Urban inland inundation simulation method and urban inland inundation emergency scheme selection method

Abstract

The invention discloses an urban inland inundation simulation method and an urban inland inundation emergency scheme selection method, and belongs to the technical field of intelligent water conservancy, and the method comprises the steps: obtaining historical social media data related to urban inland inundation; obtaining a first waterlogging feature set based on historical water level data collected by a water level monitor; obtaining a second waterlogging feature set based on the historical social media data and the historical water level data; the second waterlogging feature set and the first waterlogging feature set are adopted to perform parameter calibration on the urban waterlogging model under different rainfall capacities, and a parameter corresponding relation is obtained; obtaining target rainfall data, real-time social media data and real-time water level data; determining a target parameter of the urban waterlogging model based on the target rainfall data and a parameter corresponding relation; and substituting the target parameter into the urban inland inundation model to obtain a first urban inland inundation simulation result. The method can effectively improve the accuracy degree of urban inland inundation simulation.

Description

Technical Field

[0001] The present disclosure relates to the field of smart water conservancy technology, and in particular to a method for simulating urban waterlogging and a method for selecting an emergency plan for urban waterlogging. Background Art

[0002] Urban waterlogging simulation, prediction and emergency response are important ways to reduce losses caused by waterlogging disasters.

[0003] The urban waterlogging simulation method in the related art includes: obtaining historical water level data collected by a water level monitor, establishing an urban waterlogging model based on the historical water level data collected by the water level monitor, and simulating and predicting urban waterlogging based on the urban waterlogging model.

[0004] However, due to urban planning constraints, some cities have fewer water level monitors, leaving large areas without them. For such cities, the accuracy of urban flooding models built based on historical water level data collected by these monitors is low, leading to lower accuracy in urban flooding simulations. Summary of the Invention

[0005] This disclosure provides a method for simulating urban waterlogging and a method for selecting an emergency plan for urban waterlogging, which can effectively improve the accuracy of urban waterlogging simulation. The technical solution includes at least the following solutions: In a first aspect, a method for simulating urban waterlogging is provided, comprising: obtaining historical social media data related to urban waterlogging; obtaining a first waterlogging feature set based on historical water level data collected by a water level monitor; obtaining a second waterlogging feature set based on the historical social media data and the historical water level data; using the second waterlogging feature set and the first waterlogging feature set to calibrate parameters of an urban waterlogging model under different rainfall amounts to obtain a parameter correspondence, wherein the parameter correspondence includes parameters of the urban waterlogging model corresponding to multiple rainfall amounts; obtaining target rainfall data, real-time social media data, and real-time water level data, wherein the target rainfall data is determined based on the real-time rainfall and the forecast rainfall; determining target parameters of the urban waterlogging model based on the target rainfall data and the parameter correspondence; substituting the target parameters into the urban waterlogging model, and using the urban waterlogging model to perform deduction based on the waterlogging features corresponding to the real-time social media data and the waterlogging features corresponding to the real-time water level data to obtain a first urban waterlogging simulation result.

[0006] Optionally, the social media data is multimodal data, which includes text data, image data and video data. The second waterlogging feature set is obtained based on the historical social media data, including: extracting waterlogging features after segmenting the text data in the historical social media data using a word segmentation tool; extracting waterlogging features after converting the video data in the historical social media data into a single-frame image using a pre-trained YOLO v8 model; extracting waterlogging features from the image data in the historical social media data using a pre-trained YOLO v8 model; and filtering the waterlogging features extracted from the historical social media data using water level error and correlation coefficient to obtain the second waterlogging feature set; wherein the second waterlogging feature set includes multiple waterlogging water levels, and the location and time corresponding to each waterlogging water level.

[0007] Optionally, the urban flooded area is divided into a plurality of grids, and the water level error and the correlation coefficient are calculated with the grid as the smallest unit. The water level error of a first grid is calculated using the following formula, where the first grid is any one of the plurality of grids.

[0008] in, Indicates the water level error, For the The mean value of the water level in the waterlogging feature corresponding to the historical social media data in the first grid of the time step, For the The mean value of the waterlogging water level in the historical water level data corresponding to the water level monitor in the first grid in the time step; The correlation coefficient of the first grid is calculated using the following formula:

[0009] in, represents the mean value of the water level in the waterlogging feature corresponding to the historical social media data at all time steps in the first grid, represents the mean value of the water level in the historical water level data corresponding to the water level monitor at all time steps in the first grid, is the correlation coefficient, Indicates the calculation of covariance, Indicates the calculation of variance.

[0010] Optionally, the water level of the urban flooding corresponding to the kth rainfall in the first urban flooding feature set and the water level of the urban flooding corresponding to the kth rainfall in the second urban flooding feature set are divided into a training set and a validation set according to a set ratio, and the following objective function is used to determine the urban waterlogging model parameters corresponding to the rainfall indicated by the kth rainfall:

[0011] in, is the objective function, is the maximum time step of the k-th rainfall, is the total number of flooding water levels belonging to the second feature set in the training set, is the total number of waterlogging water levels in the training set that belong to the first waterlogging feature set, represents the sum of squared errors, For the The first feature of the second waterlogging feature set in the training set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level, For the The first waterlogging feature set in the training set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level.

[0012] In a second aspect, a method for selecting an urban waterlogging emergency plan is provided, comprising: obtaining an emergency plan set, the emergency plan set including a plurality of emergency plans for managing urban waterlogging; obtaining an urban waterlogging simulation result corresponding to each emergency plan based on an urban waterlogging model, the urban waterlogging model being obtained based on the urban waterlogging simulation method according to any one of claims 1 to 4; calculating a total score of each emergency plan based on the urban waterlogging simulation result corresponding to each emergency plan; and adopting the emergency plan with the highest total score to manage urban waterlogging.

[0013] Optionally, the score of each emergency plan is calculated based on an indicator set, wherein the indicator set includes the maximum water depth reduction value of each waterlogging point, the maximum waterlogging area reduction value of each waterlogging point, the average maximum water depth reduction value of all waterlogging points, the sum of the large waterlogging area reduction values ​​of all waterlogging points, the shortening value of the waterlogging receding time of each waterlogging point, the average of the shortening value of the waterlogging receding time of all waterlogging points, and the cost of implementing the plan.

[0014] In a third aspect, an urban waterlogging simulation device is provided, which includes a first acquisition module for acquiring historical social media data related to urban waterlogging; a second acquisition module for acquiring a first waterlogging feature set based on historical water level data collected by a water level monitor; a third acquisition module for acquiring a second waterlogging feature set based on the historical social media data and the historical water level data; a parameter calibration module for calibrating parameters of an urban waterlogging model under different rainfall amounts using the second waterlogging feature set and the first waterlogging feature set to obtain a parameter correspondence, wherein the parameter correspondence includes a plurality of rainfall corresponding parameters. Parameters of the urban waterlogging model; a fourth acquisition module, used to obtain target rainfall data, real-time social media data and real-time water level data, wherein the target rainfall data is determined based on the real-time rainfall and the forecast rainfall; a target parameter determination module, used to determine the target parameters of the urban waterlogging model based on the correspondence between the target rainfall data and the parameters; an urban waterlogging simulation module, used to substitute the target parameters into the urban waterlogging model, and based on the waterlogging characteristics corresponding to the real-time social media data and the waterlogging characteristics corresponding to the real-time water level data, use the urban waterlogging model to perform deduction to obtain a first urban waterlogging simulation result.

