Method and system for generating potential capacity regulation and targeted intervention strategies for an irrigation oasis

Abstract

The application discloses an irrigation oasis regulation and control potential power and targeted intervention strategy generation method and system, the method comprises the following steps: dividing a target irrigation oasis region into a plurality of management units, obtaining multi-source data of all the management units, and calculating the coupling coordination degree of each management unit; taking the management unit with a coupling coordination degree less than the regulation lower limit as a regulation unit, and taking the difference between the coupling coordination degree of the regulation unit and the regulation lower limit or the regulation upper limit as the regulation potential; constructing and training a Bayesian causal network, and according to the coupling coordination degree and the index of the regulation unit, using the probability output by the trained Bayesian causal network, calculating the mutual information or belief variance of each driving node of the regulation unit to the target node; taking the driving node with the highest mutual information or belief variance as the main control factor of the corresponding regulation unit, and searching for the targeted intervention strategy of the regulation unit in the factor-strategy knowledge base by using the main control factor.

Description

TECHNICAL FIELD

[0001] The present application relates to oasis irrigation technology, in particular to an irrigation oasis regulation potential quantification and targeted intervention strategy generation method and system. BACKGROUND

[0002] Irrigated oasis is an important agricultural production base and human settlement in arid areas, and water resources are the lifeline of its survival and development. Under the background of increasing water scarcity and tightening ecological environment constraints, achieving efficient use of water resources, ensuring stable grain production and maintaining ecological health have become the core challenges for the sustainable development of irrigated oasis; precise irrigation and fertilization are key management measures to improve resource utilization efficiency.

[0003] The existing related technologies mainly include:

[0004] (1) Irrigation control technology based on soil moisture monitoring: real-time acquisition of soil moisture information through field sensors, and control of irrigation start and stop according to preset threshold. Disadvantages: This method often ignores the dynamic water demand differences of crops at different growth stages, future changes in weather conditions, and the interaction of water-fertilizer-salt and other environmental factors, which can easily lead to one-sidedness and lag in irrigation decision-making; at the same time, the density, cost and maintenance and calibration of sensors limit the precision and feasibility of their application in a large area.

[0005] (2) Irrigation decision-making technology based on weather forecast and crop model: combining weather forecast and crop growth model (such as FAO-56 method, DSSAT, etc.) to predict crop water demand and develop irrigation plans. Disadvantages: The accuracy of the model is highly dependent on the accuracy of weather forecasts and the localization of crop model parameters, and there is uncertainty; less consideration of actual soil conditions and overall coordination of regional water-crop-ecosystem, which may lead to a disconnect between decision-making and actual needs.

[0006] (3) Irrigation management technology based on remote sensing monitoring: using satellite or unmanned aerial vehicle remote sensing to obtain vegetation index, land surface temperature and other information, to invert crop water stress conditions or evapotranspiration, to guide irrigation. Disadvantages: Remote sensing data is easily affected by cloudy and rainy weather, and the temporal and spatial resolution may not meet the real-time decision-making needs; the data inversion model is complex, and the evaluation results are also less directly translated into precise intervention measures for specific management units considering multiple factor constraints.

[0007] Therefore, the existing technology has obvious deficiencies in dynamically quantifying the system coordination improvement potential of specific management units, accurately diagnosing limiting factors, and generating targeted intervention strategies. SUMMARY

[0008] In view of the above deficiencies in the prior art, the irrigation oasis regulation potential quantification and targeted intervention strategy generation method and system provided by the present application can accurately locate the main control factor of the management unit that needs to be regulated and managed, thereby providing an accurate intervention strategy.

