A heavy rain flood water level line early warning monitoring method

Abstract

The present application relates to the technical field of hydro-meteorological monitoring, in particular to a heavy rain and flood water level line early warning monitoring method. The present application ensures the consistency and timeliness of real-time water level data from different data sources through multi-source hydro-meteorological data collection and time synchronization technology. Through accurate time synchronization processing, the deviation caused by data time lag is avoided, ensuring the accurate docking and fusion of various water level data, making the subsequent flood risk assessment and early warning more reliable. Especially under sudden weather conditions such as heavy rain, it can quickly respond and provide accurate risk assessment; the present application combines the combination preprocessing algorithm of moving average filtering and Kalman filtering, effectively removes the noise in the water level data of each data source, and weights the weight of different data sources according to the prediction error, so as to obtain a comprehensive and accurate water level data. This fusion method improves the smoothness and accuracy of the data, providing high-quality input data for subsequent risk assessment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydro-meteorological monitoring, in particular to a heavy rain and flood water level line early warning monitoring method. BACKGROUND

[0002] With the intensification of global climate change, the frequency and intensity of extreme weather events such as heavy rain and flood are increasing year by year, which poses a great threat to people's life and property safety, agricultural production, urban infrastructure, etc. In the flood warning system, real-time and accurate water level monitoring and prediction are the key to ensuring timely and effective disaster prevention and mitigation measures. Therefore, it is of great practical significance to develop intelligent and real-time flood water level line early warning monitoring technology.

[0003] Existing flood warning systems mostly rely on a single data source (such as water level data from hydrological stations, precipitation data from meteorological stations, etc.), and these data often face problems such as poor timeliness, insufficient accuracy, limited coverage, etc. In addition, although there are some multi-source data fusion technologies, the existing fusion methods mostly focus on static data analysis and cannot fully consider the time series variation characteristics between data, resulting in insufficient prediction accuracy and timeliness. Therefore, how to efficiently integrate real-time and predicted data from different monitoring devices (such as meteorological stations, radars, satellite remote sensing, etc.) to solve the timeliness, accuracy, and multi-source data collaboration problems in existing technologies has become a technical challenge that needs to be addressed.

[0004] Existing flood warning systems mainly rely on traditional single data sources such as hydrological stations and meteorological stations. These systems face the following problems:

[0005] Firstly, there is a time delay in collecting hydrological and meteorological data, especially in extreme weather conditions such as heavy rain, the frequency and real-time nature of data updates cannot meet the demand for rapid response.

[0006] Secondly, traditional hydrological and meteorological monitoring stations have low distribution density and are difficult to comprehensively cover real-time data such as water levels and precipitation in large areas.

[0007] In view of the above problems, it is necessary to propose a heavy rain and flood water level line early warning monitoring method. SUMMARY

[0008] The present application aims to solve the problems in the background art and proposes a heavy rain and flood water level line early warning monitoring method.

[0009] The object of the present application can be achieved by the following technical solutions:

[0010] A heavy rain and flood water level line early warning monitoring method, comprising the following steps:

[0011] Step 1, multi-source hydro-meteorological data collection and time synchronization;

[0012] Collecting multi-source hydro-meteorological data, including real-time meteorological data of weather stations, water level and flow data of hydrological stations, precipitation monitoring data, radar monitoring data and satellite remote sensing data. And convert it into a standard format suitable for processing, through a serial communication interface to the central processing server. Time synchronization processing is carried out during the collection of multi-source data to ensure the consistency and timeliness of the water level data.

[0013] Accessing real-time meteorological data of weather stations, water level and flow data of hydrological stations, precipitation monitoring data, radar monitoring data and satellite remote sensing data, collecting real-time water level data Hi(t) from each data source i, where i is the data source number, and i = 1, 2, 3, 4, 5; Corresponding to weather stations, hydrological stations, precipitation monitoring, radar monitoring and satellite remote sensing; Where t is the collection time of real-time water level data.

[0014] As a preferred mode of the present application, the real-time water level data collected from each data source is subjected to time synchronization processing, and the specific process is as follows:

[0015] When it is identified that the collection time ti of the water level data Hi(t) is inconsistent with the target synchronization time t0, the water level data Hi(t) is mapped to the target time through a linear interpolation algorithm, and the linear interpolation algorithm is as follows:

[0016] Obtaining the water level data Hi(t') collected last time from the data source i, where t' is the last data collection time, and the formula is as follows:

[0017] Calculating the linear mapping value Hi(t0) of the water level data Hi(t) from the current time t to the target synchronization time t0.

