Method and device for determining crop irrigation strategy and storage medium

Abstract

The invention discloses a method and device for determining a crop irrigation strategy and a storage medium. The method comprises the steps of obtaining historical planting data of a target planting area; inputting each first soil moisture content threshold group comprising a plurality of soil moisture content thresholds in different growth stages and historical planting data into a crop moisture production model to obtain a target function of crop yield and irrigation volume and a function value corresponding to each first soil moisture content threshold group, the target function is the highest crop yield and the minimum irrigation volume of the target planting area; and determining a first target soil moisture content threshold group corresponding to the target planting area according to the function values corresponding to all the first soil moisture content threshold groups, and taking the first target soil moisture content threshold group as an irrigation strategy of the target planting area. According to the scheme, the soil moisture content threshold values are set and optimized respectively according to the moisture requirements of the crops in different growth stages, the water utilization efficiency is improved through refined moisture management, and the optimal growth conditions of the crops in each growth period are met.

Description

Technical Field

[0001] This application relates to the field of agricultural intelligent management technology, and specifically to a method, apparatus and storage medium for determining crop irrigation strategies. Background Technology

[0002] In agricultural production, irrigation management is crucial for crop yield and quality. Traditional irrigation methods often rely on experience or pre-set plans, which cannot accurately respond to changes in crop water requirements and variable weather conditions, frequently leading to water waste or shortages. Furthermore, inappropriate irrigation strategies can cause soil salinization, groundwater level decline, and water quality problems, thereby impacting the health and sustainability of agricultural ecosystems. Summary of the Invention

[0003] The purpose of this application is to provide a method, apparatus, and storage medium for determining crop irrigation strategies, in order to solve the technical problem of inappropriate irrigation strategies affecting agricultural ecosystems in the prior art.

[0004] To achieve the above objectives, the first aspect of this application provides a method for determining crop irrigation strategies, the method comprising:

[0005] An initial population is generated, which includes multiple first moisture threshold groups, each of which includes multiple moisture thresholds for different reproductive stages.

[0006] Obtain historical planting data for the target planting area;

[0007] Each first soil moisture threshold group and historical planting data are input into the crop water production model to obtain the objective functions of crop yield and irrigation amount and the function values ​​corresponding to each first soil moisture threshold group. The function values ​​include the first crop yield and the first irrigation amount of the target planting area. The objective function is to maximize the crop yield and minimize the irrigation amount of the target planting area.

[0008] Based on the function values ​​corresponding to all the first moisture threshold groups, determine the first target moisture threshold group corresponding to the target planting area, and then determine the irrigation strategy for the target planting area based on the first target moisture threshold group and the irrigation amount corresponding to the first target moisture threshold group.

[0009] In embodiments of this application, the method further includes: determining the non-dominated relationships among the function values ​​corresponding to all first moisture threshold groups to determine a second moisture threshold group corresponding to each non-dominated level in all first moisture threshold groups; determining the crowding distance corresponding to each second moisture threshold group for each non-dominated level; selecting a second target moisture threshold group from all first moisture threshold groups based on the non-dominated level and crowding distance, and using the second target moisture threshold group as the parent generation moisture threshold group; performing crossover mutation operations on multiple moisture thresholds in the parent generation moisture threshold group to generate a offspring moisture threshold group corresponding to the parent generation moisture threshold group; and merging the parent generation moisture threshold group and the offspring generation moisture threshold group to update the initial population.

[0010] In embodiments of this application, the method further includes: obtaining a preset prior distribution of each model parameter corresponding to the crop water production model; sampling the preset prior distribution of each model parameter to generate a candidate parameter set for each iteration based on the parameter values ​​of all sampled model parameters; obtaining measured crop data of the target planting area over historical time; determining the acceptance probability corresponding to each candidate parameter set based on the measured crop data; for each iteration, determining whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set to determine a first target parameter set corresponding to each iteration; inputting historical planting data into the crop water production model corresponding to each first target parameter set to output simulated crop data corresponding to each first target parameter set through the crop water production model corresponding to each first target parameter set; and selecting a second target parameter set corresponding to the crop water production model from all first target parameter sets based on the measured crop data and all simulated crop data.

[0011] In the embodiments of this application, sampling is performed on a preset prior distribution of each model parameter to generate a candidate parameter set for each iteration based on the parameter values ​​of all sampled model parameters. This includes: sampling is performed on a preset prior distribution of each model parameter to generate an initial parameter set based on the parameter values ​​of all sampled model parameters; and a candidate parameter set corresponding to each iteration is generated based on the preset distribution and the initial parameter set.

[0012] In the embodiments of this application, generating a candidate parameter set corresponding to each iteration based on a preset distribution and an initial parameter set includes: determining the covariance matrix of the preset distribution; and generating a candidate parameter set corresponding to each iteration based on the covariance matrix and the initial parameter set.

