Mountain torrent disaster intelligent monitoring and early warning system and method

Abstract

The invention discloses a mountain torrent disaster intelligent monitoring and early warning system and method, and relates to the field of disaster monitoring and early warning. The system comprises a data acquisition module and an edge calculation module, the data acquisition module is used for acquiring rainfall and soil humidity of a target area at the current moment; the edge calculation module is used for simulating a process of collecting runoff in a target area to a section by adopting a water level prediction model based on the rainfall and the soil humidity at the current moment, and calculating the section water level at the current moment; based on the rainfall, the soil humidity, the section water level and the environmental characteristics at the current moment, adopting a mountain torrent risk prediction model to predict the mountain torrent occurrence probability of the target area; and sending an early warning signal according to the section water level and the mountain torrent occurrence probability at the current moment. According to the invention, the response speed and accuracy of early warning can be improved.

Description

Technical Field

[0001] This application relates to the field of disaster monitoring and early warning, and in particular to an intelligent monitoring and early warning system and method for flash flood disasters. Background Technology

[0002] Flash floods are one of the natural disasters that pose a serious threat to the lives and property of people in hilly areas. Traditional flash flood monitoring and early warning systems typically rely on automatic monitoring stations to collect data and report it to monitoring and early warning systems at various levels for processing and analysis before finally issuing early warning information. However, current technologies have the following two problems.

[0003] First, it suffers from poor real-time performance: relying on centralized data processing results in a slow response time when disasters occur, making it difficult to achieve rapid early warning; second, it suffers from delayed early warning: it is difficult to conduct timely risk assessments and accurate early warnings before disasters occur, leading to tension and disorder in the response after disasters occur. Summary of the Invention

[0004] The purpose of this application is to provide an intelligent monitoring and early warning system and method for flash flood disasters, which can improve the response speed and accuracy of early warnings.

[0005] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides an intelligent monitoring and early warning system for flash floods, comprising: a data acquisition module and an edge computing module; the data acquisition module is connected to the edge computing module; The data acquisition module is used for: Collect the current rainfall and soil moisture in the target area; The edge computing module is used for: Based on the current rainfall and soil moisture, a water level prediction model is used to simulate the process of runoff converging at the cross-section in the target area, and the cross-sectional water level at the current moment is calculated. The water level prediction model is obtained by calibrating the runoff generation and confluence model using historical data. The historical data includes: measured water level, rainfall and soil moisture in the target area at historical moments. Based on the current rainfall, soil moisture, cross-sectional water level, and environmental characteristics, a flash flood risk prediction model is used to predict the probability of flash floods in the target area. The flash flood risk prediction model is obtained by training a machine learning model using training data. The training data includes: measured water level, rainfall, soil moisture, environmental characteristics, and flash flood occurrence records of the target area at historical times. A warning signal is issued based on the current cross-sectional water level and the probability of flash floods.

[0006] In one embodiment, the data acquisition module is further configured to acquire historical data of the target area; The edge computing module includes: a runoff generation and confluence model calibration unit, a cross-sectional water level prediction unit, a machine learning model training unit, a disaster prediction unit, and a disaster early warning unit; The flow generation and confluence model calibration unit is used for: A runoff generation model and a runoff collection model are constructed; the runoff generation model is used to simulate the process of rainfall being converted into surface runoff; the runoff collection model is used to simulate the flow process of runoff in the river channel. The production flow model and the confluence flow model are integrated to obtain the production-confluence flow model; The rainfall and soil moisture from the historical data are input into the runoff generation and confluence model. The runoff generation and confluence model is calibrated based on the comparison between the cross-sectional water level output by the runoff generation and confluence model and the measured water level in the historical data to determine the model parameters. The runoff generation and confluence model with the determined model parameters is used as the water level prediction model. The model parameters include: parameters in the runoff generation model and parameters in the confluence model; the parameters in the runoff generation model include at least: infiltration rate; the parameters in the confluence model include at least: roughness. The cross-sectional water level prediction unit is used for: Input the current rainfall and soil moisture into the water level prediction model to obtain the current cross-sectional water level. The machine learning model training unit is used for: The training data is input into the machine learning model for training, and the trained machine learning model is used as a flash flood risk prediction model. The disaster prediction unit is used for: Input the current rainfall, soil moisture, cross-sectional water level and environmental characteristics into the flash flood risk prediction model to predict the probability of flash floods in the target area; The disaster early warning unit is used for: Calculate the difference between the current cross-sectional water level and the warning water level; A warning signal is issued based on the difference and the probability of flash floods.

