A digital twin scenario-based system for water and salt dynamics in irrigation areas

Abstract

The present invention relates to a digital twin scenario-based system for the water-salinization dynamics of an irrigation area, which solves the technical problems of complex systems and low precision. The system comprises a twin data acquisition module; a twin space modeling module for mapping the collected physical parameters into twin entities in a virtual space to construct an irrigation area model, including an engineering layout model, a water-salinity dynamic distribution model, and a water control system model; a salinization analysis and prediction module for analyzing and predicting the salinization of the irrigation area; a salinization control module comprising a player strategy supply module, a strategy evaluation module constructed based on expert knowledge and a historical database, and a distributed expert manual evaluation module. The player strategy supply module is a technical solution for manipulating a variety of control strategies for a virtual space irrigation area water control system, which is obtained through crowdfunding by the general public in a twin virtual space based on the water-salinization dynamic parameters of a certain irrigation area. The system effectively solves the problem and can be used for irrigation area salinization control.

Description

Technical Field

[0001] The present invention relates to the field of irrigation area salinization control, and in particular to a dynamic digital twin scenario-based system for water and salt in irrigation area salinization. Background Art

[0002] In recent years, with the gradual utilization of developable land, people have turned their attention to arid and semi-arid areas. Due to the scarcity of water resources in these areas, agricultural water conservation has become an important issue in water resource management. Moreover, due to the alternating effects of rainfall, irrigation and evaporation, salt continues to accumulate in the unsaturated soil, forming secondary salinization. Soil salinization and secondary salinization have now become a global environmental problem faced by the whole world. Soil salinization involves resource issues and ecological environmental issues. It directly affects food production and is a major limiting condition and obstacle to agricultural development and sustainable agricultural development.

[0003] The existing technology has technical problems such as complex system and low precision. The present invention provides a new irrigation area salinization water and salt dynamic digital twin scenario system to solve the above technical problems. Summary of the Invention

[0004] The present invention aims to address the complex and inaccurate systems found in existing technologies. This paper proposes a new scenario-based digital twin system for analyzing the dynamics of salinized water and salt in irrigation areas. This system is simple, highly accurate, and can effectively analyze and simulate the dynamic parameters of salinized water and salt in irrigation areas, as well as simulate and evaluate treatment methods.

[0005] In order to solve the above technical problems, the technical solutions adopted are as follows:

[0006] A digital twin scenario-based system for the dynamics of water and salt in salinization of irrigation areas, comprising:

[0007] Twin data acquisition module, used to collect irrigation area project layout parameters, irrigation area salinization water and salt dynamic parameters, and irrigation area water system control parameters;

[0008] The twin space modeling module is used to map the collected physical parameters into twin entities in virtual space and build an irrigation district model, including an engineering layout model, a water and salt dynamic distribution model, and a water control system model.

[0009] Salinization analysis and prediction module, used to analyze and predict the salinization situation in irrigation areas;

[0010] The salinization control module is used to generate irrigation management strategies for the irrigation area in the virtual twin space, evaluate and screen the irrigation management strategies, and control the irrigation area water control system in the physical world based on the selected optimal irrigation management strategy;

[0011] The salinization control module includes a player strategy supply module, a strategy evaluation module built based on expert knowledge and historical database, and a distributed expert manual evaluation module;

[0012] The player strategy supply module is a variety of governance strategies for manipulating the water control system of a virtual space irrigation area, which are crowdfunded by the general public in the twin virtual space based on the dynamic parameters of salinization water and salt in a certain irrigation area;

[0013] The strategy evaluation module is used to evaluate the availability and feasibility of the various governance strategies provided by the player strategy supply module by calling the historical database and using algorithms;

[0014] The distributed expert manual evaluation mode is to invite experts to conduct manual supplementary evaluation of the governance strategy in a distributed manner when the strategy evaluation module cannot effectively perform automatic evaluation.