[0013] In the embodiments of this application, for each iteration, determining whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set, in order to determine the first target parameter set corresponding to each iteration, includes: generating a random number corresponding to each iteration, the random number following a uniform distribution; for each iteration, if the acceptance probability corresponding to the candidate parameter set is greater than or equal to the random number, determining to perform an update operation based on the candidate parameter set, so as to determine the candidate parameter set as the first target parameter set corresponding to this iteration; for each iteration, if the acceptance probability corresponding to the candidate parameter set is less than the random number, prohibiting the update operation based on the candidate parameter set, and determining the first target parameter set corresponding to the previous iteration as the first target parameter set corresponding to this iteration.

[0014] In the embodiments of this application, determining the acceptance probability corresponding to each candidate parameter set based on measured crop data includes: for each iteration, determining a first likelihood function corresponding to the candidate parameter set and a second likelihood function corresponding to the first target parameter set corresponding to the previous iteration based on measured crop data; for each candidate parameter set, determining the acceptance probability corresponding to the candidate parameter set based on the first likelihood value and the second likelihood value.

[0015] In the embodiments of this application, for each iteration, it is determined whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set, so as to determine the first target parameter set corresponding to each iteration, wherein the acceptance probability corresponding to the candidate parameter set is calculated according to the following formula (1):

[0016]

[0017] Where t refers to the t-th iteration, t-1 refers to the first target parameter set corresponding to the (t-1)-th iteration, t-1 refers to the candidate parameter set corresponding to the t-th iteration, t-1 refers to the acceptance probability corresponding to the candidate parameter set corresponding to the t-th iteration, D refers to the measured crop data, t-1 refers to the first likelihood function corresponding to the candidate parameter set corresponding to the t-th iteration, t-2 refers to the second likelihood function corresponding to the first target parameter set corresponding to the (t-1)-th iteration, t-1 refers to the prior probability distribution corresponding to the candidate parameter set corresponding to the t-th iteration, and t-1 refers to the prior probability distribution corresponding to the first target parameter set corresponding to the (t-1)-th iteration.

[0018] In the embodiments of this application, determining the first target moisture threshold group corresponding to the target planting area based on the function values ​​corresponding to all first moisture threshold groups includes: determining the sales revenue and irrigation cost of the irrigation strategy corresponding to each first moisture threshold group; sorting all first moisture threshold groups according to sales revenue; determining the profit growth corresponding to each sorted first moisture threshold group according to irrigation cost; and determining the first moisture threshold group whose profit growth meets the preset conditions as the first target moisture threshold group corresponding to the target planting area.

[0019] A second aspect of this application provides an apparatus for determining a crop irrigation strategy, comprising:

[0020] The memory is configured to store instructions;

[0021] The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the method described above for determining a crop irrigation strategy.

[0022] A third aspect of this application provides a machine-readable storage medium storing instructions for causing a machine to perform the method described above for determining a crop irrigation strategy.

[0023] The above-mentioned approach sets and optimizes soil moisture thresholds for different growth stages of crops, rather than using a uniform threshold. This refined water management improves water use efficiency, meets the optimal growth conditions for crops at each growth stage, and provides a scientifically accurate irrigation strategy.

[0024] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0025] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings:

[0026] Figure 1 The illustration shows a flowchart of a method for determining a crop irrigation strategy according to an embodiment of this application;

[0027] Figure 2 The illustration shows a flowchart of a method for determining a crop irrigation strategy according to a specific embodiment of this application;

[0028] Figure 3 The diagram schematically illustrates a structural block diagram of an apparatus for determining a crop irrigation strategy according to an embodiment of this application;

[0029] Figure 4 The illustration shows a schematic diagram of the structure of a computer device according to an embodiment of the present application. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0031] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0032] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0033] Figure 1 The illustration schematically shows a flowchart of a method for determining a crop irrigation strategy according to an embodiment of this application. Figure 1 As shown in the figure, this application provides a method for determining a crop irrigation strategy, which may include the following steps.

[0034] S102, Generate an initial population. The initial population includes multiple first moisture threshold groups, and each first moisture threshold group includes multiple moisture thresholds for different reproductive stages.

[0035] S104, Obtain historical planting data for the target planting area.

[0036] S106, input each first soil moisture threshold group and historical planting data into the crop water production model to obtain the objective function of crop yield and irrigation amount and the function value corresponding to each first soil moisture threshold group. The function value includes the first crop yield and the first irrigation amount of the target planting area. The objective function is to maximize the crop yield and minimize the irrigation amount of the target planting area.

[0037] S108, determine the first target moisture threshold group corresponding to the target planting area based on the function values ​​corresponding to all the first moisture threshold groups, and determine the irrigation strategy for the target planting area based on the first target moisture threshold group and the irrigation amount corresponding to the first target moisture threshold group.