[0009] In order to achieve the above-mentioned purposes, the technical scheme adopted by the present application is:

[0010] In the first aspect, an irrigation oasis regulation potential quantification and targeted intervention strategy generation method is provided, which comprises the following steps:

[0011] S1, dividing a target irrigation oasis region into a plurality of management units and obtaining multi-source data of all the management units, and then calculating the coupling coordination degree of each management unit based on the multi-source data;

[0012] S2, taking the management unit with a coupling coordination degree less than the regulation lower limit as a regulation unit, and taking the difference between the coupling coordination degree of the regulation unit and the regulation lower limit or the regulation upper limit as the regulation potential;

[0013] S3, constructing and training a Bayesian causal network, wherein the target node in the network is the regulation potential of the regulation unit, and the driving nodes are controllable management factors and indicators of the regulation unit;

[0014] S4, calculating the mutual information or belief variance of each driving node to the target node of the regulation unit according to the coupling coordination degree and the indicators of the regulation unit, and using the probability output by the trained Bayesian causal network;

[0015] S5, taking the driving node with the highest mutual information or belief variance as the main control factor of the corresponding regulation unit, and searching for the targeted intervention strategy of the regulation unit in the factor-strategy knowledge base using the main control factor.

[0016] Further, the method for calculating the coupling coordination degree of each management unit based on multi-source data comprises:

[0017] S11, using the multi-source data to calculate the indicators of the water resource subsystem, the crop growth subsystem and the ecological environment subsystem of each management unit at a selected time step;

[0018] S12, calculating the weight of each indicator of the management unit, and calculating the comprehensive scores of the three subsystems of each management unit according to the normalized indicators and weights;

[0019] S13, calculating the coupling coordination degree of each management unit at a set time step according to the comprehensive scores of the three subsystems.

[0020] Further, the indexes of the water resource subsystem include maximum temperature, average humidity, ET0 and root zone soil moisture; the indexes of the crop growth subsystem include chlorophyll index GI, leaf area index LAI, normalized vegetation index NDVI and crop actual evapotranspiration ETc; the indexes of the ecological environment subsystem include normalized salinity index SI, atmospheric resistance vegetation index ARVI, salinity index S2 and groundwater level change; the controllable management factors include actual irrigation water and fertilizer application amount.

[0021] Further, in step S12, the calculation method of the comprehensive score of each subsystem in the management unit includes the following steps:

[0022] S121, standardize the indexes of each subsystem and calculate the proportion of the standardized indexes:

[0023]

[0024] wherein, is the jth standardized index in the ith year; m is the total number of evaluation years; is the proportion of .

[0025] S122, according to the index proportion, calculate the information entropy of each index in all evaluation years:

[0026]

[0027] wherein, n is the total number of indexes of each subsystem; is the information entropy of the jth index in the ith year, 0≤ ≤1;

[0028] S123, according to the information entropy, calculate the index weight:

[0029]

[0030] wherein, is the index weight in the ith year;

[0031] S124, according to the standardized index and the corresponding index weight, calculate the comprehensive score of each subsystem:

[0032]

[0033] wherein, S is the comprehensive score of the subsystem, and the comprehensive scores of the water resource subsystem, the crop growth subsystem and the ecological environment subsystem are represented by , and respectively.

[0034] Further, step S13 further includes:

[0035] According to the comprehensive scores of the three subsystems, the coupling degree and the coupling coordination degree of each management unit at the selected time step are calculated:

[0036] ,

[0037] Wherein, C is the coupling degree; T is the comprehensive coordination index; α, β and γ are the weights of the water resource subsystem, the crop growth subsystem and the ecological environment subsystem respectively;

[0038] According to the coupling degree and the comprehensive coordination index of each management unit, the coupling coordination degree of the management unit is calculated:

[0039]

[0040] Wherein, D is the coupling coordination degree.

[0041] Further, the method for obtaining the lower limit and the upper limit of the regulation of each management unit comprises:

[0042] S21, normality test is performed on the coupling coordination degrees of all the management units, if the normal distribution is not met, then quantile standardization is adopted to adjust all the coupling coordination degrees until the normal distribution is met;

[0043] S22, when the normal distribution is met, the coupling coordination degrees are fitted into a normal distribution curve, and the confidence interval is calculated;

[0044] S23, the upper and lower limit values of the confidence interval are taken as the lower limit and the upper limit of the regulation of the management unit respectively.