[0015] Working principle of the invention: The invention aims to solve the problem that the operating status of existing irrigation and drainage projects for salinization control in irrigation areas is unclear, resulting in the inability to timely grasp the effects of water saving, salt control and drainage. The invention constructs a base database of irrigation and drainage project layout including regional channels, machine wells, drainage ditches, underground pipe facilities, etc., salinization distribution, crop planting structure, crop growth and yield, groundwater depth and water quality, irrigation and drainage water volume and water quality, meteorological information, soil moisture, salinity, etc., to realize two-way real-time mapping and interaction between physical irrigation area saline-alkali land and groundwater depth distribution and digital twin water and salt migration scenario system, ensuring the real-time and accuracy of data. Construct a water-salt dynamic change analysis model using cloud computing; make full use of water-salt dynamic physical models, historical water diversion and drainage volume and water quality, salinization distribution, groundwater depth changes and other data, integrate water conservancy, agriculture, geographic information, mathematics, and computer science, and build a digital twin system for salinization and water-salt migration in irrigation areas to simulate and analyze the water-salt migration laws under different irrigation and drainage scenarios, simulate and deduce the water-salt migration status of saline-alkali land in irrigation areas, predict future development trends, and monitor and warn the water-salt migration status of irrigation areas through data analysis and technical processing.

[0016] This invention innovatively incorporates a player strategy supply module, allowing players to crowdsource various management strategies for controlling the virtual irrigation area's water control system within the twin virtual space based on the dynamic parameters of salinization water and salt in a specific irrigation area. This system then automatically or through expert evaluation of these strategies, selecting effective strategies that match the current level of salinization. This overcomes the incompleteness of expert-designed strategies and the limitations of automatically generated strategies that require historical data as support.

[0017] In the above scheme, for optimization, the salinization analysis and prediction module is further equipped with a water-salt dynamic parameter analysis and prediction algorithm, including:

[0018] Step 1: Collect 1 satellite remote sensing image data as global data, and collect s in real time m UAV remote sensing image data is used as local data, and s m The coordinates of the UAV remote sensing image data are calibrated and associated with the coordinates of the satellite remote sensing image data to construct a two-layer data layer, where s m is a positive integer not less than 1;

[0019] Step 2: preset the water-salt dynamic analysis algorithm library, which has built-in s n algorithm model; input the global data of step 1 into the hth i An algorithm model is used to obtain the global output result;

[0020] Step 3, change s m The local data corresponds to the input j i 、j i +1, ..., j i +m-1 algorithm models, get s m Local output results, j i is a positive integer not less than 1;

[0021] Step 4: m The calculation error is evaluated internally for each local output result to determine h≤s m The hth positive local output result with an error value less than the predefined threshold is matched and settled with the global output result, and the number of positive local output results with an error value lower than the predefined threshold is calculated. If the number is lower than the predefined threshold, the hth positive local output result is matched and settled with the global output result. i This algorithm model is defined as the global matching algorithm basic model, which is used for water and salt dynamic parameter analysis and prediction; otherwise, h is defined i =h i +1, return to step 2.

[0022] The present invention integrates satellite remote sensing data and local UAV data as the basis for analyzing and predicting salinization algorithms, and comprehensively considers the matching degree of cost and accuracy by controlling the number and layout of UAVs.

[0023] Furthermore, the method further includes step 5:

[0024] Perform grid labeling on the global grid, and mark the h positive local output results in the grid label as positive local output nodes;

[0025] Select h r (h r =1,2,...,h) positive local output nodes as the starting point, define the diffusion radius R, select a grid label node as the diffusion node, and define the diffusion node using the same algorithm model as the starting point;

[0026] Traverse h positive local output nodes, define the diffusion nodes corresponding to the h positive local output nodes, and determine the algorithm model of each diffusion node;

[0027] Based on the basic model of the global matching algorithm, the algorithm model of the positive local output node and the algorithm model of the diffusion node are replaced to optimize the global matching algorithm model combination for the analysis and prediction of water and salt dynamic parameters.

[0028] Among them, node h r Select node h s The probability of being accessed as a diffusion node is P k (r, s), the diffusion node may be visited by multiple positive local output nodes, and the positive local output node corresponding to the maximum probability is used;

[0029]

[0030]

[0031] Among them, L k (r) represents the node h where ant k is located r The set of all next nodes to be visited, node h v ∈L k (r);

[0032] H k (r) represents the node h where ant k is located r The set of nodes that have not been visited, node h u ∈Z k (r);

[0033] δ(r,s) is the node h r With node h s The degree of selectivity between

[0034] η(r,s)=1 / d(r,s); d(r,s) represents the node h r and node h s the distance between them;

[0035] is the importance coefficient of the predefined selectivity;

[0036] β is the relative importance coefficient of the predefined visibility;

[0037] E(s) indicates the next neighbor node h to be visited s The remaining energy;

[0038] represents the sum of the remaining energies of the next accessible neighbor nodes;

[0039] R is the diffusion radius.

