Smart water conservancy management system and method based on Internet of Things

Abstract

The present application provides a smart water management system and method based on the Internet of Things, which relates to the field of water management. It collects the time queue of the water level values ​​of the river sections through water level sensors deployed in each key river section of the basin, and uses data analysis technology based on deep learning to perform the time series correlation of the water level values ​​of each river section, so as to automatically evaluate the probability level of flood occurrence based on the inference features obtained by the spatiotemporal transmission between the time series correlation features of the water level values ​​of each river section, and determine whether to generate a flood early warning prompt. In this way, the mutual influence between different river sections can be utilized to more comprehensively evaluate the possibility of flood occurrence, so as to make up for the limitation of traditional methods that only focus on a single river section and improve the accuracy of the evaluation results.

Description

Technical Field

[0001] The present application relates to the field of water management, and more specifically, to a smart water management system and method based on the Internet of Things. Background Art

[0002] As global climate change leads to frequent extreme weather events, rainstorms and heavy precipitation are increasing, and with the acceleration of urbanization, a large number of green spaces have been replaced by hardened roads, reducing the natural infiltration of rainwater, making cities and surrounding areas more vulnerable to floods. These changes have not only led to a significant increase in the frequency of floods, but also increased the intensity and destructive power of disasters, posing severe challenges to people's safety of life and property and social and economic development. Therefore, through effective water management, disasters caused by floods can be prevented and mitigated, and floods can be effectively controlled to reduce casualties and economic losses.

[0003] However, traditional water management methods often only focus on water level changes in a single river section, while ignoring the spatiotemporal correlation between different river sections. In fact, water level changes in rivers are a complex spatiotemporal dynamic process. The rising water level in the upstream river section will be transmitted along the river channel, affecting the water level in the downstream river section. This spatiotemporal transmission effect means that water level changes in one place may trigger a chain reaction, causing the water level in the downstream area to rise rapidly, increasing the risk of floods. However, if we only focus on the changes in a single river section, we cannot fully assess the possibility of floods, which will affect the accuracy of the assessment results.

[0004] Therefore, an optimized water management solution is desired. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a smart water management system and method based on the Internet of Things, which collects the time queue of the water level values ​​of the river sections through water level sensors deployed in each key river section of the basin, and uses data analysis technology based on deep learning to perform time series correlation of the water level values ​​of each river section, so as to automatically evaluate the probability level of flood occurrence according to the inference features obtained by the spatiotemporal transfer between the time series correlation features of the water level values ​​of each river section, and determine whether to generate a flood early warning prompt. In this way, the mutual influence between different river sections can be utilized to more comprehensively evaluate the possibility of flood occurrence, so as to make up for the limitation of traditional methods that only focus on a single river section and improve the accuracy of the evaluation results.

[0006] According to one aspect of the present application, a smart water management system based on the Internet of Things is provided, which includes:

[0007] The river section water level collection module is used to collect the time queue of river section water level values ​​through water level sensors deployed in each key river section of the basin;

[0008] A river section water level time series encoding module is used to perform river section water level sequence encoding on the time queue of the river section water level values ​​of each key river section to obtain a set of implicitly associated feature vectors of the river section water level time series;

[0009] A river water level spatiotemporal transfer module is used to perform spatiotemporal transfer correlation of river water level on the set of implicit correlation feature vectors of the river water level time series to obtain a spatiotemporal transfer correlation representation of the river water level;

[0010] A flood warning module, used to obtain a warning result based on the spatiotemporal transmission correlation representation of the river water level, and determine whether to generate a flood early warning prompt based on the warning result;

[0011] Wherein, the river water level spatiotemporal transmission module includes:

[0012] A current historical river section water level characteristic determination unit, used to determine a sequence of the current river section water level time series implicit correlation characteristic vector and the historical river section water level time series implicit correlation characteristic vector based on the set of the river section water level time series implicit correlation characteristic vector;

[0013] A historical river section water level static energy value calculation unit, used to calculate the river section water level static energy value of each historical river section water level time series implicitly associated characteristic vector in the sequence of the historical river section water level time series implicitly associated characteristic vectors;

[0014] The historical current river section water level spatiotemporal transfer reasoning unit is used to perform characteristic message propagation reasoning on the sequence of historical river section water level time series implicit correlation feature vectors and the current river section water level time series implicit correlation feature vectors based on the river section water level static energy value of the historical river section water level time series implicit correlation feature vectors to obtain the river water level spatiotemporal transfer correlation representation.

[0015] In the above-mentioned smart water management system based on the Internet of Things, the river section water level time series encoding module is used to: input the time queue of the river section water level values ​​of each key river section into the river section water level sequence encoder based on the Bi-LSTM model to obtain a set of implicitly associated feature vectors of the river section water level time series.

[0016] In the above-mentioned smart water management system based on the Internet of Things, the current historical river section water level feature determination unit is used to: extract the current river section water level time series implicit correlation feature vector from the set of the river section water level time series implicit correlation feature vectors, and define other river section water level time series implicit correlation feature vectors in the set of the river section water level time series implicit correlation feature vectors as historical river section water level time series implicit correlation feature vectors to obtain a sequence of the historical river section water level time series implicit correlation feature vectors.

[0017] In the above-mentioned smart water conservancy management system based on the Internet of Things, the historical river section water level static energy value calculation unit is used to: extract the maximum value of the historical river section water level time series implicit correlation feature vector to obtain the maximum value of the historical river section water level time series implicit correlation feature; respectively calculate the mean and variance of the historical river section water level time series implicit correlation feature vector to obtain the mean value of the historical river section water level time series implicit correlation feature and the variance of the historical river section water level time series implicit correlation feature; compare the historical river section water level time series implicit correlation feature vector with the historical river section water level time series implicit correlation feature mean by position Subtract, calculate the fourth power of the characteristic value of each position in the obtained historical river section water level time series difference vector to obtain the historical river section water level time series difference modulation vector; calculate the expected value of the historical river section water level time series difference modulation vector, and divide the obtained expected value by the square of the variance of the historical river section water level time series implicit correlation feature to obtain the historical river section water level time series energy value; add the maximum value of the historical river section water level time series implicit correlation feature to the historical river section water level time series energy value to obtain the static energy value of the river section water level corresponding to the historical river section water level time series implicit correlation feature vector.

