limitations. For example, land use interpretation data describes the spatial extent of population distribution but struggles to reveal differences in population density within the same land type. Nighttime light remote sensing data can, to some extent, distinguish the heterogeneity of population distribution and reflect population density; however, light spillover effects caused by light sources unrelated to population distribution, such as streetlights, construction sites, and reflected light, can lead to misallocation of the population. It is evident that population spatialization methods based on remote sensing and GIS primarily employ multi-source data fusion. However, the main auxiliary data sources used in these studies, such as land use and nighttime light data, have low spatial resolution and exhibit homogeneity issues at smaller scale units, limiting their ability to depict detailed differences in population distribution. Therefore, it is necessary to introduce auxiliary data sources with finer spatial granularity for population spatialization. Fine-grained social perception data, especially points of interest (POIs), possess precise locations and rich spatial semantics, and are widely available and easily accessible. They can reflect potential human activities in or around these POIs, making them a core data source for refined population modeling.

[0005] It is worth noting that there is a complex nonlinear relationship between population distribution and various potential influencing factors, so there is still a great challenge in simulating the spatial distribution of population. Summary of the Invention

[0006] The technical problem to be solved by this application is to provide a refined method and medium for estimating the population buried in earthquakes. This method can integrate various complex nonlinear potential influencing factors to obtain a more refined spatial distribution of the population, thus providing a more scientific and effective basis for decision-making in earthquake emergency response.

[0007] In a first aspect, one embodiment provides a refined method for estimating the population buried in earthquakes, including:

[0008] Identify the functional types of buildings in the target area based on remote sensing data;

[0009] Obtain the population weights for buildings of various functional types;

[0010] For each building, calculate the building-scale population based on its function type, building area, and population weight;

[0011] For each building, obtain its collapse rate and occupancy rate at the time of the earthquake;

[0012] For any given building, the number of people buried is estimated based on its building-scale population, the collapse rate at the time of the earthquake, and the number of people inside the building.

[0013] In a second aspect, one embodiment provides a computer-readable storage medium storing a program that can be loaded and executed by a processor, the refined earthquake-buried population estimation method.

[0014] The beneficial effects of this invention are:

[0015] First, the functional types of buildings in the target area are identified based on remote sensing data, thereby obtaining the population weight of each functional type of building. This allows rescue forces to accurately locate each high-risk building after an earthquake, rather than a vague area, laying the core spatial foundation for the precise deployment of rescue resources. For each building, based on its functional type, building area, and population weight, the building-scale population is calculated, and the collapse rate and occupancy rate at the time of the earthquake are obtained. This allows for a dynamic estimation of the number of people buried. By integrating various complex nonlinear potential influencing factors, a more refined spatial distribution of the population can be obtained, providing a more scientific and effective decision-making basis for earthquake emergency response. Attached Figure Description

[0016] Figure 1 This is a schematic flowchart of a refined earthquake-buried population estimation method according to an embodiment of this application;

[0017] Figure 2 This is a schematic flowchart of a method for identifying the functional type of a building with an unknown functional type based on kernel density and NED value according to an embodiment of this application;

[0018] Figure 3 This application Figure 1 A schematic diagram of the method flow for one embodiment of step S30. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of this application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to this application are not shown or described in the specification. This is to avoid obscuring the core parts of this application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0020] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0021] The serial numbers assigned to components in this article, such as "first" and "second", are used only to distinguish the objects being described and have no sequential or technical meaning.

[0022] In view of the shortcomings of existing technologies, this application provides a refined method and medium for estimating the population buried in earthquakes. First, based on remote sensing data, the functional types of buildings in the target area are identified, thereby obtaining the population weights for each functional type of building. This allows rescue forces to accurately locate each high-risk building after an earthquake, rather than a vague area, laying the core spatial foundation for the precise deployment of rescue resources. For each building, based on its functional type, building area, and population weight, the building-scale population is calculated, and the collapse rate and occupancy rate at the time of the earthquake are obtained. This allows for a dynamic estimation of the number of people buried. By integrating various complex nonlinear potential influencing factors, a more refined spatial distribution of the population can be obtained, providing a more scientific and effective decision-making basis for earthquake emergency response.

