Flood disaster early warning method, equipment and medium

Abstract

The invention discloses a flood disaster early warning method and device, a medium, and the technical field of Internet of Things and artificial intelligence, and the method comprises the steps that a cloud platform collects key hydrological elements corresponding to a to-be-predicted region through a monitoring device, and obtains multivariate data corresponding to the key hydrological elements; key features in the historical multivariate data are extracted through an analysis algorithm, the historical multivariate data are divided into a training set and a verification set according to the key features, and a hydrological forecasting model constructed through multiple neural networks is trained based on the training set; acquiring real-time weather forecast data, and outputting forecast flow and water level through the trained hydrological forecast model; according to the forecast flow and the water level, the flood occurrence probability is obtained through the pre-trained risk assessment model, and corresponding early warning information is automatically generated based on the flood occurrence probability. Through the hydrological forecasting model, the flow of the key water body can be quickly predicted, and early warning information can be generated in a short time.

Description

Technical Field

[0001] The present application relates to the field of Internet of Things and artificial intelligence technology, and specifically to a flood disaster early warning method, equipment and medium. Background Art

[0002] Due to the complex and diverse topography and landforms in mountainous areas, as well as the numerous rivers and lakes, flood sources are widespread. In the rainy and typhoon seasons, regional heavy rainfall is prone to cause flood disasters, especially in mountainous areas with fragile geological conditions, where mountain torrents and geological disasters occur frequently, further increasing the risk of flood disasters.

[0003] With the advancement of science and technology, especially the rapid development of remote sensing technology, geographic information system (GIS), big data and artificial intelligence, the application of flood disaster warning is becoming more and more extensive. The traditional method of flood disaster warning is usually implemented using a prediction model based on static assumptions. However, the natural environment in mountainous areas is complex. For example, factors such as steep mountain terrain, land use changes, and uneven precipitation intensity will lead to differences between the actual environment and model assumptions. Models created based on static assumptions are difficult to adapt to complex and changeable natural environments and extreme weather events. Summary of the invention

[0004] In order to solve the above problems, this application proposes a flood disaster early warning method, including:

[0005] The cloud platform collects key hydrological elements corresponding to the predicted area through monitoring equipment to obtain multivariate data corresponding to the key hydrological elements;

[0006] Extracting key features from historical multivariate data through an analysis algorithm, and dividing the historical multivariate data into a training set and a validation set according to the key features, so as to train a hydrological forecast model constructed by multiple neural networks based on the training set;

[0007] Acquire real-time weather forecast data, and output the forecast flow and water level corresponding to the key water bodies in the area to be predicted through the trained hydrological forecast model;

[0008] According to the forecast flow and the water level, the probability of flood occurrence in the area to be predicted is obtained through a pre-trained risk assessment model, and corresponding warning information is automatically generated based on the probability of flood occurrence.

[0009] In one implementation of the present application, the monitoring equipment includes intelligent buoys, drones, satellite remote sensing and ground sensors; key water bodies in the area to be predicted are determined, and the key water bodies include rivers, lakes, reservoirs, and groundwater; based on various types of monitoring equipment set up in the key water bodies, a multivariate perception network is formed, and key hydrological elements in the key water bodies are collected at a regular interval through the multivariate perception network to obtain multivariate data corresponding to the key hydrological elements.

[0010] In one implementation of the present application, encrypted historical multivariate data is obtained through a secure transmission protocol, and the historical multivariate data is preprocessed; time series patterns and spatial distribution patterns in the preprocessed historical multivariate data are extracted through a time series analysis algorithm and a spatial analysis algorithm.

[0011] In one implementation of the present application, based on the spatial distribution pattern, the hydrological characteristics corresponding to different locations in the historical multivariate data are determined; based on the time series pattern, the time series in the historical multivariate data is determined; based on the time series and the hydrological characteristics, the multivariate data is divided into a training set and a validation set.