[0007] In one embodiment, the intelligent monitoring and early warning system for flash floods further includes: a wireless communication module; The wireless communication module is connected to the edge computing module; the wireless communication module is used to remotely transmit the warning signal.

[0008] In one embodiment, the intelligent monitoring and early warning system for flash floods further includes: a local propagation module; The local propagation module is connected to the edge computing module; the local propagation module is used to propagate the warning signal locally.

[0009] In one embodiment, the intelligent monitoring and early warning system for flash floods further includes: a power supply module; The power supply module is connected to the data acquisition module, the edge computing module, and the wireless communication module, respectively.

[0010] In one embodiment, the intelligent monitoring and early warning system for flash floods further includes: an intelligent power management module; The intelligent power management module is connected to the power supply module; the intelligent power management module is used to monitor the power level of the power supply module and switch the operating mode according to the power level; the operating mode includes: normal mode and low power mode; The normal mode includes: when the power level is greater than or equal to the set power level, the data acquisition module, the edge computing module, and the wireless communication module maintain full-function operation; The low-power mode includes: when the power level is less than a set power level, reducing the sampling frequency of the data acquisition module, pausing the model training process in the edge computing module, and reducing the communication frequency of the wireless communication module.

[0011] In one embodiment, the local propagation module includes: a display screen and / or a speaker; the display screen is used to display the warning signal; and the speaker is used to verbally play the warning signal.

[0012] In one embodiment, the machine learning model is a logistic regression model, a random forest model, a support vector machine, or a neural network model.

[0013] In one embodiment, the data acquisition module includes: a water level sensor, a rainfall sensor, and a soil moisture sensor.

[0014] Secondly, this application provides a method for intelligent monitoring and early warning of flash flood disasters, including: Obtain the current rainfall and soil moisture in the target area; Based on the current rainfall and soil moisture, a water level prediction model is used to simulate the process of runoff converging at the cross-section in the target area, and the cross-sectional water level at the current moment is calculated. The water level prediction model is obtained by calibrating the runoff generation and confluence model using historical data. The historical data includes: measured water level, rainfall and soil moisture in the target area at historical moments. Based on the current rainfall, soil moisture, cross-sectional water level, and environmental characteristics, a flash flood risk prediction model is used to predict the probability of flash floods in the target area. The flash flood risk prediction model is obtained by training a machine learning model using training data. The training data includes: measured water level, rainfall, soil moisture, environmental characteristics, and flash flood occurrence records of the target area at historical times. A warning signal is issued based on the current cross-sectional water level and the probability of flash floods.

[0015] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides an intelligent monitoring and early warning system and method for flash flood disasters. By introducing edge computing technology through an edge computing module, the data processing capability is brought down to the small watershed at the monitoring source. Water level prediction model and flash flood risk prediction model are used to perform real-time analysis and early warning at the data collection source, thereby improving the response speed and accuracy of the early warning. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the structure of an intelligent monitoring and early warning system for flash floods provided in an embodiment of this application; Figure 2 This is a flowchart illustrating an intelligent monitoring and early warning method for flash floods provided in an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] The intelligent monitoring and early warning system and method for flash floods provided in this application realizes intelligent monitoring and early warning of flash floods based on edge computing. It can perform real-time monitoring and early warning before flash floods occur, solving the problems of long monitoring data processing links and delayed early warning in the existing technology.

[0021] In one exemplary embodiment, such as Figure 1 As shown, a smart monitoring and early warning system for flash flood disasters is provided, including: a data acquisition module and an edge computing module; the data acquisition module is connected to the edge computing module.

[0022] The data acquisition module is used to collect the current rainfall and soil moisture in the target area.

[0023] The edge computing module is used for: Based on the current rainfall and soil moisture, a water level prediction model is used to simulate the process of runoff converging at the cross section in the target area, and the cross section water level at the current moment is calculated. The water level prediction model is obtained by calibrating the runoff generation and confluence model using historical data. The historical data includes: measured water level, rainfall and soil moisture in the target area at historical moments.

[0024] Based on the current rainfall, soil moisture, cross-sectional water level, and environmental characteristics, a flash flood risk prediction model is used to predict the probability of flash floods in the target area. The flash flood risk prediction model is obtained by training a machine learning model using training data. The training data includes: measured water level, rainfall, soil moisture, environmental characteristics, and flash flood occurrence records of the target area at historical times.

[0025] A warning signal is issued based on the current cross-sectional water level and the probability of flash floods.