[0023] This application provides a refined method for estimating the population buried by earthquakes. Please refer to [link / reference]. Figure 1 It can include:

[0024] Step S10: Identify the functional types of buildings in the target area based on remote sensing data.

[0025] The target area can be a designated administrative region. In the embodiments of this application, the target area will be described using an administrative village as a unit. Those skilled in the art will understand that the technical teachings of this application can also be applied to other types of regional divisions.

[0026] Building function types are categorized based on the building's core purpose, design objectives, and the activities it undertakes, focusing on the building's "use attributes." These can include residential, accommodation services, commercial services, administrative offices, and healthcare facilities. Points of Interest (POI) types are categorized based on the core services or attributes offered by the point of interest, offering a more granular classification, such as restaurants, supermarkets, schools, cinemas, and parking lots. Therefore, a building may have only one function type, but it can include multiple POI types and POI points.

[0027] In current methods for identifying the functional type of buildings, the functional type of a building and the POI type points it contains can be identified based on existing data. For example, the location of a specific building can be obtained based on remote sensing data, and the functional type of the current building and the POI type points it contains can be obtained by combining land use planning and existing building types and POI type identifiers.

[0028] If the functional type of any building cannot be identified based on existing data, it is possible to check whether there is a known POI type within a preset first distance threshold range (e.g., 10m). If it exists, the functional type of any building is determined based on the known POI type. If it does not exist, the shape outline of any building and the shape outline of known building types within a preset second distance threshold range (e.g., 100m) are obtained, and the similarity of the shape outlines is calculated. The functional type of the building with the highest similarity value that reaches the preset similarity threshold is taken as the functional type of any building.

[0029] When determining the functional type of a building based on a known POI type, if multiple known functional types exist within the first distance threshold range, the determination can be made based on the POI type priority principle and the nearest principle. If there is a preset priority class of POI type points, the functional type of the building with the highest priority POI type is determined as the functional type of the unknown functional type building based on the importance priority. If there is no priority class of POI type points, the functional type of the building with the nearest POI type point is determined as the functional type of the unknown functional type building based on the nearest principle.

[0030] In one feasible embodiment of a method for determining function type based on shape and contour similarity, spatial feature vectors of identified and unidentified buildings can be obtained. The cosine value of the two feature vectors is used to determine the similarity between the two buildings. Specifically, the larger the absolute value of the cosine value, the higher the similarity.

[0031] However, the applicant found in the study that although associating POI type with overlapping buildings is a very effective way to determine the function of buildings, there is not always POI data inside or near buildings, and there are still many buildings that cannot be matched with neighboring buildings through spatial similarity.

[0032] In view of this, this application provides a novel method for identifying building function types. This method identifies the function type of buildings with unknown function types based on kernel density and NED (Normalized Euclidean Distance) values. In one embodiment of this application, if the function type of a building still cannot be identified based on the above method, the function type of the building with unknown function types is identified based on kernel density and NED values.

[0033] Please refer to Figure 2 In one embodiment, the method for identifying the functional type of a building with an unknown functional type based on kernel density and NED value may include:

[0034] Step S100: Obtain the grid covered by each building. For any grid in the covered grid, take the center point of the grid as the center, find the POI type points contained within the preset third distance threshold range, and calculate the kernel density of each POI type. Add the kernel densities of the POI types corresponding to all grids in the covered grid to obtain the comprehensive kernel density of each POI type.

[0035] For example, if the preset third distance threshold is 500m, then POI type points within a 500m radius centered on the grid's center point can be found. If there are multiple types of POI points, then the kernel density for each POI type needs to be calculated.

[0036] In one embodiment, the kernel density of any POI type can be calculated as follows:

[0037]

[0038] in, Here, represents the kernel density, h represents the preset third distance threshold, d represents the spatial dimension, i represents the index of a point of a certain POI type, and I represents the total number of points of a certain POI type. This represents the kernel function, where x represents the coordinates of the center point of the grid. Let i represent the coordinates of the i-th point of a certain POI type, where 1 ≤ i ≤ I.

[0039] If the spatial dimension is two-dimensional, then d=2; if the spatial dimension is three-dimensional, then d=3. For a given POI type, if its type points are 3, then I=3. Kernel function You can choose based on your needs, which will not be elaborated here.