[0045] Further, the expression for calculating mutual information and belief variance is:

[0046]

[0047]

[0048] Wherein, MI and VB are mutual information and belief variance respectively; H is entropy; Q is the target node; F is the driving node; q and f are the states of Q and F respectively; is the real value corresponding to the state q; is the probability predicted by the Bayesian causal network when the target node Q takes the state q; is the probability predicted by the Bayesian causal network when the driving node F takes the state f; is the joint probability distribution of the state q and the state f; is the probability of the state q occurring under the condition that the state f occurs.

[0049] Further, the irrigation oasis regulation potential quantification and targeted intervention strategy generation method further comprises visualizing the regulation potential, the main control factor and the targeted intervention strategy of the unit to be regulated to display to the management personnel of the irrigation oasis region.

[0050] In a second aspect, a system applied to the irrigation oasis regulation potential quantification and targeted intervention strategy generation method is provided, comprising:

[0051] A coordination degree determination module is configured to divide the target irrigation oasis region into a plurality of management units, and acquire multi-source data of all the management units, and calculate the coupling coordination degree of each management unit based on the multi-source data.

[0052] A regulation potential quantification module is configured to take the management unit with a coupling coordination degree less than the regulation lower limit as the unit to be regulated, and take the difference between the coupling coordination degree of the unit to be regulated and the regulation lower limit or the regulation upper limit of the unit to be regulated as the regulation potential.

[0053] A main control factor diagnosis module is configured to construct and train a Bayesian causal network, wherein the target node in the network is the regulation potential of the unit to be regulated, and the driving nodes are the controllable management factors and the indexes of the unit to be regulated; and the mutual information or the belief variance of each driving node to the target node of the unit to be regulated is calculated according to the coupling coordination degree and the indexes of the unit to be regulated, and according to the probability output by the trained Bayesian causal network.

[0054] A targeted strategy generation module is configured to take the driving node with the highest mutual information or belief variance as the main control factor of the corresponding unit to be regulated, and search for the targeted intervention strategy of the unit to be regulated in the factor-strategy knowledge base by using the main control factor.

[0055] Further, the targeted intervention strategy generation system further comprises a user interaction and visualization module configured to visualize the regulation potential, the main control factor and the targeted intervention strategy of the unit to be regulated to display to the management personnel of the irrigation oasis region.

[0056] The present application has the following advantages:

[0057] 1) Precise quantification of regulation potential: The present application overcomes the shortcomings of the prior art that it is difficult to quantitatively evaluate the improvement space of a specific management unit, and can objectively and accurately quantify the "regulation potential" of each management unit to be improved based on the dynamic coordination degree statistical distribution method, thereby providing a clear target and basis for resource optimization allocation.

[0058] 2) Targeted diagnosis of bottleneck factors: Unlike the general evaluation or correlation-based analysis of existing technologies, the present application uses Bayesian causal networks to specifically diagnose the "master control factors" that cause the low coordination level (potential exists) of a specific management unit, achieving deep and causal problem positioning from "where is not good" to "why is not good".

[0059] 3) Generation of spatiotemporal specific strategies: For the diagnosed master control factors, the present application can generate highly customized, spatiotemporal specific targeted intervention strategies (especially water and fertilizer management recommendations) through the factor-strategy knowledge base, which is significantly superior to the general recommendations of existing technologies, greatly improving the precision and effectiveness of management measures.

[0060] 4) Improve system coordination efficiency: By accurately identifying potential, diagnosing bottlenecks, and implementing targeted interventions, the present application can more effectively guide management measures to act on key links, thereby accelerating the improvement of the coupling and coordination level of the entire irrigated oasis water-crop-ecosystem, promoting efficient use of resources and improvement of the ecological environment.