[0038] It can be understood that the soil moisture threshold refers to the maximum possible soil moisture level suitable for plant growth and development; that is, when soil moisture falls below the threshold, irrigation is triggered. The target planting area refers to the planting area where technicians conduct irrigation studies. Historical planting data includes, but is not limited to, historical meteorological data, soil baseline data, and field management data for the target planting area. Meteorological data includes, but is not limited to, maximum and minimum temperatures, rainfall, solar radiation, maximum and minimum relative humidity (RHmin), dew point temperature, and average wind speed. The crop water production model refers to the AquaCrop model. The AquaCrop model can accurately simulate the water requirements and consumption during crop growth, helping agricultural workers develop more scientific and efficient irrigation strategies. Through precise calculations of crop water requirements and supply, AquaCrop can improve water resource utilization efficiency and optimize crop yield and quality. To accurately predict irrigation demand in target planting areas and reduce water waste, irrigation decisions can be based on soil moisture. A multi-objective optimization model can be constructed with the objectives of maximizing crop yield and minimizing irrigation volume to derive the soil moisture thresholds and irrigation volumes guiding irrigation in the target planting area. Specifically, the initial population refers to a set of initial solutions generated randomly. Each individual in the initial population consists of multiple decision variables, which can be soil moisture thresholds at different growth stages. Specifically, these can be the seedling stage soil moisture threshold SMT1, the canopy growth stage soil moisture threshold SMT2, the maximum canopy stage soil moisture threshold SMT3, and the senescence stage soil moisture threshold SMT4. Further, each first soil moisture threshold group and historical planting data are input into the crop water production model. The crop water production model is used to simulate the crop growth process for each first soil moisture threshold group to obtain the objective function and the function value corresponding to each first soil moisture threshold group. The function value includes the first crop yield and the first irrigation volume of the target planting area. It can be understood that the objective function is to maximize crop yield and minimize irrigation volume in the target planting area. Therefore, the function values ​​corresponding to all the first soil moisture threshold groups can be compared, and the first soil moisture threshold group corresponding to the smaller first irrigation amount and the higher first crop yield can be selected as the first target soil moisture threshold group to guide irrigation. Then, the irrigation strategy for the target planting area can be determined based on the first target soil moisture threshold group and the corresponding irrigation amount. For example, if the soil moisture is lower than the seedling stage moisture threshold SMT1, irrigation can be triggered for the crop during the seedling stage.

[0039] The above approach sets and optimizes soil moisture thresholds for different water requirements at different growth stages of crops, rather than using a uniform threshold. This refined water management improves water use efficiency and meets the optimal growth conditions for crops at each growth stage.

[0040] In embodiments of this application, the method further includes: determining the non-dominated relationships among the function values ​​corresponding to all first moisture threshold groups to determine a second moisture threshold group corresponding to each non-dominated level in all first moisture threshold groups; determining the crowding distance corresponding to each second moisture threshold group for each non-dominated level; selecting a second target moisture threshold group from all first moisture threshold groups based on the non-dominated level and crowding distance, and using the second target moisture threshold group as the parent generation moisture threshold group; performing crossover mutation operations on multiple moisture thresholds in the parent generation moisture threshold group to generate a offspring moisture threshold group corresponding to the parent generation moisture threshold group; and merging the parent generation moisture threshold group and the offspring generation moisture threshold group to update the initial population.

[0041] It can be understood that non-dominance relationship refers to the superiority relationship between function values ​​corresponding to the first moisture threshold group. Specifically, if solution A is not inferior to solution B on all objectives, and is superior to solution B on at least one objective, then solution A is said to dominate solution B, denoted as A < B. If there is no such dominance relationship between two solutions, then they are non-dominant, denoted as A < dB. Based on the first moisture threshold group in the initial population, the population is stratified according to non-dominance relationships, with each stratum corresponding to a non-dominant level, and each stratum includes a set of non-dominant solutions. Specifically, after simulating the growth process corresponding to each first moisture threshold group in the initial population using a crop water production model, and calculating the corresponding first crop yield and first irrigation amount, the non-dominance relationship is determined by comparing the function values ​​corresponding to the first moisture threshold group, and the population is sorted based on the non-dominance relationship to obtain the second moisture threshold group corresponding to each non-dominant level in all first moisture threshold groups. Within the same non-dominant level, the distribution density of individuals in the target space can be assessed by the crowding distance within each second moisture threshold group within its non-dominant level. Then, a second target moisture threshold group can be selected from all the first moisture threshold groups based on the non-dominated level and crowding distance. Specifically, individuals with lower non-dominated levels can be selected, and within the same non-dominated level, individuals with larger crowding distances can be selected to form the second target moisture threshold group as the parent generation moisture threshold group. Further, crossover and mutation operations are performed on multiple moisture thresholds in the parent generation moisture threshold group to generate offspring moisture threshold groups corresponding to the parent generation moisture threshold groups. For example, the simulated binary crossover (SBX) method can be used to recombine the decision variables of the parent individuals, i.e., crossover and mutation are performed on SMT1 to SMT4 and then recombine to generate new moisture threshold groups. The parent generation moisture threshold groups and offspring generation moisture threshold groups are then merged to update the initial population. Therefore, the updated initial population can be input into the crop water production model to obtain the corresponding function value. Then, based on the function value, the non-dominant level and crowding distance are determined to select a new second target soil moisture threshold group to iterate the next generation population. The above steps are repeated until the preset number of iterations is reached or the convergence condition is met.