[0040] For any building, if it covers 6 grids (e.g., each grid is 50m in length and width), the kernel density of the POI type for each of the 6 grids can be calculated. If 4 of the 6 grids have a kernel density of POI type 1, then the kernel densities of the 4 POI type 1 grids are added together to obtain the comprehensive kernel density of POI type 1 for the building. And so on, the comprehensive kernel density of all covered POI types can be obtained.

[0041] Current solutions, while spatially dividing the data into grid cells, lack joint analysis of the grid cell data, thus failing to meet the requirements for refined assessment. Therefore, this application's embodiments combine NED values ​​with joint analysis of grid cells to improve the refinement of population data assessment, thereby enabling more accurate assessment of population data within individual buildings.

[0042] Step S200: For each building with a known functional type, calculate the average and standard deviation of the comprehensive kernel density of all corresponding POI types.

[0043] Step S300: For any building with an unknown functional type, calculate the NED value of the building with an unknown functional type and that of each building with a known functional type based on the average and standard deviation of the comprehensive kernel density of each POI type and the comprehensive kernel density of each building with a known functional type.

[0044] In one embodiment, step S300 can be represented as:

[0045]

[0046] in, This represents the NED value between any building of unknown functional type and the k-th building of known functional type. This represents the overall density of the j-th POI type for any building with an unknown functional type, where J represents the number of POI types for any building with an unknown functional type, 1≤j≤J. and Let represent the average and standard deviation of the overall density of the k-th building with known functional type, respectively.

[0047] The NED value is a similarity metric; the smaller the value, the higher the similarity between the two.

[0048] Step S400: The functional type of the known functional type building corresponding to the minimum NED value is taken as the functional type of any unknown functional type building.

[0049] Through steps S100 to S400, the kernel density calculation method converts the points of interest into a continuous density surface, making the points of interest exhibit distribution characteristics. Then, combined with the NED value, the unknown building type is determined based on the distribution characteristics of the POI type, thereby identifying the functional type of buildings that cannot be matched with neighboring buildings through spatial similarity.

[0050] Step S20: Obtain the population weight of each functional type of building.

[0051] To quantify the impact of different functional building types on the total population and thus determine the population weight of each functional building type, this application embodiment introduces a Random Forest machine learning algorithm to construct an impact weight model, thereby obtaining the population weight of each functional building type.

[0052] In one embodiment of this application, the method for calculating the population weight of buildings of various functional types may include:

[0053] Step S1000: Construct a sample dataset, which includes multiple target area sample unit data after data preprocessing. Each target area sample unit data includes the total population and the building area of ​​each functional type of building. In each target area sample unit data, the building area of ​​each functional type of building is used as the independent variable, and the total population is used as the dependent variable.

[0054] In one embodiment, a sample dataset is constructed using all administrative villages within the study area as sample units. The total population of each administrative village is used as the dependent variable, directly reflecting the population size of the village and serving as a core indicator for measuring population distribution. The building area of ​​various functional types of buildings within the administrative villages, including residential, commercial, factory / warehouse, and medical / health facilities, is used as the independent variable. All data for the independent variables are derived from the statistical results at the administrative village scale, ensuring data consistency and accuracy.

[0055] The variable preprocessing stage may include: standardizing all independent and dependent variables (using the Z-score standardization method) to eliminate the interference of differences in the scales of different variables on model training.

[0056] Step S2000: Construct a random forest regression model. Input the sample dataset into the constructed random forest regression model for training and parameter optimization, and extract the feature importance of each independent variable.

[0057] In one embodiment, the random forest regression model is constructed using Python's Scikit-learn library. The core parameters of the model include the number of decision trees (n_estimators), the maximum depth of the decision trees (max_depth), and the minimum number of samples required for node splits (min_samples_split). To improve model performance, parameter optimization can be performed using grid search combined with 5-fold cross-validation: the search range for n_estimators is set to 100-500, the search range for max_depth is 5-20, and the search range for min_samples_split is 2-10. The optimal parameter combination is determined by minimizing the root mean square error (RMSE) of cross-validation.

