A method and system for reverse deduction of water cycle in a river basin based on artificial intelligence

Abstract

The present invention discloses a method and system for reverse deduction of water cycle in a watershed based on artificial intelligence. The method includes: data collection, expansion of watershed perception range, construction of a probability analysis model, precise time-space matching, and result evaluation and verification. The present invention belongs to the field of intelligent water conservancy technology, and specifically refers to a method and system for reverse deduction of water cycle in a watershed based on artificial intelligence. This solution adopts the method of expanding the perception range of the watershed, specifically building a basic architecture of the perception network, clarifying the information transmission path and node connection, calculating the effective coverage probability of sensor nodes and target points, and dynamically adjusting the distribution of sensor nodes according to the calculation results to optimize the perception network; adopts precise time-space matching, specifically extending the spatial data fusion framework to the spatiotemporal domain, using a constant additive deviation to correct the system error, calculating the latent process to reflect the dependence of space and time, and finally establishing a visual display panel to present the fused spatiotemporal data and dynamically update it.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent water conservancy technology, and specifically is a basin water cycle reverse deduction method and system based on artificial intelligence. Background Art

[0002] The basin water cycle reverse deduction method is a method that uses artificial intelligence, conditional probability analysis and reverse probability analysis to simulate and deduce the process of water and drought disasters in the basin and reversely deduce the influencing factors. At the same time, it combines the big data platform to perform logical deduction and judgment of artificial condition settings.

[0003] However, the existing methods for reverse deduction of river basin water cycle lack large-scale river basin perception data, resulting in inaccurate and incomplete deduction results, making it difficult to truly reflect the actual situation of the river basin water cycle, and affecting the scientific planning of water resources management and water environment protection. There are also technical problems such as difficulty in matching time and space dimensions, and inability to accurately reflect the characteristics and changes of the river basin water cycle at different times and spaces, which limits the construction of digital twin river basins. Summary of the Invention

[0004] In response to the above situation, in order to overcome the defects of the existing technology, the present invention provides a basin water cycle reverse deduction method and system based on artificial intelligence. In order to solve the technical problem that there is a lack of large-scale basin perception data, which leads to inaccurate and incomplete deduction results, making it difficult to truly reflect the actual situation of the basin water cycle, and affecting the scientific planning of water resources management and water environment protection, the basin perception range is expanded. Specifically, the basic architecture of the perception network is constructed, the information transmission path and node connection are clarified, the effective coverage probability of sensor nodes and target points is calculated, and the distribution of sensor nodes is dynamically adjusted according to the calculation results to optimize the perception network; in order to solve the technical problem that it is difficult to match the time and space dimensions, and it is impossible to accurately reflect the characteristics and changes of the basin water cycle in different time and space, which limits the construction of digital twin basins, precise time and space matching is adopted. Specifically, the spatial data fusion framework is extended to the time and space domain, and a constant additive deviation is used to correct the system error. The latent process is calculated to reflect the dependence of space and time. Finally, a visual display panel is established to present the fused time and space data and dynamically update it.

[0005] The technical solution adopted by the present invention is as follows: The present invention provides an artificial intelligence-based water cycle reverse deduction method for a watershed, which includes the following steps:

[0006] Step S1: Data collection, specifically collecting data and performing data preprocessing to obtain a watershed water cycle dataset;

[0007] Step S2: Expand the sensing range of the watershed, specifically by building the basic architecture of the sensing network, clarifying the information transmission path and node connections, calculating the effective coverage probability of sensor nodes and target points, and dynamically adjusting the distribution of sensor nodes based on the calculation results to optimize the sensing network;

[0008] Step S3: Construct a probability analysis model, specifically by setting initial conditions and performing forward calculations. When floods and droughts occur, use reverse probability analysis to infer the impact factors.

