Regional crop rolling correction intelligent irrigation decision-making method

Abstract

The invention relates to the technical field of agricultural irrigation, and discloses a regional crop rolling correction intelligent irrigation decision-making method, which comprises the following steps of S1, collecting basic data and dividing an irrigation decision-making region; s2, crop growth model parameter sensitivity analysis; s3, localizing parameters of the crop growth model; s4, a regional space long-term irrigation decision scheme is formulated; s5, establishing a farmland information database based on remote sensing and Internet of Things technologies; and S6, constructing a regional irrigation system rolling correction system. The invention provides a regional crop rolling correction intelligent irrigation decision-making method. The method comprises the following steps: analyzing a mutual relation and an evolution trend of each element of an irrigation decision-making model; a regional grain and cotton crop irrigation system based on crop water shortage information perception-process recognition-efficient decision is constructed, and irrigation decisions are dynamically adjusted in real time according to future climate changes and field management changes, so that the balance between the water resource utilization efficiency and the crop yield is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural irrigation, and in particular to a regional crop rolling correction intelligent irrigation decision method. Background Art

[0002] Fixed irrigation systems usually have the problem of low water resource utilization, and given that different crops have different water requirements during their respective growth cycles, fixed irrigation systems are difficult to effectively meet the water needs of crops.

[0003] Changes in meteorological factors such as precipitation and temperature drive changes in soil factors, which in turn affect the water requirements of crops.

[0004] Time lags and data acquisition challenges exist in the irrigation process, making it critical to accurately manage the timing and amount of irrigation for crop growing conditions.

[0005] Precise and intelligent irrigation decisions require refined and intelligent decision-making based on the actual needs of crops, which involves precise planning of irrigation water volume and irrigation dates. With the advancement of information technology and intelligent algorithms, the decision-making process can now be intelligentized through remote sensing, meteorological, hydrological and other data. However, there is currently no system or method that can dynamically adjust irrigation decisions in real time based on future climate change to achieve a new balance between water resource utilization efficiency and crop yields, thereby optimizing benefits. Summary of the invention

[0006] In order to overcome or alleviate one or more of the above technical problems, the present invention aims to provide a regional crop rolling correction intelligent irrigation decision method.

[0007] The present invention provides the following technical solutions:

[0008] The present invention provides a regional crop rolling correction intelligent irrigation decision method, which comprises the following steps:

[0009] S1: Basic data collection and irrigation decision-making area division;

[0010] Use remote sensing technology, meteorological stations and field soil monitoring instruments to collect crop, soil, hydrological, meteorological and field management data in the irrigation decision area; divide the decision area into n sub-areas according to the soil type and field management method of the decision area; and rasterize each sub-area;

[0011] S2: sensitivity analysis of crop growth model parameters;

[0012] The Sobol method was used to conduct sensitivity analysis on crop growth model parameters, and non-sensitive parameters and sensitive parameters were screened out. Non-sensitive parameters used the system default values, and sensitive parameters needed to be calibrated, including crop parameters, soil parameters, and meteorological parameters.

[0013] S3: Localization of crop growth model parameters;

[0014] After each sub-region determines the model parameters that need to be calibrated, the NSGA-Ⅱ algorithm is combined with the crop model to calibrate the model parameters, set the minimum error target between the simulated value and the measured value, realize automatic calibration of the parameters in the crop growth model, obtain the crop parameters, soil parameters and meteorological parameters after the localization of the crop growth model, and establish a localized crop growth model.

[0015] S4: Develop regional spatial long-term irrigation decision-making plan;

[0016] Based on the localized crop growth model, the relevant data of the planning year are input into the crop growth model, including meteorological data, soil data, crop data and management data, to form a comprehensive crop growth model suitable for the planning year; the crop growth model is coupled with the multi-objective optimization model to construct a regional grid-based dynamic irrigation decision-making technology, set the maximum yield target and the minimum irrigation target, and assign weights to multiple targets through the TOPSIS comprehensive evaluation method to obtain different regional spatial grid long-term irrigation systems;

[0017] S5: Establish a farmland information database based on remote sensing and Internet of Things technology;

[0018] Including crop data, meteorological data, soil data, hydrological data and field management data; remote sensing images of the irrigation decision area are obtained through satellite platforms or drones, and aboveground biomass and leaf area index crop data are obtained by inversion with regression models; real-time soil data are obtained by arranging soil moisture monitors in the irrigation decision area, hydrological data are obtained through a series of runoff monitoring data from hydrological stations, and real-time meteorological data are obtained by setting up small meteorological stations in the irrigation decision area; field management data are obtained through real-time input of agricultural management measures by field managers, and agricultural management measures include but are not limited to: single irrigation amount, irrigation date, fertilization amount, fertilization date, ridge height;

[0019] S6: Construct a rolling revision system for regional irrigation regimes;

[0020] The obtained regional spatial grid long-term irrigation system is used as the reference irrigation system and applied to the irrigation decision-making area. The farmland information database is read into the system in real time. According to the actual irrigation effect and the predicted effect, the adaptive control algorithm is used to make short-term rolling corrections to the reference irrigation system to quickly adjust the irrigation decision.