[0061] 5) Strong operability of decision support: The present application provides a complete process from evaluation, quantification, diagnosis to strategy generation, and presents the results in a clear and visualized manner, providing intuitive and operable decision support for managers, reducing the difficulty of complex system management. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 Flowchart of the irrigated oasis regulation potential quantification and targeted intervention strategy generation method.

[0063] Figure 2 Network structure of the Bayesian causal network.

[0064] Figure 3 Principle block diagram of the irrigated oasis regulation potential quantification and targeted intervention strategy generation system. DETAILED DESCRIPTION

[0065] The specific embodiments of the present application are described below to facilitate understanding of the present application by those skilled in the art, but it should be clear that the present application is not limited to the scope of the specific embodiments, and for those skilled in the art, it is obvious that various changes are within the spirit and scope of the present application as defined and determined by the appended claims, and all applications utilizing the concept of the present application are within the scope of protection.

[0066] REFERENCES Figure 1 , Figure 1 The flowchart of the irrigated oasis regulation potential quantification and targeted intervention strategy generation method is shown; as shown in Figure 1 the method S includes steps S1-S5.

[0067] In step S1, the target irrigation oasis area is divided into several management units (for example, divided by a 1050m*1352m grid), and multi-source data of all management units is obtained, and then based on the multi-source data, the coupling coordination degree of each management unit is calculated.

[0068] The multi-source data used in the scheme includes but is not limited to the following data:

[0069] 1) Remote sensing data: multispectral, thermal infrared images obtained by satellites such as Landsat, MODIS, Sentinel, etc., used to extract vegetation index (NDVI, GI), salt index (SI, S2), land surface temperature, albedo, etc.

[0070] 2) Meteorological data: including temperature (maximum, average), humidity, precipitation, wind speed, radiation, etc. data of each meteorological station or grid, used to calculate reference evapotranspiration (ET0).

[0071] 3) Hydrological / water resources data: groundwater level monitoring data, irrigation canal water supply data, soil moisture monitoring data (sensors or model simulation / remote sensing inversion).

[0072] 4) Agricultural statistics and basic data: crop type distribution, soil physical and chemical properties, irrigation quota, fertilizer amount, pesticide usage, etc. historical or current data.

[0073] In an embodiment of the present application, the method for calculating the coupling coordination degree of each management unit based on multi-source data includes:

[0074] S11, using multi-source data to calculate the indexes of the water resource subsystem, crop growth subsystem and ecological environment subsystem of each management unit at the selected time step (such as monthly), which can be calculated by using existing mature calculation formula, such as NDVI, SI, ET0, ETc, root zone soil moisture estimation model, etc.

[0075] In implementation, the indexes of the water resource subsystem preferably include maximum temperature, average humidity, ET0 and root zone soil moisture; the indexes of the crop growth subsystem preferably include chlorophyll index GI, leaf area index LAI, normalized vegetation index NDVI, crop actual evapotranspiration ETc; the indexes of the ecological environment system preferably include normalized salt index SI, atmospheric resistance vegetation index ARVI, salt index S2 and groundwater level change; the controllable management factors preferably include actual irrigation water quantity and fertilizer amount.

[0076] The indexes of the three subsystems constitute a comprehensive evaluation index system, and these indexes can be statistically high-frequency indexes by using the literature frequency method, and can be comprehensively determined by using the expert scoring method.

[0077] S12, calculate the weight of each index of the management unit, and calculate the comprehensive score of each management unit according to the normalized index and the weight;

[0078] In an embodiment of the present application, in step S12, the method for calculating the comprehensive score of each subsystem in the management unit comprises the following steps:

[0079] S121, standardize the index of each subsystem, and calculate the proportion of the standardized index:

[0080]

[0081] Wherein, is the jth standardized index in the ith year; m is the total number of evaluation years; is the proportion of

[0082] Because the dimensions, orders of magnitude and positive and negative orientations of the indexes are different, the initial data needs to be standardized before step S121. For the positive and negative indexes, the standardization method is as follows:

[0083] For positive indexes: ;

[0084] For negative indexes:

[0085] Wherein, is the jth index in all indexes.