[0040] Furthermore, the characteristic parameters of the salinization water-salt dynamic parameters include vegetation parameters and salinity parameters.

[0041] Furthermore, sub-parameters of the characteristic parameters of all salinization water-salt dynamic parameters are collected, the sub-parameters are evaluated and ranked for relevance, and then fused into characteristic parameters, including:

[0042] Step A1: perform correlation evaluation and sorting on all feature parameter fusion values ​​in units of feature sub-parameters, including:

[0043] A1.1 Build an ant colony classification model by training random samples and define the deviation coefficient as:

[0044]

[0045] Among them, c x is the category of the sub-parameter classification of the characteristic parameter, tx is the number of algorithm steps of the ant colony separation model, p x is the relative probability of cx; ci and cj are the pixel values ​​of coordinates (x, y) and (x+△x, y+△y) on the image grayscale feature matrix; V ci，cj is the matrix element, N is the number of matrix elements;

[0046] A1.2 Calculate the deviation coefficient of the outward diameter data of each worker ant for the sub-parameter X, denoted as

[0047] A1.3 For sub-parameter X, add existing Gaussian noise interference to the out-of-path data sample to reconstruct the new out-of-path data sample and estimate the deviation coefficient of the new out-of-path data sample

[0048] A1.4 Calculate the correlation coefficient value of sub-parameter X It is used to characterize the correlation evaluation of the sub-parameter X, NS is the number of worker ants in the ant classification model;

[0049] A2, based on the sorting results of A1, eliminates sub-parameters whose correlation is greater than a predefined threshold, and defines weighted fusion to calculate feature parameters.

[0050] Furthermore, step A2 includes:

[0051] A2.1 Sort the NN sub-parameter features in the feature data sub-parameter set according to the value of the correlation coefficient NR. The sorting result is {z1, z2, ...zi..., zNN};

[0052] A2.2. J feature parameter sub-parameters whose correlation coefficient values ​​NR are less than a predefined threshold are fused into correlation features;

[0053]

[0054] Among them, the features (b1, b2, ...bj) belong to {z1, z2, ...zi..., zNR}, xfc(·) is the covariance function, and fc(·) is the deviation function;

[0055] A2.3 fuses j feature parameter sub-parameters whose correlation coefficient values ​​NR are greater than a predefined threshold into correlation features, and

[0056]

[0057]

[0058] is the measurement variance of the ni-th feature parameter sub-parameter

[0059] y ni =H·x ni +e;x ni is the estimated value of the ni-th characteristic parameter sub-parameter;

[0060] H=[1,1,....1] T , H is the NN-j-dimensional vector;

[0061] A2.4 The final fused features in steps A2.2 and A2.3 are defined as fused feature A and melted feature B, respectively. Calculate the Euclidean distance from a point in fused feature A to all points in fused feature B, take the minimum value, and traverse all points in fused feature A to take the average of the minimum values ​​as min a1.

[0062] Calculate the Euclidean distance from a point in fusion feature B to all points in fusion feature A, take the minimum value, and traverse all points in fusion feature B to take the average value of the minimum value as min b1;

[0063] A2.5 The fusion weight of the predefined fusion feature A is α a , the fusion weight of the melt feature B is β b , calculate T AB =max(min a1, min b1) is the fusion feature threshold, and the predefined feature fusion algorithm is called to finally complete the fusion of fusion feature A and fusion feature B.

[0064] Furthermore, the salinization control module performs the following steps:

[0065] Step s1: Visually presenting the results of the salinization analysis and prediction module to multiple public players, who then generate player strategies in the virtual twin space to manage salinization risks in the virtual twin space;

[0066] Step s2: Virtually control the water control system model in the virtual twin space according to the player's strategy to generate virtual watering parameters;

[0067] Step s3: Grid-label the entire irrigation area, divide the virtual irrigation parameters into multiple local virtual irrigation parameters, and search the historical database for historical data that is consistent with the local virtual irrigation data. If the historical data exists, the corresponding local virtual irrigation parameter is defined as a positive local virtual irrigation parameter; otherwise, it is defined as a negative local virtual irrigation parameter. If the grids of the positive local virtual irrigation parameters can be combined into a global label for the irrigation area, step s4 is executed; otherwise, a distributed expert manual evaluation module is called to require multiple distributed experts to perform manual evaluation, and step s5 is executed.