[0012] In one implementation of the present application, the initial flow, rainfall, and evaporation in the training set are taken as input, and the corresponding time series and spatial forecast flow are taken as output to train the hydrological forecast model; and the training results of the hydrological forecast model are verified based on the verification set through a time series cross-validation method.

[0013] In one implementation of the present application, historical hydrological data corresponding to the occurrence of floods and current hydrological data are obtained; based on the historical hydrological data and the current hydrological data, the warning thresholds of different warning levels in the risk assessment model are adjusted.

[0014] In one implementation of the present application, based on the geospatial information data in the multivariate data, a digital twin model of the water conservancy project in the predicted area is constructed through three-dimensional modeling technology; based on the predicted flow, dynamic simulation is performed through the digital twin model.

[0015] In one implementation of the present application, through a flood forecast analysis model, based on the forecast flow, the water level, flow velocity and number of confluences in the upstream of the key river in the predicted area are calculated to determine the inflow flow of the water conservancy project; through a flood discharge scheduling model, based on the inflow flow, the changes in the flood discharge curve of the water conservancy project are scheduled and analyzed to obtain the flood discharge situation in the predicted area; the downstream inundation analysis model is called to predict the inundation range and duration of the downstream of the key river based on the flood discharge situation; through the digital twin model, multiple user images are set in the preset personnel transfer route, and based on the inundation range and inundation duration, the downstream inundation process and the transfer of personnel during the inundation process are dynamically simulated.

[0016] On the other hand, the present application also proposes a flood disaster warning device, comprising:

[0017] at least one processor; and,

[0018] a memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a flood disaster warning method as described in the above example.

[0020] On the other hand, the present application also proposes a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured as: a flood disaster warning method described in the above example.

[0021] The present application proposes a flood disaster early warning method which can bring the following beneficial effects:

[0022] By integrating real-time weather forecast data and hydrological forecast models, this method can quickly predict the flow of key water bodies and generate warning information in a short time, improving the timeliness of warnings. This efficient data processing and prediction capability ensures the rapid transmission of warning information, providing the public and relevant agencies with sufficient time to take countermeasures. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0024] Figure 1 A schematic diagram of a flood disaster warning method in an embodiment of the present application;

[0025] Figure 2This is a schematic diagram of a flood disaster warning device in an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0027] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0028] It should be noted that the Internet of Things technology is a huge network formed by combining various information sensing devices, such as sensors, radio frequency identification (RFID) technology, global positioning system, infrared sensors, laser scanners, gas sensors and other devices and technologies, to collect any objects or processes that need to be monitored, connected, and interacted in real time, and collect various required information such as sound, light, heat, electricity, mechanics, chemistry, biology, location, etc. Its purpose is to achieve the connection between things and things, things and people, and all objects and the network, so as to facilitate identification, management and control.

[0029] like Figure 1 As shown, the embodiment of the present application provides a flood disaster early warning method, including:

[0030] S101: The cloud platform collects key hydrological elements corresponding to the predicted area through monitoring equipment to obtain multivariate data corresponding to the key hydrological elements.

[0031] Specifically, a cloud platform is constructed to store and process the collected multivariate data, such as meteorological data, remote sensing image data, socio-economic data, business application data, real-time monitoring data, GIS spatial data, etc. The cloud platform collects key hydrological elements corresponding to the predicted area through monitoring equipment, obtains multivariate data corresponding to the key hydrological elements, encrypts the collected multivariate data, and transmits it using a high-speed communication network (such as a fourth-generation mobile communication network, a fifth-generation mobile communication network, etc.).

[0032] Among them, monitoring equipment includes smart buoys, drones, satellite remote sensing and ground sensors.

[0033] It should be noted that a smart buoy is a buoy equipped with sensors and data collection devices, which are used to monitor various environmental parameters of lakes, such as water temperature, water depth, water quality (such as pH value, dissolved oxygen content), etc. A drone is an aircraft that flies by remote control or preset program, and can carry various sensors to collect air and ground data. Satellite remote sensing uses data obtained from satellites to monitor various parameters of the earth's surface, including the type and location of water bodies in the road. Ground sensors are used to monitor and record surface and underground environmental parameters in real time, such as changes in soil moisture in farmland. A Geographic Information System (GIS) is a computer system used to input, store, query, analyze and display geographic data.