[0018] In the above-mentioned smart water management system based on the Internet of Things, the historical current river section water level spatiotemporal transfer reasoning unit includes: a river section water level dynamic energy spatiotemporal transfer value calculation subunit, which is used to calculate the river section water level dynamic energy spatiotemporal transfer value of the river section water level of each historical river section water level time series implicitly associated characteristic vector relative to the current river section water level time series implicitly associated characteristic vector based on the river section water level static energy value of the historical river section water level time series implicitly associated characteristic vector; a river section water level dynamic energy spatiotemporal transfer weight calculation subunit, which is used to input the river section water level dynamic energy spatiotemporal transfer value of the river section water level of each historical river section water level time series implicitly associated characteristic vector into the energy based on the mask function. A transmission gating module is used to obtain a sequence of spatiotemporal transfer weights of dynamic energy of river section water level; a historical river section water level time series transfer aggregation modulation vector generation subunit is used to calculate the weighted sum of the sequence of historical river section water level time series implicit correlation feature vectors based on the sequence of spatiotemporal transfer weights of dynamic energy of river section water level to obtain the historical river section water level time series transfer aggregation modulation vector; a river water level spatiotemporal transfer association representation generation subunit is used to calculate the positional sum of the historical river section water level time series transfer aggregation modulation vector and the current river section water level time series implicit correlation feature vector to obtain the river water level spatiotemporal transfer association inference representation vector as the river water level spatiotemporal transfer association representation.

[0019] In the above-mentioned smart water management system based on the Internet of Things, the river section water level dynamic energy spatiotemporal transfer value calculation subunit includes: a river section water level energy spatial attenuation factor calculation secondary subunit, which is used to determine the river section water level energy spatial attenuation factor based on the spatial span coefficient between the implicit correlation feature vectors of the time series of each historical river section water level and the implicit correlation feature vector of the time series of the current river section water level; a river section water level energy time attenuation factor calculation secondary subunit, which is used to calculate the river section water level energy spatial attenuation factor based on the spatial span coefficient between the implicit correlation feature vectors of the time series of each historical river section water level and the implicit correlation feature vector of the time series of the current river section water level; The time span coefficient between the implicitly associated characteristic vectors of the historical river section water level time series is used to determine the time decay factor of the river section water level energy; the spatiotemporal transfer value result of the dynamic energy of the river section water level generates a secondary sub-unit, which is used to perform spatiotemporal modulation on the static energy value of the river section water level of each historical river section water level time series implicitly associated characteristic vector based on the spatial attenuation factor of the river section water level energy and the time decay factor of the river section water level energy to obtain the spatiotemporal transfer value of the dynamic energy of the river section water level of each river section water level time series implicitly associated characteristic vector relative to the current river section water level time series implicitly associated characteristic vector.

[0020] In the above-mentioned smart water management system based on the Internet of Things, the secondary sub-unit for calculating the spatial attenuation factor of the river section water level energy is used to: calculate the square of the spatial span value between the historical river section water level time series implicit correlation feature vector and the current river section water level time series implicit correlation feature vector to obtain the river section water level spatial span coefficient; divide the river section water level static energy value corresponding to the historical river section water level time series implicit correlation feature vector by the river section water level spatial span coefficient to obtain the river section water level energy spatial attenuation factor.

[0021] In the above-mentioned smart water management system based on the Internet of Things, the secondary sub-unit for calculating the river section water level energy time decay factor is used to: subtract the river section water level timestamp of the current river section water level time series implicitly associated characteristic vector from the river section water level timestamp of the historical river section water level time series implicitly associated characteristic vector, and then round down to obtain the river section water level time span value; divide the historical river section water level time span value by the time decay parameter to obtain the river section water level time series span energy decay coefficient; use the river section water level time series span energy decay coefficient as the exponential power, calculate the natural exponential function value with the natural constant e as the base to obtain the river section water level time series span energy decay exponential coefficient; divide the river section water level static energy value corresponding to the historical river section water level time series implicitly associated characteristic vector by the river section water level time series span energy decay exponential coefficient to obtain the river section water level energy time decay factor.

[0022] In the above-mentioned smart water management system based on the Internet of Things, the flood warning module is used to:

[0023] The river water level spatiotemporal transfer association reasoning representation vector is input into a classifier-based flood warning module to obtain the warning result, and the warning result is used to represent the probability level of a flood; in response to the warning result being that the probability level of a flood is high, the flood early warning prompt is generated.

[0024] According to another aspect of the present application, a smart water management method based on the Internet of Things is provided, which includes:

[0025] The time sequence of river water level values ​​is collected through water level sensors deployed at each key river section in the basin;

[0026] Performing river section water level sequence encoding on the time queue of the river section water level values ​​of each key river section to obtain a set of river section water level time series implicit correlation feature vectors;

[0027] Performing spatiotemporal transfer correlation of river section water level on the set of implicit correlation feature vectors of river section water level time series to obtain a spatiotemporal transfer correlation representation of river water level;

[0028] Based on the spatiotemporal transmission correlation representation of the river water level, an early warning result is obtained, and based on the early warning result, it is determined whether to generate a flood early warning prompt;

[0029] The process of performing spatiotemporal transfer correlation of the river section water level on the set of implicitly associated feature vectors of the river section water level time series to obtain a spatiotemporal transfer correlation representation of the river water level includes:

[0030] Based on the set of implicitly associated characteristic vectors of the river section water level time series, determining a sequence of implicitly associated characteristic vectors of the current river section water level time series and implicitly associated characteristic vectors of the historical river section water level time series;

[0031] Calculating the static energy value of the river section water level of each historical river section water level time series implicitly associated characteristic vector in the sequence of the historical river section water level time series implicitly associated characteristic vectors;

[0032] Based on the static energy value of the river section water level of each historical river section water level time series implicit correlation feature vector, characteristic message propagation inference is performed on the sequence of the historical river section water level time series implicit correlation feature vector and the current river section water level time series implicit correlation feature vector to obtain the spatiotemporal transmission correlation representation of the river water level.