[0068] Step s4: calling the salinization analysis and prediction module, calling the corresponding salinization water-salt dynamic parameter characteristic parameters in the historical data based on the global irrigation virtual twin data, and analyzing the salinization situation;

[0069] Step s5, evaluate the treatment effect of the salinization risk situation. If the treatment effect reaches a predefined threshold, define the player strategy expert strategy, store it in the expert strategy library, and control the irrigation area water control system in the physical world according to the expert strategy to deal with the salinization risk of the irrigation area. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] The present invention will be further described below with reference to the accompanying drawings and examples.

[0071] Figure 1 , Schematic diagram of the digital twin scenario system for the dynamic water-salt digital twin of the irrigation area salinization in Example 1. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0073] Example 1

[0074] This embodiment provides a scenario-based digital twin system for water and salt dynamics in irrigation areas, such as Figure 1 The irrigation area salinization water and salt dynamic digital twin scenario system includes:

[0075] Twin data acquisition module, used to collect irrigation area project layout parameters, irrigation area salinization water and salt dynamic parameters, and irrigation area water system control parameters;

[0076] The twin space modeling module is used to map the collected physical parameters into twin entities in virtual space and build an irrigation district model, including an engineering layout model, a water and salt dynamic distribution model, and a water control system model.

[0077] Salinization analysis and prediction module, used to analyze and predict the salinization situation in irrigation areas;

[0078] The salinization control module is used to generate irrigation management strategies for the irrigation area in the virtual twin space, evaluate and screen the irrigation management strategies, and control the irrigation area water control system in the physical world based on the selected optimal irrigation management strategy;

[0079] The salinization control module includes a player strategy supply module, a strategy evaluation module built based on expert knowledge and historical database, and a distributed expert manual evaluation module;

[0080] The player strategy supply module is a variety of governance strategies for manipulating the water control system of a virtual space irrigation area, which are crowdfunded by the general public in the twin virtual space based on the dynamic parameters of salinization water and salt in a certain irrigation area;

[0081] The strategy evaluation module is used to evaluate the availability and feasibility of the various governance strategies provided by the player strategy supply module by calling the historical database and using algorithms;

[0082] The distributed expert manual evaluation mode is to invite experts to conduct manual supplementary evaluation of the governance strategy in a distributed manner when the strategy evaluation module cannot effectively perform automatic evaluation.

[0083] This embodiment addresses the problem that the operating status of existing irrigation and drainage projects for salinization control in irrigation areas is unclear, resulting in the inability to timely grasp the effects of water-saving, salt control, and drainage. It constructs a database of irrigation and drainage project layouts including regional channels, machine wells, drainage ditches, and underground pipe facilities, as well as underlying information on the dynamic water and salt conditions in irrigation areas, including salinization distribution, crop planting structure, crop growth and yield, groundwater depth and quality, irrigation and drainage volume and quality, meteorological information, soil moisture conditions, and salinity. It achieves two-way real-time mapping and interaction between saline-alkali land in physical irrigation areas and groundwater depth distribution and digital twin water and salt migration scenario systems, ensuring the real-time and accuracy of data. Construct a water-salt dynamic change analysis model using cloud computing; make full use of water-salt dynamic physical models, historical water diversion and drainage volume and water quality, salinization distribution, groundwater depth changes and other data, integrate water conservancy, agriculture, geographic information, mathematics, and computer science, and build a digital twin system for salinization and water-salt migration in irrigation areas to simulate and analyze the water-salt migration laws under different irrigation and drainage scenarios, simulate and deduce the water-salt migration status of saline-alkali land in irrigation areas, predict future development trends, and monitor and warn the water-salt migration status of irrigation areas through data analysis and technical processing.

[0084] This embodiment uses a player strategy supply module, which is a crowdfunding platform for players in the twin virtual space to manipulate the water control system of the virtual space irrigation area based on the dynamic parameters of salinization water and salt in a certain irrigation area. The community's governance strategies are evaluated automatically or by experts, and effective governance strategies that match the current salinization level are selected, which overcomes the incompleteness of expert-designed strategies and the lack of historical data support for automatically generated strategies.