[0034] Furthermore, the key water bodies and their corresponding locations in the area to be predicted are determined, wherein the key water bodies include rivers, lakes, reservoirs, groundwater, etc. Based on various types of monitoring equipment set up in the key water bodies, a multivariate sensing network is formed, and the multivariate sensing network covers all the key water bodies in the area to be predicted. The key hydrological elements in the key water bodies are collected regularly through the multivariate sensing network to obtain the multivariate data corresponding to the key hydrological elements.

[0035] It should be noted that the multi-sensing network is a sensing system covering key water bodies, built using a variety of different sensors and technical means, such as smart buoys, drones, satellite remote sensing and ground sensors. Through the multi-sensing network, all-weather, multi-scale and high-frequency data collection of key water bodies in the forecast area is achieved, realizing comprehensive, accurate and efficient monitoring and management of hydrological elements.

[0036] S102: extracting key features from historical multivariate data through an analysis algorithm, and dividing the historical multivariate data into a training set and a validation set according to the key features, so as to train a hydrological forecasting model constructed by multiple neural networks based on the training set.

[0037] Specifically, the cloud platform extracts historical multivariate data, wherein the historical multivariate data is the multivariate data corresponding to the area to be predicted at the historical collection time, including multivariate data of different spaces corresponding to different seasons and different time periods of each day.

[0038] Furthermore, the time series analysis algorithm and spatial analysis algorithm are used to extract the time series characteristics and spatial distribution characteristics of historical multivariate data. Based on the spatial distribution characteristics, the hydrological characteristics corresponding to different locations in the historical multivariate data are determined. Based on the time series characteristics, the time series in the historical multivariate data is determined. The time series pattern and spatial distribution pattern are used as key features, and the historical multivariate data are divided into training and validation sets according to the key features.

[0039] More specifically, in the process of extracting historical multivariate data from the cloud platform, encrypted historical multivariate data are obtained through a secure transmission protocol, and the historical multivariate data are preprocessed, including data fusion using advanced data fusion algorithms, data cleaning, and standardization.

[0040] In an embodiment of the present application, the historical multivariate data may be arranged in chronological order and divided according to geographical regions, with the first eighty percent of the data being divided into a training set and the remaining data being divided into a validation set.

[0041] It should be noted that the time series analysis algorithm is an algorithm that analyzes a series of data points listed in chronological order. These data points are usually sequences at continuous equidistant time points. The time series analysis algorithm can extract data features in different time series in historical multivariate data, such as key hydrological elements in different seasons such as summer and winter. The spatial analysis algorithm analyzes geographic spatial data and its corresponding theories, methods and techniques to obtain the relationship between geographic elements, thereby revealing the inherent laws and mechanisms of geographic features and processes, and realizing the cognition, interpretation, prediction and regulation of geographic spatial information. For example, in different geographical locations of the model to be predicted, there are differences in factors such as terrain, vegetation cover, and soil type. The time series analysis algorithm mainly focuses on the law of data change and prediction in the time dimension, while the spatial analysis algorithm focuses on the differences in geographic spatial data.

[0042] The hydrological forecast model is constructed by a variety of neural networks, including long short-term memory networks and deep reinforcement learning algorithm convolutional neural networks. The network structure of the combined multiple neural networks includes the input layer, convolution layer, pooling layer of the convolutional neural network, and the LSTM layer and output layer of the long short-term memory network.