[0009] Step S4: precise spatiotemporal matching, specifically, extending the spatial data fusion framework to the spatiotemporal domain, using a constant additive bias to correct the systematic error, calculating the latent process to reflect the spatial and temporal dependencies, and finally establishing a visual display panel to present the fused spatiotemporal data and dynamically update it;

[0010] Step S5: Result evaluation and verification, specifically comparing the deduction results with the actual basin water cycle related data in detail, analyzing the error distribution and verifying the results.

[0011] Furthermore, in step S1, the data collection includes the following steps:

[0012] Step S11: Collect meteorological data, obtain precipitation, temperature, wind speed, and humidity information through meteorological stations and satellites, and use them to analyze precipitation input and evaporation processes in the water cycle;

[0013] Step S12: Collect hydrological data, obtain flow, water level, and water quality data from river monitoring stations and lake monitoring points to understand the flow and changes of water in the basin;

[0014] Step S13: Collect topographic data, using topographic mapping and remote sensing technology to obtain the topographic relief, slope and altitude of the watershed, which have a significant impact on the flow path and potential energy distribution of water;

[0015] Step S14: collecting soil data, including soil texture, moisture content, and permeability, to reflect the soil's ability to retain and transmit water;

[0016] Step S15: Collecting human water use activity data, including domestic water use, industrial water use, drainage data, and water project operation data;

[0017] Step S16: Summarize meteorological data, hydrological data, topographic data, soil data, and human water use activity data, and perform data preprocessing to obtain a watershed water cycle dataset.

[0018] Furthermore, in step S2, the expansion of the watershed sensing range includes the following steps:

[0019] Step S21: Construct the basic architecture of the perception network to clarify the information transmission path and the connection between nodes, laying the foundation for the perception range. The cluster is composed of sensor nodes. Each individual sensor node has the ability to communicate directly with the corresponding cluster head. The spatial interval between the sensor node and the target point is quantified. The formula used is as follows:

[0020] ;

[0021] Where, represents the spatial separation distance between the sensor node and the target point, represents the i-th sensor node, represents the jth target point, represents the horizontal coordinate of the i-th sensor node, represents the horizontal coordinate of the j-th target point, represents the ordinate of the i-th sensor node, Indicates the ordinate of the j-th target point;

[0022] Step S22: Calculate the probability that a single sensor covers the target point. The formula used is as follows:

[0023] ;

[0024] Where, represents the probability of the sensor node covering the target point, γ represents the induction attenuation coefficient, r i Represents the maximum ideal sensing radius of the sensor node, which means that the node has a greater probability of sensing coverage within this radius. e represents the sensing error inherent in a given sensor node;

[0025] Step S23: Calculate the effective coverage probability of the target point. Based on the probability of a single sensor node covering the target point, the probability of the target point being effectively covered is obtained by calculating the complement of the product of the coverage probabilities of all sensor nodes for the target point. The formula used is as follows:

[0026] ;

[0027] Where, represents the probability that the target point is effectively covered, and n represents the total number of sensor nodes;

[0028] Step S24: Optimize the perception network, dynamically adjust the distribution of sensor nodes based on the calculated effective coverage efficiency of the target points, optimize the coverage effect and accuracy of the perception network, continuously monitor and evaluate the performance of the perception network, and further improve the perception network based on feedback information.

[0029] Furthermore, in step S3, constructing the probability analysis model includes the following steps:

[0030] Step S31: Influencing factor analysis, establishing a mathematical model that can analyze the probability of flood and drought disasters and related influencing factors, comprehensively considering various factors that may affect the occurrence of flood and drought disasters, including rainfall, temperature, soil moisture and river flow;

[0031] Step S32: setting initial conditions to clarify the parameter values ​​and states when the model starts calculating, including the initial rainfall value and temperature range within the specified time period;

[0032] Step S33: Perform forward calculations to calculate the probability of flood and drought disasters under the initial conditions and the specific circumstances that may arise based on the set initial conditions and the model rules. Through forward calculations, the probability of disasters occurring under specific conditions and the possible consequences are understood.