[0085] Preferably, the salinization analysis and prediction module has a built-in water-salt dynamic parameter analysis and prediction algorithm, including:

[0086] Step 1: Collect 1 satellite remote sensing image data as global data, and collect s in real time m UAV remote sensing image data is used as local data, and s m The coordinates of the UAV remote sensing image data are calibrated and associated with the coordinates of the satellite remote sensing image data to construct a two-layer data layer, where s m is a positive integer not less than 1;

[0087] Step 2: preset the water-salt dynamic analysis algorithm library, which has built-in s n algorithm model; input the global data of step 1 into the hth i An algorithm model is used to obtain the global output result;

[0088] Step 3, change s m The local data corresponds to the input j i 、j i +1, ..., j i +m-1 algorithm models, get s m Local output results, j i is a positive integer not less than 1;

[0089] Step 4: m The calculation error is evaluated internally for each local output result to determine h≤s m The hth positive local output result with an error value less than the predefined threshold is matched and settled with the global output result, and the number of positive local output results with an error value lower than the predefined threshold is calculated. If the number is lower than the predefined threshold, the hth positive local output result is matched and settled with the global output result. i This algorithm model is defined as the global matching algorithm basic model, which is used for water and salt dynamic parameter analysis and prediction; otherwise, h is defined i =h i +1, return to step 2.

[0090] The embodiment of the present invention integrates satellite remote sensing data and local drone data as the basis for analyzing and predicting salinization algorithms, and comprehensively considers the matching degree of cost and accuracy by controlling the number and layout of drones.

[0091] Specifically, it also includes step 5:

[0092] Perform grid labeling on the global grid, and mark the h positive local output results in the grid label as positive local output nodes;

[0093] Select h r (h r =1,2,...,h) positive local output nodes as the starting point, define the diffusion radius R, select a grid label node as the diffusion node, and define the diffusion node using the same algorithm model as the starting point;

[0094] Traverse h positive local output nodes, define the diffusion nodes corresponding to the h positive local output nodes, and determine the algorithm model of each diffusion node;

[0095] Based on the basic model of the global matching algorithm, the algorithm model of the positive local output node and the algorithm model of the diffusion node are replaced to optimize the global matching algorithm model combination for the analysis and prediction of water and salt dynamic parameters.

[0096] Among them, node h r Select node h s The probability of being accessed as a diffusion node is P k (r, s), the diffusion node may be visited by multiple positive local output nodes, and the positive local output node corresponding to the maximum probability is used;

[0097]

[0098]

[0099] Among them, L k (r) represents the node h where ant k is located r The set of all next nodes to be visited, node h v ∈L k (r);

[0100] H k (r) represents the node h where ant k is located r The set of nodes that have not been visited, node h u ∈Z k (r);

[0101] δ(r,s) is the node h r With node h s The degree of selectivity between

[0102] η(r,s)=1 / d(r,s); d(r,s) represents the node h r and node h s the distance between them;

[0103] is the importance coefficient of the predefined selectivity;

[0104] β is the relative importance coefficient of the predefined visibility;

[0105] E(s) indicates the next neighbor node h to be visited s The remaining energy;

[0106] represents the sum of the remaining energies of the next accessible neighbor nodes;

[0107] R is the diffusion radius.

[0108] Specifically, the characteristic parameters of the salinization water-salt dynamic parameters include vegetation parameters and salinity parameters.

[0109] Preferably, sub-parameters are collected for the characteristic parameters of all salinization water-salt dynamic parameters, and the sub-parameters are evaluated and ranked for relevance and fused into characteristic parameters, including:

[0110] Step A1: perform correlation evaluation and sorting on all feature parameter fusion values ​​in units of feature sub-parameters, including:

[0111] A1.1 Build an ant colony classification model by training random samples and define the deviation coefficient as:

[0112]

[0113] Among them, c x is the category of the sub-parameter classification of the characteristic parameter, tx is the number of algorithm steps of the ant colony separation model, p x is the relative probability of cx; ci and cj are the pixel values ​​of coordinates (x, y) and (x+△x, y+△y) on the image grayscale feature matrix; V ci，cj is the matrix element, N is the number of matrix elements;

[0114] A1.2 Calculate the deviation coefficient of the outward diameter data of each worker ant for the sub-parameter X, denoted as

[0115] A1.3 For sub-parameter X, add existing Gaussian noise interference to the out-of-path data sample to reconstruct the new out-of-path data sample and estimate the deviation coefficient of the new out-of-path data sample

[0116] A1.4 Calculate the correlation coefficient value of sub-parameter X It is used to characterize the correlation evaluation of the sub-parameter X, NS is the number of worker ants in the ant classification model;

[0117] A2, based on the sorting results of A1, eliminates sub-parameters whose correlation is greater than a predefined threshold, and defines weighted fusion to calculate feature parameters.