[0015] Optionally, the social media data is multimodal data, which includes text data, image data and video data. The third acquisition module is further used to extract waterlogging features after segmenting the text data in the historical social media data using a word segmentation tool; for the video data in the historical social media data, after converting the video data into a single-frame image, the pre-trained YOLO v8 model is used to extract waterlogging features; for the image data in the historical social media data, the pre-trained YOLO v8 model is used to extract waterlogging features; the water level error and correlation coefficient are used to screen the waterlogging features extracted from the historical social media data to obtain the second waterlogging feature set; wherein the second waterlogging feature set includes multiple waterlogging water levels, and the location and time corresponding to each waterlogging water level.

[0016] Optionally, in the third acquisition module, the urban waterlogging area is divided into a plurality of grids, and the water level error and the correlation coefficient are calculated with the grid as the smallest unit. The water level error of the first grid is calculated using the following formula, where the first grid is any one of the plurality of grids:

[0017] in, Indicates the water level error, For the The mean value of the water level in the waterlogging feature corresponding to the historical social media data in the first grid of the time step, For the The mean value of the waterlogging water level in the historical water level data corresponding to the water level monitor in the first grid in the time step; The correlation coefficient of the first grid is calculated using the following formula:

[0018] in, represents the mean value of the water level in the waterlogging feature corresponding to the historical social media data at all time steps in the first grid, represents the mean value of the water level in the historical water level data corresponding to the water level monitor at all time steps in the first grid, is the correlation coefficient, Indicates the calculation of covariance, Indicates the calculation of variance.

[0019] Optionally, in the parameter calibration module, the water level of the urban flooding corresponding to the kth rainfall in the first urban flooding feature set and the water level of the urban flooding corresponding to the kth rainfall in the second urban flooding feature set are divided into a training set and a validation set according to a set ratio, and the following objective function is used to determine the urban waterlogging model parameters corresponding to the rainfall indicated by the kth rainfall:

[0020] in, is the objective function, is the maximum time step of the k-th rainfall, is the total number of flooding water levels belonging to the second feature set in the training set, is the total number of waterlogging water levels in the training set that belong to the first waterlogging feature set, represents the sum of squared errors, For the The first feature of the second waterlogging feature set in the training set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level, For the The first waterlogging feature set in the training set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level.

[0021] In a fourth aspect, a device for selecting an urban waterlogging emergency plan is provided, comprising: an acquisition module for acquiring an emergency plan set, the emergency plan set including multiple emergency plans for managing urban waterlogging; a waterlogging simulation module for acquiring an urban waterlogging simulation result corresponding to each emergency plan based on an urban waterlogging model, the urban waterlogging model being obtained based on the urban waterlogging simulation method described in any one of claims 1 to 4; a scoring module for calculating a total score of each emergency plan based on the urban waterlogging simulation result corresponding to each emergency plan; and a waterlogging management module for adopting the emergency plan with the highest total score to manage urban waterlogging.

[0022] Optionally, in the scoring module, the score of each emergency plan is calculated based on an indicator set, and the indicator set includes the maximum waterlogging depth reduction value of each waterlogging point, the maximum waterlogging area reduction value of each waterlogging point, the average maximum water depth reduction value of all waterlogging points, the sum of the large waterlogging area reduction values ​​of all waterlogging points, the shortening value of the waterlogging receding time of each waterlogging point, the average of the shortening values ​​of the waterlogging receding time of all waterlogging points, and the cost of implementing the plan.

[0023] In a fifth aspect, a computer device is also provided, comprising: a memory and a processor, wherein the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor, thereby executing the urban waterlogging simulation method and the urban waterlogging emergency plan selection method described in the above embodiments.

[0024] In the sixth aspect, a computer-readable storage medium is also provided, in which at least one computer program is stored. The at least one computer program is loaded and executed by a processor, thereby executing the urban waterlogging simulation method and the urban waterlogging emergency plan selection method described in the above embodiments.

[0025] In a seventh aspect, a computer program product is provided, comprising a computer program / instruction, which implements the method described in the first aspect when executed by a processor.

[0026] The beneficial effects of the technical solutions provided by the embodiments of the present disclosure include at least: In the disclosed embodiment, a first waterlogging feature set is obtained through a water level monitor, and a second waterlogging feature set is obtained through social media data. Compared to existing technologies that rely solely on a single data source (such as using only a water level monitor), relevant information about waterlogging events can be obtained more comprehensively and in real time. By using the first waterlogging feature set and the second waterlogging feature set to jointly calibrate the parameters of the urban waterlogging model, the second waterlogging feature set can effectively compensate for the waterlogging characteristics in blank areas without water level monitors. This effectively improves the accuracy of the parameter calibration of the urban waterlogging model, thereby improving the reliability of the urban waterlogging simulation. By calibrating the parameters of the urban waterlogging model under different rainfall amounts, a parameter correspondence is obtained, and then based on the target rainfall data and the parameter correspondence, the target parameters of the urban waterlogging model are determined. Parameters of the urban waterlogging model that are more suitable for the target rainfall data can be selected, further improving the reliability of the urban waterlogging simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0028] Figure 1 A flow chart of an urban waterlogging simulation method provided by an exemplary embodiment of the present disclosure is shown; Figure 2 A flow chart of a method for selecting an urban flood emergency plan provided by an exemplary embodiment of the present disclosure is shown; Figure 3 A schematic structural diagram of an urban waterlogging simulation device provided by an exemplary embodiment of the present disclosure is shown; Figure 4 A schematic structural diagram of an urban flood emergency plan selection device provided by an exemplary embodiment of the present disclosure is shown; Figure 5 It is a structural diagram of a computer device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0029] Unless otherwise defined, the technical or scientific terms used herein shall have the usual meanings understood by persons of ordinary skill in the field to which the present disclosure belongs. The words “first”, “second”, “third” and similar terms used in the patent application specification and claims of the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as “one” or “a” do not indicate a quantity limitation, but rather indicate the presence of at least one. Words such as “include” or “comprising” and similar terms mean that the elements or objects appearing before “include” or “comprising” cover the elements or objects listed after “include” or “comprising” and their equivalents, and do not exclude other elements or objects.

[0030] In order to make the objectives, technical solutions and advantages of the present disclosure more clear, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.

[0031] Figure 1 A flow chart of an urban flooding simulation method provided by an exemplary embodiment of the present disclosure is shown. The method can be executed by a computer device. Figure 1 , the method comprising: In step 101, historical social media data related to urban waterlogging is obtained.

[0032] In the embodiment of the present disclosure, social media data is multimodal data, which includes text data, image data, and video data.