[0042] Specifically, this application provides a multi-objective optimization algorithm based on a genetic algorithm. The genetic algorithm can use the second-generation non-dominated sorting genetic algorithm (NSGA-II) to solve the multi-objective optimization model, and specifically includes the following steps.

[0043] 11. Population Initialization: In the initial population, multiple candidate solutions are randomly generated. Each individual consists of four decision variables, namely, soil moisture thresholds for different growth stages: seedling stage moisture threshold SMT1, canopy growth stage moisture threshold SMT2, maximum canopy stage moisture threshold SMT3, and senescence stage moisture threshold SMT4. Specifically, in the embodiments of this application, the upper and lower limits of the soil moisture threshold can be set to 30%–90%, and the initial population is generated using a random number generation function to ensure population diversity and solution coverage.

[0044] 12. Objective Function Setting: A multi-objective optimization model was established with the dual objectives of maximizing crop yield and minimizing irrigation amount. The specific objective function and its calculation formula are as follows:

[0045] (1) Objective function 1: Maximize crop yield

[0046] maxf1(SMT1,SMT2,SMT3,SMT4)=-Y(SMT1,SMT2,SMT3,SMT4)

[0047] Wherein, Y(SMTi) is the crop yield obtained by simulation by the AquaCrop model based on the current soil moisture threshold combination SMT, and the unit is tons per hectare (tonne / ha).

[0048] (2) Objective function 2: Minimize total irrigation amount

[0049] maxf2(SMT1,SMT2,SMT3,SMT4)=I(SMT1,SMT2,SMT3,SMT4)

[0050] Wherein, I(SMTi) is the total irrigation amount set based on the current soil moisture threshold combination SMT and simulated by the AquaCrop model, in millimeters (mm).

[0051] (3) Constraints: Soil moisture threshold range: SMT min <SMT i <SMT max i = 1, 2, 3, 4, SMT min Set to 30%, SMT max Set to 90%.

[0052] 13. Non-dominated ranking: For each individual's decision variables, combined with acquired meteorological, soil, and field management data, the data are input into the Aquacrop model to simulate the crop growth process. Based on the objective function, the first crop yield and first irrigation amount corresponding to each first soil moisture threshold group are calculated. The first irrigation amount refers to the total irrigation amount simulated throughout the entire growth period by the crop water production model. According to the Pareto optimization principle, the function values ​​among individuals are compared, and the individuals in the population are non-dominatedly ranked to determine the rank of each individual, thereby determining the second soil moisture threshold group corresponding to each non-dominated rank.

[0053] 14. Calculation of Crowding Distance: For the second soil moisture threshold group within the same non-dominated level, sort them in ascending order according to the first crop yield and the first irrigation amount. The crowding distance of boundary individuals is set to infinite to ensure that they are selected first. For individuals within the same non-dominated level, calculate the normalized distance between the two function values ​​as the crowding distance. The crowding distance can be used to assess the distribution density of individuals in the target space within the same non-dominated level and maintain population diversity. Specifically, the crowding distance can be calculated according to the following formula (2):

[0054]

[0055] Among them, f m,i Let be the value of the i-th individual on the m-th objective function, and M be the number of objective functions (M=2). Crowding Distance i It refers to the crowding distance of the i-th individual.

[0056] 15. Selection Mechanism: Based on non-dominance level and crowding distance, suitable individuals are selected as parents to participate in subsequent genetic operations. Specifically, individuals with lower non-dominance levels are preferentially selected, and within the same level, individuals with larger crowding distances are selected. A tournament selection method is used, where winners are selected from randomly selected individuals according to the above strategy as the parent soil moisture threshold group.

[0057] 16. Crossover and Mutation: The simulated binary crossover (SBX) method is used to recombine the decision variables of the parent individuals. Crossover is performed on SMT1 to SMT4, and a multinomial mutation method is used to slightly perturb the decision variables of the offspring with a certain mutation probability (e.g., 0.1), generating a set of soil moisture thresholds for the offspring, thereby avoiding the population from getting trapped in local optima. This scheme, by performing crossover and mutation operations on selected parent individuals to generate new offspring individuals, allows for the exploration of better irrigation strategies.

[0058] 17. Recombination and Selection: Merge the parent and offspring soil moisture threshold groups to form a new population with a size of 2N. Repeat the non-dominated sorting and crowding distance calculation. Select the top N individuals from the merged population to form the next generation, allowing individuals with higher fitness to enter the next generation. Repeat the above steps until the preset number of iterations (maximum number of generations) is reached or the convergence condition is met.