[0033] Step S34: Use reverse probability analysis to infer the influencing factors. When a flood or drought disaster has occurred, the probability model is used to infer the influencing factors that caused the disaster. For example, if a severe drought occurs, reverse analysis can be used to find out whether it was caused by too little rainfall in the early stage, too high a temperature, and too fast evaporation of soil moisture. Reverse analysis helps us better understand the causes and mechanisms of the disaster, thereby providing a basis for formulating response strategies and preventive measures.

[0034] Furthermore, in step S4, the precise time-space matching includes the following steps:

[0035] Step S41: Extend the spatial data fusion framework to the spatiotemporal domain so that the model can process information in both spatial and temporal dimensions simultaneously, thereby analyzing and fusing data more comprehensively and accurately. The format used is as follows:

[0036] ;

[0037] Where, It represents the spatiotemporal data result after fusion at spatial position c and time t. c represents the index in the spatial dimension, which is used to distinguish different spatial positions. t represents the specific moment in the time dimension. a represents the deviation term that remains unchanged in space and time, which is understood as a systematic and fixed deviation factor. It is a hidden spatiotemporal process that changes with space and time, that is, a latent process, which contains spatiotemporal dynamic information and laws. It represents the measurement error of remote sensing and is used to reflect the uncertainty of measurement;

[0038] Step S42: A constant additive bias is used to correct the systematic error in the satellite image and improve the data accuracy. The formula used is as follows:

[0039] ;

[0040] Where, It represents the fused spatiotemporal data result after correction at spatial position c1 and time t. It represents the latent process at spatial position c1 and time t after correction when the error is not considered. represents the measurement error of the original data, that is, the constant additive bias;

[0041] Step S43: Calculate the latent process to reflect the dependency between space and time. The formula used is as follows:

[0042] ;

[0043] Where, represents the latent process, taking into account the spatial and temporal dependencies, represents variables related to time and space, represents the spatiotemporal variable factor, Represents spatially correlated variables that follow a Gaussian distribution;

[0044] Step S44: Visual display, establish a visualization display panel for spatiotemporal data, present the fused spatiotemporal data in an intuitive form, dynamically update and calibrate the spatiotemporal data in real time to ensure the timeliness and accuracy of the data.

[0045] Furthermore, in step S5, the result evaluation and verification includes the following steps:

[0046] Step S51: Compare the actual observation results and compare the deduced results with the actual basin water cycle related data in detail to check the consistency and deviation degree;

[0047] The water cycle related data of the basin include water level, flow and water quality;

[0048] Step S52: Analyze the error distribution to determine the distribution of the error of the deduction results in different time, space and situations to understand the stability of the model;

[0049] Step S53: Result verification, cross-validation of data from different time periods and different regions to ensure the adaptability of the model under different conditions, and backtesting of the model using historical data to determine whether it can accurately reproduce the past water cycle conditions.

[0050] The artificial intelligence-based watershed water cycle reverse deduction system provided by the present invention includes a data acquisition module, a watershed perception range expansion module, a probability analysis model construction module, a time-space precise matching module, and a result evaluation and verification module;

[0051] The data acquisition module specifically collects data and performs data preprocessing to obtain a watershed water cycle dataset;

[0052] The module for expanding the watershed perception range specifically builds the basic architecture of the perception network, clarifies the information transmission path and node connections, calculates the effective coverage probability of sensor nodes and target points, and dynamically adjusts the distribution of sensor nodes based on the calculation results to optimize the perception network;

[0053] The probability analysis model module is constructed to set initial conditions, perform forward calculations, and use reverse probability analysis to infer the impact factors when floods and droughts occur;

[0054] The spatiotemporal precise matching module specifically extends the spatial data fusion framework to the spatiotemporal domain, uses a constant additive bias to correct systematic errors, calculates the latent process to reflect spatial and temporal dependencies, and finally establishes a visual display panel to present the fused spatiotemporal data and dynamically updates it.

[0055] The result evaluation and verification module specifically compares the deduction results with the actual basin water cycle related data in detail, analyzes the error distribution and verifies the results.