[0033] Therefore, when obtaining historical social media data related to urban waterlogging, it is also necessary to obtain historical multimodal data related to urban waterlogging, such as historical text data, historical image data, and historical video data related to urban waterlogging.

[0034] During implementation, social media APIs (Application Programming Interfaces) and crawler technology can be incorporated to capture text, images, and videos related to urban flooding within social media data. For example, pre-defined search keywords related to urban flooding can be used, such as: 1) Event-related keywords: urban flooding, flood, waterlogging, rainstorm, flooding, waterlogging, heavy rainfall, and downpour; 2) Location-related keywords: road, street, urban district, residential area, and building name; 3) Water depth-related keywords: water depth, waterlogging, submergence, and flooding; and 4) Time range: 24 hours before the start of rainfall to 72 hours after the end of rainfall. These keywords can be used to retrieve historical social media data related to urban flooding. It is important to note that when using crawler technology to capture information, it is important to understand and adhere to the user regulations of each social platform in advance to avoid infringing user privacy and platform policies.

[0035] In the disclosed embodiment, the historical social media data related to urban waterlogging is social media data related to each rainfall event within a historical time range, wherein the historical time range can be set as required.

[0036] In step 102, a first waterlogging feature set is obtained based on historical water level data collected by a water level monitor.

[0037] A water level monitor is a device installed within a city. It collects water level data in real time, allowing continuous flooding levels to be obtained. When considering obtaining historical water level data collected by a water level monitor, the number of rainfall events within a historical timeframe, as well as the start and end times of each rainfall event, can be counted to obtain the flooding levels for the corresponding rainfall events from the water level data collected by the water level monitor. Furthermore, once installed, the water level monitor will not move, and its location is known.

[0038] Based on this, the first waterlogging feature set can be obtained. Each waterlogging feature in the first waterlogging feature set is essentially a triple that includes the water level (collected by the water level monitor), the time of occurrence of the waterlogging water level (obtained by the time when the water level monitor uploads the waterlogging water level), and the location of occurrence of the waterlogging water level (obtained by the installation location of the water level monitor).

[0039] In step 103 , a second waterlogging feature set is obtained based on historical social media data and historical water level data.

[0040] Optionally, the second flooding feature set includes multiple flooding water levels, as well as the location and time corresponding to each flooding water level. For each flooding water level in the second flooding feature set, the time and location at which the level occurred can also be determined (information on social media platforms includes the time and location of transmission). Therefore, each flooding feature in the second flooding feature set is essentially a triple consisting of the flooding water level, the time of occurrence, and the location of the occurrence.

[0041] Since historical social media is multimodal data, different modal data need to be processed in different ways. In this case, step 103 may optionally include the following steps ad.

[0042] In step a, the text data in the historical social media data is segmented using a word segmentation tool and then the waterlogging features are extracted.

[0043] Before executing step a, the text data in the historical social media data needs to be preprocessed, and irrelevant information styles, such as advertisements, punctuation marks, special characters, irrelevant content, stop words, etc., are preset. These irrelevant words and irrelevant information are removed using keyword matching rules and Chinese stop word lists, and the preprocessed text data is stored in a unified encoding format (such as UTF-8 format) to facilitate the extraction of waterlogging features.

[0044] For the preprocessed text data in historical social media data, a word segmentation tool can be used to segment and tag the uniformly encoded text, identifying and parsing place names, time information, and flood depth information within the text. For example, the Jieba word segmentation tool can be used to implement step a. The Jieba word segmentation tool is a Chinese text processing tool developed based on natural language processing (NLP) technology and statistical machine learning algorithms. Through efficient part-of-speech tagging and semantic parsing, it accurately segments Chinese text and extracts keywords. This tool boasts robust algorithms, high processing efficiency, and wide open-source adaptability, making it particularly suitable for feature extraction from social media text data in urban waterlogging scenarios.

[0045] For example, consider a social media post that reads, "Five minutes ago, floodwater at the intersection of Jiefang Avenue and Yongqing Road in a certain city reached a depth of approximately 0.5 meters, causing traffic congestion." The structured format of the extracted information is: 1) Location: Intersection of Jiefang Avenue and Yongqing Road; 2) Time: Five minutes ago (converted to 10:30 AM on February 27, 2025, based on the time the message was posted); 3) Floodwater level: 0.5 meters. Next, use a map tool to retrieve the latitude and longitude coordinates of the location in the text, and use geographic information system software or a coordinate conversion tool to convert it to the coordinate system used in step two. This ultimately yields a flooding feature.

[0046] In step b, for the video data in the historical social media data, the video data is converted into single-frame images, and the pre-trained YOLO v8 model is used to extract waterlogging features.

[0047] Similarly, video and image data from historical social media data also requires preprocessing. The size and contrast of each video and image data must be standardized, and images and videos with clarity below a specified resolution threshold must be removed. For any image or video, the latitude and longitude coordinates of the location can be extracted based on the published geotag and mapped using mapping tools. These extracted latitude and longitude coordinates can be stored alongside the image or video data as supplementary information.

[0048] When preprocessing the video data, the video data needs to be converted into single-frame images. For example, single-frame images can be extracted from the video at fixed time intervals or fixed frame number intervals.

[0049] In the disclosed embodiments, a pre-trained YOLO v8 (You Only Look Once v8) model is used to identify flooding levels in images. During implementation, the YOLO v8 model is trained using a publicly labeled waterlogging dataset (such as a road flooding detection dataset). The training objective is to minimize the YOLO v8 model's built-in combined loss function. A threshold for the training objective is pre-set. Model training is complete when the target loss function value calculated by iterative model calculations falls below the specified threshold. The trained YOLO v8 model is now capable of identifying flooding levels in images.

[0050] After obtaining the pre-trained YOLO v8 model, the single-frame image corresponding to each video data can be input into the pre-trained YOLO v8 model to obtain the waterlogging features corresponding to each video data.

[0051] In step c, the pre-trained YOLO v8 model is used to extract waterlogging features from the image data in the historical social media data.

[0052] After obtaining the pre-trained YOLO v8 model, each image data can be input into the pre-trained YOLO v8 model to obtain the waterlogging features corresponding to each image data.

[0053] In step d, water level error and correlation coefficient are used to filter the waterlogging features extracted from historical social media data to obtain a second waterlogging feature set.

[0054] In this disclosed embodiment, the urban flooding area is divided into multiple grids, and the water level error and correlation coefficient are calculated using the grid as the smallest unit. Here, the urban flooding area refers to the area where urban flooding simulation is required. The historical social media data collected above also refers to the historical social media data within the urban flooding area.

[0055] Here, the water level error and correlation coefficient are calculated based on historical water level data and historical social media data.

[0056] Optionally, the water level error of the first grid is calculated using formula (1), wherein the first grid is any one of the multiple grids in the urban flooded area.