[0018] Replacing the linear mapping value Hi(t0) as the accurate value of the water level data Hi(t) at time t = t0.

[0019] Step two, edge fusion preprocessing and feature extraction;

[0020] The collected real-time water level data is preprocessed and feature extracted. The data may be disturbed by noise, which needs to be denoised by filtering algorithm, and the feature information closely related to flood risk is extracted, including precipitation change trend, water level fluctuation amplitude, flow rate and future water level prediction.

[0021] The water level data collected through each data source is subjected to combined filtering through a combination of moving average filtering and Kalman filtering preprocessing algorithm, and the specific process is as follows:

[0022] Through a preset formula The water level data H1(t) of each data source i is subjected to sliding equalization filtering to obtain filtered water level height data

[0023] Wherein, H1(t+k) is the water level data collected by the data source i at time t+k, N is a preset sliding window size, and k is a time window index.

[0024] The result of the water level data collected by each data source i is subjected to sliding equalization filtering by pre-Kalman filtering Further fusion is performed to obtain comprehensive water level data H(t). The Kalman filtering formula is:

[0025]

[0026] Wherein, K1(t) is the Kalman gain, which determines the reference weight of each data source i, P1(t|t-1) is the covariance of the current time prediction error, R1(t) is the variance of the observation noise, which represents the water level observation uncertainty of the i-th data source, and R1(t) obeys a Gaussian distribution.

[0027] The covariance P1(t|t-1) of the current time prediction error is obtained through the covariance update formula: P1(t|t-1) = F P1(t-1) F T + Q. Wherein, F is a state transition matrix, which describes the state number from time t-1 to t. Q is a process noise covariance matrix, which represents the model uncertainty. F T is the transpose matrix of the state transition matrix. Wherein, ΔH1(t), and are the filtered water level height data the number change, the change speed and the change acceleration from the last time t-1 to the current time t.

[0028] As a preferred mode of the present application, the fluctuation amplitude and the change trend are extracted based on time series analysis to predict the future water level, and the specific process is as follows:

[0029] An ARIMA autoregressive integrated moving average model is constructed, and the model formula is

[0030]

[0031] Wherein, Yt is the predicted value of the water level height data at future time t, c is a preset constant bias term, j1 and j2 are autoregressive parameter indexes corresponding to j1 time and j2 time in the past, P and q are preset autoregressive thresholds, representing the reference range of the past historical data; and φ j1φj1is the autoregressive term parameter of the historical data at the past j1time, representing the influence coefficient of the historical data at the past j1time on the predicted value of the water level data at the future time t; φj2is the sliding average term parameter, representing the influence coefficient of the prediction error at the past j2time on the predicted value of the water level data at the future time t. Yj1is the historical water level data at the past j1time of the current time t. ΔYj1is the difference between the predicted value and the real water level at the past j2time of the current time t, i.e. the historical prediction error. j2 t-j1 φj1is the autoregressive term parameter of the historical data at the past j1time, representing the influence coefficient of the historical data at the past j1time on the predicted value of the water level data at the future time t; φj2is the sliding average term parameter, representing the influence coefficient of the prediction error at the past j2time on the predicted value of the water level data at the future time t. Yj1is the historical water level data at the past j1time of the current time t. ΔYj1is the difference between the predicted value and the real water level at the past j2time of the current time t, i.e. the historical prediction error. t-j2

[0032] H(t) is the comprehensive water level data at each time, which is the historical water level data Yj1at the past j1time of the current time t. t-j1

[0033] Step three, cloud flood risk assessment and threshold adaptive early warning;

[0034] Collect the comprehensive real-time water level data obtained by the combination of sliding average filtering and Kalman filtering for real-time flood risk assessment. Collect the fluctuation amplitude and change trend extracted based on time series analysis for future flood risk assessment.

[0035] If the comprehensive water level data H(t) at the current time exceeds the preset threshold, an alarm factor corresponding to the time t is output. If K1 consecutive alarm factors are continuously identified, a real-time flood risk alarm signal is output.