[0026] In another exemplary embodiment of this application, the data acquisition module includes: a water level sensor, a rainfall sensor, and a soil moisture sensor.

[0027] In another exemplary embodiment of this application, the data acquisition module is also used to acquire historical data of the target area.

[0028] The edge computing module includes: a runoff generation and confluence model calibration unit, a cross-sectional water level prediction unit, a machine learning model training unit, a disaster prediction unit, and a disaster early warning unit.

[0029] The flow generation and confluence model calibration unit is used for: A runoff generation model and a runoff confluence model are constructed; the runoff generation model is used to simulate the process of rainfall being converted into surface runoff; the runoff confluence model is used to simulate the flow process of runoff in the river channel.

[0030] The production flow model and the confluence flow model are integrated to obtain the production-confluence flow model.

[0031] The rainfall and soil moisture from the historical data are input into the runoff generation and confluence model. The runoff generation and confluence model is calibrated based on the comparison between the cross-sectional water level output by the runoff generation and confluence model and the measured water level in the historical data to determine the model parameters. The runoff generation and confluence model with the determined model parameters is used as the water level prediction model.

[0032] The model parameters include: parameters in the runoff generation model and parameters in the confluence model; the parameters in the runoff generation model include at least: infiltration rate; the parameters in the confluence model include at least: roughness rate.

[0033] The cross-sectional water level prediction unit is used to input the current rainfall and soil moisture into the water level prediction model to obtain the current cross-sectional water level.

[0034] The machine learning model training unit is used to: input the training data into the machine learning model for training, and use the trained machine learning model as a flash flood risk prediction model. The machine learning model can be a logistic regression model, a random forest model, a support vector machine, or a neural network model, etc.

[0035] The disaster prediction unit is used to input the current rainfall, soil moisture, cross-sectional water level and environmental characteristics into the flash flood risk prediction model to predict the probability of flash floods occurring in the target area.

[0036] The disaster early warning unit is used to: calculate the difference between the current cross-sectional water level and the warning water level; and issue an early warning signal based on the difference and the probability of flash flood occurrence. Specifically: If the water level at the cross-section is lower than the warning level, the difference is greater than the set water level value, and the probability of flash floods is less than the first set probability value, a low-risk warning signal will be issued.

[0037] If the cross-sectional water level is lower than the warning water level, the difference is less than or equal to the set water level value, and the probability of flash floods is between the first set probability value and the second set probability value, a medium-risk warning signal will be issued.

[0038] If the cross-sectional water level is higher than the warning water level, the difference is less than or equal to the set water level value, and the probability of flash floods is between the second and third set probability values, a high-risk warning signal will be issued.

[0039] If the cross-sectional water level is higher than the warning level, the difference is greater than the set water level value, and the probability of flash floods is greater than the third set probability value, then an extremely high risk warning signal will be issued.

[0040] In practical applications, the first set probability value can be 0.3, the second set probability value can be 0.6, and the third set probability value can be 0.8.

[0041] In another exemplary embodiment of this application, see [link to embodiment]. Figure 1 The intelligent monitoring and early warning system for flash floods further includes: a wireless communication module; the wireless communication module is connected to the edge computing module; the wireless communication module is used to remotely transmit the early warning signal.

[0042] The wireless communication module can be a LoRa module or a 4G / 5G module.

[0043] In another exemplary embodiment of this application, see [link to embodiment]. Figure 1 The intelligent monitoring and early warning system for flash floods further includes: a local propagation module; the local propagation module is connected to the edge computing module; the local propagation module is used to propagate the early warning signal locally.

[0044] In another exemplary embodiment of this application, see [link to embodiment]. Figure 1 The intelligent monitoring and early warning system for flash floods further includes a power supply module; the power supply module is connected to the data acquisition module, the edge computing module and the wireless communication module respectively.

[0045] The power supply module includes a solar panel and a supercapacitor. The solar panel collects electrical energy during the day and stores it in the supercapacitor, which then powers other devices at night or on cloudy days.

[0046] In another exemplary embodiment of this application, see [link to embodiment]. Figure 1 The intelligent monitoring and early warning system for flash floods also includes an intelligent power management module.

[0047] The intelligent power management module is connected to the power supply module; the intelligent power management module is used to monitor the power level of the power supply module and switch the operating mode according to the power level; the operating mode includes: normal mode and low power mode.

[0048] The normal mode includes: when the power level is greater than or equal to a set power level, the data acquisition module, the edge computing module, and the wireless communication module maintain full-function operation; the low-power mode includes: when the power level is less than a set power level, reducing the sampling frequency of the data acquisition module, pausing the model training process in the edge computing module, and reducing the communication frequency of the wireless communication module.