[0033] Compared with the prior art, the IoT-based smart water management system and method provided by the present application collects the time queue of the water level values ​​of the river sections through water level sensors deployed in each key river section of the basin, and uses data analysis technology based on deep learning to perform time series correlation of the water level values ​​of each river section, so as to automatically evaluate the probability level of flood occurrence based on the inference features obtained by the spatiotemporal transmission between the time series correlation features of the water level values ​​of each river section, and determine whether to generate a flood early warning prompt. In this way, the mutual influence between different river sections can be utilized to more comprehensively evaluate the possibility of flood occurrence, so as to make up for the limitation of traditional methods that only focus on a single river section and improve the accuracy of the evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0035] Figure 1 This is a system block diagram of an IoT-based smart water management system according to an embodiment of the present application.

[0036] Figure 2 This is a data flow diagram of an IoT-based smart water management system according to an embodiment of the present application.

[0037] Figure 3 It is a block diagram of the spatiotemporal transmission module of the river water level in the smart water management system based on the Internet of Things according to the embodiment of the present application.

[0038] Figure 4 It is a block diagram of the spatiotemporal transfer reasoning unit of the historical current river section water level in the smart water management system based on the Internet of Things according to an embodiment of the present application.

[0039] Figure 5 This is a block diagram of a subunit for calculating the spatiotemporal transfer value of dynamic energy of a river section water level in an IoT-based smart water management system according to an embodiment of the present application.

[0040] Figure 6 This is a flow chart of a smart water management method based on the Internet of Things according to an embodiment of the present application. DETAILED DESCRIPTION

[0041] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0042] Due to global climate change, there are more rainstorms and heavy rainfall in extreme weather, and the urbanization process has led to the replacement of green spaces with hardened roads, which in turn has led to a decrease in the natural infiltration of rainwater, making cities and their surrounding areas more vulnerable to floods. Floods will bring severe challenges to the safety of life and property and social and economic development. Effective water management can prevent and reduce flood disasters, reduce casualties and economic losses. However, traditional water management methods usually only focus on the water level changes in a single river section, and often ignore the spatiotemporal correlation between different river sections. In fact, the change of river water level is a complex spatiotemporal dynamic process. The rise of water levels in the upstream will affect the downstream. This spatiotemporal transmission effect will cause a water level change in one place to trigger a chain reaction. Only focusing on the changes in a single river section cannot fully assess the possibility of floods, which will affect the accuracy of the assessment.

[0043] In response to the above technical problems, the technical concept of this application is to collect the time queue of the water level values ​​of the river sections through the water level sensors deployed in each key river section of the basin, and use data analysis and processing technology based on deep learning to perform the time series correlation of the water level values ​​of each river section, so as to automatically evaluate the probability level of flood occurrence according to the inference characteristics obtained by the spatiotemporal transmission between the time series correlation characteristics of the water level values ​​of each river section, and determine whether to generate a flood early warning prompt. In this way, the mutual influence between different river sections can be utilized to more comprehensively evaluate the possibility of flood occurrence, so as to make up for the limitation of traditional methods that only focus on a single river section and improve the accuracy of the evaluation results. At the same time, the dynamic propagation process of the water level time series changes can be revealed, so as to better predict the development trend of floods and realize the intelligentization of water conservancy management.

[0044] Figure 1 This is a system block diagram of an IoT-based smart water management system according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the smart water management system based on the Internet of Things according to an embodiment of the present application. Figure 1 and Figure 2 As shown, in the smart water management system 100 based on the Internet of Things, it includes: a river section water level acquisition module 110, which is used to collect the time queue of river section water level values ​​through water level sensors deployed in each key river section of the basin; a river section water level time series encoding module 120, which is used to perform river section water level sequence encoding on the time queue of river section water level values ​​of each key river section to obtain a set of river section water level time series implicit correlation feature vectors; a river water level spatiotemporal transfer module 130, which is used to perform river section water level spatiotemporal transfer association on the set of river section water level time series implicit correlation feature vectors to obtain a river water level spatiotemporal transfer association representation; a flood warning module 140, which is used to obtain a warning result based on the river water level spatiotemporal transfer association representation, and based on the warning result, determine whether to generate a flood early warning prompt.

[0045] In the embodiment of the present application, the river section water level acquisition module 110 is used to collect the time queue of the river section water level value through the water level sensors deployed in each key river section of the basin. It should be understood that by collecting the time queue data of the river section water level value, the water level changes in different river sections of the basin can be monitored in real time, which is conducive to timely detection of water level abnormalities and early identification of flood risks. In a specific implementation method of the embodiment of the present application, the water level sensors deployed in each key river section of the basin through the Internet of Things technology can collect water level data of different river sections in real time. These water level sensors are usually equipped with wireless communication modules, which can transmit the collected water level data to the local server in real time. Such sensor data obtained through the Internet of Things technology is usually centrally managed and analyzed in the server.

[0046] In an embodiment of the present application, the river section water level time series encoding module 120 is used to perform river section water level sequence encoding on the time queue of the river section water level values ​​of each key river section to obtain a set of implicitly associated feature vectors of the river section water level time series. Specifically, in an embodiment of the present application, the river section water level time series encoding module is used to: input the time queue of the river section water level values ​​of each key river section into the river section water level sequence encoder based on the Bi-LSTM model to obtain a set of implicitly associated feature vectors of the river section water level time series. It should be understood that the river section water level values ​​of each key river section have different time series characteristics at different time scales, and the different time series characteristics reflect the potential time series patterns and trends, which is crucial for understanding and predicting the occurrence of floods. Taking into account that the Bi-LSTM model can simultaneously consider the past and future information of time series data, and thus more comprehensively capture the temporal characteristics of the input time series data, based on this, in the technical solution of the present application, the time queue of the river section water level values ​​of each key river section is input into the river section water level sequence encoder based on the Bi-LSTM model to obtain a set of implicitly associated feature vectors of the river section water level time series.

[0047] In an embodiment of the present application, the river water level spatiotemporal transfer module 130 is used to perform spatiotemporal transfer association of river section water level on the set of implicitly associated characteristic vectors of the river section water level time series to obtain a spatiotemporal transfer association representation of the river water level. Accordingly, considering that the water level change of a river is not only affected by the current river section, but also by the water level change of the upstream river section, at the same time, the water level change of a river section may also have an impact on the downstream river section. That is, the water level change of each river section is often affected by the adjacent river section. Based on this, in the technical solution of the present application, the river section water level spatiotemporal transfer association is performed on the set of implicitly associated characteristic vectors of the river section water level time series to obtain a river water level spatiotemporal transfer association inference representation vector as a river water level spatiotemporal transfer association representation.