[0058] During model training, the sample dataset is randomly divided into a training set (70%) and a test set (30%) in a 7:3 ratio. The training set is used for model parameter learning, and the test set is used for validating the model's generalization ability. The determination coefficient (R²) of the test set is calculated. 2 The model's fit and prediction accuracy are evaluated using metrics such as mean absolute error (MAE) and RMSE. 2 The results showed that the model was >0.8 and the RMSE and MAE were at low levels, indicating that the model could capture the nonlinear relationship between the building area of ​​each functional building and the total population, and could be used to extract the subsequent influence weights.

[0059] After the random forest model is trained, the influence weights of different functional building types on the total population are determined by extracting the feature importance of each independent variable in the model. The calculation of feature importance is based on the information gain at the time of decision tree node splitting; the higher the feature importance value of the building area of ​​a certain functional building, the greater the influence of that functional building on the total population.

[0060] Step S3000: Normalize the feature importance to obtain the population weight of buildings of each functional type.

[0061] In one embodiment, step S3000 can be represented as:

[0062]

[0063] in, This represents the population weight of the nth functional type building. This represents the feature importance value of the nth functional type building, where N represents the total number of functional types, and 1 ≤ n ≤ N.

[0064] The normalized weights satisfy Furthermore, the larger the weight value, the more significant the role of this type of building in population carrying capacity.

[0065] Step S30: For each building, calculate the building-scale population based on its function type, building area, and population weight.

[0066] Please refer to Figure 3 Step S30 may include:

[0067] Step S301: Calculate the population density of each functional type of building based on the population weight of each functional type of building and the total population of the target area.

[0068] In one embodiment, step S301 can be represented as:

[0069]

[0070] in, This represents the population density per unit area of ​​buildings of the nth functional type. This represents the total population of the target area (e.g., an administrative village). Let n be the area of ​​the building of the nth functional type.

[0071] This calculation method takes into account the differentiated impact of different functional buildings on population carrying capacity, while ensuring that the sum of the population of all functional buildings is consistent with the overall population of the administrative village.

[0072] Step S302: Based on the population density of buildings of various functional types and the building area of ​​each building, calculate the building-scale population of each building.

[0073] In one embodiment, step S302 can be represented as:

[0074]

[0075] in, This represents the building size and population of the m-th building. This represents the building area of ​​the m-th building.

[0076] In one embodiment, when calculating the building-scale population, the method further includes: summarizing the building-scale population of all buildings to obtain the total building-scale population of the target area, and calculating the relative error with the actual total population of the target area. If the relative error is greater than or equal to a preset error threshold, the functional type identification and population weight acquisition process of the buildings are traced back, and the building-scale population and the total building-scale population of each building are recalculated after correction until the relative error is less than the preset error threshold.

[0077] In some embodiments, the error threshold can be set to 5%, and the relative error can be expressed as:

[0078]

[0079] in, This represents the relative error, where M represents the total number of buildings within the target area (e.g., an administrative village). This represents the total actual population within the target area.

[0080] Step S40: For each building, obtain its collapse rate and occupancy rate at the time of the earthquake.

[0081] The occupancy rate is related to various factors such as the time of earthquake occurrence, weekdays versus weekends, and building function. In this embodiment, by analyzing the behavioral characteristics of people at different time periods, the patterns of social activities in various industries and the occupancy rate of functional buildings are summarized. Determining the proportion of residents inside buildings at different times of the day is crucial for assessing the actual distribution of the population within buildings over time. Furthermore, the occupancy rate varies among buildings with different functions within the same time period. The number of people inside a building during an earthquake is influenced by factors such as holidays, the time of the earthquake, building function, indoor population structure, and occupation.

[0082] In summary, in the embodiments of this application, the method for obtaining the occupancy rate of each building includes: for each type of building, dividing each day into weekdays and non-weekdays, and obtaining the occupancy rate of each time period on weekdays and each time period on non-weekdays; each time period on weekdays and each time period on non-weekdays includes sleep time, morning peak commuting time, morning work time, lunch break time, afternoon work time, evening peak commuting time, and evening rest time.

[0083] In one embodiment, the sleep period is from 22:00 to 7:00, the morning rush hour is from 7:00 to 9:00, the morning work period is from 9:00 to 12:00, the lunch break is from 12:00 to 14:00, the afternoon work period is from 14:00 to 18:00, the evening rush hour is from 18:00 to 20:00, and the evening rest period is from 20:00 to 22:00.