[0021] According to some embodiments, the specific process of the Sobol method in step S2 includes:

[0022] S21: Determine input parameters and target variables:

[0023] Select the model input parameters and their ranges to be analyzed, including crop parameters, soil parameters and meteorological parameters; determine the key target variables of the model, which are the leaf area index and aboveground biomass changes output by time series and the crop yield output by non-time series;

[0024] S22: Generate input parameter samples:

[0025] Use the Sobol sequence to generate a sample set of input parameters with low variance;

[0026] S23: Run model simulation:

[0027] Use the generated input parameter samples to run the model simulation one by one; each sample corresponds to a model simulation and the corresponding output results are recorded;

[0028] S24: Calculate Sobol index:

[0029] The Sobol index is a quantitative indicator used to measure the contribution of each input parameter to the model output; the first-order sensitivity index and the full-order sensitivity index are calculated by programming software, and the formula is as follows:

[0030] First-order sensitivity index:

[0031]

[0032] In the formula, S i is the first-order sensitivity index, V i The input parameter x i V(Y) is the model variance caused by a single change, V(Y) is the total output variance of the model, m is the number of input parameters, V ij is the parameter x i With parameter x j The variance due to interaction, V 1…,m The variance of the output caused by the interaction of all parameters;

[0033] Full-order sensitivity index:

[0034]

[0035] In the formula, ST i is the full-order sensitivity index, V -i is the variance of all parameters except the i-th parameter;

[0036] S25: Analyze and interpret the results:

[0037] The sensitivity index obtained by analysis is used to identify the sensitivity of input variables to crop yield or other target variables, and input parameters with larger sensitivity index are localized.

[0038] According to some implementations, after the model parameters to be calibrated are determined in step S2, step S3 combines the NSGA-II algorithm with the crop growth model to perform parameter calibration, and the specific steps are as follows:

[0039] S31: Initialization: Set the initial population size, define the maximum number of algorithm iterations, set the crossover probability and mutation probability, and randomly generate the initial population. Each individual is a set of model parameter combinations.

[0040] S32: Evaluation: According to the objective function, the fitness of each individual is evaluated, that is, the minimum error between the simulated value and the measured value;

[0041] S33: Selection: Classify individuals by non-dominated sorting and calculate crowding distance to maintain diversity, and give priority to individuals with low non-dominated rank and large crowding distance;

[0042] S34: Crossover: Perform a crossover operation on the selected parent individuals to generate new individuals; the crossover operation determines whether to perform a crossover by the crossover probability, and generates a new individual by combining the genetic information of the two parent individuals; mutation: Perform a mutation operation on the individuals after the crossover to increase population diversity and avoid falling into a local optimum; the mutation operation determines whether to perform a mutation by the mutation probability, and generates a new individual by randomly changing the genetic information of the individual;

[0043] S35: Evaluation: Evaluate the fitness of the newly generated individuals according to the objective function, calculate their performance in the optimization objective, and use it for sorting and screening;

[0044] S36: Replacement: Select parents and offspring through non-dominated sorting and crowding distance, optimize and retain high-quality individuals, and generate the next generation population;

[0045] S37: Termination condition: the algorithm is terminated when the maximum number of iterations is reached or the fitness converges;

[0046] S38: Output result: Output the optimal solution set, that is, the optimal combination of model parameters that minimizes the error between the simulated value and the measured value.

[0047] According to some embodiments, the objective function of the parameter localization optimization model in step S3 is as follows:

[0048]

[0049] Where LAI iobs and LAI i sim are the measured and simulated values ​​of leaf area index, respectively. i obs and Bio i sim are the measured and simulated values ​​of aboveground biomass, Y i obs and Y i sim are the measured and simulated values ​​of crop yield, respectively; n is the sampling times of canopy coverage and biomass in the measured sampling or experiment, and m is the sampling times of crop yield.

[0050] According to some implementations, the formulation of the regional spatial long-term irrigation decision plan in step S4 requires coupling the crop model with the multi-objective optimization model, and the specific steps are as follows:

[0051] S41: Data input: input the data required for the crop model in the planned year into the crop model, including meteorological data, soil data, field management data and localized model parameters, to create a crop model suitable for the planned year; determine the objective function: determine the optimization goal, i.e., maximizing crop yield and minimizing irrigation water;

[0052] S42: Setting constraints: To reflect the limitation of available resources in actual production and ensure the normal growth of crops, constraints need to be set;

[0053] S43: Model coupling: coupling the crop model and the optimization model, passing the decision variables of the optimization model as input to the crop model, and feeding back the crop yield and irrigation amount generated by the crop model to the optimization model;

[0054] S44: Model solution: Use NSGA-II algorithm to solve the optimization model and output the Pareto solution set of crop yield and irrigation amount;

[0055] S45: TOPSIS comprehensive evaluation method selection: Pareto solution set selection, and use TOPSIS comprehensive evaluation method to further select the Pareto solution set;

[0056] The specific steps of the TOPSIS comprehensive evaluation method are as follows:

[0057] S451: Determine the evaluation index: clarify the decision-making goal and select relevant evaluation indexes, i.e. the larger the crop yield value, the better, and the smaller the irrigation value, the better;

[0058] S452: Construct a normalized decision matrix: organize and normalize the original data, organize the crop yield and irrigation data in the Pareto solution set into a decision matrix Di×j, i represents the number of Pareto solution sets, j represents the number of decision targets, and each row of the matrix is ​​an alternative plan. The linear normalization method is used to convert the data into a unified scale; determine the weight of the evaluation index: according to the decision maker's preference, different weight combinations are made for each evaluation index, which can be divided into a spatial long-term irrigation system plan for a high water efficiency area, a spatial long-term irrigation system plan for a high-yield area, and a spatial long-term irrigation system plan for a water-saving area;

[0059] S453: Calculate positive and negative ideal solutions: calculate the distance from each alternative solution to the positive and negative ideal solutions;

[0060] S454: Sorting: Calculate the relative proximity of each alternative plan to the positive and negative ideal solutions, and sort the alternative plans according to the relative proximity. The one with the largest value is the optimal regional spatial long-term irrigation decision plan under the preference angle.