[0059] The above scheme aims to maximize crop yield and minimize total irrigation throughout the entire growth period, optimizing all SMT combinations simultaneously rather than iterating each SMT separately. The iterative process of the NSGA-II algorithm, including non-dominated sorting, crowding distance calculation, selection, crossover, and mutation, is used. The AquaCrop model is employed to simulate the crop growth process of each SMT combination, calculating the corresponding yield and total irrigation to evaluate the scheme's merits. A fast non-dominated sorting algorithm significantly reduces the algorithm's complexity, and an elitist strategy is introduced to expand the sampling space. Furthermore, crowding degree and crowding degree comparison operators are used to ensure rapid and uniform convergence to the Pareto optimal solution set. Each generation of the population is randomly generated under the set constraints. Specifically, considering both optimization accuracy and speed, the genetic algorithm's parameter settings are shown in Table 1 below.

[0060] Table 1. Parameter settings for the genetic algorithm

[0061]

[0062] In embodiments of this application, the method further includes: determining the first target moisture threshold group corresponding to the target planting area based on the function values ​​corresponding to all first moisture threshold groups, including: determining the sales revenue and irrigation cost of the irrigation strategy corresponding to each first moisture threshold group; sorting all first moisture threshold groups according to sales revenue; determining the profit growth corresponding to each sorted first moisture threshold group according to irrigation cost; and determining the first moisture threshold group whose profit growth meets the preset conditions as the first target moisture threshold group corresponding to the target planting area.

[0063] Specifically, economic indicators, including sales revenue and irrigation costs, can be calculated for each irrigation strategy corresponding to the first soil moisture threshold group.

[0064] Sales revenue (Sale) = Y(SMTi) × Sale_cotton, where Sale_cotton represents the selling price of the crop. Taking cotton as an example, the unit is yuan / ton.

[0065] Irrigation cost Pay_irrigation = I(SMTi) × Wat_irrigation × 10 -3 ×10 4Pay_irrigation represents the total cost required for irrigation per hectare of land, and Wat_irrigation represents the irrigation water fee, in yuan / ton.

[0066] Optimal solution selection: Sort all solutions by sales revenue from low to high, and select the solution with the lowest sales revenue as the initial optimal solution. Traverse the solutions and calculate profit growth: Starting from the second solution in the sorted solution list, compare and calculate the profit growth in turn. The profit growth is shown in (3) below:

[0067] profit_growth = (Sale current -Sale best )-(Pay_irrigation crrrent -Pay_irrigation best (3)

[0068] Among them, Sale current For the sales revenue of the current plan, Sale best Pay_irrigatio represents the sales revenue of the current optimal solution. current Pay_irrigation is the irrigation cost for the current scheme. best This represents the current optimal irrigation scheme. If profit_growth≥0, then the optimal solution is updated to the current scheme, and the corresponding optimal soil moisture threshold set can be output.

[0069] This scheme utilizes the NSGA-II algorithm to optimize soil moisture thresholds (SMT1-SMT4) at different growth stages, establishing a multi-objective optimization model with the dual objectives of maximizing crop yield and minimizing irrigation volume. Through steps such as population initialization, non-dominated ordination, crowding distance calculation, selection, crossover, and mutation, the AquaCrop model is used to simulate the crop growth process and evaluate the effectiveness of various irrigation strategies. In post-processing, economic indicators such as sales revenue and irrigation costs are calculated, and the optimal solution is selected based on the profit growth rate, ultimately outputting the best combination of soil moisture thresholds to guide irrigation decisions in actual production. An optimal solution selection mechanism based on profit growth rate is adopted, rather than relying solely on single indicators such as yield or irrigation volume. This scheme comprehensively considers sales revenue and irrigation costs, ensuring that the selected scheme maximizes yield while minimizing irrigation costs, thereby improving overall economic efficiency.

[0070] In embodiments of this application, the method further includes: obtaining a preset prior distribution of each model parameter corresponding to the crop water production model; sampling the preset prior distribution of each model parameter to generate a candidate parameter set for each iteration based on the parameter values ​​of all sampled model parameters; obtaining measured crop data of the target planting area over historical time; determining the acceptance probability corresponding to each candidate parameter set based on the measured crop data; for each iteration, determining whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set to determine a first target parameter set corresponding to each iteration; inputting historical planting data into the crop water production model corresponding to each first target parameter set to output simulated crop data corresponding to each first target parameter set through the crop water production model corresponding to each first target parameter set; and selecting a second target parameter set corresponding to the crop water production model from all first target parameter sets based on the measured crop data and all simulated crop data.

[0071] It is understandable that crop water production models require localization and calibration before application, meaning the model parameters need to be calibrated. Model parameters refer to the varietal genetic parameters of the crop water production model. These parameters include, but are not limited to, Tbase (base temperature), Tupper (upper limit temperature), CC0 (initial canopy cover), CGC (canopy growth coefficient), CCx (maximum canopy cover), CDC (canopy decay coefficient), and Kcb (crop coefficient). Specifically, a preset prior distribution for each model parameter can be obtained. This preset prior distribution can be a reasonable prior distribution set by technicians based on literature data for calibrating the AquaCrop model. By sampling a parameter value from the preset prior distribution for each model parameter, a candidate parameter set corresponding to each sampling can be generated based on the sampled parameter values. It is understood that the parameter values ​​in the candidate parameter set can be the sampled parameter values ​​or values ​​generated further based on the sampled parameter values.