[0048] Specifically, firstly, the current node and the historical node are extracted from the set of implicitly associated feature vectors of the water level time series of the river section to obtain the sequence of implicitly associated feature vectors of the water level time series of the current river section and the implicitly associated feature vectors of the water level time series of the historical river section. That is, by extracting the implicitly associated feature vectors of the water level time series of the current river section, the water level status at the current moment and its relationship with the surrounding environment can be obtained, which is crucial for real-time monitoring of the water level status of the river section and can help to detect abnormalities in time and respond. The sequence of implicitly associated feature vectors of the water level time series of the historical river section provides information about the changes in water levels over a period of time in the past, and can identify the trend, periodicity and other factors of the water level changes, providing a basis for subsequent time series reasoning. Then, the static energy value of each historical feature is calculated, and in particular, the static energy value is related to the mean, variance and maximum value of each historical feature. That is, by calculating the static energy value, the importance of each historical feature can be evaluated, and the static energy value can be used to measure the significance of a feature in the overall data set space, which helps to screen out the most valuable historical features for the current analysis and helps to understand the static energy value of each node in the spatial and temporal dimensions. Then, the energy space attenuation factor is determined based on the spatial span coefficient between the feature vectors of each historical node and the feature vector of the current node, and the energy time attenuation factor is determined based on the time span coefficient between the feature vectors of each historical node and the feature vector of the current node. It can be understood that by introducing the energy time attenuation factor and the energy space attenuation factor, the model can adjust the weight of information exchange according to the distance and time difference between nodes, thereby enhancing the model's sensitivity to geographical proximity and time urgency, and ensuring that the information transfer process is closer to the actual hydrological dynamics. Subsequently, based on the energy space attenuation factor and the energy time attenuation factor, the static energy value of the feature vector of each historical node is spatiotemporally modulated to obtain the dynamic energy spatiotemporal transfer value of the river section water level. That is, the spatial span coefficient reflects the spatial distance between the current river section and its historical nodes. By introducing the spatial attenuation factor, the weight of the historical node feature vector in the current node feature analysis can be adjusted. The historical nodes farther away have less influence on the current node, so their contribution should be weakened accordingly; the time span coefficient reflects the time interval between the current time node and its historical nodes. As time goes by, the correlation of historical data gradually weakens. By introducing the time attenuation factor, it can ensure that more recent historical data has a greater impact on the current analysis, while older data gradually weakens its influence, thereby strengthening the model's ability to capture dynamic changes in time series. Furthermore, in order to more accurately determine the importance of each spatiotemporal transfer value, the dynamic energy spatiotemporal transfer value of each historical node feature vector is input into the energy transfer gating module model based on the mask function for masking so that the model focuses on the historical node features that have the greatest impact on the current river section water level change, thereby reducing the interference of irrelevant information.Afterwards, each masked value is used as a weight to transfer and aggregate each historical node feature to integrate the information of different historical nodes, and the historical river section water level time series transfer aggregation modulation vector is obtained. The historical aggregation modulation vector and the current node feature are added by position to the river water level spatiotemporal transfer association reasoning representation vector to obtain the river water level spatiotemporal transfer association reasoning representation vector that integrates the current and historical water level information.

[0049] Specifically, Figure 3 FIG. 1 is a block diagram of a river water level spatiotemporal transmission module in an IoT-based smart water management system according to an embodiment of the present application. Figure 3 As shown, the river water level spatiotemporal transfer module 130 includes: a current historical river section water level feature determination unit 131, which is used to determine the current river section water level time series implicitly associated feature vector and the sequence of historical river section water level time series implicitly associated feature vectors based on the set of the river section water level time series implicitly associated feature vectors; a historical river section water level static energy value calculation unit 132, which is used to calculate the river section water level static energy value of each historical river section water level time series implicitly associated feature vector in the sequence of the historical river section water level time series implicitly associated feature vectors; a historical current river section water level spatiotemporal transfer inference unit 133, which is used to perform feature message propagation inference on the sequence of the historical river section water level time series implicitly associated feature vectors and the current river section water level time series implicitly associated feature vector based on the river section water level static energy value of each historical river section water level time series implicitly associated feature vector to obtain the river water level spatiotemporal transfer association representation.

[0050] More specifically, in an embodiment of the present application, the current historical river section water level feature determination unit is used to: extract the current river section water level timing implicit correlation feature vector from the set of the river section water level timing implicit correlation feature vectors, and define other river section water level timing implicit correlation feature vectors in the set of the river section water level timing implicit correlation feature vectors as historical river section water level timing implicit correlation feature vectors to obtain a sequence of the historical river section water level timing implicit correlation feature vectors.

[0051] More specifically, in the embodiment of the present application, the historical river section water level static energy value calculation unit is used to: extract the maximum value of the historical river section water level time series implicit correlation feature vector to obtain the maximum value of the historical river section water level time series implicit correlation feature; respectively calculate the mean and variance of the historical river section water level time series implicit correlation feature vector to obtain the mean value of the historical river section water level time series implicit correlation feature and the variance of the historical river section water level time series implicit correlation feature; subtract the historical river section water level time series implicit correlation feature vector from the historical river section water level time series implicit correlation feature mean by position , calculate the fourth power of the characteristic value of each position in the obtained historical river section water level time series difference vector to obtain the historical river section water level time series difference modulation vector; calculate the expected value of the historical river section water level time series difference modulation vector, and divide the obtained expected value by the square of the variance of the historical river section water level time series implicit correlation feature to obtain the historical river section water level time series energy value; add the maximum value of the historical river section water level time series implicit correlation feature to the historical river section water level time series energy value to obtain the river section water level static energy value corresponding to the historical river section water level time series implicit correlation feature vector.