[0061] According to some embodiments, the objective function of the regional spatial long-term irrigation decision optimization model is as follows:

[0062] Obj2=maxY (5)

[0063] Obj3=minI (6)

[0064] In the formula, Y is the crop yield per unit area; I is the irrigation water per unit area;

[0065] The constraints of the regional spatial long-term irrigation decision optimization model are as follows:

[0066] Irrigation water quantity constraint: the irrigation water quantity should be less than the available water quantity in the irrigation area:

[0067] I≤R (7)

[0068] In the formula, R is the amount of water available for irrigation per unit area;

[0069] Yield constraints: the production output of irrigation areas must meet food demand:

[0070] ∑Y·A≥Y dem (8)

[0071] Where Y dem is the food demand in the area where the irrigation district is located;

[0072] Soil moisture content / water layer depth constraints: To ensure normal growth of crops, soil moisture content and water layer depth should be kept within an appropriate range:

[0073] S min ≤S≤S max (9)

[0074] h min ≤h≤h max (10)

[0075] In the formula, S is the soil moisture content, h is the depth of the water layer; S max With S min are the upper and lower limits of soil moisture suitable for crops; h max With h min They are the upper and lower limits of suitable water layer depth respectively.

[0076] According to some embodiments, the regional irrigation schedule rolling correction system in step S6 includes:

[0077] Data reading module, used to read farmland information database into the system in real time, including but not limited to crop soil moisture status, precipitation, temperature and irrigation water volume data;

[0078] Data processing and storage module, used to process and analyze the read data, including preprocessing of missing data, outlier detection and format standardization, and then store the data, back up the data regularly to prevent data loss, and have a recovery mechanism;

[0079] The reference irrigation system selection module is used by decision makers to select one of the long-term irrigation systems in the spatial grid of high water efficiency, high yield and water-saving areas as a reference irrigation system for application in the irrigation decision area according to their preference perspective;

[0080] Intelligent prediction module, which is used to predict future crop water demand and water supply based on relevant historical data provided by the data processing and storage module using machine learning algorithms;

[0081] The crop growth module is used to import relevant data from the data processing and storage modules. The crop growth module is responsible for simulating the biomass and crop water consumption during the period;

[0082] The rolling correction module is used to determine whether the climate prediction results of the intelligent prediction module and the irrigation effect of the crop growth module meet the rolling correction constraints. When the rolling correction constraints are met, there is no need to roll back the irrigation system and the reference irrigation system is still used. When the rolling correction constraints are not met, an early warning message is generated and the irrigation system should be re-decided to update the irrigation water volume and irrigation date in the short term to obtain a dynamic and high-water-efficiency irrigation system.

[0083] The user interface and interaction module is used to provide a user interface for users to operate and monitor. The interface includes mobile applications, web pages and control panels. Users can view farmland environmental data, crop growth conditions, irrigation plans and early warning information in real time through the interface, and perform operations and provide feedback.

[0084] The system management and maintenance module is responsible for the management and maintenance of the system, including permission settings, performance monitoring, log management, software updates and troubleshooting; in addition, the system is also scalable to facilitate the addition of new functions and services.

[0085] According to some embodiments, the rolling correction constraint in step S6 is as follows:

[0086] Biomass constraint for a period, when the biomass for a period is less than the biomass under the reference irrigation system, a rolling correction is required:

[0087] α·B≥B R (11)

[0088] Where, B is the biomass of the period simulated by the crop growth module; B R is the biomass under the reference irrigation regime, α is the correction factor;

[0089] Water supply constraint: when the water supply is less than the water demand, rolling correction is required:

[0090] Q S ≥Q D (12)

[0091] In the formula, Q S is the water supply; Q D For water demand.

[0092] Compared with the prior art, the present invention has the following beneficial effects:

[0093] The regional crop rolling correction intelligent irrigation decision method provided by the present invention analyzes the interrelationships and evolution trends of the various elements of the irrigation decision model; constructs a regional grain and cotton crop irrigation system based on crop water shortage information perception-process understanding-efficient decision-making, and dynamically adjusts irrigation decisions in real time according to future climate changes and field management changes to achieve a balance between water resource utilization efficiency and crop yield. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 A flow chart of a regional crop rolling correction intelligent irrigation decision method provided in an embodiment of the present invention.

[0095] Figure 2 The sensitivity analysis results of the model parameters provided by the embodiment of the present invention.

[0096] Figure 3 A Pareto result diagram of regional average yield and average irrigation amount provided by an embodiment of the present invention.

[0097] Figure 4 This is a diagram of irrigation schedule optimization results for a certain grid in a decision area provided by an embodiment of the present invention.

[0098] Figure 5 This is a graph of rolling correction results of the irrigation schedule provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0099] The present invention is described in detail below in conjunction with the embodiments and drawings, but it should be understood that the embodiments and drawings are only used to exemplify the present invention and do not constitute any limitation on the protection scope of the present invention. All reasonable changes and combinations within the scope of the inventive concept of the present invention fall within the protection scope of the present invention.

[0100] The present invention will be further described below in conjunction with the accompanying drawings.

[0101] Example 1

[0102] This embodiment is applied to a cotton producing area in Xinjiang.