[0036] If the predicted water level data Yt at the future time t exceeds the second preset threshold, an early warning factor corresponding to the time t is output. If k2 consecutive early warning factors are continuously identified, a flood risk early warning signal is output. K1 and k2 are preset detection window thresholds.

[0037] Step four, multi-channel early warning release and automatic disposal execution;

[0038] Obtain real-time alarm characteristic values and early warning characteristic values, obtain risk assessment results through data fusion, convert the risk assessment results into specific early warning information, and release the early warning information through multiple channels.

[0039] If a real-time flood risk alarm signal is identified, an alarm process is started, and alarm information is sent to the background manager.

[0040] If a flood risk early warning signal is identified, early warning information and the corresponding predicted time t are released to guide disaster prevention and emergency response measures.

[0041] ​​​The alarm information and early warning information are rapidly published to relevant units, the public and background managers through communication channels including short messages, emails, APP push, websites and broadcasts.

[0042] Compared with the prior art, the present application has the following advantages:

[0043] 1. The present application ensures the consistency and timeliness of real-time water level data from different data sources through multi-source hydrological and meteorological data collection and time synchronization technology. Through accurate time synchronization processing, the deviation caused by data time lag is avoided, ensuring accurate docking and fusion of various water level data, making subsequent flood risk assessment and early warning more reliable. Especially under sudden weather conditions such as heavy rain, accurate risk assessment can be quickly responded to;

[0044] 2. The present application effectively removes noise in water level data from each data source by combining the combination preprocessing algorithm of moving average filtering and Kalman filtering, and weights the weights of different data sources according to the prediction error, so as to obtain a comprehensive and accurate water level data. This fusion method improves the smoothness and accuracy of the data, providing high-quality input data for subsequent risk assessment. In addition, through time series analysis (such as ARIMA model), the amplitude of water level fluctuation and the trend of change can be extracted, so that the future water level change can be accurately predicted, providing a scientific basis for early warning;

[0045] 3. The present application can automatically identify and output real-time flood risk alarm signals and predicted flood risk warning signals through real-time flood risk assessment and threshold adaptive early warning. After the risk alarm and warning signals are identified, the system can start the alarm process or issue the warning information according to the set threshold conditions, and publish it to relevant departments and the public through multiple channels, ensuring the timely transmission of warning information and the efficient execution of emergency response. This multi-channel and all-round early warning release method can effectively improve the coverage and response speed of disaster warning. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to facilitate the understanding of those skilled in the art, the present application will be further described below in conjunction with the drawings:

[0047] Figure 1 The flowchart of the method of the present application. DETAILED DESCRIPTION

[0048] The technical solutions of the present application will be described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0049] Please refer toFigure 1 As shown, a heavy rain flood water level line early warning monitoring method, comprising the following steps:

[0050] Step one, multi-source hydrological and meteorological data collection and time synchronization;

[0051] Collect multi-source hydrological and meteorological data, including real-time meteorological data of meteorological stations, water level and flow data of hydrological stations, precipitation monitoring data, radar monitoring data and satellite remote sensing data. And convert it into a standard format suitable for processing, and send it to the central processing server through the serial communication interface. Time synchronization processing is carried out during the collection of multi-source data to ensure the consistency and timeliness of the water level line data.

[0052] Access real-time meteorological data of meteorological stations, water level and flow data of hydrological stations, precipitation monitoring data, radar monitoring data and satellite remote sensing data, and collect real-time water level data Hi(t) from each data source i, where i is the data source number, and i = 1, 2, 3, 4, 5; Corresponding to meteorological station, hydrological station, precipitation monitoring, radar monitoring and satellite remote sensing; Where t is the collection time of real-time water level data.

[0053] Further, the real-time water level data collected by each data source is subjected to time synchronization processing, and the specific process is as follows:

[0054] When it is identified that the collection time ti of the water level data Hi(t) and the target synchronization time t0 are inconsistent, the water level data Hi(t) is mapped to the target time through a linear interpolation algorithm, and the linear interpolation algorithm is as follows:

[0055] Obtain the water level data Hi(t') collected by the data source i last time, where t' is the last data collection time, and the formula is as follows:

[0056] Calculate the linear mapping value Hi(t0) of the water level data Hi(t) from the current time t to the target synchronization time t0.

[0057] Replace the linear mapping value Hi(t0) as the accurate value of the water level data Hi(t) at time t = t0.