[0057] (1) In formula (1), Indicates the water level error, For the The mean value of the water level in the waterlogging feature corresponding to the historical social media data in the first grid of the time step, .in, is the total number of waterlogging features in the first grid in the second waterlogging feature set, Indicates the The i-th waterlogging feature in the first grid in the second waterlogging feature set at time steps. For the The mean value of the flooding water level in the historical water level data corresponding to the water level monitor in the first grid in the time step, .in, is the total number of waterlogging features in the first grid in the first waterlogging feature set, Indicates the The i-th waterlogging feature in the first grid in the first waterlogging feature set at time steps.

[0058] The correlation coefficient of the first grid is calculated using formula (2).

[0059] (2) In formula (2), represents the mean value of the water level in the waterlogging feature corresponding to the historical social media data of all time steps in the first grid, , is the maximum time step, represents the mean value of the flooding water level in the historical water level data corresponding to the water level monitors at all time steps in the first grid. , is the correlation coefficient, Indicates the calculation of covariance, The meanings of other parameters in formula (2) are the same as those in formula (1), and their detailed description is omitted here.

[0060] The water level error threshold and correlation coefficient threshold are pre-set. Only when the water level error of the first grid meets the water level error threshold and the correlation coefficient of the first grid meets the correlation coefficient threshold, the data in the first grid can be retained. For example, when the water level error threshold is 20% and the correlation coefficient threshold is 0.7, only when the water level error of the first grid meets the water level error threshold and the correlation coefficient of the first grid meets the correlation coefficient threshold, the data in the first grid can be retained. When the correlation coefficient threshold of the first grid is greater than or equal to 0.7, the waterlogging features corresponding to the historical social media data in the first grid are retained in the second waterlogging feature set.

[0061] Depending on whether there are water level monitors and social media data in the first grid, it can be divided into the following four situations.

[0062] (1) If both water level monitors and historical social media data are present in the first grid, the water level error and correlation coefficient of the first grid are evaluated according to the set water level error threshold and correlation coefficient threshold. If the threshold requirements are met, the waterlogging features corresponding to the historical social media data are retained in the second waterlogging feature set. If the threshold requirements are not met, the waterlogging features corresponding to the historical social media data in the first grid are removed.

[0063] (2) If there is only historical social media data in the first grid but no water level monitor, the water level error and correlation coefficient of the adjacent grid (the adjacent grid is the grid closest to the first grid that meets condition (1)) are calculated. If the historical social media data of the adjacent grid meets the requirements, it is considered that the historical social media data of the first grid also meets the requirements, and the waterlogging features corresponding to the historical social media data in the first grid are retained in the second waterlogging feature set.

[0064] (3) If there is only a water level monitor in the first grid and no historical social media data, the water level error and correlation coefficient are not calculated for the first grid, because the waterlogging feature corresponding to the water level monitor will be stored in the first waterlogging feature set.

[0065] (4) If there is neither a water level monitor nor historical social media data in the first grid, the first grid does not calculate the water level error and correlation coefficient, nor does it participate in the screening process of the second waterlogging feature set.

[0066] For other grids in the urban waterlogging area except the first grid, calculations can also be performed in a similar manner to the first grid, and finally a second waterlogging feature set can be obtained by screening.

[0067] In step 104, the second waterlogging feature set and the first waterlogging feature set are used to calibrate parameters of the urban waterlogging model under different rainfall amounts to obtain a parameter correspondence.

[0068] The parameter correspondence relationship includes parameters of the urban waterlogging model corresponding to multiple rainfall amounts.

[0069] Before executing step 104, an urban waterlogging model can be first established. The urban waterlogging model is a two-dimensional model. During implementation, the scope of the model can be determined based on the preset urban waterlogging area, and then the city parameters corresponding to the urban waterlogging area can be obtained. The urban parameters include but are not limited to: the latest high-precision terrain data of the city, high-definition image maps, river system direction and basin boundaries, drainage network vector data, drainage pump station, urban land attribute vector data, etc. The above city parameters need to be preprocessed, which includes the following two steps.

[0070] The first step is to remove outliers in city parameters.

[0071] You can first determine the range of values ​​for various urban parameters based on empirical values, and then eliminate outliers within these ranges. For example, the normal range of drainage flow at a drainage pump station is 0-400 cubic meters per second, and the normal range of drainage pipe diameter is greater than or equal to 300 mm.

[0072] The second step is to review and calibrate the city parameters.

[0073] For each type of city parameter collected, 30-50% of the data is randomly selected for review. For example, the accuracy of drainage network data is reviewed using closed-circuit television (CCTV) and field measurements, with a focus on verifying network node coordinates, orientation, pipe bottom elevation, internal pipe diameter, pipe material, and connection status. Field measurements are also used to verify the rationality and accuracy of the calculated base surface and existing topographic data for each water level sensor. If the proportion of accurate data after review is greater than or equal to a set percentage (e.g., 80% or 90%) of the total reviewed data, the original city parameters are highly accurate and can be used directly. If the proportion of accurate data after review is less than the set percentage, other high-quality data will be required to replace the city parameters.

[0074] After preprocessing urban parameters, they can be integrated with urban terrain data and imported into appropriate modeling software to create an urban flooding model. For example, using geographic information system software to unify the coordinate systems of data from different data sources, existing urban land use data, river systems, drainage pumping stations, drainage network vector data, and image data can be overlaid on the urban terrain data based on coordinates and elevations. The overlaid and integrated data can then be carefully verified. Finally, the verified data can be imported into appropriate modeling software to create an urban flooding model.

[0075] Among them, the following two aspects need to be focused on during detailed verification.

[0076] (1) Review the upper surface elevation of urban rainwater grates, the top elevation of the drainage network and the connection with the urban road surface. Generally speaking, the upper surface elevation of the rainwater grate and the top elevation of the drainage network should be lower than the ground elevation at the location.

[0077] (2) The consistency between the urban land vector boundary and the urban terrain boundary, especially the terrain accuracy of special plots such as community walls, river embankments, urban sunken bridges, and overpasses.

[0078] When obtaining the first urban waterlogging feature set and the second urban waterlogging feature set, the number of rainfall events within the historical time range has been counted. At this time, the multiple rainfall events within the historical time range can be sorted according to the rainfall amount, and then the features in the first urban waterlogging feature set and the features in the second urban waterlogging feature set corresponding to any rainfall event can be substituted into the urban waterlogging model to obtain the parameters of the urban waterlogging model corresponding to the rainfall amount of the rainfall event.

[0079] For example, the amount of rainfall in the kth rainfall event is the kth rainfall amount, and the parameters of the urban waterlogging model corresponding to the kth rainfall amount are calibrated using the following four steps.

[0080] In the first step, the water level corresponding to the kth rainfall in the first waterlogging feature set and the water level corresponding to the kth rainfall in the second waterlogging feature set are divided into a training set and a validation set according to a set ratio.

[0081] For example, the division ratio can be 7 to 3, but is not limited thereto. Taking the division ratio of 7 to 3 as an example, 70% of the water levels corresponding to the k-th rainfall in the first waterlogging feature set are used as the training set, and 30% of the water levels corresponding to the k-th rainfall in the first waterlogging feature set are used as the validation set; 70% of the water levels corresponding to the k-th rainfall in the second waterlogging feature set are used as the training set, and 30% of the water levels corresponding to the k-th rainfall in the second waterlogging feature set are used as the validation set.