[0086] S122, calculate the information entropy of each index in all evaluation years according to the index proportion:

[0087]

[0088] Wherein, n is the total number of indexes of each subsystem; is the information entropy of the jth index in the ith year, 0≤ ≤1;

[0089] S123, calculate the index weight according to the information entropy:

[0090]

[0091] Wherein, is the index weight in the ith year;

[0092] S124, calculate the comprehensive score of each subsystem according to the standardized index and the corresponding index weight:

[0093] ​

[0094] wherein S is the comprehensive score of the subsystem, and is respectively represented by , and comprehensive score of the water resource subsystem, the crop growth subsystem and the ecological environment subsystem.

[0095] S13, according to the comprehensive scores of the three subsystems, calculating the coupling coordination degree of each management unit at the set time step; step S13 further comprises:

[0096] According to the comprehensive scores of the three subsystems, calculating the coupling degree and the coupling coordination degree of each management unit at the selected time step:

[0097] ,

[0098] wherein C is the coupling degree; T is the comprehensive coordination index; and α, β and γ are respectively the weights of the water resource subsystem, the crop growth subsystem and the ecological environment subsystem.

[0099] According to the coupling degree and the comprehensive coordination index of each management unit, calculating the coupling coordination degree of the management unit:

[0100]

[0101] wherein D is the coupling coordination degree.

[0102] In step S2, the management unit with the coupling coordination degree less than the lower limit of the regulation is taken as the management unit to be regulated, and the difference between the coupling coordination degree of the management unit to be regulated and the lower limit or the upper limit of the regulation corresponding to the management unit to be regulated is taken as the regulation potential.

[0103] In implementation, the preferred method for obtaining the lower limit and the upper limit of the regulation of each management unit comprises:

[0104] S21, performing normality test on the coupling coordination degrees of all the management units, and if the normal distribution is not met, adjusting all the coupling coordination degrees by using quantile standardization until the normal distribution is met;

[0105] When the data does not meet the normal distribution, the following methods can be used to adjust the analysis strategy to ensure the robustness and rationality of the results. Mild skewness: logarithmic transformation: x'=ln(x+1) (applicable to right-skewed data such as income, GDP). Square root transformation: x'=x (applicable to mild right-skewed data). Severe skewness: Box-Cox transformation: determine the optimal transformation parameter λ by maximum likelihood estimation. Inverse transformation: x'=1 / x (applicable to right-skewed data containing positive values). Quantile standardization: replace the mean and standard deviation with the median and interquartile range for standardization.

[0106] S22, when the normal distribution is met, fitting all coupling coordination degrees to a normal distribution curve and calculating the confidence interval thereof;

[0107] S23, taking the upper and lower limit values of the confidence interval as the regulation lower limit and regulation upper limit of the management unit respectively.

[0108] In step S3, a Bayesian causal network is constructed and trained, the target node in the network being the regulation potential of the unit to be regulated, and the driving nodes being the controllable management factors and the indicators of the unit to be regulated; Figure 2 The architecture diagram of the Bayesian causal network of the present scheme is shown, in which the regulation potential is the dependent variable of the network, the three subsystems are the direct factors, and all the impact factors, i.e. the driving nodes, are the indirect factors.

[0109] The present scheme uses the data of the unit to be regulated (indicator data and regulation potential) to perform parameter learning (calculate the conditional probability table) on the Bayesian network, which can be realized by using software such as Netica, and the weights of prior knowledge and data are adjusted in combination with experience (Experience) and training times (Degree).