[0049] In another exemplary embodiment of this application, see [link to embodiment]. Figure 1 The intelligent monitoring and early warning system for flash floods further includes a data caching module; the data caching module is connected to the data acquisition module and the edge computing module.

[0050] The data caching module includes a built-in storage unit. When the network is interrupted or there is no network, the built-in storage unit caches data (such as rainfall, soil moisture, cross-sectional water level, probability of flash floods, historical data, training data, etc.) to avoid data loss when the network is interrupted. When the network is restored, the cached data is uploaded to the cloud or management platform.

[0051] In another exemplary embodiment of this application, the local propagation module includes: a display screen and / or a speaker; the display screen is used to display the warning signal; and the speaker is used to voice-play the warning signal.

[0052] The following provides a more specific embodiment to further describe the above-mentioned intelligent monitoring and early warning system for flash floods.

[0053] The mountain flood disaster intelligent monitoring and early warning system of this embodiment includes a data acquisition module, an edge computing module, a power supply module, an intelligent power management module, a data caching module, a wireless communication module, a display screen, and a speaker.

[0054] (1) Data acquisition module.

[0055] The data acquisition module includes a water level sensor, a rainfall sensor, and a soil moisture sensor, which are used to collect water level, rainfall, and soil moisture data in real time and transmit the data to the edge computing module.

[0056] The data acquisition module adopts a multi-functional sensor integration design, integrating data from multiple sensors. These sensors collect environmental data in real time and transmit the data to the edge computing module for real-time processing via wireless communication protocols (such as LoRa).

[0057] (2) Edge computing module.

[0058] The edge computing module is used to filter and denoise the data collected by various sensors to ensure data quality. Then, based on real-time monitoring data such as rainfall and soil moisture, it calculates the cross-sectional water level, compares it with the measured water level, and then calibrates the runoff generation and confluence model in real time. Based on historical and real-time data, it uses a machine learning model to assess the probability of flash floods, identify potential dangers, and conduct risk assessments. Based on the real-time analysis results, it automatically judges the severity of the disaster and issues early warning signals through the equipment.

[0059] The calibration process is as follows: ① Construction of the production and confluence model.

[0060] Constructing a runoff generation model: A runoff generation model is used to simulate the process of rainfall being converted into surface runoff, taking into account factors such as soil moisture content, infiltration capacity, and evaporation. Suitable models include the SCS-CN model or the Green-Ampt infiltration model. The parameters in the runoff generation model are determined based on watershed soil type, land use, and historical data.

[0061] Constructing a confluence model: The confluence model is used to simulate the flow process of runoff in the river channel. The hydraulic equation (Saint-Venant equation) or a simplified method (unit line method) is used. The parameters in the confluence model include the river channel slope, roughness, cross-sectional shape, etc.

[0062] Integrated construction: The runoff generation model and the runoff confluence model are combined to form a complete runoff generation and confluence model, and initial calibration and verification are performed using historical hydrological data (rainfall, water level, etc.).

[0063] The purpose of constructing runoff generation and confluence models is to predict changes in river levels and flow rates caused by rainfall events, thereby assessing the risk of flash floods. It can simulate the hydrological response of small watersheds, providing a basis for water level predictions for early warning systems.

[0064] ② Calculation of cross-sectional water level.

[0065] Cross-sectional water level refers to the water level at a specific cross-section of a river or stream. Its calculation is based on real-time monitored data such as rainfall and soil moisture content, and is simulated using a runoff generation and confluence model. Specific calculation method: Input data: real-time monitored rainfall (rainfall intensity, cumulative rainfall), soil moisture (reflecting soil saturation), and watershed characteristic data (watershed area, slope, channel roughness, etc.).

[0066] Calculation Process: The runoff generation model in the runoff generation and confluence model is used. Based on the rainfall-runoff relationship, rainfall and soil moisture are converted into runoff volume. Then, the cross-sectional water level is calculated using a confluence model (Saint-Venant equation or simplified kinematic wave model). If the unit hydrograph method or distributed hydrological model is used, the process of runoff converging at the cross-section is simulated, and the cross-sectional water level is output. The calculated cross-sectional water level is a theoretical value, which is used to compare with the measured water level (from water level sensors) to verify and calibrate the model.