[0082] The second step is to establish the objective function of the urban waterlogging model.

[0083] (3) In formula (3), is the objective function, is the maximum time step of the k-th rainfall, is the total number of flooding water levels belonging to the second feature set in the training set, is the total number of waterlogging levels in the training set that belong to the first waterlogging feature set, represents the sum of squared errors, For the The first feature in the training set belonging to the second flooding feature set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level, For the The first waterlogging feature set in the training set under the time step The flood water level, The urban flooding model predicts The corresponding flood water level.

[0084] In formula (3), Calculated using formula (4), It is calculated using formula (5).

[0085] (4) (5) The meanings of the parameters in formula (4) and formula (5) are the same as those in formula (3), and their detailed description is omitted here.

[0086] It should be noted that water level monitoring data is usually continuous monitoring data, and there is no data loss under normal circumstances; however, this is not always the case with data extracted from social media, and there may be no data in some time steps. Therefore, when calculating the objective function, if does not exist at the time step , then it is considered that the time step in formula (4) is 0.

[0087] In the third step, the urban flooding model is used to predict the rainfall of the kth rainfall at each time step.

[0088] Through the second step, the rainfall of the kth rainfall at each time step predicted by the urban flooding model can be obtained, from which the corresponding and .

[0089] The fourth step is to use the global optimization algorithm to solve the objective function according to the rainfall of the kth rainfall at each time step predicted by the urban waterlogging model, and obtain the parameters of the urban waterlogging model corresponding to the kth rainfall.

[0090] For example, the SCEM-UA (Shuffled Complex Evolution-University of Arizona) algorithm is used to solve the objective function. The algorithm terminates when the GR (Gauge Repeatability) value within the optimization algorithm falls below an optimization threshold (e.g., 1.2). After solving the objective function, the parameters of the calibrated urban flooding model are obtained. These parameters are then evaluated and verified using a validation set.

[0091] For example, NSE (Nash-Sutcliffe efficiency coefficient) and RMSE (Root Mean Squared Error) are used to evaluate and verify the parameter. NSE is calculated using formula (6) and RMSE is calculated using formula (7).

[0092] (6) In formula (6), is the calculated Nash efficiency coefficient, For the The validation set in the time step The flood water level, Output of urban flooding model The corresponding flood water level, represents the mean of the flood water level at all time steps in the training set, The meanings of other parameters in formula (6) are the same as those in formula (3), and their detailed description is omitted here.

[0093] (7) In formula (7), is the calculated root mean square error. The meanings of other parameters in formula (7) are the same as those in formula (6), and their detailed description is omitted here.

[0094] The NSE threshold and RMSE threshold can be pre-set. A larger NSE value or a smaller RMSE value indicates a better model prediction; conversely, a lower value indicates a worse prediction. Therefore, if the NSE calculated for the urban waterlogging model parameters corresponding to the k-th rainfall event is greater than or equal to the NSE threshold and less than or equal to the RMSE threshold, the urban waterlogging model parameters corresponding to the k-th rainfall event are reasonable and can be retained. If the NSE and RMSE threshold requirements are not met, the input conditions must be modified and the urban waterlogging model parameters recalibrated. Once the NSE and RMSE threshold requirements for the urban waterlogging model corresponding to the k-th rainfall event are met, the parameters for the urban waterlogging model corresponding to the rainfall amount of the k-th rainfall event in the parameter correspondence can be obtained.

[0095] For the parameters of the urban waterlogging model corresponding to the rainfall of other rainfall events, the parameters of the urban waterlogging model can also be calibrated in a manner similar to that of the kth rainfall event, thereby obtaining the parameter correspondence.

[0096] In step 105 , target rainfall data, real-time social media data, and real-time water level data are obtained.

[0097] The target rainfall data is determined based on the real-time rainfall and the forecast rainfall.

[0098] For example, if the current real-time rainfall is 10mm during a rainfall event and the weather forecast predicts a remaining rainfall of 10mm, the target rainfall is 10mm + 10mm, or 20mm. In other words, the target rainfall is the total rainfall for the current rainfall event.

[0099] The real-time social media data and the real-time water level data also need to be processed in a similar manner to steps 101 to 103 .

[0100] In step 106, target parameters of the urban waterlogging model are determined based on the target rainfall data and the parameter correspondence.

[0101] Substituting the target rainfall into the parameter correspondence, the parameter corresponding to the target rainfall can be found, and this parameter is the target parameter. If the target rainfall does not exist in the parameter correspondence, the parameter corresponding to the rainfall closest to the target rainfall in the parameter correspondence is used as the target parameter.

[0102] In step 107, the target parameters are substituted into the urban waterlogging model, and based on the waterlogging characteristics corresponding to the real-time social media data and the waterlogging characteristics corresponding to the real-time water level data, the urban waterlogging model is used for deduction to obtain the first urban waterlogging simulation result.

[0103] Based on the results of the first urban flooding simulation, staff can assess the extent of urban flooding caused by the current rainfall event and determine which emergency response plans to implement to address the affected areas. For example, if the first urban flooding simulation indicates that water levels in certain areas of the city have exceeded the designated level, these areas will be considered affected and require emergency response plans.

[0104] In the disclosed embodiment, a first waterlogging feature set is obtained through a water level monitor, and a second waterlogging feature set is obtained through social media data. Compared to existing technologies that rely solely on a single data source (such as using only a water level monitor), relevant information about waterlogging events can be obtained more comprehensively and in real time. By using the first waterlogging feature set and the second waterlogging feature set to jointly calibrate the parameters of the urban waterlogging model, the second waterlogging feature set can effectively compensate for the waterlogging characteristics in blank areas without water level monitors. This effectively improves the accuracy of the parameter calibration of the urban waterlogging model, thereby improving the reliability of the urban waterlogging simulation. By calibrating the parameters of the urban waterlogging model under different rainfall amounts, a parameter correspondence is obtained, and then based on the target rainfall data and the parameter correspondence, the target parameters of the urban waterlogging model are determined. Parameters of the urban waterlogging model that are more suitable for the target rainfall data can be selected, further improving the reliability of the urban waterlogging simulation.

[0105] Figure 2 A flowchart of a method for selecting an urban flood emergency plan provided by an exemplary embodiment of the present disclosure is shown. The method can be executed by a computer device. Figure 2 , the method comprising: In step 201, a set of emergency plans is obtained.

[0106] The emergency plan set includes multiple emergency plans for managing urban waterlogging.

[0107] In step 202, based on the urban waterlogging model, the urban waterlogging simulation results corresponding to each emergency plan are obtained.