[0052] More specifically, Figure 4 FIG. 1 is a block diagram of a spatiotemporal transfer reasoning unit for the historical current river section water level in an IoT-based smart water management system according to an embodiment of the present application. Figure 4 As shown, the historical current river section water level spatiotemporal transfer reasoning unit 133 includes: a river section water level dynamic energy spatiotemporal transfer value calculation subunit 1331, which is used to calculate the river section water level dynamic energy spatiotemporal transfer value of the historical river section water level time series implicitly associated characteristic vector relative to the current river section water level time series implicitly associated characteristic vector based on the river section water level static energy value of the historical river section water level time series implicitly associated characteristic vector; a river section water level dynamic energy spatiotemporal transfer weight calculation subunit 1332, which is used to input the river section water level dynamic energy spatiotemporal transfer value of the historical river section water level time series implicitly associated characteristic vector into the energy transfer gating module based on the mask function To obtain a sequence of spatiotemporal transfer weights of dynamic energy of river section water level; a historical river section water level time series transfer aggregate modulation vector generating subunit 1333 is used to calculate the weighted sum of the sequence of historical river section water level time series implicit correlation feature vectors based on the sequence of spatiotemporal transfer weights of dynamic energy of river section water level to obtain the historical river section water level time series transfer aggregate modulation vector; a river water level spatiotemporal transfer association representation generating subunit 1334 is used to calculate the positional sum of the historical river section water level time series transfer aggregate modulation vector and the current river section water level time series implicit correlation feature vector to obtain the river water level spatiotemporal transfer association inference representation vector as the river water level spatiotemporal transfer association representation.

[0053] More specifically, Figure 5The block diagram is a subunit for calculating the dynamic energy spatiotemporal transfer value of the river water level in the smart water conservancy management system based on the Internet of Things according to an embodiment of the present application. Figure 5 As shown, the river section water level dynamic energy spatiotemporal transfer value calculation subunit 1331 includes: a river section water level energy spatial attenuation factor calculation secondary subunit 13311, which is used to determine the river section water level energy spatial attenuation factor based on the spatial span coefficient between the implicit correlation feature vectors of the water level time series of each historical river section and the implicit correlation feature vector of the water level time series of the current river section; a river section water level energy time attenuation factor calculation secondary subunit 13312, which is used to calculate the river section water level energy spatial attenuation factor based on the spatial span coefficient between the implicit correlation feature vectors of the water level time series of each historical river section and the implicit correlation feature vector of the water level time series of the current river section The time span coefficient between the associated characteristic vectors determines the time attenuation factor of the river section water level energy; the river section water level dynamic energy spatiotemporal transfer value result generates a secondary sub-unit 13313, which is used to perform spatiotemporal modulation on the river section water level static energy value of each historical river section water level time series implicitly associated characteristic vector based on the river section water level spatial attenuation factor and the river section water level energy time attenuation factor to obtain the river section water level dynamic energy spatiotemporal transfer value of each river section water level time series implicitly associated characteristic vector relative to the current river section water level time series implicitly associated characteristic vector.

[0054] More specifically, in an embodiment of the present application, the secondary sub-unit for calculating the spatial attenuation factor of the river section water level energy is used to: calculate the square of the spatial span value between the historical river section water level time series implicit correlation feature vector and the current river section water level time series implicit correlation feature vector to obtain the river section water level spatial span coefficient; divide the river section water level static energy value corresponding to the historical river section water level time series implicit correlation feature vector by the river section water level spatial span coefficient to obtain the river section water level energy spatial attenuation factor.

[0055] More specifically, in an embodiment of the present application, the secondary sub-unit for calculating the river section water level energy time decay factor is used to: subtract the river section water level timestamp of the current river section water level time series implicitly associated characteristic vector from the river section water level timestamp of the historical river section water level time series implicitly associated characteristic vector, and then round down to obtain the river section water level time span value; divide the historical river section water level time span value by the time decay parameter to obtain the river section water level time series span energy decay coefficient; use the river section water level time series span energy decay coefficient as the exponential power, calculate the natural exponential function value with the natural constant e as the base to obtain the river section water level time series span energy decay exponential coefficient; divide the river section water level static energy value corresponding to the historical river section water level time series implicitly associated characteristic vector by the river section water level time series span energy decay exponential coefficient to obtain the river section water level energy time decay factor.

[0056] More specifically, in an embodiment of the present application, the river section water level dynamic energy spatiotemporal transfer value result generates a secondary sub-unit, which is used to: multiply the river section water level energy spatial attenuation factor and the river section water level energy temporal attenuation factor of the historical river section water level time series implicitly associated characteristic vector with the first modulation parameter and the second modulation parameter respectively, and add the obtained river section water level energy spatial attenuation modulation factor and river section water level energy temporal attenuation modulation factor to obtain the river section water level dynamic energy spatiotemporal transfer value.

[0057] More specifically, in an embodiment of the present application, the river section water level dynamic energy spatiotemporal transfer weight calculation subunit is used to: in response to the river section water level dynamic energy spatiotemporal transfer value being greater than a predetermined threshold, input the river section water level dynamic energy spatiotemporal transfer value into a sigmoid function to obtain a value as the river section water level dynamic energy spatiotemporal transfer weight.

[0058] In the embodiment of the present application, specifically, the river water level spatiotemporal transfer module is used to: perform spatiotemporal transfer association of the river section water level on the set of implicitly associated feature vectors of the river section water level time series according to the following formula; wherein the formula is expressed as:

[0059] ;

[0060] in, is the set of implicitly associated feature vectors of the water level time series of the river section, are the first, second, ..., and third implicitly associated feature vectors of the water level time series of the river section. ,..., The implicit correlation feature vector of the water level time series of each river section, It is The eigenvalues ​​of each position in the implicitly associated eigenvector of the water level time series of each river section, To calculate the expected value, and They are The mean and variance of For extraction The maximum value in yes The corresponding static energy value of the river section water level, express and The space span value between yes and The spatial attenuation factor of the water level energy in the river section between and Distribution represents and The timestamp of the implicitly associated eigenvector of the water level time series of each river section, To round down, is the time decay parameter, represents the exponential function value with the natural constant e as the base, yes and The time attenuation factor of the water level energy in the river section between and is the corresponding modulation parameter, yes and The dynamic energy spatiotemporal transfer value of the river water level between For masking, yes function, is a predetermined threshold, is the number of vectors in the set of implicitly associated eigenvectors of the river section water level time series, It is the representation vector of the spatiotemporal transmission association reasoning of the river water level.