[0103] like Figure 1 This embodiment provides a regional crop rolling correction intelligent irrigation decision method, the regional crop is a cotton producing area in Xinjiang, and the irrigation decision method includes the following steps:

[0104] S1: Basic data collection and irrigation decision-making area division;

[0105] Using remote sensing technology, meteorological stations and field soil monitoring instruments to collect crop, soil, hydrological, meteorological and field management data in the irrigation decision area; dividing the decision area into n sub-areas according to the soil type and field management method of the decision area. In this embodiment, the study area is divided into 4 sub-areas, and each sub-area is rasterized;

[0106] S2: sensitivity analysis of crop growth model parameters;

[0107] The Sobol method was used to conduct sensitivity analysis on crop growth model parameters, and non-sensitive parameters and sensitive parameters were screened out. The non-sensitive parameters used the system default values, and the sensitive parameters needed to be calibrated, including crop parameters, soil parameters and meteorological parameters.

[0108] The process of Sobol method is as follows:

[0109] 1. Determine input parameters and target variables:

[0110] Select the model input parameters and their ranges to be analyzed, including crop parameters, soil parameters and meteorological parameters. Determine the key target variables of the model, which are the leaf area index and aboveground biomass changes of time series output and the crop yield of non-time series output.

[0111] 2. Generate input parameter samples:

[0112] Use the Sobol sequence to generate a sample set of input parameters with low variance.

[0113] 3. Run the model simulation:

[0114] Run the model simulation one by one using the generated input parameter samples. Each sample corresponds to a model simulation and the corresponding output results are recorded.

[0115] 4. Calculate the Sobol index:

[0116] The Sobol index is a quantitative indicator used to measure the contribution of each input parameter to the model output. The first-order sensitivity index and the full-order sensitivity index are calculated through programming software. The formula is as follows:

[0117] First-order sensitivity index:

[0118]

[0119] In the formula, S i is the first-order sensitivity index, V i The input parameter x i V(Y) is the model variance caused by a single change, V(Y) is the total output variance of the model, m is the number of input parameters, V ij is the parameter x i With parameter x j The variance due to interaction, V 1…,m The variance of the output is caused by the interaction of all parameters.

[0120] Full-order sensitivity index:

[0121]

[0122] In the formula, ST i is the full-order sensitivity index, V -i is the variance of all parameters except the i-th parameter.

[0123] 5. Analyze and interpret the results:

[0124] The sensitivity index obtained by analysis is used to identify the sensitivity of input variables to crop yield or other target variables, and input parameters with larger sensitivity index are localized.

[0125] The results of the sensitivity analysis of the model parameters in this embodiment are as follows: Figure 2As shown in the figure, due to spatial heterogeneity, the sensitivity of cotton yield to model parameters is different in each sub-region. The parameter CVO (efficiency of assimilates converted to storage organs) is a sensitive parameter in all sub-regions because it directly affects the efficiency of cotton converting aboveground biomass into yield, so it is a sensitive parameter in all sub-regions. However, the parameter AMAXTB1.3 (maximum leaf CO2 assimilation rate as a function of crop development stage) is not a sensitive parameter in sub-regions 1 and 4 because of soil heterogeneity and different management measures in the sub-regions.

[0126] S3: localization of crop growth model parameters;

[0127] After each sub-region determines the model parameters that need to be calibrated, the NSGA-Ⅱ algorithm is combined with the crop model to calibrate the model parameters, and the goal of minimizing the error between the simulated value and the measured value is set to achieve automatic calibration of the parameters in the crop growth model, obtain the crop parameters, soil parameters and meteorological parameters after localization of the crop growth model, and establish a localized crop growth model.

[0128] After the model parameters to be calibrated are determined in step S2, step S3 needs to combine the NSGA-II algorithm with the crop growth model for parameter calibration. The specific steps are as follows:

[0129] Initialization: Set the initial population size, define the maximum number of algorithm iterations, set the crossover probability and mutation probability, and randomly generate the initial population. Each individual is a set of model parameter combinations.

[0130] Evaluation: According to the objective function, the fitness of each individual is evaluated, that is, the minimum error between the simulated value and the measured value;

[0131] Selection: Individuals are ranked by non-dominated sorting and crowding distance is calculated to maintain diversity, and individuals with low non-dominated rank and large crowding distance are preferred;

[0132] Crossover: Perform a crossover operation on the selected parent individuals to generate new individuals. The crossover operation determines whether to perform a crossover based on the crossover probability, and generates a new individual by combining the genetic information of the two parent individuals.

[0133] Mutation: Mutate the individuals after crossover to increase population diversity and avoid falling into local optimality. The mutation operation determines whether to mutate based on the mutation probability and generates new individuals by randomly changing the individual gene information.

[0134] Evaluation: Evaluate the fitness of the newly generated individuals according to the objective function and calculate their performance in the optimization objective for sorting and screening;

[0135] Replacement: Select parents and offspring through non-dominated sorting and crowding distance, optimize and retain high-quality individuals, and generate the next generation population;

[0136] Termination condition: the algorithm is terminated when the maximum number of iterations is reached or the fitness converges;

[0137] Output results: Output the optimal solution set, that is, the optimal combination of model parameters that minimizes the error between the simulated value and the measured value;

[0138] The objective function of the parameter localization optimization model is as follows:

[0139]

[0140] Where LAI i obs and LAI i sim are the measured and simulated values ​​of leaf area index, respectively. i obs and Bio i sim are the measured and simulated values ​​of aboveground biomass, Y i obs and Y i sim are the measured and simulated values ​​of crop yield, respectively. n is the sampling times of canopy coverage and biomass in the measured sampling or experiment, and m is the sampling times of crop yield.