[0110] The present scheme can mine high water efficiency ecological and other constraint factors by establishing a Bayesian causal network of the irrigation district. In the process of constructing the Bayesian causal network, firstly, the preliminary modeling of the Bayesian causal network is performed, and the expectation and maximum likelihood estimation of the parameters are iteratively calculated by using the EM algorithm (Expectation-Maxmization algorithm). In this case, the Bayesian network is realized by using the Netica software, which is widely used in the research of Bayesian network modeling. In Netica, the "experience" variable is used to represent the credibility of the observed data behind the prior knowledge, and the "degree" variable is used to represent the training times of the observed data, and the combination of the two can dynamically adjust the weights of the prior knowledge and the observed data in the determination of the probability distribution.

[0111] Experience:

[0112] Degree:

[0113] In the formula, P (X | Y) represents the probability, N (X | Y) represents the network, O (X | Y) represents each observation data, O (X | Y) represents the observation data set.

[0114] For Bayesian network model parameterization, input data, and parameterization are realized; finally, based on the entropy reduction (Entropy Reduction) mutual information MI (Mutual Information) and variance reduction (Variance Reduction) belief variance VB (Variance of Belief) are used to analyze the sensitivity of the Bayesian network model, and the calculation formula is as follows:

[0115]

[0116]

[0117] Where MI and VB are mutual information and belief variance, respectively; H is entropy; Q is the target node; F is the driving node; q and f are the states of Q and F, respectively; is the true value corresponding to the state q; is the probability predicted by the Bayesian causal network when the target node Q takes state q; is the probability predicted by the Bayesian causal network when the driving node F takes state f; is the joint probability distribution of state q and state f; is the probability of state q occurring under the condition that state f occurs. In order to facilitate comparison, mutual information and belief variance are normalized to the range of 0-100%, resulting in mutual information ratio and belief variance ratio.

[0118] In step S4, according to the coupling coordination degree and the index of the unit to be controlled, the probability output by the trained Bayesian causal network is used to calculate the mutual information or belief variance of each driving node of the unit to be controlled to the target node.

[0119] In step S5, the driving node with the highest mutual information or belief variance is taken as the master control factor of the corresponding unit to be controlled, and the master control factor is used to search for the targeted intervention strategy of the unit to be controlled in the factor-strategy knowledge base.

[0120] The method of establishing the "factor-strategy" knowledge base is: a mapping relationship library is established in advance, which associates the possible diagnosed master control factors with the recommended intervention measures (especially water and fertilizer management measures). For example:

[0121] Master control factor: low root zone soil moisture -> strategy: increase irrigation water amount / frequency.

[0122] Master control factor: low crop GI / LAI + low soil nutrient index -> strategy: supplement specific type of fertilizer.

[0123] Master control factor: high SI / S2 -> strategy: implement leaching irrigation, improve soil, adjust irrigation water quality.

[0124] Master factor: Irrigation quota is unreasonable (inferred based on historical data) -> Strategy: Adjust irrigation system.

[0125] Master factor: Agricultural fertilizer application amount is too high (inferred based on historical data) -> Strategy: Reduce fertilizer use and optimize fertilizer ratio.

[0126] Generate targeted strategies: For each unit that needs to be controlled, according to the diagnosed master factor, match and generate specific, quantitative (if possible) intervention suggestions from the knowledge base. For example, the suggestion is:

[0127] Spatial specificity: Grid division is performed for the monitored land plot to obtain specific management units (grids).

[0128] Temporal specificity: Based on the diagnosis results of the current time step (month).

[0129] Targetedness: Directly targeting the main bottleneck problem diagnosed.

[0130] In implementation, the irrigation oasis regulation potential quantification and targeted intervention strategy generation method preferably further includes visualizing the regulation potential of the unit that needs to be controlled, the master factor, and the targeted intervention strategy to display to the management personnel of the irrigation oasis area.

[0131] As shown in Figure 3 The present application also provides a system for applying the irrigation oasis regulation potential quantification and targeted intervention strategy generation method, which comprises:

[0132] A coordination degree determination module for dividing the target irrigation oasis area into a plurality of management units and obtaining multi-source data of all management units, and calculating the coupling coordination degree of each management unit based on the multi-source data.