[0058] Step two, edge fusion preprocessing and feature extraction;

[0059] The collected real-time water level data is preprocessed and feature extracted. The data may be disturbed by noise, which needs to be denoised by filtering algorithm, and the feature information closely related to flood risk is extracted, including precipitation change trend, water level fluctuation amplitude, flow rate and future water level prediction.

[0060] The water level data collected by each data source is combined and filtered by a combination preprocessing algorithm of sliding average filtering and Kalman filtering, and the specific process is as follows:

[0061] The water level data Hi(t) of each data source i is filtered by a sliding average filtering algorithm, and the filtered water level height data H'(t) is obtained.

[0062] Wherein, Hi(t+k) is the water level data collected by data source i at time t+k, N is a preset sliding window size, and k is a time window index.

[0063] The result of the water level data collected by each data source i is filtered by a sliding average filtering algorithm. Further fusion is performed to obtain the comprehensive water level data H(t). The Kalman filtering formula is as follows:

[0064]

[0065] Wherein, Ki(t) is the Kalman gain, which determines the reference weight of each data source i, Pi(t|t-1) is the covariance of the prediction error at the current time, Ri(t) is the variance of the observation noise, which represents the water level observation uncertainty of the i th data source, and Ri(t) obeys Gaussian distribution.

[0066] The covariance of the prediction error at the current time Pi(t|t-1) is obtained by the covariance update formula: Pi(t|t-1) = FPi(t-1)F T +Q. Wherein, F is the state transition matrix, which describes the state number from time t-1 to t. Q is the process noise covariance matrix, which represents the uncertainty of the model itself. T F is the transpose matrix of the state transition matrix. Wherein, ΔHi(t), And The filtered water level height data H'(t) is obtained. The quantity change, change speed and change acceleration from the last time t-1 to the current time t.

[0067] ​It should be noted that the combination of moving average filtering and Kalman filtering preprocessing method has significant advantages in water level data processing. Moving average filtering can effectively remove short-term noise and smooth data fluctuations, improving data stability; while Kalman filtering can provide more accurate water level estimates by weighting the fusion of multi-source data and dynamically adjusting the weight according to the noise level and uncertainty of each data source. The combination of the two can not only eliminate irrelevant noise, but also improve the reliability, accuracy and real-time performance of the data, and enhance the prediction ability of future water levels. Ultimately, this preprocessing method can effectively extract features closely related to flood risk, providing more accurate data support for risk assessment and emergency response.

[0068] Further, based on time series analysis to extract fluctuation amplitude and change trend, the future water level is predicted, and the specific process is:

[0069] An ARIMA autoregressive integrated moving average model is constructed, and the model formula is

[0070]

[0071] where Yt is the predicted value of the water level height data at future time t, c is a preset constant bias term, j1 and j2 are autoregressive parameter indexes corresponding to j1 and j2 time points in the past, P and q are preset autoregressive thresholds representing the reference range of past historical data; φ j1 is the autoregressive term parameter of the past j1 time point historical data, representing the influence coefficient of the past j1 time point historical data on the predicted value of the water level height data at future time t; φ j2 is the sliding weight term parameter, representing the influence coefficient of the prediction error at the past j2 time point on the predicted value of the water level height data at future time t. Y t-j1 is the historical water level height data at the j1 time point before the current time t. ΔY t-j2 is the difference between the predicted value of the water level at the j2 time point before the current time t and the actual water level height, i.e. the historical prediction error.

[0072] The comprehensive water level data H(t) at each time point is input into the autoregressive integrated moving average model as the historical water level height data Y t-j1 at the j1 time point before the current time t, and the predicted water level data Yt at future time is obtained.

[0073] It should be noted that the autoregressive integrated moving average model is a statistical model widely used in time series prediction. It consists of three main parts: autoregression, difference and moving average. The combination of these parts makes the ARIMA model very effective in processing non-stationary time series data.

[0074] Step three, cloud flood risk assessment and threshold adaptive early warning;

[0075] The comprehensive real-time water level data obtained by the combination of the sliding average filtering and Kalman filtering is collected for real-time flood risk assessment. The fluctuation amplitude and change trend extracted based on time series analysis are collected for future flood risk assessment.

[0076] If the comprehensive water level data H(t) at the current time exceeds the preset threshold, an alarm factor corresponding to time t is output. If K1 consecutive alarm factors are continuously identified, a real-time flood risk alarm signal is output.