[0072] Furthermore, historically measured crop data for the target planting area is obtained. This data includes, but is not limited to, the crop's growth period, aboveground biomass, and leaf area index. The time period for collecting the measured crop data is the same as the time period for collecting historical planting data. Then, the acceptance probability for each candidate parameter set is determined based on the measured crop data. The acceptance probability refers to the probability of deciding whether to accept the current candidate parameter set to update the result of the previous iteration. For each candidate parameter set, the acceptance probability can be used to determine whether to perform an update operation based on the current candidate parameter set. Specifically, the update operation refers to replacing the first target parameter set of the previous iteration with the current candidate parameter set. It can be understood that the first target parameter set is the result of each iteration of the parameter set; if the acceptance probability is high, the update can be accepted; if the acceptance probability is low, the update cannot be accepted. If an update is confirmed, the current candidate parameter set can be determined as the first target parameter set for this iteration; otherwise, the first target parameter set of the previous iteration can be determined as the first target parameter set for this iteration. In other words, if the acceptance probability in this iteration is very small, the first target parameter set remains unchanged; if the acceptance probability is large, the first target parameter set can be updated. Thus, the first target parameter set corresponding to each iteration can be determined based on the acceptance probability.

[0073] Furthermore, historical planting data is input into the crop water production model corresponding to each first target parameter set, so that the crop water production model corresponding to each first target parameter set outputs simulated crop data corresponding to each first target parameter set. Simulated crop data is crop data obtained by the crop water production model simulating the growth process based on the planting data. That is, crop water production models with different parameter values ​​will output deviations after inputting the same data. Therefore, the measured crop data collected at the same time can be compared with the simulated crop data to determine the deviation between the two. Then, for the simulated crop data with the smaller deviation selected from all the simulated crop data, its corresponding first target parameter set is the second target parameter set corresponding to the crop water production model. It can be understood that the second target parameter set refers to the parameter set of the model parameters when using the crop water production model.

[0074] In the embodiments of this application, sampling is performed on a preset prior distribution of each model parameter to generate a candidate parameter set for each iteration based on the parameter values ​​of all sampled model parameters. This includes: sampling is performed on a preset prior distribution of each model parameter to generate an initial parameter set based on the parameter values ​​of all sampled model parameters; and a candidate parameter set corresponding to each iteration is generated based on the preset distribution and the initial parameter set.

[0075] It is understandable that parameter calibration includes mainstream methods such as manual trial and error, particle swarm optimization, Monte Carlo simulation, and Bayesian optimization. Since obtaining the marginal probability P(D) of the observed data is very difficult and complex, the Markov Chain Monte Carlo (MCMC) method is used to simulate the posterior distribution, addressing the challenge of modeling complex posterior distributions in Bayesian inference. Because the MCMC method possesses powerful global search capabilities and can avoid local optima, it is chosen for model parameter calibration. Specifically, samples are taken from a pre-defined prior distribution for each model parameter. For each different model parameter, a parameter value is selected from its corresponding pre-defined prior distribution. An initial parameter set is formed based on the sampled parameter values ​​of all model parameters. It is understood that the initial parameter set is directly composed of the sampled parameter values. Furthermore, for the sampled initial parameter set, a candidate parameter set corresponding to each iteration can be generated based on the pre-defined distribution.

[0076] In the embodiments of this application, generating a candidate parameter set corresponding to each iteration based on a preset distribution and an initial parameter set includes: determining the covariance matrix of the preset distribution; and generating a candidate parameter set corresponding to each iteration based on the covariance matrix and the initial parameter set.

[0077] Specifically, the preset distribution is usually a symmetrical distribution, such as the normal distribution. From the preset distribution q(θ) * |θ t-1 Data is extracted from the data to generate a corresponding set of candidate parameters. The set of candidate parameters can be calculated according to the following formula (4):

[0078] θ * =θ t-1 +∈ (4)

[0079] Where, θ t-1 It refers to the set of the third parameters corresponding to the t-th iteration, θ * It refers to the third parameter set θ t-1 The corresponding set of candidate parameters, ∈~N(0,∑), where ∑ is the covariance matrix.

[0080] In the embodiments of this application, determining the acceptance probability corresponding to each candidate parameter set based on measured crop data includes: for each candidate parameter set, determining a first likelihood function corresponding to the candidate parameter set and a second likelihood function corresponding to the third parameter set based on measured crop data; and for each candidate parameter set, determining the acceptance probability corresponding to the candidate parameter set based on the first likelihood value and the second likelihood value.