[0061] In the embodiment of the present application, the flood warning module 140 is used to obtain a warning result based on the spatiotemporal transfer association representation of the river water level, and determine whether to generate a flood early warning prompt based on the warning result. Specifically, in the embodiment of the present application, the flood warning module is used to: input the spatiotemporal transfer association reasoning representation vector of the river water level into the flood warning module based on the classifier to obtain the warning result, and the warning result is used to represent the probability level of the occurrence of a flood; in response to the warning result that the probability level of the occurrence of a flood is a high probability, the flood early warning prompt is generated. That is, the spatiotemporal transfer association representation of the river water level obtained by spatiotemporal transfer association using the set of implicit association feature vectors of the river water level time series is classified and processed, so as to automatically evaluate the probability level of the occurrence of a flood and determine whether to generate a flood early warning prompt. In this way, the mutual influence between different river sections can be captured, so as to more comprehensively evaluate the possibility of flood occurrence, so as to make up for the limitation of the traditional method that only focuses on a single river section, and improve the accuracy of the evaluation result. At the same time, the dynamic propagation process of water level changes can be revealed, so as to better predict the development trend of floods and realize the intelligentization of water conservancy management.

[0062] In particular, considering that each of the implicit temporal association feature vectors of the water level of the river section in the set of the implicit temporal association feature vectors of the water level of the river section respectively represents the long-range and short-range bidirectional temporal semantic association features of the water level value of the river section of each key river section. When the set of the implicit temporal association feature vectors of the water level of the river section is input into the characteristic message propagation network based on the spatiotemporal transmission characteristics of the nodes, the spatial topological distribution between each key river section is fitted by the spatial span between the implicit temporal association feature vectors of each river section in the set of the implicit temporal association feature vectors of the water level of the river section. Although this introduces spatial factors, it is different from the spatial topological distribution between the key river sections in the real world. At the same time, the arrangement of the implicit temporal association feature vectors of each river section in the set of the implicit temporal association feature vectors of the water level of the river section affects the accuracy of spatiotemporal transmission. In this way, the spatiotemporal transmission association reasoning representation vector of the river water level has the aggregation propagation difference of the spatiotemporal transmission significance difference, so it is expected to improve the detail temporal semantic aggregation propagation expression effect of the spatiotemporal transmission association reasoning representation vector of the river water level.

[0063] Preferably, inputting the river water level spatiotemporal transfer association reasoning representation vector into a classifier-based flood warning module to obtain a warning result includes: calculating the sum of the absolute values ​​of each eigenvalue of the river water level spatiotemporal transfer association reasoning representation vector to obtain a first river water level spatiotemporal transfer association reasoning and modulation value, and calculating the square root of the sum of the squares of each eigenvalue of the river water level spatiotemporal transfer association reasoning representation vector to obtain a second river water level spatiotemporal transfer association reasoning and modulation value; performing dot-wise subtraction of the river water level spatiotemporal transfer association reasoning representation vector and the second river water level spatiotemporal transfer association reasoning and modulation value, and then performing dot-wise multiplication with the number of eigenvalues ​​of the river water level spatiotemporal transfer association reasoning representation vector and the reciprocal of the first river water level spatiotemporal transfer association reasoning and modulation value, and taking the reciprocal of each eigenvalue to obtain the first river water level spatiotemporal transfer association reasoning and modulation value. transfer association reasoning phase conversion vector; after point subtraction of the river water level spatiotemporal transfer association reasoning representation vector and the first river water level spatiotemporal transfer association reasoning and modulation value, respectively perform point multiplication with the square root of the number of eigenvalues ​​of the river water level spatiotemporal transfer association reasoning representation vector and the reciprocal of the second river water level spatiotemporal transfer association reasoning and modulation value, and take the reciprocal of each eigenvalue to obtain the second river water level spatiotemporal transfer association reasoning phase conversion vector; point subtraction of the first river water level spatiotemporal transfer association reasoning phase conversion vector by the point product vector of the second river water level spatiotemporal transfer association reasoning phase conversion vector and the weighted hyperparameter to obtain the optimized river water level spatiotemporal transfer association reasoning representation vector; the optimized river water level spatiotemporal transfer association reasoning representation vector is input into the classifier-based flood warning module to obtain the warning result.

[0064] Here, the optimized representation of the spatiotemporal transfer association reasoning representation vector of the river water level is:

[0065] ;

[0066] in, represents the spatiotemporal transfer association inference representation vector of the river water level, The spatiotemporal transfer association inference representation vector of the river water level is represented by The eigenvalues ​​at the positions, represents the number of eigenvalues ​​of the spatiotemporal transfer association inference representation vector of the river water level, represents the temporal and spatial transfer correlation inference and modulation value of the first river water level, represents the spatiotemporal transfer correlation inference and modulation value of the second river water level, It means point multiplication by position. It means to subtract by position point. To take the inverse of each eigenvalue of a vector, represents the phase transformation vector of the spatiotemporal transfer association reasoning of the first river water level, represents the phase transformation vector of the spatiotemporal transfer association inference of the second river water level, represents the weighted hyperparameter, A vector representing the optimized spatiotemporal transmission association reasoning of the river water level.

[0067] In this preferred embodiment, the difference of the characteristic value of the spatiotemporal transfer association reasoning representation vector relative to the feature set of the whole vector of the spatiotemporal transfer association reasoning representation vector of the river water level is used as the semantic change intensity information, and the phase-like conversion corresponding to the position-based intensity modulation is performed through different and modulation representation forms, so that the aggregation enhancement of the semantic change phase perception can improve the axial aggregation receptive field along the feature aggregation direction by performing a spatial translation operation based on alternating stacking under the scale balance of the vector set of the spatiotemporal transfer association reasoning representation vector of the river water level, thereby improving the perception effect of the aggregation semantics of the spatiotemporal transfer association reasoning representation vector of the river water level for the semantic changes of the time series, so as to improve the expression effect of the spatiotemporal transfer association reasoning representation vector of the river water level, and improve the accuracy of the warning results obtained by the flood warning module based on the classifier. In this way, the mutual influence between different river sections can be captured, so as to more comprehensively evaluate the possibility of flood occurrence, so as to make up for the limitation of the traditional method that only focuses on a single river section, and improve the accuracy of the evaluation results. At the same time, the dynamic propagation process of water level changes can be revealed, so as to better predict the development trend of floods and realize the intelligentization of water conservancy management.