[0108] During implementation, each emergency plan can be generalized and substituted into the urban flooding model along with the target parameters. Based on the flooding characteristics corresponding to real-time social media data and real-time water level data, the urban flooding model is used for deduction, thereby obtaining urban flooding simulation results corresponding to each emergency plan. Because the simulation calculations of the flooding model require certain computing power from the underlying computing platform, it is recommended that the calculations be performed using supercomputing platforms, high-performance computers, cloud computing platforms, etc., during step 202. Simultaneously, each plan can be run in parallel to accelerate the calculations, thereby reserving sufficient time for on-site emergency response to the flooding incident.

[0109] The urban flooding simulation results for each emergency plan differ from the initial urban flooding simulation results. The initial urban flooding simulation results represent natural flooding without any human intervention, while the urban flooding simulation results for each emergency plan represent flooding after implementing the emergency plan. Compared to the initial urban flooding simulation results, the urban flooding simulation results for each emergency plan can better reflect the effectiveness of the emergency plan in managing urban flooding. Some emergency plans are shown in Table 1 below.

[0110] Table 1: Schematic diagram of emergency plan set.

[0111]

[0112] In step 203 , based on the urban flooding simulation results corresponding to each emergency plan, a total score of each emergency plan is calculated.

[0113] Optionally, the score of each emergency plan is calculated based on an indicator set, which includes the maximum waterlogging depth reduction value of each waterlogging point, the maximum waterlogging area reduction value of each waterlogging point, the average maximum water depth reduction value of all waterlogging points, the sum of the large waterlogging area reduction values ​​of all waterlogging points, the shortening value of the waterlogging receding time value of each waterlogging point, the average of the shortening value of the waterlogging receding time value of all waterlogging points, and the cost of implementing the plan.

[0114] The indicators in the indicator set can be divided into three categories: hydraulic effect, time effect, and solution implementation cost. The hydraulic effect category includes four indicators: the reduction in maximum waterlogging depth at each waterlogging point, the reduction in maximum waterlogging area at each waterlogging point, the average reduction in maximum waterlogging depth across all waterlogging points, and the sum of the reductions in maximum waterlogging area across all waterlogging points. The time effect category includes two indicators: the reduction in waterlogging time at each waterlogging point and the average reduction in waterlogging time across all waterlogging points. The solution implementation cost category includes one indicator: solution implementation cost.

[0115] The seven indicators in the indicator set are calculated using formulas (8) to (14) respectively.

[0116] (8) (9) (10) (11) (12) (13) (14) In formulas (8) to (14), For the An emergency plan, For reference scheme, , is the total number of emergency plans in the emergency plan set, and the reference plan is the case where no emergency plan is adopted. The waterlogging points are determined based on the waterlogging simulation results of the first city.

[0117] For the The waterlogging point is The maximum water depth reduction value under each emergency plan is: For the The first emergency plan The maximum water depth of each flood point, For the reference scheme The maximum water depth at each flood point. For the The waterlogging point is The maximum reduction in flooded area under each emergency plan is: For the The first emergency plan The maximum waterlogging area of ​​each flood point, For the reference scheme The maximum waterlogging area of ​​a flood point. is the average maximum water depth reduction value of all waterlogging points, The total reduction in large waterlogged areas at all waterlogged points.

[0118] For the The waterlogging point is The shortened time for waterlogging to disappear under each emergency plan is: For the The waterlogging point is The time it takes for the accumulated water to recede under each emergency plan. For the The time it takes for water to recede at each flood point under the reference plan. It is the average of the shortened time for waterlogging to recede at all waterlogging points.

[0119] The implementation cost of the program, is the equipment cost, For labor costs, is the material cost, For energy costs, To remove to Other costs besides.

[0120] Optionally, a weighted scoring method is used to obtain the total score of each emergency plan.

[0121] The main principles of the weighted scoring method are explained below.

[0122] In the weighted scoring method, different types of indicators have different weights. In the embodiment of the present disclosure, the weight sum of the hydraulic effect indicators is , the weight sum of time effect indicators is , the weights of program implementation indicators are . .

[0123] The weights of the four indicators in the hydraulic effect category are equal. ,in and The calculation of these two indicators is related to the specific waterlogging points Therefore, the weight of each waterlogged point in these two items is .

[0124] Similarly, the weights of the indicators in the time effect category are equal, that is, the weights of the two indicators in the time effect category are equal. ,in Calculation and specific waterlogging points Therefore, the weight of each flood point in this item is .

[0125] There is only one indicator in the implementation cost category , whose weight is .

[0126] The dimensions of each calculation indicator are different and cannot be directly compared, so it is necessary to unify the dimensions of the indicators of each emergency plan and assign points according to the ranking.

[0127] Each emergency plan can be calculated using the seven indicators above. For the nth indicator, sort each emergency plan from highest to lowest according to the value of the nth indicator to obtain the ranking corresponding to the nth indicator. The seven indicators above can be ranked similarly to the nth indicator, with each indicator corresponding to a ranking, meaning there are a total of seven rankings.

[0128] The score of each emergency plan can be determined based on its ranking in these 7 rankings and the weight of each indicator itself.

[0129] The full score is 100 points. Take the corresponding ranking as an example. In this ranking, the score of the first emergency plan is , the score of any waterlogging point in this scheme is The score of the emergency plan of the i-th place is , the score of any waterlogging point in this scheme is The last person's contingency plan scores , the score of any waterlogging point in this scheme is In this way, we can calculate The score of each emergency plan in the corresponding ranking.

[0130] The emergency plans in any of the seven rankings can be scored using similar schemes, and finally the scores of each emergency plan on the seven indicators can be obtained, so that the formula (15) can be used to score the emergency plan. The emergency plans are weighted and scored to obtain the The total score of the emergency plan.

[0131] (15) In formula (15), For the The emergency plan is in The ratings on the indicators, For the The total score of the emergency plan.

[0132] In step 204, the emergency plan with the highest total score is adopted for urban waterlogging management.

[0133] If an emergency plan has the highest total score, it means that the emergency plan is relatively balanced in all aspects and meets the requirements for selecting urban waterlogging emergency plans. The emergency plan can be used for urban waterlogging management.

[0134] By quantifying and constructing a flooding model that takes into account different emergency response plans, simulating the expected effects of each emergency plan, and quantitatively comparing the treatment effects of each plan using a weighted scoring method, we ensure the objectivity and accuracy of the selection of the optimal emergency response plan.

[0135] The following are device embodiments of the present application. For details not described in detail in the device embodiments, reference may be made to the above method embodiments.

[0136] Figure 3 FIG2 shows a schematic diagram of the structure of an urban flooding simulation device provided by an exemplary embodiment of the present disclosure. Figure 3 The urban waterlogging simulation device 300 includes a first acquisition module 301 , a second acquisition module 302 , a third acquisition module 303 , a parameter calibration module 304 , a fourth acquisition module 305 , a target parameter determination module 306 and a waterlogging simulation module 307 .

[0137] The first acquisition module 301 is used to acquire historical social media data related to urban waterlogging.

[0138] The second acquisition module 302 is used to acquire a first waterlogging feature set based on historical water level data collected by the water level monitor.

[0139] The third acquisition module 303 is configured to acquire a second waterlogging feature set based on the historical social media data and the historical water level data.