[0118] Furthermore, step A2 includes:

[0119] A2.1 Sort the NN sub-parameter features in the feature data sub-parameter set according to the value of the correlation coefficient NR. The sorting result is {z1, z2, ...zi..., zNN};

[0120] A2.2. J feature parameter sub-parameters whose correlation coefficient values ​​NR are less than a predefined threshold are fused into correlation features;

[0121]

[0122] Among them, the features (b1, b2, ...bj) belong to {z1, z2, ...zi..., zNR}, xfc(·) is the covariance function, and fc(·) is the deviation function;

[0123] A2.3 fuses j feature parameter sub-parameters whose correlation coefficient values ​​NR are greater than a predefined threshold into correlation features, and

[0124]

[0125]

[0126] is the measurement variance of the ni-th feature parameter sub-parameter

[0127] y ni =H·x ni +e;x ni is the estimated value of the ni-th characteristic parameter sub-parameter;

[0128] H=[1,1,....1] T , H is the NN-j-dimensional vector;

[0129] A2.4 The final fused features in steps A2.2 and A2.3 are defined as fused feature A and melted feature B, respectively. Calculate the Euclidean distance from a point in fused feature A to all points in fused feature B, take the minimum value, and traverse all points in fused feature A to take the average of the minimum values ​​as min a1.

[0130] Calculate the Euclidean distance from a point in fusion feature B to all points in fusion feature A, take the minimum value, and traverse all points in fusion feature B to take the average value of the minimum value as min b1;

[0131] A2.5 The fusion weight of the predefined fusion feature A is α a , the fusion weight of the melt feature B is β b , calculate T AB =max(min a1, min b1) is the fusion feature threshold, and the predefined feature fusion algorithm is called to finally complete the fusion of fusion feature A and fusion feature B.

[0132] Preferably, the salinization control module performs the following steps:

[0133] Step s1: Visually presenting the results of the salinization analysis and prediction module to multiple public players, who then generate player strategies in the virtual twin space to manage salinization risks in the virtual twin space;

[0134] Step s2: Virtually control the water control system model in the virtual twin space according to the player's strategy to generate virtual watering parameters;

[0135] Step s3: Grid-label the entire irrigation area, divide the virtual irrigation parameters into multiple local virtual irrigation parameters, and search the historical database for historical data that is consistent with the local virtual irrigation data. If the historical data exists, the corresponding local virtual irrigation parameter is defined as a positive local virtual irrigation parameter; otherwise, it is defined as a negative local virtual irrigation parameter. If the grids of the positive local virtual irrigation parameters can be combined into a global label for the irrigation area, step s4 is executed; otherwise, a distributed expert manual evaluation module is called to require multiple distributed experts to perform manual evaluation, and step s5 is executed.

[0136] Step s4: calling the salinization analysis and prediction module, calling the corresponding salinization water-salt dynamic parameter characteristic parameters in the historical data based on the global irrigation virtual twin data, and analyzing the salinization situation;

[0137] Step s5, evaluate the treatment effect of the salinization risk situation. If the treatment effect reaches a predefined threshold, define the player strategy expert strategy, store it in the expert strategy library, and control the irrigation area water control system in the physical world according to the expert strategy to deal with the salinization risk of the irrigation area.

[0138] The contents not described in this embodiment may adopt the techniques and methods in the prior art, and will not be described in detail in this embodiment.

[0139] Although the above describes the illustrative specific embodiments of the present invention so that those skilled in the art can understand the present invention, the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, all inventions and creations based on the concepts of the present invention are protected.