[0133] The coordination degree determination module includes a data management and preprocessing module and a dynamic coordination degree evaluation module. The data management and preprocessing module is used to divide the target irrigation oasis area into a plurality of management units and obtain multi-source data of all management units, and to access, store, clean and fuse the multi-source data. The dynamic coordination degree evaluation module is used to calculate the coupling coordination degree of each management unit based on the multi-source data.

[0134] A regulation potential quantification module for taking the management unit with a coupling coordination degree less than its regulation lower limit as a unit that needs to be controlled, and taking the difference between the coupling coordination degree of the unit that needs to be controlled and its corresponding regulation lower limit or regulation upper limit as the regulation potential.

[0135] The master factor diagnosis module is configured to construct and train a Bayesian causal network, wherein a target node in the network is a regulation potential of the unit to be regulated, and a driving node is a controllable management factor and an index of the unit to be regulated; according to the coupling coordination degree and the index of the unit to be regulated, the mutual information or the belief variance of each driving node of the unit to be regulated to the target node is calculated by using the probability output by the trained Bayesian causal network;

[0136] The target strategy generation module is configured to take the driving node with the highest mutual information or belief variance as the master factor of the corresponding unit to be regulated, and search for a targeted intervention strategy of the unit to be regulated in the factor-strategy knowledge base by using the master factor.

[0137] The targeted intervention strategy generation system further comprises a user interaction and visualization module configured to visually process the regulation potential of the unit to be regulated, the master factor and the targeted intervention strategy, and display them to the management personnel of the irrigated oasis area.

Claims

1. An irrigation oasis regulation potential force quantification and targeted intervention strategy generation method, characterized in that, The method comprises the steps of: S1, dividing a target irrigation oasis area into a plurality of management units and obtaining multi-source data of all the management units, and then calculating a coupling coordination degree of each management unit based on the multi-source data; S2, taking a management unit with a coupling coordination degree less than a lower limit of regulation as a regulation unit, and taking a difference between the coupling coordination degree of the regulation unit and a lower limit or an upper limit of regulation corresponding to the regulation unit as a regulation potential; S3, constructing and training a Bayesian causal network, wherein a target node in the network is the regulation potential of the regulation unit, and driving nodes are controllable management factors and indexes of the regulation unit; S4, calculating mutual information or belief variance of each driving node to the target node of the regulation unit according to the coupling coordination degree and the indexes of the regulation unit and using a probability output by the trained Bayesian causal network; S5, taking a driving node with the highest mutual information or belief variance as a master control factor of the corresponding regulation unit, and searching for a targeted intervention strategy of the regulation unit in a factor-strategy knowledge base by using the master control factor.

2. The method of claim 1, wherein the method is characterized by, The method for calculating the coupling coordination degree of each management unit based on the multi-source data comprises: S11, calculating indexes of a water resource subsystem, a crop growth subsystem and an ecological environment subsystem of each management unit at a selected time step by using the multi-source data; S12, calculating weights of each index of the management unit, and calculating comprehensive scores of the three subsystems of each management unit according to the normalized indexes and the weights; S13, calculating the coupling coordination degree of each management unit at the set time step according to the comprehensive scores of the three subsystems.

3. The method of claim 2, wherein the method is characterized by, The indexes of the water resource subsystem include maximum temperature, average humidity, ET0 and root zone soil moisture; the indexes of the crop growth subsystem include chlorophyll index GI, leaf area index LAI, normalized vegetation index NDVI and crop actual evapotranspiration ETc; the indexes of the ecological environment subsystem include normalized salinity index SI, atmospheric resistance vegetation index ARVI, salinity index S2 and groundwater level change; and the controllable management factors include actual irrigation water quantity and fertilizer quantity.