[0068] In summary, the IoT-based smart water management system 100 according to the embodiment of the present application is explained, which collects the time queue of the water level values ​​of the river sections through the water level sensors deployed in each key river section of the basin, and uses the data analysis technology based on deep learning to perform the time series correlation of the water level values ​​of each river section, so as to automatically evaluate the probability level of flood occurrence according to the inference features obtained by the spatiotemporal transmission between the time series correlation features of the water level values ​​of each river section, and determine whether to generate a flood early warning prompt. In this way, the mutual influence between different river sections can be utilized to more comprehensively evaluate the possibility of flood occurrence, so as to make up for the limitation of the traditional method that only focuses on a single river section and improve the accuracy of the evaluation results.

[0069] As described above, the smart water management system 100 based on the Internet of Things according to the embodiment of the present application can be implemented in various terminal devices, such as a server for smart water management based on the Internet of Things. In one example, the smart water management system 100 based on the Internet of Things according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the smart water management system 100 based on the Internet of Things can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the smart water management system 100 based on the Internet of Things can also be one of the many hardware modules of the terminal device.

[0070] Alternatively, in another example, the IoT-based smart water management system 100 and the terminal device may also be separate devices, and the IoT-based smart water management system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0071] Figure 6 FIG. 1 is a flow chart of a smart water management method based on the Internet of Things according to an embodiment of the present application. Figure 6As shown, in the smart water management method based on the Internet of Things, it includes: S110, collecting the time queue of river section water level values ​​through water level sensors deployed in each key river section of the basin; S120, encoding the time queue of river section water level values ​​of each key river section to obtain a set of implicitly associated feature vectors of river section water level time series; S130, performing spatiotemporal transfer association of river section water level on the set of implicitly associated feature vectors of river section water level time series to obtain a spatiotemporal transfer association representation of river water level; S140, obtaining an early warning result based on the spatiotemporal transfer association representation of river water level, and determining whether to generate a flood early warning prompt based on the early warning result; wherein, Step S130 includes: determining the sequence of the implicitly associated characteristic vectors of the water level time series of the current river section and the implicitly associated characteristic vectors of the water level time series of the historical river section based on the set of implicitly associated characteristic vectors of the water level time series of the river section; calculating the static energy value of the water level of the river section of each historical river section water level time series implicitly associated characteristic vector in the sequence of the water level time series implicitly associated characteristic vectors of the water level; and performing characteristic message propagation inference on the sequence of the implicitly associated characteristic vectors of the water level time series of the historical river section and the implicitly associated characteristic vector of the water level time series of the current river section based on the static energy value of the water level of each historical river section water level time series implicitly associated characteristic vector to obtain the spatiotemporal transmission association representation of the river water level.

[0072] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned smart water conservancy management method based on the Internet of Things have been referred to above. Figures 1 to 5 The invention has been introduced in detail in the description of the smart water management system based on the Internet of Things, and therefore, its repeated description will be omitted.

[0073] In summary, the IoT-based smart water management method based on the embodiment of the present application is explained, which collects the time queue of the water level values ​​of the river sections through water level sensors deployed in each key river section of the basin, and uses data analysis technology based on deep learning to perform time series correlation of the water level values ​​of each river section, so as to automatically evaluate the probability level of flood occurrence based on the inference features obtained by the spatiotemporal transmission between the time series correlation features of the water level values ​​of each river section, and determine whether to generate a flood early warning prompt. In this way, the mutual influence between different river sections can be utilized to more comprehensively evaluate the possibility of flood occurrence, so as to make up for the limitation of traditional methods that only focus on a single river section and improve the accuracy of the evaluation results.

[0074] In the several embodiments provided in this application, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0075] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0076] In addition, each functional module in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0077] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the spirit or essential features of the present application.