[0141] In the embodiment, the model parameters are calibrated based on the sensitivity analysis results to establish a localized crop growth model. The calibration results of the model parameters for sub-regions 1 to 4 are shown in Table 1:

[0142] Table 1

[0143]

[0144] S4: Develop regional spatial long-term irrigation decision-making plan;

[0145] Based on the localized crop growth model, relevant data for the planning year are input into the crop growth model, including meteorological data, soil data, crop data and management data, to form a comprehensive crop growth model suitable for the planning year.

[0146] The crop growth model and the multi-objective optimization model are coupled to construct a regional grid-based dynamic irrigation decision-making technology. The maximum yield target and the minimum irrigation target are set. The NSGA-II algorithm is used to solve the optimization model to obtain the Pareto result diagram of the regional average yield and the average irrigation amount, as shown in Figure 2. Figure 3As shown. The TOPSIS comprehensive evaluation method is used to assign weights to multiple objectives and obtain different regional spatial grid long-term irrigation systems. A regional spatial long-term irrigation decision plan is formulated based on the localization of the crop growth model, and the Pareto result diagram of the regional average yield and average irrigation volume is obtained, as shown in Figure 3 As shown in the figure, the obtained regional spatial long-term irrigation decision-making plan is applied as a reference irrigation system in the planning year, and the irrigation system is revised on a rolling basis based on changes in meteorological and field management elements.

[0147] The process of TOPSIS comprehensive evaluation method includes: determining evaluation indicators, constructing a normalized decision matrix, determining the weights of evaluation indicators, calculating positive and negative ideal solutions, and ranking;

[0148] The coupling crop model and the multi-objective optimization model are as follows:

[0149] Data input: Input the data required for the crop model in the planned year into the crop model, including meteorological data, soil data, field management data and localized model parameters, to create a crop model suitable for the planned year;

[0150] Determine the objective function: Determine the optimization goal, i.e., maximizing crop yield and minimizing irrigation water;

[0151] Setting constraints: To reflect the limitations of available resources in actual production and to ensure the normal growth of crops, constraints need to be set;

[0152] Model coupling: coupling the crop model with the optimization model, passing the decision variables of the optimization model as input to the crop model, and feeding the crop yield and irrigation amount generated by the crop model into the optimization model;

[0153] Model solution: The NSGA-II algorithm is used to solve the optimization model and output the Pareto solution set of crop yield and irrigation amount. The specific steps of the NSGA-II algorithm are shown in S3;

[0154] TOPSIS comprehensive evaluation method selection: Pareto solution set selection. Since the Pareto solution set obtained by the NSGA-II algorithm can only give several solutions with good effects, the TOPSIS comprehensive evaluation method is used to further select the Pareto solution set. The specific steps of the TOPSIS comprehensive evaluation method are as follows:

[0155] Determine the evaluation indicators: clarify the decision-making goals and select relevant evaluation indicators, that is, the larger the crop yield value, the better, and the smaller the irrigation value, the better;

[0156] Construct a normalized decision matrix: The original data is sorted and normalized, and the crop yield and irrigation data in the Pareto solution set are sorted into a decision matrix Di×j, where i represents the number of Pareto solution sets and j represents the number of decision targets. Each row of the matrix is ​​an alternative plan, and the linear normalization method is used to convert the data into a unified scale.

[0157] Determine the weights of evaluation indicators: According to the decision maker's preference, different weight combinations are made for each evaluation indicator, which can be divided into the spatial long-term irrigation system plan for high water efficiency areas, the spatial long-term irrigation system plan for high-yield areas, and the spatial long-term irrigation system plan for water-saving areas;

[0158] Calculate positive and negative ideal solutions: Calculate the distance from each alternative to the positive and negative ideal solutions;

[0159] Sorting: Calculate the relative proximity of each alternative plan to the positive and negative ideal solutions, and sort the alternative plans according to the relative proximity. The one with the largest value is the optimal regional spatial long-term irrigation decision plan under this preference angle.

[0160] The objective function of the regional spatial long-term irrigation decision optimization model is as follows:

[0161] Obj2=maxY (5)

[0162] Obj3=minI (6)

[0163] Where Y is the crop yield per unit area; I is the irrigation water volume per unit area.

[0164] The constraints of the regional spatial long-term irrigation decision optimization model are as follows:

[0165] Irrigation water quantity constraint: the irrigation water quantity should be less than the available water quantity in the irrigation area:

[0166] I≤R (7)

[0167] Where R is the amount of water available for irrigation per unit area.

[0168] Yield constraints: the production output of irrigation areas must meet food demand:

[0169] ∑Y·A≥Y dem (8)

[0170] Where Y dem It is the grain demand in the area where the irrigation district is located.

[0171] Soil moisture content / water layer depth constraints: To ensure normal growth of crops, soil moisture content and water layer depth should be kept within an appropriate range:

[0172] S min ≤S≤S max (9)

[0173] h min ≤h≤h max (10)

[0174] In the formula, S is the soil moisture content, h is the depth of the water layer; S max With S min are the upper and lower limits of soil moisture suitable for crops; h max With h min They are the upper and lower limits of suitable water layer depth respectively.

[0175] Figure 4 This is the irrigation schedule optimization result for a certain grid in the decision area of ​​this embodiment.

[0176] S5: Establish a farmland information database based on remote sensing and Internet of Things technology; to build a dynamic irrigation system intelligent decision-making system based on crop growth monitoring, real-time decision-making and precision irrigation. Including establishing a farmland information database based on remote sensing and Internet of Things technology, and building a regional irrigation system rolling correction system.