4. The method of claim 3, wherein the method is characterized by, In step S12, the method for calculating the comprehensive score of each subsystem in the management unit comprises the following steps: S121, standardizing the indexes of each subsystem and calculating proportions of the standardized indexes; Wherein, is the normalized index of the ith year and jth standard; m is the total number of evaluation years; is the proportion of the proportion of S122, calculating information entropy of each index in all evaluation years according to the index proportions; wherein n is the total number of indicators for each subsystem; H (i, j) is the information entropy of the ith year and jth indicator, 0≤ ≤1; S123, calculating the index weights according to the information entropy; wherein, wi is the weight of the indicator for the i-th year; S124, calculating the comprehensive score of each subsystem according to the standardized indexes and the corresponding index weights; Wherein, S is the comprehensive score of the subsystem, respectively represented by , and represent the comprehensive score of the water resource subsystem, the crop growth subsystem and the ecological environment subsystem.

5. The method of claim 4, wherein the method is characterized by, Step S13 further comprises: calculating the coupling degree and the coupling coordination degree of each management unit at the selected time step according to the comprehensive scores of the three subsystems: ， wherein C is the coupling degree, T is the comprehensive coordination index, and α, β and γ are weights of the water resource subsystem, the crop growth subsystem and the ecological environment subsystem, respectively; calculating the coupling coordination degree of the management unit according to the coupling degree and the comprehensive coordination index of each management unit: wherein D is the coupling coordination degree.

6. The method of claim 1, wherein the method is characterized by, The method for obtaining the lower limit and the upper limit of regulation of each management unit comprises: S21, normality test is performed on the coupling coordination degrees of all management units, if the normal distribution is not met, then quantile standardization is adopted to adjust all coupling coordination degrees until the normal distribution is met; S22, when the normal distribution is met, the coupling coordination degrees are fitted into a normal distribution curve, and the confidence interval is calculated; S23, the upper and lower limits of the confidence interval are taken as the lower limit and upper limit of the management unit regulation, respectively.

7. The method of claim 1, wherein the method is characterized by, The expressions for calculating mutual information and belief variance are: where MI and VB are mutual information and belief variance, respectively; H is entropy; Q is the target node; F is the driver node; q and f are the states of Q and F, respectively; is the true value corresponding to the state q; is the probability predicted by the Bayesian causal network for the target node Q to take state q; is the probability predicted by the Bayesian causal network for the driver node F to take state f; is the joint probability distribution of state q and state f; is the probability of state q occurring given that state f occurs.

8. The method of claim 1, wherein the method is characterized by, It also includes visual processing of the regulation potential, main control factor and targeted intervention strategy of the unit to be regulated, to show the management personnel of the irrigated oasis area.

9. A system for applying the method of generating a strategy for regulating the potential force of an irrigation oasis and targeted intervention according to any one of claims 1 to 8, characterized in that, It includes: The coordination degree determination module is used to divide the target irrigated oasis area into a plurality of management units, and to obtain multi-source data of all management units, and to calculate the coupling coordination degree of each management unit based on the multi-source data; The regulation potential visualization module is used to take the management unit with a coupling coordination degree less than the regulation lower limit as the unit to be regulated, and to take the difference between the coupling coordination degree of the unit to be regulated and the corresponding regulation lower limit or regulation upper limit as the regulation potential; The main control factor diagnosis module is used to construct and train a Bayesian causal network, in which the target node is the regulation potential of the unit to be regulated, and the driving node is the controllable management factor and the index of the unit to be regulated; According to the coupling coordination degree and the index of the unit to be regulated, the probability output by the trained Bayesian causal network is used to calculate the mutual information or belief variance of each driving node to the target node of the unit to be regulated; The targeted strategy generation module is used to take the driving node with the highest mutual information or belief variance as the main control factor of the corresponding unit to be regulated, and to search for the targeted intervention strategy of the unit to be regulated in the factor-strategy knowledge base by using the main control factor.

10. The system of claim 9, wherein, It also includes a user interaction and visualization module for visual processing of the regulation potential, main control factor and targeted intervention strategy of the unit to be regulated, to show the management personnel of the irrigated oasis area.

Images