[0084] In one embodiment, following the above time period sequence, the occupancy rates of different functional types of buildings in the city at different times can be referenced as follows: For residential buildings, the occupancy rates for each time period on weekdays are 0.90, 0.40, 0.20, 0.30, 0.20, 0.40, and 0.60, respectively; and for each time period on non-working days, the occupancy rates are 0.95, 0.70, 0.60, 0.60, 0.60, 0.70, and 0.80, respectively. For accommodation service buildings, the occupancy rates for each time period on weekdays and non-working days are 1.00, 0.40, 0.10, 0.30, 0.10, 0.30, and 0.40, respectively. The attendance rates for administrative office buildings during different time periods on weekdays were 0.05, 0.30, 0.85, 0.70, 0.85, 0.40, and 0.20, respectively, while the attendance rates for different time periods on non-working days were 0.02, 0.03, 0.10, 0.10, 0.10, 0.03, and 0.02, respectively. The attendance rates for commercial service buildings during different time periods on weekdays were 0.05, 0.20, 0.55, 0.65, 0.55, 0.75, and 0.30, respectively, while the attendance rates for different time periods on non-working days were 0.05, 0.20, 0.85, 0.90, 0.85, 0.85, and 0.40, respectively. The attendance rates for higher education buildings during different time periods on weekdays were 0.95, 0.70, 1.00, 0.85, 1.00, 0.95, and 0.95, respectively, and the attendance rates for different time periods on non-weekdays were 0.95, 0.70, 0.45, 0.55, 0.45, 0.75, and 0.95, respectively. The attendance rates for other school buildings during different time periods on weekdays were 0.01, 0.50, 1.00, 0.70, 1.00, 0.40, and 0.05, respectively, and the attendance rates for different time periods on non-weekdays were 0.01, 0.02, 0.05, 0.05, 0.05, 0.03, and 0.02, respectively. The attendance rates for medical and health buildings during different time periods on weekdays were 0.15, 0.40, 1.00, 0.70, 1.00, 0.40, and 0.15, respectively, and the attendance rates for different time periods on non-working days were 0.15, 0.20, 0.50, 0.35, 0.50, 0.20, and 0.15, respectively. The attendance rates for factory and warehouse buildings during different time periods on weekdays were 0.05, 0.30, 0.95, 0.70, 0.95, 0.30, and 0.05, respectively, and the attendance rates for different time periods on non-working days were 0.05, 0.05, 0.05, 0.05, 0.05, 0.05, and 0.05, respectively.The attendance rates of visitors to sports and recreational buildings during different time periods on weekdays were 0.02, 0.15, 0.45, 0.55, 0.45, 0.15, and 0.03, respectively, while the attendance rates during different time periods on non-weekdays were 0.02, 0.20, 0.85, 0.90, 0.85, 0.25, and 0.02, respectively.

[0085] The occupancy rates of different functional types of buildings in rural areas at different times can be referenced as follows: For residential buildings, the occupancy rates for each time period on weekdays and non-weekdays are 0.95, 0.60, 0.30, 0.45, 0.30, 0.70, and 0.85, respectively. For non-residential buildings, the occupancy rates for each time period on weekdays and non-weekdays are 0.05, 0.30, 0.95, 0.70, 0.95, 0.30, and 0.05, respectively. For other school buildings, the occupancy rates for each time period on weekdays are 0.01, 0.50, 1.00, 0.70, 1.00, 0.40, and 0.05, respectively, and for each time period on non-weekdays are 0.01, 0.02, 0.05, 0.05, 0.05, 0.03, and 0.02, respectively.

[0086] Based on the analysis of the occupancy rate of people in different time periods, the dynamic temporal fluctuations of the occupancy rate can be reflected.

[0087] The following describes the collapse rate of each building during an earthquake.

[0088] Due to regional differences across the globe, assessing earthquake casualties requires locally oriented seismic vulnerability curves or building damage probability matrices. Generally, building damage can be categorized into five levels: good, minor, moderate, severe, and critical. Due to limitations in historical earthquake data, accurately determining the collapse rate for different building damage levels is extremely difficult. Preliminary investigations have found that being trapped in earthquakes is primarily due to severely damaged buildings. Severely damaged buildings mean that most load-bearing components are severely damaged, the building structure is nearing or has already collapsed, and is beyond repair. Therefore, this study considers the severe damage rate of different building structures under a specific earthquake intensity as their collapse rate.