[0177] Farmland information database based on remote sensing and Internet of Things technology, including crop data, meteorological data, soil data, hydrological data and field management data. Remote sensing images of the irrigation decision area are obtained through satellite platforms or drones, and crop data such as aboveground biomass and leaf area index are obtained by inversion with regression models; real-time soil data is obtained by arranging soil moisture monitors in the irrigation decision area, hydrological data is obtained through a series of runoff monitoring data from hydrological stations, and real-time meteorological data is obtained by setting up small meteorological stations in the irrigation decision area; field management data is obtained through real-time input of agricultural management measures by field managers, and agricultural management measures include but are not limited to: single irrigation amount, irrigation date, fertilization amount, fertilization date, ridge height, etc.

[0178] S6: Construct a rolling correction system for regional irrigation systems; use the obtained regional spatial grid long-term irrigation system as a reference irrigation system and apply it to the irrigation decision area. Read the farmland information database to the system in real time. According to the actual irrigation effect and the predicted effect, use an adaptive control algorithm to make short-term rolling corrections to the reference irrigation system and quickly adjust the irrigation decision.

[0179] Among them, the regional irrigation system rolling correction system includes:

[0180] Data reading module: real-time reading of farmland information database into the system, including but not limited to crop soil moisture conditions, precipitation, temperature and irrigation water volume data.

[0181] Data processing and storage module: Process and analyze the read data, including preprocessing work such as missing value processing, outlier detection and format standardization, and then store the data, back up the data regularly to prevent data loss, and have a recovery mechanism.

[0182] Reference irrigation system selection module: Decision makers select one of the long-term irrigation systems in the spatial grid of high water efficiency, high yield and water-saving areas as a reference irrigation system for application in the irrigation decision area based on their preference.

[0183] Intelligent prediction module: Based on the relevant historical data provided by the data processing and storage module, the machine learning algorithm is used to predict future crop water demand and water supply.

[0184] Crop growth module: imports relevant data from the data processing and storage module. The crop growth module is responsible for simulating the biomass and crop water consumption during the period.

[0185] Rolling correction module: determines whether the climate forecast results of the intelligent prediction module and the irrigation effect of the crop growth module meet the rolling correction constraints. When the rolling correction constraints are met, there is no need to roll back the irrigation system and the reference irrigation system is still used. When the rolling correction constraints are not met, early warning information is generated and the irrigation system should be re-decided to update the irrigation water volume and irrigation date in the short term to obtain a dynamic and high-efficiency irrigation system.

[0186] Figure 5 This is the rolling correction result map of the irrigation system. Irrigation management is carried out according to the reference irrigation system. If the rolling correction constraints are not met, the irrigation plan is corrected through an intelligent algorithm. Figure 5 In the first 29 days of the cotton growth period, the reference irrigation system is used for irrigation. If the actual aboveground biomass meets the rolling correction constraint, no rolling correction will be performed. However, according to the crop growth module simulation and prediction of crop growth status, if the reference irrigation system is still implemented in the next 15 days, the aboveground biomass will not reach the expected threshold and will not meet the rolling correction constraint. In this case, rolling correction is required. The irrigation plan is corrected through the intelligent algorithm, and the irrigation amount is updated to 30 mm, and the irrigation date is the 30th day.

[0187] User interface and interaction module: Provides a user interface for users to operate and monitor, including mobile applications, web pages and control panels. Users can view farmland environmental data, crop growth, irrigation plans and early warning information in real time through the interface, and perform operations and provide feedback.

[0188] System management and maintenance module: responsible for the management and maintenance of the system, including permission settings, performance monitoring, log management, software updates and troubleshooting. In addition, the system is also scalable to facilitate the addition of new functions and services.

[0189] The rolling correction constraints are as follows:

[0190] Biomass constraint for a period, when the biomass for a period is less than the biomass under the reference irrigation system, a rolling correction is required:

[0191] α·B≥B R (11)

[0192] Where, B is the biomass of the period simulated by the crop growth module; B R is the biomass under the reference irrigation system, and α is the correction factor.

[0193] Water supply constraint: when the water supply is less than the water demand, rolling correction is required:

[0194] Q S ≥Q D (12)

[0195] In the formula, Q S is the water supply; Q D For water demand.