[0056] The beneficial results achieved by the present invention using the above scheme are as follows:

[0057] (1) In order to solve the technical problem that the lack of large-scale watershed sensing data has resulted in inaccurate and incomplete deduction results, making it difficult to truly reflect the actual situation of the water cycle in the watershed and affecting the scientific planning of water resources management and water environment protection, the sensing range of the watershed is expanded. Specifically, the basic architecture of the sensing network is constructed, the information transmission path and node connection are clarified, the effective coverage probability of sensor nodes and target points is calculated, and the distribution of sensor nodes is dynamically adjusted according to the calculation results to optimize the sensing network;

[0058] (2) In order to solve the technical problems that are difficult to match in time and space dimensions and cannot accurately reflect the characteristics and changes of the water cycle in the basin at different times and spaces, which limits the construction of digital twin basins, precise time and space matching is adopted. Specifically, the spatial data fusion framework is extended to the time and space domain, and a constant additive bias is used to correct the system error. The latent process is calculated to reflect the dependence of space and time. Finally, a visual display panel is established to present the fused time and space data and dynamically update it. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A schematic diagram of the flow chart of the basin water cycle reverse deduction method based on artificial intelligence provided by the present invention;

[0060] Figure 2A schematic diagram of the artificial intelligence-based water cycle reverse deduction system provided by the present invention;

[0061] Figure 3 Schematic diagram of the process of step S2;

[0062] Figure 4 Schematic diagram of the process of step S3;

[0063] Figure 5 Schematic diagram of the process of step S4.

[0064] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0065] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0066] In the description of the present invention, it should be understood that terms such as "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting the present invention.

[0067] Example 1, see Figure 1 The present invention provides a method for reverse deduction of water cycle in a watershed based on artificial intelligence, which comprises the following steps:

[0068] Step S1: Data collection, specifically collecting data and performing data preprocessing to obtain a watershed water cycle dataset;

[0069] Step S2: Expand the sensing range of the watershed, specifically by building the basic architecture of the sensing network, clarifying the information transmission path and node connections, calculating the effective coverage probability of sensor nodes and target points, and dynamically adjusting the distribution of sensor nodes based on the calculation results to optimize the sensing network;

[0070] Step S3: Construct a probability analysis model, specifically by setting initial conditions and performing forward calculations. When floods and droughts occur, use reverse probability analysis to infer the impact factors.

[0071] Step S4: precise spatiotemporal matching, specifically, extending the spatial data fusion framework to the spatiotemporal domain, using a constant additive bias to correct the systematic error, calculating the latent process to reflect the spatial and temporal dependencies, and finally establishing a visual display panel to present the fused spatiotemporal data and dynamically update it;

[0072] Step S5: Result evaluation and verification, specifically comparing the deduction results with the actual basin water cycle related data in detail, analyzing the error distribution and verifying the results.

[0073] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the data collection includes the following steps:

[0074] Step S11: Collect meteorological data, obtain precipitation, temperature, wind speed, and humidity information through meteorological stations and satellites, and use them to analyze precipitation input and evaporation processes in the water cycle;

[0075] Step S12: Collect hydrological data, obtain flow, water level, and water quality data from river monitoring stations and lake monitoring points to understand the flow and changes of water in the basin;

[0076] Step S13: Collect topographic data, using topographic mapping and remote sensing technology to obtain the topographic relief, slope and altitude of the watershed, which have a significant impact on the flow path and potential energy distribution of water;

[0077] Step S14: collecting soil data, including soil texture, moisture content, and permeability, to reflect the soil's ability to retain and transmit water;

[0078] Step S15: Collecting human water use activity data, including domestic water use, industrial water use, drainage data, and water project operation data;

[0079] Step S16: Summarize meteorological data, hydrological data, topographic data, soil data, and human water use activity data, and perform data preprocessing to obtain a watershed water cycle dataset.