[0140] The parameter calibration module 304 is used to calibrate the parameters of the urban waterlogging model under different rainfall amounts using the second waterlogging feature set and the first waterlogging feature set to obtain a parameter correspondence, which includes parameters of the urban waterlogging model corresponding to multiple rainfall amounts.

[0141] The fourth acquisition module 305 is used to acquire target rainfall data, real-time social media data, and real-time water level data, wherein the target rainfall data is determined based on the real-time rainfall and the forecast rainfall.

[0142] The target parameter determination module 306 is used to determine the target parameters of the urban waterlogging model based on the target rainfall data and the parameter correspondence.

[0143] The urban flooding simulation module 307 is used to substitute the target parameters into the urban flooding model, and based on the urban flooding characteristics corresponding to the real-time social media data and the urban flooding characteristics corresponding to the real-time water level data, use the urban flooding model to perform deduction to obtain a first urban flooding simulation result.

[0144] Optionally, the social media data is multimodal data, which includes text data, image data and video data. The third acquisition module 303 is further used to extract waterlogging features after segmenting the text data in the historical social media data using a word segmentation tool; extract waterlogging features by using a pre-trained YOLO v8 model after converting the video data in the historical social media data into a single-frame image; extract waterlogging features by using a pre-trained YOLO v8 model for the image data in the historical social media data; and filter the waterlogging features extracted from the historical social media data using water level error and correlation coefficient to obtain the second waterlogging feature set; wherein the second waterlogging feature set includes multiple waterlogging water levels, and the location and time corresponding to each waterlogging water level.

[0145] Optionally, in the third acquisition module 303, the urban waterlogging area is divided into a plurality of grids, and the water level error and the correlation coefficient are calculated with the grid as the smallest unit. The water level error of the first grid is calculated using the following formula, where the first grid is any one of the plurality of grids:

[0146] in, Indicates the water level error, For the The mean value of the water level in the waterlogging feature corresponding to the historical social media data in the first grid of the time step, For the The mean value of the waterlogging water level in the historical water level data corresponding to the water level monitor in the first grid in the time step; The correlation coefficient of the first grid is calculated using the following formula:

[0147] in, represents the mean value of the water level in the waterlogging feature corresponding to the historical social media data at all time steps in the first grid, represents the mean value of the water level in the historical water level data corresponding to the water level monitor at all time steps in the first grid, is the correlation coefficient, Indicates the calculation of covariance, Indicates the calculation of variance.

[0148] Optionally, in the parameter calibration module 304, the urban waterlogging water level corresponding to the k-th rainfall in the first urban waterlogging feature set and the urban waterlogging water level corresponding to the k-th rainfall in the second urban waterlogging feature set are divided into a training set and a validation set according to a set ratio, and the following objective function is used to determine the urban waterlogging model parameters corresponding to the rainfall indicated by the k-th rainfall:

[0149] in, is the objective function, is the maximum time step of the k-th rainfall, is the total number of flooding water levels belonging to the second feature set in the training set, is the total number of waterlogging water levels in the training set that belong to the first waterlogging feature set, represents the sum of squared errors, For the The first feature of the second waterlogging feature set in the training set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level, For the The first waterlogging feature set in the training set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level.

[0150] Figure 4 FIG2 shows a schematic diagram of the structure of an urban flood emergency plan selection device provided by an exemplary embodiment of the present disclosure. Figure 4 The urban waterlogging emergency plan selection device 400 includes: an acquisition module 401, a waterlogging simulation module 402, a scoring module 403 and a waterlogging management module 404.

[0151] The acquisition module 401 is used to acquire an emergency plan set, wherein the emergency plan set includes multiple emergency plans for managing urban waterlogging; The waterlogging simulation module 402 is used to obtain an urban waterlogging simulation result corresponding to each emergency plan based on an urban waterlogging model, wherein the urban waterlogging model is obtained based on the urban waterlogging simulation method according to any one of claims 1 to 4; The scoring module 403 is configured to calculate a total score of each emergency plan based on the urban flooding simulation result corresponding to each emergency plan; The waterlogging management module 404 is used to adopt the emergency plan with the highest total score to manage urban waterlogging.

[0152] Optionally, in the scoring module 403, the score of each emergency plan is calculated based on an indicator set, and the indicator set includes the maximum water depth reduction value of each waterlogging point, the maximum waterlogging area reduction value of each waterlogging point, the average maximum water depth reduction value of all waterlogging points, the sum of the large waterlogging area reduction values ​​of all waterlogging points, the shortening value of the waterlogging receding time of each waterlogging point, the average of the shortening values ​​of the waterlogging receding time of all waterlogging points, and the cost of implementing the plan.

[0153] It should be noted that when the urban waterlogging simulation device provided in the above embodiment performs urban waterlogging simulation, or when the urban waterlogging emergency plan selection device provided in the above embodiment performs urban waterlogging emergency plan selection, the division of the above functional modules is only used as an example to illustrate. In actual applications, the above functional distribution can be completed by different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the urban waterlogging simulation device provided in the above embodiment and the urban waterlogging simulation method embodiment belong to the same concept, and the urban waterlogging emergency plan selection device provided in the above embodiment and the urban waterlogging emergency plan selection method embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0154] The division of modules in the embodiments of the present disclosure is illustrative and represents only a logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of the present disclosure may be integrated into a single processor, exist physically as separate modules, or be integrated into a single module. The integrated modules may be implemented in either hardware or software functional modules.

[0155] If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a terminal device (which can be a personal computer, mobile phone, or communication device, etc.) or a processor to execute all or part of the steps of the method of each embodiment of the present disclosure. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program code.

[0156] Figure 5Schematic diagram of the structure of the computer device provided by the embodiment of the present disclosure. Figure 5 As shown, the computer device 500 includes a processor 501 and a memory 502 .

[0157] Processor 501 may include one or more processing cores, such as a quad-core processor or an octa-core processor. Processor 501 may be implemented in hardware using at least one of the following: a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), or a PLA (Programmable Logic Array). Processor 501 may also include a main processor and a coprocessor. The main processor is used to process data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 501 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing content displayed on the display screen. In some embodiments, processor 501 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0158] The memory 502 may include one or more computer-readable storage media, which may be non-transitory. The memory 502 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 502 is used to store at least one instruction, which is executed by the processor 501 to implement the urban waterlogging simulation method and urban waterlogging emergency plan selection method provided in the embodiments of the present disclosure.

[0159] Those skilled in the art will understand that Figure 5 The structure shown in the figure does not constitute a limitation on the computer device 500, and the computer device 500 may include more or fewer components than shown in the figure, or combine some components, or adopt a different component arrangement.

[0160] The embodiments of the present disclosure also provide a non-temporary computer-readable storage medium. When the instructions in the storage medium are executed by the processor of a computer device, the computer device can execute the urban waterlogging simulation method and urban waterlogging emergency plan selection method provided in the embodiments of the present disclosure.