[0067] The purpose of real-time calibration of runoff generation and confluence models is to dynamically adjust model parameters (infiltration rate, roughness, etc.) by comparing calculated cross-sectional water levels with measured water levels to adapt to current watershed conditions (such as changes in soil moisture) and improve model prediction accuracy. Specifically, it adapts to changes: watershed conditions (such as soil saturation) change with weather, and real-time calibration ensures that model parameters are updated, reducing prediction errors; it improves reliability: through real-time feedback, the model can more accurately reflect actual hydrological processes, enhancing the timeliness and accuracy of early warnings.

[0068] The training process for the machine learning model is as follows: ① Data collection: Collect historical data, including rainfall data, soil moisture data, water level data, flash flood records, environmental characteristics (topography, vegetation cover), etc.

[0069] ② Feature engineering: Select key feature variables, such as rainfall intensity, rainfall duration, soil moisture content, and previous water level change rate.

[0070] ③ Model Training: Use machine learning algorithms (such as logistic regression, random forest, support vector machine, or neural network) to train a classification or regression model, using the occurrence of flash floods as a label, to learn the association between features and disasters. The model is optimized through cross-validation and performance metrics (such as accuracy and recall).

[0071] ④ Deployment and Updates: The trained machine learning model is integrated into the edge computing module as a flash flood risk prediction model, and is retrained regularly with new data to maintain the model's self-learning ability.

[0072] The specific process of assessing the probability of flash floods using a flash flood risk prediction model is as follows: ① Real-time data input: Real-time monitored data such as rainfall and soil moisture content are input as features into the flash flood risk prediction model.

[0073] ② Probability Calculation: The model outputs a probability value between 0 and 1, representing the likelihood of a flash flood occurring. For example, when using a logistic regression model, the probability is calculated using the sigmoid function; when using a neural network, the probability is obtained through the activation function of the output layer.

[0074] ③ Risk Identification: A probability threshold (e.g., 0.7) is set. When the output probability exceeds the threshold, it is identified as a potential hazard, triggering further early warning analysis. Flash flood risk prediction models may also incorporate time series analysis to consider trend changes.

[0075] The real-time analysis process is as follows: The real-time analysis results integrate the analysis results of the water level prediction model and the flash flood risk prediction model: the water level prediction model provides the predicted value of the cross-sectional water level, which is used to assess the severity of the flood situation (such as whether the water level exceeds the warning line); the flash flood risk prediction model provides the probability of flash floods, which is used to assess the likelihood of disasters. The two are combined to form a complete real-time analysis result, which is used to determine the severity of disasters. For example, water level prediction and probability assessment are jointly input into the decision-making logic.

[0076] The specific method for judging the severity of a disaster based on real-time analysis results using a multi-indicator fusion strategy is as follows: ①Indicator extraction: Obtain the predicted cross-sectional water level from the water level prediction model, calculate the difference between it and the warning water level, and obtain the probability of flash flood occurrence from the flash flood risk prediction model; at the same time, auxiliary indicators such as rainfall intensity and soil saturation are considered.

[0077] ② Risk Level Classification: Define warning levels, such as: Low risk: probability < 0.3, water level is below the warning line; Medium risk: probability 0.3-0.6, water level is close to the warning line; High risk: probability 0.6-0.8, water level exceeds the warning line; Very high risk: probability > 0.8, water level significantly exceeds the warning line.

[0078] Automatic decision-making: The edge computing module automatically outputs the severity level based on preset rules or fuzzy logic and triggers corresponding warnings.

[0079] In summary, the edge computing module of this embodiment has built-in high-efficiency computing units and machine learning algorithms, which are responsible for real-time analysis and processing of the data collected from various sensors, and realize the following functions.

[0080] Data preprocessing: Sensor data is filtered and denoised to ensure data quality; Model calibration: Based on real-time monitoring data such as rainfall and soil moisture content, the cross-sectional water level is calculated and compared with the measured water level, and then the runoff generation and confluence model is calibrated in real time; Risk assessment: Based on historical and real-time data, the probability of flash floods is assessed through machine learning models to identify potential hazards; Intelligent decision-making: Based on real-time analysis results, the edge computing module can automatically determine the severity of the disaster and issue early warning signals through devices; The edge computing module can quickly calibrate the model and make decisions before a disaster occurs, improving the real-time performance and accuracy of early warnings.

[0081] To adapt to environments with insufficient power, such as mountainous areas, the edge computing module uses a low-power AI chip, which has the following characteristics.