[0077] If the predicted water level data Yt at the predicted future time t exceeds the second preset threshold, an early warning factor corresponding to time t is output. If k2 consecutive early warning factors are continuously identified, a flood risk early warning signal is output. K1 and k2 are preset detection window thresholds.

[0078] Step four, multi-channel early warning release and automatic disposal execution;

[0079] Real-time alarm characteristic values and early warning characteristic values are obtained, risk assessment results are obtained through data fusion, the risk assessment results are converted into specific early warning information, and the early warning information is released through multiple channels.

[0080] If a real-time flood risk alarm signal is identified, an alarm process is started, and alarm information is sent to the background manager.

[0081] If a flood risk early warning signal is identified, early warning information and the corresponding predicted time t are released to guide disaster prevention and emergency response measures.

[0082] The alarm information and early warning information are quickly released to relevant units, the public, and the background manager through communication channels including short messages, emails, APP push, websites, and broadcasts.

[0083] It should be understood that the terms "comprise" and "include" used in the specification and claims of the present disclosure indicate the presence of the described features, integers, steps, operations, elements, and / or components, but do not exclude one or more other features, integers, steps, operations, elements, components, and / or collections thereof.

[0084] It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used in this disclosure and the claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "and / or," as used herein, refer to and set forth any and all possible combinations of one or more of the associated listed items, including permutations of such combinations;

[0085] The preferred embodiments of the application disclosed above are only to help explain the principles of the present application. The preferred embodiments do not limit the present application to only the specific embodiments described. It is apparent that many modifications and variations of this application are possible in light of this disclosure. The preferred embodiments are chosen and described in order to best explain the principles of the application and its practical application to thereby enable others skilled in the art to best utilize the application. The present application is limited only by the following claims and their equivalents.

Claims

1. A method for early warning and monitoring of rainstorm and flood water levels, characterized in that, Includes the following steps; Step 1: Multi-source hydrological and meteorological data acquisition and time synchronization; Collect hydrological and meteorological data from multiple sources, including real-time meteorological data from meteorological stations, water level and flow data from hydrological stations, precipitation monitoring data, radar monitoring data, and satellite remote sensing data; convert the data into a standard format suitable for processing and send it to the central processing server via a serial communication interface; perform time synchronization processing during the multi-source data collection process to ensure the consistency and timeliness of water level data; Step 2: Edge fusion preprocessing and feature extraction; The collected real-time water level data is preprocessed and features are extracted to obtain features closely related to flood risk, including precipitation change trends, water level fluctuation amplitude, flow rate change rate, and future water level prediction. A combined preprocessing algorithm of moving average filtering and Kalman filtering is used to filter water level data collected from various data sources. Moving average filtering is used to filter water level data from independent sources, and Kalman filtering is used to fuse water level data from multiple sources. Based on time series analysis, fluctuation amplitude and change trends are extracted to predict future water levels. Step 3: Cloud-based flood risk assessment and threshold-adaptive early warning; Real-time flood risk assessment is conducted by collecting comprehensive real-time water level data obtained through a combination of moving average filtering and Kalman filtering preprocessing algorithms. Data on fluctuation amplitude and trends extracted from time series analysis are collected to assess flood risk at future times. Step 4: Multi-channel early warning issuance and automatic response execution; The system acquires real-time alarm feature values ​​and early warning feature values, obtains risk assessment results through data fusion, transforms the risk assessment results into specific early warning information, and disseminates the information through multiple channels.

2. The method for early warning and monitoring of rainstorm and flood water levels according to claim 1, characterized in that, The specific process for collecting hydrological and meteorological data from multiple sources is as follows: Access real-time meteorological data from meteorological stations, water level and flow data from hydrological stations, precipitation monitoring data, radar monitoring data, and satellite remote sensing data. Collect real-time water level data Hi(t) from each data source i, where i is the data source number and i=1, 2, 3, 4, 5; corresponding to meteorological stations, hydrological stations, precipitation monitoring, radar monitoring, and satellite remote sensing, respectively; and t is the time of data collection for the real-time water level data.