[0080] Example 3, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S2, the expansion of the watershed sensing range includes the following steps:

[0081] Step S21: Construct the basic architecture of the perception network to clarify the information transmission path and the connection between nodes, laying the foundation for the perception range. The cluster is composed of sensor nodes. Each individual sensor node has the ability to communicate directly with the corresponding cluster head. The spatial interval between sensor node i and target point j is quantified using the following formula:

[0082] ;

[0083] Where, represents the spatial separation distance between the sensor node and the target point, represents the i-th sensor node, represents the jth target point, represents the horizontal coordinate of the i-th sensor node, represents the horizontal coordinate of the j-th target point, represents the ordinate of the i-th sensor node, Indicates the ordinate of the j-th target point;

[0084] Step S22: Calculate the probability that a single sensor covers the target point. The formula used is as follows:

[0085] ;

[0086] Where, represents the probability of the sensor node covering the target point, γ represents the induction attenuation coefficient, r i Represents the maximum ideal sensing radius of the sensor node, which means that the node has a greater probability of sensing coverage within this radius. e represents the sensing error inherent in a given sensor node;

[0087] Step S23: Calculate the effective coverage probability of the target point. Based on the probability of a single sensor node covering the target point, the probability of the target point being effectively covered is obtained by calculating the complement of the product of the coverage probabilities of all sensor nodes for the target point. The formula used is as follows:

[0088] ;

[0089] Where, represents the probability that the target point is effectively covered, and n represents the total number of sensor nodes;

[0090] Step S24: Optimize the perception network, dynamically adjust the distribution of sensor nodes based on the calculated effective coverage efficiency of the target points, optimize the coverage effect and accuracy of the perception network, continuously monitor and evaluate the performance of the perception network, and further improve the perception network based on feedback information.

[0091] By performing the above operations, the basin perception range is expanded. Specifically, the basic architecture of the perception network is constructed, the information transmission path and node connection are clarified, the effective coverage probability of sensor nodes and target points is calculated, and the distribution of sensor nodes is dynamically adjusted according to the calculation results to optimize the perception network. This solves the technical problem that the lack of large-scale basin perception data leads to inaccurate and incomplete deduction results, making it difficult to truly reflect the actual situation of the basin water cycle, and affecting the scientific planning of water resources management and water environment protection.

[0092] Example 4, see Figure 1 and Figure 4 This embodiment is based on the above embodiment. In step S3, the construction of the probability analysis model includes the following steps:

[0093] Step S31: Influencing factor analysis, establishing a mathematical model that can analyze the probability of flood and drought disasters and related influencing factors, comprehensively considering various factors that may affect the occurrence of flood and drought disasters, including rainfall, temperature, soil moisture and river flow;

[0094] Step S32: setting initial conditions to clarify the parameter values ​​and states when the model starts calculating, including the initial rainfall value and temperature range within the specified time period;

[0095] Step S33: Perform forward calculations to calculate the probability of flood and drought disasters under the initial conditions and the specific circumstances that may arise based on the set initial conditions and the model rules. Through forward calculations, the probability of disasters occurring under specific conditions and the possible consequences are understood.

[0096] Step S34: Use reverse probability analysis to infer the influencing factors. When a flood or drought disaster has occurred, the probability model is used to infer the influencing factors that caused the disaster. For example, if a severe drought occurs, reverse analysis can be used to find out whether it was caused by too little rainfall in the early stage, too high a temperature, and too fast evaporation of soil moisture. Reverse analysis helps us better understand the causes and mechanisms of the disaster, thereby providing a basis for formulating response strategies and preventive measures.

[0097] Example 5, see Figure 1 and Figure 5 This embodiment is based on the above embodiment. In step S4, the time-space precise matching includes the following steps:

[0098] Step S41: Extend the spatial data fusion framework to the spatiotemporal domain so that the model can process information in both spatial and temporal dimensions simultaneously, thereby analyzing and fusing data more comprehensively and accurately. The format used is as follows:

[0099] ;

[0100] Where, It represents the spatiotemporal data result after fusion at spatial position c and time t. c represents the index in the spatial dimension, which is used to distinguish different spatial positions. t represents the specific moment in the time dimension. a represents the deviation term that remains unchanged in space and time, which is understood as a systematic and fixed deviation factor. It is a hidden spatiotemporal process that changes with space and time, that is, a latent process, which contains spatiotemporal dynamic information and laws. It represents the measurement error of remote sensing and is used to reflect the uncertainty of measurement;