[0043] Among them, the input layer, convolution layer, and pooling layer are the feature extraction parts after combining multiple neural networks. The input layer receives input data, namely time series data such as initial flow, rainfall, and evaporation; the convolution layer performs convolution operations on the input data to extract local features of the input data in time and space, and uses activation functions (such as ReLU (Rectified Linear Unit) functions, etc.) to introduce nonlinearity to improve the expression ability of the model; the pooling layer reduces the feature dimension through pooling operations to reduce the amount of calculation and reduce the risk of overfitting. The LSTM layer is a time series processing layer that receives the features extracted by the feature extraction part as input and further processes these features to capture the long-term dependencies in the time series. The LSTM layer controls the flow of information through its internal forget gate, input gate, and output gate, thereby effectively solving the long-term dependency problem in RNN. The LSTM layer processes the input of each time step in turn and maintains a hidden state to capture the information of the previous time step. The output layer maps the output of the LSTM layer to the final prediction result for output.

[0044] The initial flow, rainfall and evaporation in the training set are taken as input, and the corresponding time series and spatial forecast flow are taken as output to train the hydrological forecast model. The training results of the hydrological forecast model are verified based on the validation set through the time series cross-validation method.

[0045] It should be noted that the time series cross-validation method divides the entire time series data set into multiple continuous time periods. In each window, the previous data is used as the training set, and the following data is used as the validation set or test set. As the window slides, the model will continuously be trained with new training data and evaluated with new validation sets or test sets.

[0046] S103: Acquire real-time weather forecast data, and output the forecast flow and water level corresponding to the key water bodies in the area to be predicted through the trained hydrological forecast model.

[0047] Specifically, weather forecast data from the meteorological station is obtained in real time. In an embodiment of the present application, the data source of the weather forecast data may also be satellite images or ground observation stations, wherein the weather forecast data includes information such as temperature, humidity, wind speed, wind direction, expected rainfall, and expected evaporation.

[0048] Furthermore, the expected rainfall and expected evaporation in the weather forecast data, as well as the initial flow of key water bodies, are input into the trained hydrological forecast model. The input layer in the hydrological forecast model receives the input data and sends it to the convolution layer in the hydrological forecast model. The input data is convolved through multiple convolution kernels (also called filters) in the convolution layer to extract the temporal and spatial features corresponding to the expected rainfall, expected evaporation and initial flow, and the extracted features are reduced in dimension through the pooling layer to reduce the amount of calculation.

[0049] The features output by the pooling layer are fully connected through the fully connected layer to obtain the output results of the hydrological forecast model. Finally, the output layer outputs the output results to obtain the forecast flow and water level corresponding to the key water body.

[0050] It should be noted that by using real-time weather forecast data, the hydrological forecast model can quickly respond to the latest meteorological changes and output the latest forecast results in a timely manner to make flood prevention preparations and reduce the losses caused by flood disasters.

[0051] S104: According to the predicted flow and water level, the probability of flood occurrence in the area to be predicted is obtained through a pre-trained risk assessment model, and corresponding warning information is automatically generated based on the probability of flood occurrence.

[0052] Specifically, based on the forecast flow, a pre-trained risk assessment model is used to evaluate the probability of flood disasters under given flow conditions. One or more thresholds for the probability of flood occurrence are preset, such as 0.5, to determine whether an early warning is needed. If the probability of flood occurrence exceeds the preset threshold, an early warning information is issued and released to the public and relevant agencies through various channels (such as mobile phone text messages, social media, TV broadcasts, official websites, etc.) so that they can take timely response measures to reduce the losses caused by flood disasters.

[0053] It should be noted that the risk assessment model is a pre-trained model, and the model structure is data input layer, data processing layer, risk assessment algorithm layer, and result output layer. Among them, the risk assessment algorithm layer combines multiple indicators for comprehensive assessment, such as the analytic hierarchy process (AHP), fuzzy comprehensive evaluation method, etc., and integrates multiple indicators into a comprehensive risk assessment result by setting the weights and scoring standards of different indicators.

[0054] Before obtaining the probability of flood occurrence, the historical hydrological data corresponding to the flood occurrence and the current hydrological data are obtained, and the warning thresholds of different warning levels in the risk assessment model are adjusted according to the historical hydrological data and the current hydrological data.