[0089] Under the same earthquake intensity, the damage rate of buildings with different structural types varies. To accurately reflect the damage characteristics of buildings in the XX region, this study used a vulnerability matrix of buildings to determine the collapse rate of each structural type during earthquakes. Through a sampling survey of various buildings in the region, the seismic resistance index and damage matrix of individual buildings were obtained. Then, through statistical analysis of the damage proportions of buildings in three different areas of the region (Area 1, Area 2, and Area 3 in Table 1), a vulnerability matrix of buildings under different earthquake intensities was presented. Table 1 illustrates the collapse rate based on the data source of the buildings, and the building structural types are divided into four categories: frame structure, brick-concrete structure, brick-timber structure, and civil engineering structure.

[0090] Table 1. Collapse rate (%) of buildings with different structures in the XX region

[0091]

[0092] The term "persons buried under buildings during an earthquake" is defined as all those who died or were injured while buried inside buildings. In this embodiment, an assessment of the number of people buried under buildings during an earthquake is constructed based on building scale, taking into account ground seismic intensity, the number of buildings affected at a specific intensity, structural damage rate, the function of a single building, and indoor population density during the earthquake. Based on this, step S50 can be expressed as:

[0093]

[0094] in, This represents the number of people buried in the m-th collapsed building. This represents the population density per unit area of ​​the m-th building as the n-th functional type. Let m represent the total area of ​​the m-th building with structure type q. This represents the occupancy rate of the m-th building as a damaged building at time t. This represents the collapse rate of the qth structural type building under a set intensity u; the structural types include civil engineering structures, brick-timber structures, brick-concrete structures, and frame structures.

[0095] Based on the above embodiments, the functional types of buildings in the target area are first identified using remote sensing data, thereby obtaining the population weight of each functional type of building. This allows rescue forces to accurately locate each high-risk building after an earthquake, rather than a vague area, laying the core spatial foundation for the precise deployment of rescue resources. For each building, based on its functional type, building area, and population weight, the building-scale population is calculated, and the collapse rate and occupancy rate at the time of the earthquake are obtained. This allows for a dynamic estimation of the number of people buried. In this way, a more refined spatial distribution of the population can be obtained by integrating various complex nonlinear potential influencing factors, providing a more scientific and effective decision-making basis for earthquake emergency response.

[0096] One embodiment of this application provides a computer-readable storage medium storing a program, the stored program including methods that can be loaded by a processor and processed in any of the above embodiments.

[0097] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0098] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. A refined method for estimating the population buried by earthquakes, characterized in that, include: Identify the functional types of buildings in the target area based on remote sensing data; Obtain the population weights for buildings of various functional types; For each building, calculate the building-scale population based on its function type, building area, and population weight; For each building, obtain its collapse rate and occupancy rate at the time of the earthquake; For any given building, estimate the number of people buried based on its building-scale population, the collapse rate at the time of the earthquake, and the number of people inside the building; The aforementioned identification of building function types based on remote sensing data includes: identifying the function type of buildings with unknown function types based on kernel density and NED values, including: Obtain the grid covered by each building. For any grid in the covered grid, take the center point of the grid as the center, find the POI type points contained within the preset third distance threshold range, and calculate the kernel density of each POI type. Add the kernel densities of the POI types corresponding to all grids in the covered grid to obtain the comprehensive kernel density of each POI type. For each building with a known functional type, calculate the mean and standard deviation of the comprehensive kernel density for all corresponding POI types; For any building with an unknown function type, calculate the NED value of the building with an unknown function type and that of each building with a known function type based on the average and standard deviation of the comprehensive kernel density of each POI type and the comprehensive kernel density of each building with a known function type. The functional type of the known functional type building corresponding to the minimum NED value shall be taken as the functional type of any unknown functional type building; The process of finding POI type points within a preset third distance threshold range and calculating the kernel density for each POI type includes: in, Here, represents the kernel density, h represents the preset third distance threshold, d represents the spatial dimension, i represents the index of a point of a certain POI type, and I represents the total number of points of a certain POI type. This represents the kernel function, where x represents the coordinates of the center point of the grid. Let i represent the coordinates of the i-th point of a certain POI type, where 1 ≤ i ≤ I; For any building with an unknown functional type, the method for calculating the NED value of that building with an unknown functional type and that of any building with a known functional type includes: in, This represents the NED value between any building of unknown functional type and the k-th building of known functional type. This represents the overall density of the j-th POI type for any building with an unknown functional type, where J represents the number of POI types for any building with an unknown functional type, 1≤j≤J. and Let represent the average and standard deviation of the overall density of the k-th building with known functional type, respectively.