Claims

1. A smart water management system based on the Internet of Things, characterized in that: include: The river section water level collection module is used to collect the time queue of river section water level values ​​through water level sensors deployed in each key river section of the basin; A river section water level time series encoding module is used to perform river section water level sequence encoding on the time queue of the river section water level values ​​of each key river section to obtain a set of implicitly associated feature vectors of the river section water level time series; A river water level spatiotemporal transfer module is used to perform spatiotemporal transfer correlation of river water level on the set of implicit correlation feature vectors of the river water level time series to obtain a spatiotemporal transfer correlation representation of the river water level; A flood warning module, used to obtain a warning result based on the spatiotemporal transmission correlation representation of the river water level, and determine whether to generate a flood early warning prompt based on the warning result; Wherein, the river water level spatiotemporal transmission module includes: A current historical river section water level characteristic determination unit, used to determine a sequence of the current river section water level time series implicit correlation characteristic vector and the historical river section water level time series implicit correlation characteristic vector based on the set of the river section water level time series implicit correlation characteristic vector; A historical river section water level static energy value calculation unit, used to calculate the river section water level static energy value of each historical river section water level time series implicitly associated characteristic vector in the sequence of the historical river section water level time series implicitly associated characteristic vectors; The historical current river section water level spatiotemporal transfer reasoning unit is used to perform characteristic message propagation reasoning on the sequence of the historical river section water level time series implicitly associated characteristic vectors and the current river section water level time series implicitly associated characteristic vectors based on the river section water level static energy value of the historical river section water level time series implicitly associated characteristic vectors to obtain the river water level spatiotemporal transfer association representation; The river section water level time series encoding module is used to: input the time queue of the river section water level values ​​of each key river section into the river section water level sequence encoder based on the Bi-LSTM model to obtain a set of implicitly associated feature vectors of the river section water level time series; The current historical river section water level feature determination unit is used to: extract the current river section water level time series implicit correlation feature vector from the set of the river section water level time series implicit correlation feature vectors, and define other river section water level time series implicit correlation feature vectors in the set of the river section water level time series implicit correlation feature vectors as historical river section water level time series implicit correlation feature vectors to obtain a sequence of the historical river section water level time series implicit correlation feature vectors; The historical river section water level static energy value calculation unit is used to: Extracting the maximum value of the implicit correlation feature vector of the historical river section water level time series to obtain the maximum value of the implicit correlation feature of the historical river section water level time series; Calculating the mean and variance of the implicit correlation feature vector of the historical river section water level time series respectively to obtain the mean of the implicit correlation feature of the historical river section water level time series and the variance of the implicit correlation feature of the historical river section water level time series; Subtract the implicit correlation feature vector of the historical river section water level time series from the mean value of the implicit correlation feature of the historical river section water level time series by position, and calculate the fourth power of the feature value of each position in the obtained historical river section water level time series difference vector to obtain the historical river section water level time series difference modulation vector; Calculating the expected value of the differential modulation vector of the historical river section water level time series, and dividing the obtained expected value by the square of the implicit correlation characteristic variance of the historical river section water level time series to obtain the energy value of the historical river section water level time series; Adding the maximum value of the implicit correlation feature of the historical river section water level time series to the energy value of the historical river section water level time series to obtain the static energy value of the river section water level corresponding to the implicit correlation feature vector of the historical river section water level time series; The historical current river section water level spatiotemporal transfer reasoning unit includes: The river section water level dynamic energy spatiotemporal transfer value calculation subunit is used to calculate the river section water level dynamic energy spatiotemporal transfer value of the river section water level implicitly associated characteristic vector of each historical river section water level time series relative to the current river section water level implicitly associated characteristic vector based on the river section water level static energy value of the river section water level time series implicitly associated characteristic vector; A river section water level dynamic energy spatiotemporal transfer weight calculation subunit is used to input the river section water level dynamic energy spatiotemporal transfer value of each historical river section water level time series implicitly associated feature vector into the energy transfer gating module based on the mask function to obtain a sequence of river section water level dynamic energy spatiotemporal transfer weights; A historical river section water level time series transfer aggregation modulation vector generating subunit, used for calculating the weighted sum of the sequence of implicitly associated characteristic vectors of the historical river section water level time series based on the sequence of spatiotemporal transfer weights of the river section water level dynamic energy to obtain the historical river section water level time series transfer aggregation modulation vector; A river water level spatiotemporal transfer association representation generating subunit is used to calculate the positional sum of the historical river section water level time series transfer aggregation modulation vector and the current river section water level time series implicit association feature vector to obtain a river water level spatiotemporal transfer association inference representation vector as the river water level spatiotemporal transfer association representation; The river section water level dynamic energy spatiotemporal transfer value calculation subunit includes: The river section water level energy spatial attenuation factor calculation secondary subunit is used to determine the river section water level energy spatial attenuation factor based on the spatial span coefficient between the implicit correlation feature vectors of the water level time series of each historical river section and the implicit correlation feature vector of the water level time series of the current river section; The river section water level energy time decay factor calculation secondary subunit is used to determine the river section water level energy time decay factor based on the time span coefficient between the implicit correlation feature vectors of the water level time series of each historical river section and the implicit correlation feature vector of the water level time series of the current river section; The spatiotemporal transfer value result of the dynamic energy of the river section water level generates a secondary subunit, which is used to perform spatiotemporal modulation on the static energy value of the river section water level of each historical river section water level time series implicitly associated characteristic vector based on the spatial attenuation factor of the river section water level energy and the temporal attenuation factor of the river section water level energy to obtain the spatiotemporal transfer value of the dynamic energy of the river section water level of each historical river section water level time series implicitly associated characteristic vector relative to the current river section water level time series implicitly associated characteristic vector; The second-level subunit for calculating the spatial attenuation factor of the river section water level energy is used for: Calculate the square of the spatial span value between the implicit correlation feature vector of the historical river section water level time series and the implicit correlation feature vector of the current river section water level time series to obtain the river section water level spatial span coefficient; Dividing the static energy value of the river section water level corresponding to the implicitly associated characteristic vector of the historical river section water level time series by the spatial span coefficient of the river section water level to obtain the spatial attenuation factor of the river section water level energy; The river section water level energy time decay factor calculation secondary subunit is used for: Subtracting the river section water level timestamp of the current river section water level time series implicitly associated feature vector from the river section water level timestamp of the historical river section water level time series implicitly associated feature vector and rounding down to obtain a river section water level time span value; Dividing the historical river section water level time span value by the time decay parameter to obtain the river section water level time series span energy decay coefficient; Taking the energy attenuation coefficient of the water level time series span of the river section as the exponential power, calculating the value of the natural exponential function with the natural constant e as the base to obtain the energy attenuation exponential coefficient of the water level time series span of the river section; The static energy value of the river section water level corresponding to the implicitly associated characteristic vector of the historical river section water level time series is divided by the energy decay index coefficient of the river section water level time series span to obtain the energy time decay factor of the river section water level.

2. The smart water management system based on the Internet of Things according to claim 1 is characterized in that: The flood warning module is used to: Inputting the spatiotemporal transfer association reasoning representation vector of the river water level into a flood warning module based on a classifier to obtain the warning result, wherein the warning result is used to represent the probability level of flood occurrence; In response to the warning result that the probability level of flood occurrence is high, the flood early warning prompt is generated.

3. A smart water management method based on the Internet of Things for the smart water management system based on the Internet of Things as claimed in claim 1, characterized in that: include: The time sequence of river water level values ​​is collected through water level sensors deployed at each key river section in the basin; Performing river section water level sequence encoding on the time queue of the river section water level values ​​of each key river section to obtain a set of river section water level time series implicit correlation feature vectors; Performing spatiotemporal transfer correlation of river section water level on the set of implicit correlation feature vectors of river section water level time series to obtain a spatiotemporal transfer correlation representation of river water level; Based on the spatiotemporal transmission correlation representation of the river water level, an early warning result is obtained, and based on the early warning result, it is determined whether to generate a flood early warning prompt; The process of performing spatiotemporal transfer correlation of the river section water level on the set of implicitly associated feature vectors of the river section water level time series to obtain a spatiotemporal transfer correlation representation of the river water level includes: Based on the set of implicitly associated characteristic vectors of the river section water level time series, determining a sequence of implicitly associated characteristic vectors of the current river section water level time series and implicitly associated characteristic vectors of the historical river section water level time series; Calculating the static energy value of the river section water level of each historical river section water level time series implicitly associated characteristic vector in the sequence of the historical river section water level time series implicitly associated characteristic vectors; Based on the static energy value of the river section water level of each historical river section water level time series implicit correlation feature vector, characteristic message propagation inference is performed on the sequence of the historical river section water level time series implicit correlation feature vector and the current river section water level time series implicit correlation feature vector to obtain the spatiotemporal transmission correlation representation of the river water level.

Images