[0101] Step S42: A constant additive bias is used to correct the systematic error in the satellite image and improve the data accuracy. The formula used is as follows:

[0102] ;

[0103] Where, It represents the fused spatiotemporal data result after correction at spatial position c1 and time t. It represents the latent process at spatial position c1 and time t after correction when the error is not considered. represents the measurement error of the original data, that is, the constant additive bias;

[0104] Step S43: Calculate the latent process to reflect the dependency between space and time. The formula used is as follows:

[0105] ;

[0106] Where, represents the latent process, taking into account the spatial and temporal dependencies, represents variables related to time and space, represents the spatiotemporal variable factor, Represents spatially correlated variables that follow a Gaussian distribution;

[0107] Step S44: Visual display, establish a visualization display panel for spatiotemporal data, present the fused spatiotemporal data in an intuitive form, dynamically update and calibrate the spatiotemporal data in real time to ensure the timeliness and accuracy of the data.

[0108] By performing the above operations, precise space-time matching is adopted. Specifically, the spatial data fusion framework is extended to the space-time domain, a constant additive bias is used to correct the system error, the latent process is calculated to reflect the dependence of space and time, and finally a visual display panel is established to present the fused space-time data and dynamically update it. This solves the technical problems of difficulty in matching the time and space dimensions, inability to accurately reflect the characteristics and changes of the water cycle in the basin at different times and spaces, and limiting the construction of digital twin basins.

[0109] Example 6, see Figure 1 This embodiment is based on the above embodiment. In step S5, the result evaluation and verification includes the following steps:

[0110] Step S51: Compare the actual observation results and compare the deduced results with the actual basin water cycle related data in detail to check the consistency and deviation degree;

[0111] The water cycle related data of the basin include water level, flow and water quality;

[0112] Step S52: Analyze the error distribution to determine the distribution of the error of the deduction results in different time, space and situations to understand the stability of the model;

[0113] Step S53: Result verification, cross-validation of data from different time periods and different regions to ensure the adaptability of the model under different conditions, and backtesting of the model using historical data to determine whether it can accurately reproduce the past water cycle conditions.

[0114] Example 7, see Figure 2 This embodiment is based on the above embodiment. The artificial intelligence-based watershed water cycle reverse deduction system provided by the present invention includes a data acquisition module, a watershed perception range expansion module, a probability analysis model construction module, a time-space precise matching module, and a result evaluation and verification module;

[0115] The data acquisition module specifically collects data and performs data preprocessing to obtain a watershed water cycle dataset;

[0116] The module for expanding the watershed perception range specifically builds the basic architecture of the perception network, clarifies the information transmission path and node connections, calculates the effective coverage probability of sensor nodes and target points, and dynamically adjusts the distribution of sensor nodes based on the calculation results to optimize the perception network;

[0117] The probability analysis model module is constructed to set initial conditions, perform forward calculations, and use reverse probability analysis to infer the impact factors when floods and droughts occur;

[0118] The spatiotemporal precise matching module specifically extends the spatial data fusion framework to the spatiotemporal domain, uses a constant additive bias to correct systematic errors, calculates the latent process to reflect spatial and temporal dependencies, and finally establishes a visual display panel to present the fused spatiotemporal data and dynamically updates it.

[0119] The result evaluation and verification module specifically compares the deduction results with the actual basin water cycle related data in detail, analyzes the error distribution and verifies the results.

[0120] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0121] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

[0122] The present invention and its embodiments are described above. This description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if a person skilled in the art is inspired by this and, without departing from the purpose of the present invention, designs structures and embodiments similar to this technical solution without inventiveness, they shall fall within the scope of protection of the present invention.