Claims

1. A digital twin scenario-based system for water-salinization dynamics in irrigation areas, characterized by: The irrigation area salinization water and salt dynamic digital twin scenario system includes: Twin data acquisition module, used to collect irrigation area project layout parameters, irrigation area salinization water and salt dynamic parameters, and irrigation area water system control parameters; The twin space modeling module is used to map the collected physical parameters into twin entities in virtual space and build an irrigation district model, including an engineering layout model, a water and salt dynamic distribution model, and a water control system model. Salinization analysis and prediction module, used to analyze and predict the salinization situation in irrigation areas; The salinization control module is used to generate irrigation management strategies for the irrigation area in the virtual twin space, evaluate and screen the irrigation management strategies, and control the irrigation area water control system in the physical world based on the selected optimal irrigation management strategy; The salinization control module includes a player strategy supply module, a strategy evaluation module built based on expert knowledge and historical database, and a distributed expert manual evaluation module; The player strategy supply module is a variety of governance strategies for manipulating the water control system of a virtual space irrigation area, which are crowdfunded by the general public in the twin virtual space based on the dynamic parameters of salinization water and salt in a certain irrigation area; The strategy evaluation module is used to evaluate the availability and feasibility of the various governance strategies provided by the player strategy supply module by calling the historical database and using algorithms; The distributed expert manual evaluation mode is to invite experts to conduct manual supplementary evaluation of the governance strategy when the strategy evaluation module cannot effectively perform automatic evaluation; The salinization analysis and prediction module has a built-in water-salt dynamic parameter analysis and prediction algorithm, including: Step 1: Collect 1 satellite remote sensing image data as global data, and collect s in real time m UAV remote sensing image data is used as local data, and s m The coordinates of the UAV remote sensing image data are calibrated and associated with the coordinates of the satellite remote sensing image data to construct a two-layer data layer, where s m is a positive integer not less than 1; Step 2: preset the water-salt dynamic analysis algorithm library, which has built-in s n algorithm model; input the global data in step 1 into the hth i An algorithm model is used to obtain the global output result; Step 3, change s m The local data corresponds to the input j i 、j i +1, ..., j i +m-1 algorithm models, get s m Local output results, j i is a positive integer not less than 1; Step 4: m The calculation error is evaluated internally for each local output result to determine h≤s m The hth positive local output result with an error value less than the predefined threshold is matched and settled with the global output result, and the number of positive local output results with an error value lower than the predefined threshold is calculated. If the number is lower than the predefined threshold, the hth positive local output result is matched and settled with the global output result. i This algorithm model is defined as the global matching algorithm basic model, which is used for water and salt dynamic parameter analysis and prediction; otherwise, h is defined i =h i +1, return to step 2; Also includes step 5: Perform grid labeling on the global grid, and mark the h positive local output results in the grid label as positive local output nodes; Select h r (h r =1,2,...,h) positive local output nodes as the starting point, define the diffusion radius R, select a grid label node as the diffusion node, and define the diffusion node using the same algorithm model as the starting point; Traverse h positive local output nodes, define the diffusion nodes corresponding to the h positive local output nodes, and determine the algorithm model of each diffusion node; Based on the basic model of the global matching algorithm, the algorithm model of the positive local output node and the algorithm model of the diffusion node are replaced to optimize the global matching algorithm model combination for the analysis and prediction of water and salt dynamic parameters. Among them, node h r Select node h s The probability of being accessed as a diffusion node is P k (r, s), the diffusion node may be visited by multiple positive local output nodes, and the positive local output node corresponding to the maximum probability is used; Among them, L k (r) represents the node h where ant k is located r The set of all next nodes to be visited, node h v ∈L k (r); H k (r) represents the node h where ant k is located r The set of nodes that have not been visited, node h u ∈Z k (r); δ(r,s) is the node h r With node h s The degree of selectivity between η(r,s)=1 / d(r,s); d(r,s) represents the node h r and node h s the distance between them; is the importance coefficient of the predefined selectivity; β is the relative importance coefficient of the predefined visibility; E(s) indicates the next neighbor node h to be visited s The remaining energy; represents the sum of the remaining energies of the next accessible neighbor nodes; R is the diffusion radius.

2. The irrigation area salinization water and salt dynamic digital twin scenario-based system according to claim 1 is characterized by: The characteristic parameters of the salinization water-salt dynamic parameters include vegetation parameters and salt parameters.