[0055] Furthermore, based on the geospatial information data in the multivariate data, a digital twin model of the water conservancy project in the predicted area is constructed through three-dimensional modeling technology, and dynamic simulation is carried out through the digital twin model based on the predicted flow.

[0056] It should be noted that digital twins make full use of data such as physical models, sensor updates, and operation history, integrate multi-disciplinary, multi-physical, multi-scale, and multi-probability simulation processes, and complete mapping in virtual space to reflect the entire life cycle of the corresponding physical equipment. Digital twins are a concept that transcends reality and can be viewed as a digital mapping system of one or more important, interdependent equipment systems.

[0057] Specifically, through the flood forecast analysis model, based on the predicted flow, the water level, flow velocity and number of confluences in the upstream of the key rivers in the predicted area are calculated to determine the inflow flow of the water conservancy project. Through the flood discharge scheduling model, based on the inflow flow, the changes in the flood discharge curve of the water conservancy project are scheduled and analyzed to obtain the flood discharge situation in the predicted area. The downstream inundation analysis model is called to predict the inundation range and duration of the downstream of the key rivers based on the flood discharge situation. Through the digital twin model, multiple user images are set in the preset personnel transfer route, and the personnel transfer is simulated based on the inundation range and duration.

[0058] Furthermore, simulation results are obtained. Based on the simulation results, the best plan is automatically matched among the preset multiple hydrological disaster emergency plans, and the plan can also be dynamically adjusted and optimized. Based on the deep learning intelligent scheduling algorithm, the emergency response strategies under different disaster scenarios are simulated, and the resource allocation, action path, time window, etc. of the plan are continuously optimized to improve the practicality and flexibility of the plan. And the action plans or plans to resist different levels of floods are optimally determined, including personnel and material communication guarantees, scheduling plans, residents' evacuation routes, etc., clarify the scheduling mechanism and information flow, and organize related implementation.

[0059] For example, if a flood with a frequency of 50 years occurs in the basin, the analysis results can be viewed based on the historical plan. In the plan optimization module, according to the gate opening in the engineering scheduling model, the downstream inundation analysis results are analyzed and compared at the optimal balance point. The downstream flood inundation analysis results are compared with the historical plan flood inundation analysis results to select the result with the least impact on the downstream.

[0060] It should be noted that the hydrological forecast model has significantly improved the intelligence, automation and efficiency of hydrological flood disaster management, effectively shortened the information processing cycle of hydrological flood disasters, improved the accuracy of forecasts and warnings, enhanced the timeliness and effectiveness of disaster response, and reduced the risks and impacts of natural disasters such as floods. Through digital, networked and intelligent means, it provides strong technical support for the sustainable use of water resources and ecological environmental protection, and is of great value in promoting the development of smart water conservancy.

[0061] By integrating real-time weather forecast data and hydrological forecast models, this method can quickly predict the flow of key water bodies and generate warning information in a short time, further improving the timeliness of the warning. This efficient data processing and forecasting capability ensures the rapid transmission of warning information and provides the public and relevant agencies with sufficient time to take countermeasures.

[0062] like Figure 2 As shown, the embodiment of the present application also proposes a flood disaster warning device, including:

[0063] at least one processor; and,

[0064] a memory communicatively connected to the at least one processor; wherein,

[0065] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a flood disaster warning method as described in any of the above embodiments.

[0066] An embodiment of the present application further provides a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured to be: a flood disaster warning method as described in any of the above embodiments.

[0067] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0068] The devices and media provided in the embodiments of the present application correspond one-to-one to the methods. Therefore, the devices and media also have similar beneficial technical effects as the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.