[0082] High efficiency and low power consumption: The AI ​​chip supports complex machine learning algorithms and real-time data processing, while also featuring a low-power operating mode to ensure long-term operation even under energy-constrained conditions. Strong data processing capabilities: Supports edge computing, enabling real-time sensor data processing and risk prediction.

[0083] (3) Data transmission module (wireless communication module, display screen and speaker).

[0084] When the edge computing module detects the risk of flash floods, the data transmission module sends early warning information to the in-home alarm and relevant departments, and displays detailed early warning information on the screen to remind local residents to take timely precautions; or it plays voice warnings to loudspeakers in densely populated areas (such as villages, schools, etc.).

[0085] The wireless communication module transmits warning signals (including warning information) to relevant departments and affected areas via wireless communication. Specifically, it sends SMS or data warnings to in-home alarms and local authorities through wireless communication modules such as LoRa or 4G / 5G.

[0086] Local dissemination via display screen and speaker: Warning information is displayed on the device's own screen, or voice broadcasts are played through connected speakers, covering densely populated areas such as villages and schools.

[0087] (4) Power supply module.

[0088] The power supply module includes a solar panel and an energy storage system. The solar panel collects electrical energy during the day and stores it in the energy storage system. At night, the energy storage system or on cloudy days, the energy is used to power other devices for normal operation.

[0089] The power supply module provides continuous power to the device, and its specific design is as follows: Solar panels: High-efficiency solar panels are used to ensure that the equipment can still provide sufficient energy in mountainous areas with poor sunlight conditions.

[0090] Energy storage system: Using high-efficiency lithium batteries or supercapacitors, it can store solar energy during the day to ensure the normal operation of the equipment at night or on cloudy days.

[0091] (5) Intelligent power management module.

[0092] The intelligent power management operates in a low-power mode, further extending the device's usage time. The intelligent power management module achieves power management through the following process: ① Power monitoring: Real-time monitoring of solar panel output, supercapacitor charge, and equipment power consumption status.

[0093] ②Mode switching: Automatically switch operating modes based on battery level.

[0094] Normal mode: When the battery is fully charged, all modules (data acquisition module, edge computing module and wireless communication module) operate at full capacity.

[0095] Low power mode: When the power is low, the sampling frequency of the data acquisition module is reduced, unnecessary calculations of the edge computing module (such as model training) are paused, and the communication frequency of the wireless communication module is reduced (such as extending the data upload interval), maintaining only key functions (such as basic data acquisition and early warning).

[0096] ③ Dynamic adjustment: Using scheduling algorithms, priority is given to core tasks (such as risk assessment and early warning), supercapacitor power is used at night or on cloudy days, and the sleep-wake cycle is optimized.

[0097] ④ Remote management: Supports remote command adjustment of power consumption strategy, such as issuing energy-saving parameters through cloud platform to reduce manual intervention.

[0098] (6) Data caching module.

[0099] The data caching module includes a built-in storage unit. When the network is interrupted or there is no network, the data is cached through the built-in storage unit to avoid data loss during network interruption. When the network is restored, the cached data is uploaded to the cloud or management platform.

[0100] This embodiment constructs a complete intelligent monitoring and early warning system for flash floods through the cooperation of a data acquisition module, an edge computing module, a wireless communication module, and an intelligent power management module. Data collected by the data acquisition module is transmitted to the edge computing module via a wireless communication protocol for real-time data processing and early warning analysis. Before a disaster occurs, the system can quickly issue early warning signals and disseminate warning information to relevant personnel through wireless networks, SMS, voice broadcasts, and other means to ensure timely disaster response.

[0101] The intelligent monitoring and early warning system for flash floods in this embodiment has self-learning and self-optimization functions, local caching and data uploading functions, and disaster early warning information dissemination and emergency response functions.

[0102] Self-learning and self-optimization features: The machine learning algorithm in this embodiment has self-learning and self-optimization capabilities. The device can continuously optimize the early warning model by learning from and analyzing historical and real-time data to improve the accuracy of early warnings. Specific functions are as follows: Self-learning function: Continuously adjusts and optimizes the parameters of the early warning model based on the matching of real-time and historical data. Self-optimization function: The device can automatically adjust the model under different environments (such as seasonal changes and climate differences) to improve early warning accuracy and reduce false alarms.

[0103] Local caching and data upload functionality: The device supports local caching and data upload functions, ensuring that it can continue to operate even when the network is interrupted. Details are as follows: Local caching: The device has built-in storage that caches data, ensuring that important data is not lost even when there is no network connection. Data upload: When the network is restored, the device automatically uploads the cached data to the cloud or management platform.