Claims

1. The artificial intelligence-based reverse deduction method for watershed water cycle is characterized by: The method comprises the following steps: Step S1: Data collection, specifically collecting data and performing data preprocessing to obtain a watershed water cycle dataset; Step S2: Expand the sensing range of the watershed, specifically by building the basic architecture of the sensing network, clarifying the information transmission path and node connections, calculating the effective coverage probability of sensor nodes and target points, and dynamically adjusting the distribution of sensor nodes based on the calculation results to optimize the sensing network; Step S3: Construct a probability analysis model, specifically by setting initial conditions and performing forward calculations. When floods and droughts occur, use reverse probability analysis to infer the impact factors. Step S4: precise spatiotemporal matching, specifically, extending the spatial data fusion framework to the spatiotemporal domain, using a constant additive bias to correct the systematic error, calculating the latent process to comprehensively characterize the correlation characteristics of space and time, and finally establishing a visual display panel to present the fused spatiotemporal data and dynamically update it; In step S4, the time-space precise matching includes the following steps: Step S41: Extend the spatial data fusion framework to the spatiotemporal domain so that the model can process information in both spatial and temporal dimensions simultaneously. The format used is as follows: ； Where, It represents the spatiotemporal data result after fusion at spatial position c and time t. c represents the index in the spatial dimension, which is used to distinguish different spatial positions. t represents the specific moment in the time dimension. a represents the deviation term that remains unchanged in space and time, which is understood as a systematic and fixed deviation factor. It is a hidden spatiotemporal process that changes with space and time, that is, a latent process, which contains spatiotemporal dynamic information and laws. It represents the measurement error of remote sensing and is used to reflect the uncertainty of measurement; Step S42: A constant additive bias is used to correct the systematic error in the satellite image. The formula used is as follows: ； Where, It represents the fused spatiotemporal data result after correction at spatial position c1 and time t. It represents the latent process at spatial position c1 and time t after correction when the error is not considered. represents the measurement error of the original data, that is, the constant additive bias; Step S43: Calculate the latent process using the following formula: ； Where, represents the latent process, taking into account the spatial and temporal dependencies, represents the variables related to time and space, β represents the time and space variable factor, Represents spatially correlated variables that follow a Gaussian distribution; Step S44: Visualization display, establishing a visualization display panel for spatiotemporal data, presenting the fused spatiotemporal data in an intuitive form, and dynamically updating and calibrating the spatiotemporal data in real time; Step S5: Result evaluation and verification, specifically comparing the deduction results with the actual basin water cycle related data in detail, analyzing the error distribution and verifying the results.

2. The artificial intelligence-based water cycle reverse deduction method according to claim 1 is characterized by: In step S2, the expansion of the watershed sensing range includes the following steps: Step S21: Construct the basic architecture of the perception network to clarify the information transmission path and the connection between nodes, laying the foundation for the perception range. The cluster is composed of sensor nodes. Each individual sensor node has the ability to communicate directly with the corresponding cluster head. The spatial interval between the sensor node and the target point is quantified. The formula used is as follows: ； Where, represents the spatial separation distance between the sensor node and the target point, represents the i-th sensor node, represents the jth target point, represents the horizontal coordinate of the i-th sensor node, represents the horizontal coordinate of the j-th target point, represents the ordinate of the i-th sensor node, Indicates the ordinate of the j-th target point; Step S22: Calculate the probability that a single sensor covers the target point. The formula used is as follows: ； Where, represents the probability of the sensor node covering the target point, γ represents the induction attenuation coefficient, r i represents the maximum ideal sensing radius of the sensor node, r e represents the sensing error inherent in a given sensor node; Step S23: Calculate the effective coverage probability of the target point. Based on the probability of a single sensor node covering the target point, the probability of the target point being effectively covered is obtained by calculating the complement of the product of the coverage probabilities of all sensor nodes for the target point. The formula used is as follows: ； Where, represents the probability that the target point is effectively covered, and n represents the total number of sensor nodes; Step S24: Optimize the perception network, dynamically adjust the distribution of sensor nodes based on the calculated effective coverage efficiency of the target points, optimize the coverage effect and accuracy of the perception network, continuously monitor and evaluate the performance of the perception network, and improve the perception network based on feedback information.