[0069] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0070] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0071] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0072] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0073] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0074] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0075] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0076] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0077] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A flood disaster early warning method, characterized in that: include: The cloud platform collects key hydrological elements corresponding to the predicted area through monitoring equipment to obtain multivariate data corresponding to the key hydrological elements; Extracting key features from historical multivariate data through an analysis algorithm, and dividing the historical multivariate data into a training set and a validation set according to the key features, so as to train a hydrological forecast model constructed by multiple neural networks based on the training set; Acquire real-time weather forecast data, and output the forecast flow and water level corresponding to the key water bodies in the area to be predicted through the trained hydrological forecast model; According to the forecast flow and the water level, the probability of flood occurrence in the area to be predicted is obtained through a pre-trained risk assessment model, and corresponding warning information is automatically generated based on the probability of flood occurrence.

2. A flood disaster early warning method according to claim 1, characterized in that: The monitoring equipment includes smart buoys, drones, satellite remote sensing and ground sensors; The key hydrological elements corresponding to the predicted area are collected by monitoring equipment to obtain multivariate data corresponding to the key hydrological elements, specifically including: Determine the key water bodies in the area to be predicted, wherein the key water bodies include rivers, lakes, reservoirs, and groundwater; Based on various types of monitoring equipment installed in the key water body, a multivariate sensing network is formed. The key hydrological elements in the key water body are collected regularly through the multivariate sensing network to obtain multivariate data corresponding to the key hydrological elements.

3. A flood disaster early warning method according to claim 1, characterized in that: The key features extracted from the historical multivariate data by the analysis algorithm specifically include: Acquire encrypted historical multivariate data through a secure transmission protocol, and pre-process the historical multivariate data; Through time series analysis algorithms and spatial analysis algorithms, time series patterns and spatial distribution patterns in preprocessed historical multivariate data are extracted.

4. A flood disaster early warning method according to claim 3, characterized in that: The dividing the historical multivariate data into a training set and a validation set according to the key features specifically includes: Based on the spatial distribution pattern, determining hydrological characteristics corresponding to different locations in the historical multivariate data; Based on the time series pattern, determining a time series in the historical multivariate data; Based on the time series and the hydrological characteristics, the multivariate data is divided into a training set and a validation set.

5. A flood disaster early warning method according to claim 4, characterized in that: The training of the hydrological forecast model constructed by multiple neural networks based on the training set specifically includes: The hydrological forecast model is trained by taking the initial flow, rainfall, and evaporation in the training set as input and the corresponding time series and spatial forecast flow as output; And the training results of the hydrological forecast model are verified based on the verification set through a time series cross-validation method.

6. A flood disaster early warning method according to claim 1, characterized in that: Before obtaining the flood occurrence probability of the area to be predicted through the pre-trained risk assessment model, the method further includes: Obtain historical hydrological data corresponding to floods and current hydrological data; According to the historical hydrological data and the current hydrological data, the warning thresholds of different warning levels in the risk assessment model are adjusted.

7. A flood disaster early warning method according to claim 1, characterized in that: The method further comprises: Based on the geospatial information data in the multivariate data, a digital twin model of the water conservancy project in the area to be predicted is constructed by three-dimensional modeling technology; Based on the forecasted flow, dynamic simulation is performed through the digital twin model.

8. A flood disaster early warning method according to claim 7, characterized in that: The dynamic simulation based on the forecast flow is performed by the digital twin model, specifically including: By using a flood forecast analysis model, based on the forecast flow, the water level, flow velocity and number of confluences in the upstream of the key rivers in the area to be predicted are calculated to determine the inflow flow of the water conservancy project; Through the flood discharge scheduling model, based on the inflow flow, the change of the flood discharge curve of the water conservancy project is dispatched and analyzed to obtain the flood discharge situation of the area to be predicted; Invoking a downstream flooding analysis model to predict the flooding range and duration of the downstream of the key river based on the flood discharge situation; Through the digital twin model, multiple user images are set in the preset personnel transfer route, and based on the flooding range and flooding duration, the downstream flooding process and the transfer of personnel during the flooding process are dynamically simulated.

9. A flood disaster warning device, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a flood disaster warning method as described in any one of claims 1 to 8.

10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: The computer executable instructions are configured as: a flood disaster early warning method as described in any one of claims 1 to 8.

Images