3. The irrigation area salinization water and salt dynamic digital twin scenario system according to claim 2 is characterized by: Collect sub-parameters of the characteristic parameters of all salinization water-salt dynamic parameters, evaluate and rank the sub-parameters for relevance, and fuse them into characteristic parameters, including: Step A1: perform correlation evaluation and sorting on all feature parameter fusion values ​​in units of feature sub-parameters, including: A1.1 Build an ant colony classification model by training random samples and define the deviation coefficient as: Among them, c x is the category of the sub-parameter classification of the characteristic parameter, tx is the number of algorithm steps of the ant colony separation model, p x is the relative probability of cx; ci and cj are the pixel values ​​of coordinates (x, y) and (x+△x, y+△y) on the image grayscale feature matrix; V ci，cj is the matrix element, N is the number of matrix elements; A1.2 Calculate the deviation coefficient of the outward diameter data of each worker ant for the sub-parameter X, denoted as A1.3 For sub-parameter X, add existing Gaussian noise interference to the out-of-path data sample to reconstruct the new out-of-path data sample and estimate the deviation coefficient of the new out-of-path data sample A1.4 Calculate the correlation coefficient value of sub-parameter X It is used to characterize the correlation evaluation of the sub-parameter X, NS is the number of worker ants in the ant classification model; A2, based on the sorting results of A1, eliminates sub-parameters whose correlation is greater than a predefined threshold, and defines weighted fusion to calculate feature parameters.

4. The irrigation area salinization water and salt dynamic digital twin scenario-based system according to claim 3 is characterized by: Step A2 includes: A2.1 Sort the NN sub-parameter features in the feature data sub-parameter set according to the value of the correlation coefficient NR. The sorting result is {z1, z2, ...zi..., zNN}; A2.

2. J feature parameter sub-parameters whose correlation coefficient values ​​NR are less than a predefined threshold are fused into correlation features; Among them, the features (b1, b2, ...bj) belong to {z1, z2, ...zi..., zNR}, xfc(·) is the covariance function, and fc(·) is the deviation function; A2.3 fuses j feature parameter sub-parameters whose correlation coefficient values ​​NR are greater than a predefined threshold into correlation features, and is the measurement variance of the ni-th feature parameter sub-parameter y ni =H·x ni +e;x ni is the estimated value of the ni-th characteristic parameter sub-parameter; H=[1,1,....1] T , H is the NN-j-dimensional vector; A2.4 The final fused features in steps A2.2 and A2.3 are defined as fused feature A and melted feature B, respectively. Calculate the Euclidean distance from a point in fused feature A to all points in fused feature B, take the minimum value, and traverse all points in fused feature A to take the average of the minimum values ​​as min a1. Calculate the Euclidean distance from a point in fusion feature B to all points in fusion feature A, take the minimum value, and traverse all points in fusion feature B to take the average value of the minimum value as min b1; A2.5 The fusion weight of the predefined fusion feature A is α a , the fusion weight of the melt feature B is β b , calculate T AB =max(min a1, min b1) is the fusion feature threshold, and the predefined feature fusion algorithm is called to finally complete the fusion of fusion feature A and fusion feature B.

5. The irrigation area salinization water and salt dynamic digital twin scenario system according to claim 1 is characterized by: The salinization control module performs the following steps: Step s1: Visually presenting the results of the salinization analysis and prediction module to multiple public players, who then generate player strategies in the virtual twin space to manage salinization risks in the virtual twin space; Step s2: Virtually control the water control system model in the virtual twin space according to the player's strategy to generate virtual watering parameters; Step s3: Grid-label the entire irrigation area, divide the virtual irrigation parameters into multiple local virtual irrigation parameters, and search the historical database for historical data that is consistent with the local virtual irrigation data. If the historical data exists, the corresponding local virtual irrigation parameter is defined as a positive local virtual irrigation parameter; otherwise, it is defined as a negative local virtual irrigation parameter. If the grids of the positive local virtual irrigation parameters can be combined into a global label for the irrigation area, step s4 is executed; otherwise, a distributed expert manual evaluation module is called to require multiple distributed experts to perform manual evaluation, and step s5 is executed. Step s4: calling the salinization analysis and prediction module, calling the corresponding salinization water-salt dynamic parameter characteristic parameters in the historical data based on the global irrigation virtual twin data, and analyzing the salinization situation; Step s5, evaluate the treatment effect of the salinization risk situation. If the treatment effect reaches a predefined threshold, define the player strategy expert strategy, store it in the expert strategy library, and control the irrigation area water control system in the physical world according to the expert strategy to deal with the salinization risk of the irrigation area.

Images