3. The method for early warning and monitoring of rainstorm and flood water levels according to claim 1, characterized in that, The specific process of time synchronization processing for real-time water level data collected from various data sources is as follows: When it is detected that the acquisition time ti of the water level data Hi(t) is inconsistent with the target synchronization time t0, the water level data Hi(t) is mapped to the target time using a linear interpolation algorithm. The specific linear interpolation algorithm is as follows: Obtain the water level data Hi(t') from the last time data was collected from data source i, where t' is the time of the last data collection, using the formula: Calculate the linear mapping value Hi(t0) of the water level data Hi(t) from the current time t to the target synchronization time t0. The linear mapping value Hi(t0) is replaced with the accurate value of the water level data Hi(t) at time t=t0.

4. The method for early warning and monitoring of rainstorm and flood water levels according to claim 1, characterized in that, The specific process of moving average filtering is as follows: By preset formula A moving average weighted filter is applied to the water level data Hi(t) from each data source i to obtain the filtered water level height data. ; in, Let N be the water level data collected by data source i at time t+k, N be the preset sliding window size, and k be the time window index. The water level data collected from various data sources are further fused using Kalman filtering after passing through a sliding weighted filtering process.

5. The method for early warning and monitoring of rainstorm and flood water levels according to claim 1, characterized in that, The specific process of further fusing the results of water level data collected from various data sources using Kalman filtering and sliding weighted filtering is as follows: The result is obtained by applying a moving average filter to the water level data collected from each data source i using a pre-Kalman filter. Further fusion yields the comprehensive water level data H(t); the Kalman filter formula is: ， Ki(t) is the Kalman gain, which determines the reference weights for each data source i. Let Ri be the covariance of the prediction error at the current moment, and let Ri(t) be the variance of the observation noise, representing the uncertainty of the water level observation of the i-th data source. Ri(t) follows a Gaussian distribution. Covariance of prediction error at the current time Update using the covariance formula: We obtain: F is the state transition matrix, describing the state numbering from time t-1 to t; Q is the process noise covariance matrix, representing the uncertainty of the model itself. This is the transpose of the state transition matrix; , among them , and These are the filtered water level height data. The change in quantity, rate of change, and acceleration of change from the previous time t-1 to the current time t.

6. The method for early warning and monitoring of rainstorm and flood water levels according to claim 1, characterized in that, The specific process of predicting future water levels based on time series analysis to extract fluctuation amplitude and changing trends is as follows: Construct an ARIMA autoregressive integral moving average model, the model formula is as follows: ; Among them Let be the predicted water level height at a future time t, where c is a preset constant bias term, j1 and j2 are autoregressive parameter indices corresponding to past times j1 and j2; P and q are preset autoregressive thresholds, representing the reference range for historical data; where The parameters of the autoregressive term for historical data at time j1 represent the influence coefficients of historical data at time j1 on the predicted water level height at time t; where... Let be the moving average parameter, representing the influence coefficient of the prediction error at time j2 in the past on the predicted water level height at time t in the future; where This refers to the historical water level data from the previous j1 time points before the current time t; where... The difference between the predicted water level and the actual water level at time t (j2 seconds before the current time) is the historical prediction error. The comprehensive water level data H(t) at each moment is used as the historical water level height data j1 moments before the current moment t. Input the autoregressive integral moving average model to obtain the predicted water level data for future times. .

7. The method for early warning and monitoring of rainstorm and flood water levels according to claim 1, characterized in that, The specific process for conducting real-time flood risk assessment is as follows: If the current comprehensive water level data H(t) exceeds the preset threshold, an alarm factor corresponding to time t will be output; if K1 consecutive alarm factors are identified, a real-time flood risk alarm signal will be output.

8. The method for early warning and monitoring of rainstorm and flood water levels according to claim 1, characterized in that, The specific process for conducting a flood risk assessment for future times is as follows: When predicting the water level data at future time t If the value exceeds the second preset threshold, a warning factor corresponding to time t will be output. If k2 consecutive warning factors are identified, a flood risk warning signal will be output. K1 and k2 are preset detection window thresholds.

9. The method for early warning and monitoring of rainstorm and flood water levels according to claim 1, characterized in that, The specific process for issuing multi-channel early warnings is as follows: If a real-time flood risk alarm signal is detected, the alarm process is initiated and an alarm message is sent to the back-end administrator. If a flood risk warning signal is detected, the warning information and the corresponding forecast time t will be issued to guide disaster prevention and emergency response measures. Alarm and warning information will be rapidly disseminated to relevant units, the public, and back-end administrators through communication channels including SMS, email, app push notifications, websites, and broadcasts.

Images