3. The artificial intelligence-based water cycle reverse deduction method according to claim 1 is characterized by: In step S3, the construction of the probability analysis model includes the following steps: Step S31: Analyze influencing factors, establish a mathematical probability model to analyze the probability of flood and drought disasters and related influencing factors, and comprehensively consider various factors that affect the occurrence of flood and drought disasters; Step S32: setting initial conditions to clarify the parameter values ​​and states when the mathematical probability model starts calculating, including the initial rainfall value and temperature range within the specified time period; Step S33: Perform forward calculations based on the set initial conditions and model rules to understand the likelihood of disasters and their consequences. Step S34: Use reverse probability analysis to infer the influencing factors. When flood and drought disasters have occurred, the influencing factors that led to the disaster are inferred through mathematical probability models. This helps us better understand the causes and mechanisms of the disaster, thereby providing a basis for formulating response strategies and preventive measures.

4. The artificial intelligence-based water cycle reverse deduction method according to claim 1 is characterized by: In step S1, the data collection includes the following steps: Step S11: Collect meteorological data, obtain precipitation, temperature, wind speed, and humidity information through meteorological stations and satellites, and use them to analyze precipitation input and evaporation processes in the water cycle; Step S12: Collect hydrological data, obtain flow, water level, and water quality data from river monitoring stations and lake monitoring points to understand the flow and changes of water in the basin; Step S13: Collect topographic data, using topographic mapping and remote sensing technology to obtain the topographic relief, slope and altitude of the watershed, which have a significant impact on the flow path and potential energy distribution of water; Step S14: collecting soil data, including soil texture, moisture content, and permeability, to reflect the soil's ability to retain and transmit water; Step S15: Collecting human water use activity data, including domestic water use, industrial water use, drainage data, and water project operation data; Step S16: Summarize meteorological data, hydrological data, topographic data, soil data, and human water use activity data, and perform data preprocessing to obtain a watershed water cycle dataset.

5. The artificial intelligence-based water cycle reverse deduction method according to claim 1 is characterized by: In step S5, the result evaluation and verification includes the following steps: Step S51: Compare the actual observation results and compare the deduced results with the actual basin water cycle related data in detail to check the consistency and deviation degree; The water cycle related data of the basin include water level, flow and water quality; Step S52: Analyze the error distribution to determine the distribution of the error of the deduction results in different time, space and situations to understand the stability of the model; Step S53: Result verification, cross-validation of data from different time periods and different regions to ensure the adaptability of the model under different conditions, and backtesting of the model using historical data to determine whether it can accurately reproduce the past water cycle conditions.

6. An artificial intelligence-based watershed water cycle reverse deduction system, used to implement the artificial intelligence-based watershed water cycle reverse deduction method according to any one of claims 1 to 5, characterized in that: It includes data collection module, watershed perception range expansion module, probability analysis model construction module, time and space precise matching module and result evaluation and verification module.

7. The artificial intelligence-based watershed water cycle reverse deduction system according to claim 6 is characterized by: The data acquisition module specifically collects data and performs data preprocessing to obtain a watershed water cycle dataset; The module for expanding the watershed perception range specifically builds the basic architecture of the perception network, clarifies the information transmission path and node connections, calculates the effective coverage probability of sensor nodes and target points, and dynamically adjusts the distribution of sensor nodes based on the calculation results to optimize the perception network; The probability analysis model module is constructed to set initial conditions, perform forward calculations, and use reverse probability analysis to infer the impact factors when floods and droughts occur; The spatiotemporal precise matching module specifically extends the spatial data fusion framework to the spatiotemporal domain, uses a constant additive bias to correct systematic errors, calculates the latent process to reflect spatial and temporal dependencies, and finally establishes a visual display panel to present the fused spatiotemporal data and dynamically updates it. The result evaluation and verification module specifically compares the deduction results with the actual basin water cycle related data in detail, analyzes the error distribution and verifies the results.

Images