[0104] Disaster early warning information dissemination and emergency response functions: When the edge computing module detects a risk of flash floods, the device transmits early warning information wirelessly to relevant departments and devices in the affected areas. Specific dissemination methods include: SMS alerts: Send alert information to in-home alarm devices and relevant responsible persons (such as local government and village committees). Voice broadcasts: Play voice warnings through loudspeakers in densely populated areas (such as villages and schools). Display screens: Display detailed alert information on devices or terminal screens to remind local residents to take timely evacuation measures. The system can also automatically adjust the alert content and dissemination method according to the alert level to ensure efficient information dissemination.

[0105] The intelligent monitoring and early warning system for flash floods in this embodiment adopts a modular design, allowing each sensor and computing module to be installed independently, facilitating equipment deployment and maintenance. The installation method is as follows: Installation: The equipment can be fixed to monitoring points in mountain streams, rivers, or small watersheds using brackets. The sensor installation location should be able to reflect changes in water level, rainfall, and soil moisture to the greatest extent possible. Maintenance and Management: The equipment can be remotely monitored for status and supports online diagnostics and upgrades, reducing manual inspection and maintenance costs.

[0106] Based on the same inventive concept, this application also provides a disaster intelligent monitoring and early warning method implemented using the aforementioned intelligent monitoring and early warning system for flash floods. The solution provided by this method is similar to the implementation scheme described in the above system; therefore, the specific limitations in one or more disaster intelligent monitoring and early warning method embodiments provided below can be found in the limitations of the disaster intelligent monitoring and early warning system described above, and will not be repeated here.

[0107] The intelligent disaster monitoring and early warning system of this embodiment has the following advantages: Multifunctional sensor and edge computing terminal integrated design: Combines multiple sensors such as water level, rainfall, and soil moisture with edge computing modules to perform real-time data acquisition and processing.

[0108] Design of a self-optimizing early warning model based on machine learning: The early warning model is continuously optimized using machine learning algorithms to improve the accuracy, response speed and adaptability of the early warning.

[0109] Low-power, self-powered edge computing system design: Through low-power AI chips and a solar self-powered system, the device can operate stably in environments with limited power.

[0110] In one exemplary embodiment, such as Figure 2 As shown, a method for intelligent disaster monitoring and early warning is provided, including: Step 201: Obtain the current rainfall and soil moisture in the target area.

[0111] Step 202: Based on the current rainfall and soil moisture, use a water level prediction model to simulate the process of runoff from the target area converging at the cross section and calculate the cross section water level at the current time.

[0112] The water level prediction model is obtained by calibrating the runoff generation and confluence model using historical data; the historical data includes: measured water level, rainfall and soil moisture in the target area at historical times.

[0113] Step 203: Based on the current rainfall, soil moisture, cross-sectional water level and environmental characteristics, use the flash flood risk prediction model to predict the probability of flash floods occurring in the target area.

[0114] The flash flood risk prediction model is obtained by training a machine learning model with training data; the training data includes: measured water level, rainfall, soil moisture, environmental characteristics and flash flood occurrence records of the target area at historical times.

[0115] Step 204: Issue an early warning signal based on the current cross-sectional water level and the probability of flash flood occurrence.

[0116] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0117] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A smart monitoring and early warning system for flash flood disasters, characterized in that, The intelligent monitoring and early warning system for flash floods includes: a data acquisition module and an edge computing module; the data acquisition module is connected to the edge computing module. The data acquisition module is used for: Collect the current rainfall and soil moisture in the target area; The edge computing module is used for: Based on the current rainfall and soil moisture, a water level prediction model is used to simulate the process of runoff converging at the cross-section in the target area, and the cross-sectional water level at the current moment is calculated. The water level prediction model is obtained by calibrating the runoff generation and confluence model using historical data. The historical data includes: measured water level, rainfall and soil moisture in the target area at historical moments. Based on the current rainfall, soil moisture, cross-sectional water level, and environmental characteristics, a flash flood risk prediction model is used to predict the probability of flash floods in the target area. The flash flood risk prediction model is obtained by training a machine learning model using training data. The training data includes: measured water level, rainfall, soil moisture, environmental characteristics, and flash flood occurrence records of the target area at historical times. A warning signal is issued based on the current cross-sectional water level and the probability of flash floods.