[0196] The above embodiments are only preferred implementations of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A regional crop rolling correction intelligent irrigation decision method, characterized in that: The following steps are involved: S1: Basic data collection and irrigation decision-making area division; Use remote sensing technology, meteorological stations and field soil monitoring instruments to collect crop, soil, hydrological, meteorological and field management data in the irrigation decision area; divide the decision area into n sub-areas according to the soil type and field management method of the decision area; and rasterize each sub-area; S2: sensitivity analysis of crop growth model parameters; The Sobol method was used to conduct sensitivity analysis on crop growth model parameters, and non-sensitive parameters and sensitive parameters were screened out. Non-sensitive parameters used the system default values, and sensitive parameters needed to be calibrated, including crop parameters, soil parameters, and meteorological parameters. S3: Localization of crop growth model parameters; After each sub-region determines the model parameters that need to be calibrated, the NSGA-Ⅱ algorithm is combined with the crop model to calibrate the model parameters, set the minimum error target between the simulated value and the measured value, realize automatic calibration of the parameters in the crop growth model, obtain the crop parameters, soil parameters and meteorological parameters after the localization of the crop growth model, and establish a localized crop growth model. S4: Develop regional spatial long-term irrigation decision-making plan; Based on the localized crop growth model, the relevant data of the planning year are input into the crop growth model, including meteorological data, soil data, crop data and management data, to form a comprehensive crop growth model suitable for the planning year; the crop growth model is coupled with the multi-objective optimization model to construct a regional grid-based dynamic irrigation decision-making technology, set the maximum yield target and the minimum irrigation target, and assign weights to multiple targets through the TOPSIS comprehensive evaluation method to obtain different regional spatial grid long-term irrigation systems; S5: Establish a farmland information database based on remote sensing and Internet of Things technology; Including crop data, meteorological data, soil data, hydrological data and field management data; remote sensing images of the irrigation decision area are obtained through satellite platforms or drones, and aboveground biomass and leaf area index crop data are obtained by inversion with regression models; real-time soil data are obtained by arranging soil moisture monitors in the irrigation decision area, hydrological data are obtained through a series of runoff monitoring data from hydrological stations, and real-time meteorological data are obtained by setting up small meteorological stations in the irrigation decision area; field management data are obtained through real-time input of agricultural management measures by field managers, and agricultural management measures include but are not limited to: single irrigation amount, irrigation date, fertilization amount, fertilization date, ridge height; S6: Construct a rolling revision system for regional irrigation regimes; The obtained regional spatial grid long-term irrigation system is used as the reference irrigation system and applied to the irrigation decision-making area. The farmland information database is read into the system in real time. According to the actual irrigation effect and the predicted effect, the adaptive control algorithm is used to make short-term rolling corrections to the reference irrigation system to quickly adjust the irrigation decision.

2. According to claim 1, the method for regional crop rolling correction intelligent irrigation decision-making is characterized in that: The specific process of the Sobol method described in step S2 includes: S21: Determine input parameters and target variables: Select the model input parameters and their ranges to be analyzed, including crop parameters, soil parameters and meteorological parameters; determine the key target variables of the model, which are the leaf area index and aboveground biomass changes output by time series and the crop yield output by non-time series; S22: Generate input parameter samples: Use the Sobol sequence to generate a sample set of input parameters with low variance; S23: Run model simulation: Use the generated input parameter samples to run the model simulation one by one; each sample corresponds to a model simulation and the corresponding output results are recorded; S24: Calculate Sobol index: The Sobol index is a quantitative indicator used to measure the contribution of each input parameter to the model output; the first-order sensitivity index and the full-order sensitivity index are calculated by programming software, and the formula is as follows: First-order sensitivity index: In the formula, S i is the first-order sensitivity index, V i The input parameter x i V(Y) is the model variance caused by a single change, V(Y) is the total output variance of the model, m is the number of input parameters, V ij is the parameter x i With parameter x j The variance due to interaction, V 1…,m The variance of the output caused by the interaction of all parameters; Full-order sensitivity index: In the formula, ST i is the full-order sensitivity index, V -i is the variance of all parameters except the i-th parameter; S25: Analyze and interpret the results: The sensitivity index obtained by analysis is used to identify the sensitivity of input variables to crop yield or other target variables, and input parameters with larger sensitivity index are localized.

3. The method for regional crop rolling correction intelligent irrigation decision-making according to claim 2 is characterized in that: After the model parameters to be calibrated are determined in step S2, step S3 combines the NSGA-II algorithm with the crop growth model to calibrate the parameters. The specific steps are as follows: S31: Initialization: Set the initial population size, define the maximum number of algorithm iterations, set the crossover probability and mutation probability, and randomly generate the initial population. Each individual is a set of model parameter combinations. S32: Evaluation: According to the objective function, the fitness of each individual is evaluated, that is, the minimum error between the simulated value and the measured value; S33: Selection: Classify individuals by non-dominated sorting and calculate crowding distance to maintain diversity, and give priority to individuals with low non-dominated rank and large crowding distance; S34: Crossover: Perform a crossover operation on the selected parent individuals to generate new individuals; the crossover operation determines whether to perform a crossover by the crossover probability, and generates a new individual by combining the genetic information of the two parent individuals; Mutation: Mutate the individuals after crossover to increase population diversity and avoid falling into local optimality. The mutation operation determines whether to mutate based on the mutation probability and generates new individuals by randomly changing the individual gene information. S35: Evaluation: Evaluate the fitness of the newly generated individuals according to the objective function, calculate their performance in the optimization objective, and use it for sorting and screening; S36: Replacement: Select parents and offspring through non-dominated sorting and crowding distance, optimize and retain high-quality individuals, and generate the next generation population; S37: Termination condition: the algorithm is terminated when the maximum number of iterations is reached or the fitness converges; S38: Output result: Output the optimal solution set, that is, the optimal combination of model parameters that minimizes the error between the simulated value and the measured value.

4. The method for regional crop rolling correction intelligent irrigation decision-making according to claim 3 is characterized in that: The objective function of the parameter localization optimization model in step S3 is as follows: Where LAI i obs and LAI i sim are the measured and simulated values ​​of leaf area index, respectively. i obs and Bio i sim are the measured and simulated values ​​of aboveground biomass, Y i obs and Y i sim are the measured and simulated values ​​of crop yield, respectively; n is the sampling times of canopy coverage and biomass in the measured sampling or experiment, and m is the sampling times of crop yield.