[0081] It is understandable that the likelihood function is used to measure the difference between simulated crop data and measured crop data. Assuming that the difference follows a normal distribution, the expression of the likelihood function is shown in the following formula (5):

[0082]

[0083] Where L(θ) refers to the likelihood function, Y obs,i It refers to the i-th observation, Y sim,i (θ) refers to the i-th simulated crop data in the crop water production model under parameter θ, and σ refers to the standard deviation of the observation error.

[0084] In the embodiments of this application, for each iteration, it is determined whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set, so as to determine the first target parameter set corresponding to each iteration, wherein the acceptance probability corresponding to the candidate parameter set is calculated according to the following formula (1):

[0085]

[0086] Where t refers to the t-th iteration, t-1 refers to the first target parameter set corresponding to the (t-1)-th iteration, t-1 refers to the candidate parameter set corresponding to the t-th iteration, t-1 refers to the acceptance probability corresponding to the candidate parameter set corresponding to the t-th iteration, D refers to the measured crop data, t-1 refers to the first likelihood function corresponding to the candidate parameter set corresponding to the t-th iteration, t-2 refers to the second likelihood function corresponding to the first target parameter set corresponding to the (t-1)-th iteration, t-1 refers to the prior probability distribution corresponding to the candidate parameter set corresponding to the t-th iteration, and t-1 refers to the prior probability distribution corresponding to the first target parameter set corresponding to the (t-1)-th iteration.

[0087] In the embodiments of this application, for each iteration, determining whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set, in order to determine the first target parameter set corresponding to each iteration, includes: generating a random number corresponding to each iteration, the random number following a uniform distribution; for each iteration, if the acceptance probability corresponding to the candidate parameter set is greater than or equal to the random number, determining to perform an update operation based on the candidate parameter set, so as to determine the candidate parameter set as the first target parameter set corresponding to this iteration; for each iteration, if the acceptance probability corresponding to the candidate parameter set is less than the random number, prohibiting the update operation based on the candidate parameter set, and determining the first target parameter set corresponding to the previous iteration as the first target parameter set corresponding to this iteration.

[0088] Specifically, the random number follows a uniform distribution, which can be U(0,1). For the current iteration, if the acceptance probability is greater than or equal to the random number, the candidate parameter set is accepted; if the acceptance probability is less than the random number, the candidate parameter set is rejected, and the first target parameter set corresponding to the previous iteration remains unchanged, serving as the first target parameter set for the current iteration.

[0089] Figure 2 The illustration schematically shows a flowchart of a method for determining a crop irrigation strategy according to a specific embodiment of this application. Figure 2 As shown in the embodiments of this application, a method for determining crop irrigation strategies is provided, the method comprising:

[0090] 21. Data Acquisition: Acquire meteorological data, soil data, field management data, and actual crop data for the target planting area from the previous year. Meteorological data includes maximum temperature (Tmax), minimum temperature (Tmin), rainfall (Pre), solar radiation (Srad), maximum relative humidity (RHmax), minimum relative humidity (RHmin), dew point temperature (Tdew), and average wind speed (Wind).

[0091] 22. Construction of Aquacrop model database: The potential evapotranspiration (ET0) of crops is calculated by combining the above data with the Penman equation. The specific ET0 calculation method is based on the following formula (6):

[0092]

[0093] Where ET0 represents the evapotranspiration of the reference crop, Δ is the slope of the saturated vapor pressure curve, and R... n Let G be the net surface radiation, and e be the soil heat flux. s e is the saturated vapor pressure. a U2 is the actual vapor pressure, T is the daily average air temperature, U2 is the wind speed at 2m altitude, and γ is the humidity constant. For detailed calculation steps of various parameters, please refer to "GB T 20481-2017 Meteorological Drought Classification". The input data for the Aquacrop model includes daily maximum temperature (Tmax), daily minimum temperature (Tmin), daily precipitation (Prcp), and daily evapotranspiration (ET0). Therefore, the calculation of ET0 is a crucial step in the data input process.

[0094] 23. Calibration and evaluation of genetic parameters of model varieties: The parameters of crop water production models can be calibrated using the MCMC method to complete the localization of model parameters.

[0095] 24. Multi-objective optimization: The NSGA-II algorithm is used to construct a multi-objective optimization model with soil moisture thresholds at multiple different growth stages as decision variables and the objective functions of maximizing crop yield and minimizing total irrigation. The second-generation non-dominated sorting genetic algorithm (NSGA-II) is used to solve the multi-objective optimization model.

[0096] 25. Obtain meteorological forecast data and input the obtained forecast data into the crop water production model for model iteration simulation. At the same time, combine the optimal first target soil moisture threshold group to obtain real-time irrigation decision results, and obtain the final optimal soil moisture threshold, maximum yield, and minimum irrigation amount.

[0097] The above scheme closely integrates the NSGA-II multi-objective optimization algorithm with the AquaCrop crop water production model, simultaneously optimizing soil moisture thresholds (SMT1 to SMT4) at different growth stages. This fusion allows for the simultaneous consideration of maximizing yield and minimizing total irrigation, providing a more refined and efficient irrigation strategy.