2. The intelligent monitoring and early warning system for flash floods according to claim 1, characterized in that, The data acquisition module is also used to collect historical data of the target area; The edge computing module includes: a runoff generation and confluence model calibration unit, a cross-sectional water level prediction unit, a machine learning model training unit, a disaster prediction unit, and a disaster early warning unit; The flow generation and confluence model calibration unit is used for: A runoff generation model and a runoff collection model are constructed; the runoff generation model is used to simulate the process of rainfall being converted into surface runoff; the runoff collection model is used to simulate the flow process of runoff in the river channel. The production flow model and the confluence flow model are integrated to obtain the production-confluence flow model; The rainfall and soil moisture from the historical data are input into the runoff generation and confluence model. The runoff generation and confluence model is calibrated based on the comparison between the cross-sectional water level output by the runoff generation and confluence model and the measured water level in the historical data to determine the model parameters. The runoff generation and confluence model with the determined model parameters is used as the water level prediction model. The model parameters include: parameters in the runoff generation model and parameters in the confluence model; the parameters in the runoff generation model include at least: infiltration rate; the parameters in the confluence model include at least: roughness. The cross-sectional water level prediction unit is used for: Input the current rainfall and soil moisture into the water level prediction model to obtain the current cross-sectional water level. The machine learning model training unit is used for: The training data is input into the machine learning model for training, and the trained machine learning model is used as a flash flood risk prediction model. The disaster prediction unit is used for: Input the current rainfall, soil moisture, cross-sectional water level and environmental characteristics into the flash flood risk prediction model to predict the probability of flash floods in the target area; The disaster early warning unit is used for: Calculate the difference between the current cross-sectional water level and the warning water level; A warning signal is issued based on the difference and the probability of flash floods.

3. The intelligent monitoring and early warning system for flash floods according to claim 1, characterized in that, The intelligent monitoring and early warning system for flash floods also includes: a wireless communication module; The wireless communication module is connected to the edge computing module; the wireless communication module is used to remotely transmit the warning signal.

4. The intelligent monitoring and early warning system for flash floods according to claim 1, characterized in that, The intelligent monitoring and early warning system for flash floods also includes: a local propagation module; The local propagation module is connected to the edge computing module; the local propagation module is used to propagate the warning signal locally.

5. The intelligent monitoring and early warning system for flash floods according to claim 3, characterized in that, The intelligent monitoring and early warning system for flash floods also includes: a power supply module; The power supply module is connected to the data acquisition module, the edge computing module, and the wireless communication module, respectively.

6. The intelligent monitoring and early warning system for flash floods according to claim 5, characterized in that, The intelligent monitoring and early warning system for flash floods also includes: an intelligent power management module; The intelligent power management module is connected to the power supply module; the intelligent power management module is used to monitor the power level of the power supply module and switch the operating mode according to the power level; the operating mode includes: normal mode and low power mode; The normal mode includes: when the power level is greater than or equal to the set power level, the data acquisition module, the edge computing module, and the wireless communication module maintain full-function operation; The low-power mode includes: when the power level is less than a set power level, reducing the sampling frequency of the data acquisition module, pausing the model training process in the edge computing module, and reducing the communication frequency of the wireless communication module.

7. The intelligent monitoring and early warning system for flash floods according to claim 4, characterized in that, The local propagation module includes: a display screen and / or a speaker; the display screen is used to display the warning signal; the speaker is used to play the warning signal aloud.

8. The intelligent monitoring and early warning system for flash floods according to claim 2, characterized in that, The machine learning model is a logistic regression model, a random forest model, a support vector machine, or a neural network model.

9. The intelligent monitoring and early warning system for flash floods according to claim 1, characterized in that, The data acquisition module includes: a water level sensor, a rainfall sensor, and a soil moisture sensor.

10. A method for intelligent monitoring and early warning of flash flood disasters, characterized in that, The intelligent monitoring and early warning method for flash flood disasters includes: Obtain the current rainfall and soil moisture in the target area; Based on the current rainfall and soil moisture, a water level prediction model is used to simulate the process of runoff converging at the cross-section in the target area, and the cross-sectional water level at the current moment is calculated. The water level prediction model is obtained by calibrating the runoff generation and confluence model using historical data. The historical data includes: measured water level, rainfall and soil moisture in the target area at historical moments. Based on the current rainfall, soil moisture, cross-sectional water level, and environmental characteristics, a flash flood risk prediction model is used to predict the probability of flash floods in the target area. The flash flood risk prediction model is obtained by training a machine learning model using training data. The training data includes: measured water level, rainfall, soil moisture, environmental characteristics, and flash flood occurrence records of the target area at historical times. A warning signal is issued based on the current cross-sectional water level and the probability of flash floods.

Images