5. The method for regional crop rolling correction intelligent irrigation decision-making according to claim 4 is characterized in that: In step S4, the formulation of a regional spatial long-term irrigation decision plan requires coupling the crop model with the multi-objective optimization model. The specific steps are as follows: S41: Data input: input the data required for the crop model in the planned year into the crop model, including meteorological data, soil data, field management data and localized model parameters, to create a crop model suitable for the planned year; determine the objective function: determine the optimization goal, i.e., maximizing crop yield and minimizing irrigation water; S42: Setting constraints: To reflect the limitation of available resources in actual production and ensure the normal growth of crops, constraints need to be set; S43: Model coupling: coupling the crop model and the optimization model, passing the decision variables of the optimization model as input to the crop model, and feeding back the crop yield and irrigation amount generated by the crop model to the optimization model; S44: Model solution: Use NSGA-II algorithm to solve the optimization model and output the Pareto solution set of crop yield and irrigation amount; S45: TOPSIS comprehensive evaluation method selection: Pareto solution set selection, and use TOPSIS comprehensive evaluation method to further select the Pareto solution set; The specific steps of the TOPSIS comprehensive evaluation method are as follows: S451: Determine the evaluation index: clarify the decision-making goal and select relevant evaluation indexes, i.e. the larger the crop yield value, the better, and the smaller the irrigation value, the better; S452: Construct a normalized decision matrix: organize and normalize the original data, organize the crop yield and irrigation data in the Pareto solution set into a decision matrix Di×j, i represents the number of Pareto solution sets, j represents the number of decision targets, and each row of the matrix is ​​an alternative plan. The linear normalization method is used to convert the data into a unified scale; determine the weight of the evaluation index: according to the decision maker's preference, different weight combinations are made for each evaluation index, which can be divided into a spatial long-term irrigation system plan for high water efficiency areas, a spatial long-term irrigation system plan for high-yield areas, and a spatial long-term irrigation system plan for water-saving areas; S453: Calculate positive and negative ideal solutions: calculate the distance from each alternative solution to the positive and negative ideal solutions; S454: Sorting: Calculate the relative proximity of each alternative plan to the positive and negative ideal solutions, and sort the alternative plans according to the relative proximity. The one with the largest value is the optimal regional spatial long-term irrigation decision plan under the preference angle.

6. The method for regional crop rolling correction intelligent irrigation decision-making according to claim 5 is characterized in that: The objective function of the regional spatial long-term irrigation decision optimization model is as follows: Obj2=maxY (5) Obj3=minI (6) In the formula, Y is the crop yield per unit area; I is the irrigation water per unit area; The constraints of the regional spatial long-term irrigation decision optimization model are as follows: Irrigation water quantity constraint: the irrigation water quantity should be less than the available water quantity in the irrigation area: I≤R (7) In the formula, R is the amount of water available for irrigation per unit area; Yield constraints: the production output of irrigation areas must meet food demand: ∑Y·A≥Y dem (8) In the formula, Y dem is the food demand in the area where the irrigation district is located; Soil moisture content / water layer depth constraints: To ensure normal growth of crops, soil moisture content and water layer depth should be kept within an appropriate range: S min ≤S≤S max (9) h min ≤h≤h max (10) In the formula, S is the soil moisture content, h is the depth of the water layer; S max With S min are the upper and lower limits of soil moisture suitable for crops; h max With h min They are the upper and lower limits of suitable water layer depth respectively.

7. The method for regional crop rolling correction intelligent irrigation decision-making according to claim 6 is characterized in that: The regional irrigation schedule rolling correction system in step S6 includes: Data reading module, used to read farmland information database into the system in real time, including but not limited to crop soil moisture status, precipitation, temperature and irrigation water volume data; Data processing and storage module, used to process and analyze the read data, including preprocessing of missing data, outlier detection and format standardization, and then store the data, back up the data regularly to prevent data loss, and have a recovery mechanism; The reference irrigation system selection module is used by decision makers to select one of the long-term irrigation systems in the spatial grid of high water efficiency, high yield and water-saving areas as a reference irrigation system for application in the irrigation decision area according to their preference perspective; Intelligent prediction module, which is used to predict future crop water demand and water supply based on relevant historical data provided by the data processing and storage module using machine learning algorithms; The crop growth module is used to import relevant data from the data processing and storage modules. The crop growth module is responsible for simulating the biomass and crop water consumption during the period; The rolling correction module is used to determine whether the climate prediction results of the intelligent prediction module and the irrigation effect of the crop growth module meet the rolling correction constraints. When the rolling correction constraints are met, there is no need to roll back the irrigation system and the reference irrigation system is still used. When the rolling correction constraints are not met, an early warning message is generated and the irrigation system should be re-decided to update the irrigation water volume and irrigation date in the short term to obtain a dynamic and high-water-efficiency irrigation system. The user interface and interaction module is used to provide a user interface for users to operate and monitor. The interface includes mobile applications, web pages and control panels. Users can view farmland environmental data, crop growth conditions, irrigation plans and early warning information in real time through the interface, and perform operations and provide feedback. The system management and maintenance module is responsible for the management and maintenance of the system, including permission settings, performance monitoring, log management, software updates and troubleshooting; in addition, the system is also scalable to facilitate the addition of new functions and services.

8. The regional crop rolling correction intelligent irrigation decision method according to claim 7 is characterized in that: The rolling correction constraints in step S6 are as follows: Biomass constraint for a period, when the biomass for a period is less than the biomass under the reference irrigation system, a rolling correction is required: α·B≥B R (11) Where, B is the biomass of the period simulated by the crop growth module; B R is the biomass under the reference irrigation regime, α is the correction factor; Water supply constraint: when the water supply is less than the water demand, rolling correction is required: Q S ≥Q D (12) In the formula, Q S is the water supply; Q D For water demand.

Images