2. The refined method for estimating earthquake-buried population as described in claim 1, characterized in that, In the process of identifying the functional type of a building based on remote sensing data, before identifying the functional type of an unknown functional type building based on kernel density and NED value, the process includes: identifying the functional type of the building and the POI type points it contains based on existing data; if the functional type of any building cannot be identified based on existing data, then checking whether there is a known POI type within a preset first distance threshold range; if so, determining the functional type of any building based on the known POI type; if not, acquiring the shape outline of any building and the shape outline of known building types within a preset second distance threshold range, and calculating the similarity of the shape outlines; the functional type of the building whose shape outline similarity reaches a preset similarity threshold and has the highest similarity value is taken as the functional type of any building; if the functional type of the building still cannot be identified, then identifying the functional type of the unknown functional type building based on kernel density and NED value.

3. The refined method for estimating the population buried by earthquakes as described in claim 1, characterized in that, The methods for calculating the population weight of buildings of each functional type include: A sample dataset is constructed, which includes multiple target area sample unit data after data preprocessing. Each target area sample unit data includes the total population and the building area of ​​buildings of various functional types. The building area of ​​buildings of various functional types in each target area sample unit data is used as the independent variable, and the total population is used as the dependent variable. A random forest regression model is constructed, and the sample dataset is input into the constructed random forest regression model for training and parameter optimization, and the feature importance of each independent variable is extracted; The importance of the features is normalized to obtain the population weight of buildings of each functional type.

4. The refined method for estimating the population buried by earthquakes as described in claim 1, characterized in that, For each building, the calculation of the building-scale population based on its function type, building area, and population weight includes: Based on the population weight of buildings of each functional type and the total population of the target area, the population density of buildings of each functional type is calculated. Based on the population density of buildings of various functional types and the building area of ​​each building, calculate the building-scale population of each building.

5. The refined method for estimating the population buried by earthquakes as described in claim 4, characterized in that, Also includes: The building-scale population of all buildings is summarized to obtain the total building-scale population of the target area. The relative error between the total building-scale population and the actual total population of the target area is calculated. If the relative error is greater than or equal to the preset error threshold, the building function type identification and population weight acquisition process is traced back, and the building-scale population and total building-scale population of each building are recalculated after correction until the relative error is less than the preset error threshold.

6. The refined method for estimating earthquake-buried population as described in claim 1, characterized in that, For each building, the method for obtaining the occupancy rate includes: for each type of building, dividing each day into weekdays and non-weekdays, and obtaining the occupancy rate for each time period of the weekday and each time period of the non-weekday; each time period of the weekday and each time period of the non-weekday includes sleep time, morning peak commuting time, morning work time, lunch break time, afternoon work time, evening peak commuting time and evening rest time.

7. The refined method for estimating earthquake-buried population as described in claim 1, characterized in that, The estimation of the number of people buried under rubble for any given building, based on its building-scale population, collapse rate at the time of the earthquake, and indoor occupancy rate, includes: in, This represents the number of people buried in the m-th collapsed building. This represents the population density per unit area of ​​the m-th building as the n-th functional type. Let m represent the total area of ​​the m-th building with structure type q. This represents the occupancy rate of the m-th building as a damaged building at time t. This represents the collapse rate of the qth structural type building under a set intensity u; the structural types include civil engineering structures, brick-timber structures, brick-concrete structures, and frame structures.

8. A computer-readable storage medium, characterized in that, The medium stores a program that can be loaded by a processor and executed as the refined seismic burial population estimation method as described in any one of claims 1 to 7.

Images