[0161] The embodiments of the present disclosure further provide a computer program product, including a computer program / instruction, which, when executed by a processor, implements the urban waterlogging simulation method and the urban waterlogging emergency plan selection method provided in the embodiments of the present disclosure.

[0162] The above description is merely an optional embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the scope of protection of the present disclosure.

Claims

1. A method for simulating urban waterlogging, characterized in that: The method comprises: Obtain historical social media data related to urban waterlogging; Obtaining a first waterlogging feature set based on historical water level data collected by a water level monitor; acquiring a second waterlogging feature set based on the historical social media data and the historical water level data; calibrating parameters of an urban waterlogging model under different rainfall amounts using the second waterlogging feature set and the first waterlogging feature set to obtain a parameter correspondence, wherein the parameter correspondence includes parameters of the urban waterlogging model corresponding to a plurality of rainfall amounts; obtaining target rainfall data, real-time social media data, and real-time water level data, wherein the target rainfall data is determined based on the real-time rainfall and the forecast rainfall; Determining target parameters of the urban waterlogging model based on the target rainfall data and the parameter correspondence; The target parameters are substituted into the urban waterlogging model, and based on the waterlogging characteristics corresponding to the real-time social media data and the waterlogging characteristics corresponding to the real-time water level data, the urban waterlogging model is used for deduction to obtain a first urban waterlogging simulation result.

2. The urban flooding simulation method according to claim 1, characterized in that: Social media data is multimodal data, including text data, image data, and video data. The obtaining of a second waterlogging feature set based on the historical social media data includes: Using a word segmentation tool to segment the text data in the historical social media data and then extracting waterlogging features; For the video data in the historical social media data, the video data is converted into single-frame images, and then the waterlogging features are extracted using a pre-trained YOLO v8 model; For the image data in the historical social media data, a pre-trained YOLO v8 model is used to extract waterlogging features; Using water level errors and correlation coefficients to filter waterlogging features extracted from the historical social media data to obtain a second waterlogging feature set; The second waterlogging feature set includes multiple waterlogging water levels, and the location and time corresponding to each waterlogging water level.

3. The urban waterlogging simulation method according to claim 2, characterized in that: The urban flooded area is divided into multiple grids, and the water level error and the correlation coefficient are calculated with the grid as the smallest unit. The water level error of the first grid is calculated using the following formula, where the first grid is any one of the multiple grids: in, Indicates the water level error, For the The mean value of the water level in the waterlogging feature corresponding to the historical social media data in the first grid of the time step, For the The mean value of the waterlogging water level in the historical water level data corresponding to the water level monitor in the first grid in the time step; The correlation coefficient of the first grid is calculated using the following formula: in, represents the mean value of the water level in the waterlogging feature corresponding to the historical social media data at all time steps in the first grid, represents the mean value of the waterlogging water level in the historical water level data corresponding to the water level monitor at all time steps in the first grid, is the correlation coefficient, Indicates the calculation of covariance, Indicates the calculation of variance.

4. The urban flooding simulation method according to any one of claims 1 to 3, characterized in that: The water level of the urban flooding corresponding to the kth rainfall in the first waterlogging feature set and the water level of the urban flooding corresponding to the kth rainfall in the second waterlogging feature set are divided into a training set and a validation set according to a set ratio. The following objective function is used to determine the parameters of the urban waterlogging model corresponding to the rainfall indicated by the kth rainfall: in, is the objective function, is the maximum time step of the k-th rainfall, is the total number of flood water levels belonging to the second feature set in the training set, is the total number of waterlogging water levels in the training set that belong to the first waterlogging feature set, represents the sum of squared errors, For the The first feature of the second waterlogging feature set in the training set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level, For the The first waterlogging feature set in the training set at the time step The flood water level, The urban flooding model predicts The corresponding flood water level.

5. A method for selecting an emergency plan for urban flooding, characterized in that: The method comprises: Acquire an emergency plan set, wherein the emergency plan set includes a plurality of emergency plans for managing urban waterlogging; Obtaining an urban flooding simulation result corresponding to each emergency plan based on an urban flooding model, wherein the urban flooding model is obtained based on the urban flooding simulation method according to any one of claims 1 to 4; Calculating a total score for each emergency plan based on the urban flooding simulation results corresponding to each emergency plan; The emergency plan with the highest overall score is adopted for urban waterlogging management.

6. The method according to claim 5, characterized in that The score of each emergency plan is calculated based on an indicator set, which includes the maximum waterlogging depth reduction value of each waterlogging point, the maximum waterlogging area reduction value of each waterlogging point, the average maximum water depth reduction value of all waterlogging points, the sum of the large waterlogging area reduction values ​​of all waterlogging points, the shortening value of the waterlogging receding time value of each waterlogging point, the average of the shortening value of the waterlogging receding time value of all waterlogging points, and the cost of implementing the plan.

7. An urban flooding simulation device, characterized in that: The device comprises: The first acquisition module is used to obtain historical social media data related to urban waterlogging; A second acquisition module is configured to acquire a first waterlogging feature set based on historical water level data collected by the water level monitor; a third acquisition module, configured to acquire a second waterlogging feature set based on the historical social media data and the historical water level data; a parameter calibration module, configured to calibrate parameters of an urban waterlogging model under different rainfall amounts using the second waterlogging feature set and the first waterlogging feature set, to obtain a parameter correspondence, wherein the parameter correspondence includes parameters of the urban waterlogging model corresponding to a plurality of rainfall amounts; a fourth acquisition module, configured to acquire target rainfall data, real-time social media data, and real-time water level data, wherein the target rainfall data is determined based on the real-time rainfall and the forecast rainfall; a target parameter determination module, configured to determine the target parameters of the urban waterlogging model based on the target rainfall data and the parameter correspondence; The urban flooding simulation module is used to substitute the target parameters into the urban flooding model, and based on the urban flooding characteristics corresponding to the real-time social media data and the urban flooding characteristics corresponding to the real-time water level data, use the urban flooding model to perform deduction to obtain a first urban flooding simulation result.

8. A device for selecting an emergency plan for urban flooding, characterized in that: The device comprises: An acquisition module, configured to acquire a set of emergency plans, wherein the set of emergency plans includes a plurality of emergency plans for managing urban waterlogging; A waterlogging simulation module, configured to obtain an urban waterlogging simulation result corresponding to each emergency plan based on an urban waterlogging model, wherein the urban waterlogging model is obtained based on the urban waterlogging simulation method according to any one of claims 1 to 4; A scoring module, configured to calculate a total score of each emergency plan based on the urban flooding simulation result corresponding to each emergency plan; The waterlogging management module is used to manage urban waterlogging using the emergency plan with the highest total score.

9. A computer device, characterized in that: The computer device includes: a memory and a processor, wherein at least one computer program is stored in the memory, and the at least one computer program is loaded and executed by the processor to implement the method according to any one of claims 1 to 4 or any one of claims 5 to 6.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores at least one computer program, and the at least one computer program is loaded and executed by a processor to implement the method according to any one of claims 1 to 4 or any one of claims 5 to 6.

Images