[0098] This application provides a method for parameter calibration of a crop water production model, specifically including the following steps:

[0099] 231. Determine the prior distribution of model parameters: Based on the literature, set a reasonable prior distribution for the calibration parameters of the AquaCrop model. The prior distribution range of the genetic parameters of the varieties is shown in Table 2 below.

[0100] Table 2. Prior distribution range of genetic parameters of varieties

[0101]

[0102]

[0103]

[0104] 232. Construct a Bayesian calibration framework. The functional expression of the Bayesian calibration framework is shown in the following formula (7):

[0105]

[0106] Here, P(θ|D) is the posterior probability distribution of the parameters, P(D|θ) is the likelihood function, which represents the probability of observed data D occurring under parameter θ, P(θ) is the prior probability distribution of the parameters, and P(D) represents the marginal probability of the observed data, serving as a normalization constant. However, obtaining P(D) is very difficult and complex. Therefore, the MCMC method is used to simulate the posterior distribution, solving the problem of the difficulty in modeling complex posterior distributions in Bayesian inference.

[0107] 233. Constructing the likelihood function: The likelihood function is used to measure the difference between simulation results and measured data. Assuming that the error follows a normal distribution, the expression of the likelihood function is shown in the following formula (5):

[0108]

[0109] Where L(θ) refers to the likelihood function, Y obs,i It refers to the i-th observation, Y sim,i (θ) refers to the i-th simulated crop data in the crop water production model under parameter θ, and σ refers to the standard deviation of the observation error.

[0110] 234. Using the Metropolis-Hastings algorithm as the MCMC sampling method: The algorithm is implemented using the PyMC package in the Python library, specifically including the following steps:

[0111] (1) Initialize parameters: Randomly select a parameter value from the preset prior distribution of each model parameter as the initial parameter set θ0.

[0112] (2) Iterative sampling: For each iteration, there are t = 1, 2, ..., N, where t is the number of iterations.

[0113] (3) Generate candidate parameters θ * From the preset distribution q(θ) * |θ t-1 Data is extracted from the dataset to generate the corresponding candidate parameter set θ. * A symmetrical distribution, such as the normal distribution, is usually chosen. The candidate parameter set can be calculated according to the following formula (4):

[0114] θ * =θ t-1 +∈ (4)

[0115] Where, θ t-1 It refers to the set of the third parameters corresponding to the t-th iteration, θ * It refers to the third parameter set θ t-1 The corresponding set of candidate parameters, ∈~N(0,∑), where ∑ is the covariance matrix.

[0116] (4) Calculate the acceptance probability α. The acceptance probability corresponding to the candidate parameter set is calculated according to the following formula (1):

[0117]

[0118] Where t refers to the t-th iteration, t-1 refers to the first target parameter set corresponding to the (t-1)-th iteration, t-1 refers to the candidate parameter set corresponding to the t-th iteration, t-1 refers to the acceptance probability corresponding to the candidate parameter set corresponding to the t-th iteration, D refers to the measured crop data, t-1 refers to the first likelihood function corresponding to the candidate parameter set corresponding to the t-th iteration, t-2 refers to the second likelihood function corresponding to the first target parameter set corresponding to the (t-1)-th iteration, t-1 refers to the prior probability distribution corresponding to the candidate parameter set corresponding to the t-th iteration, and t-1 refers to the prior probability distribution corresponding to the first target parameter set corresponding to the (t-1)-th iteration.

[0119] (5) Accept or reject: Generate a uniformly distributed random number u ~ U(0,1). If u ≤ α, accept θ. * , that is, θ t =θ * Otherwise, reject θ * Let θ t =θ t-1 .

[0120] (6) Model Validation and Evaluation: The parameter set obtained from the calibration was obtained by running the Metropolis-Hastings algorithm (as shown in Table 3 below). The AquaCrop model was run to simulate the crop growth process. The simulated operation data was compared with the measured crop data, and the accuracy of the parameter calibration was evaluated by the root mean square error (RMSE) and the coefficient of determination (R2).

[0121] Table 3. Prior distribution range of genetic parameters of varieties

[0122]

[0123]

[0124] The above scheme employs the Metropolis-Hastings algorithm as the MCMC sampling method. Within the Bayesian theoretical framework, the generated Markov chain serves as the sample set of the second objective parameter set, ensuring its stationary distribution tends towards the posterior probability distribution, thus addressing the difficulty of modeling complex posterior distributions in Bayesian inference. Because the MCMC method (Markov Chain Monte Carlo) possesses powerful global search capabilities and can avoid local optima, its sampling results exhibit good statistical properties and convergence, thereby improving the accuracy of the estimation.

[0125] In the embodiments of this application, irrigation decisions in traditional crop water production models are mostly based on historical data for future prediction. To overcome this deficiency, the crop water production model, after parameter calibration, can be simulated and iterated using real-time weather forecasts to accurately reflect the environmental conditions of the current growing season. A timed task is set to acquire weather forecast data every 30 minutes, and the acquired forecast data is input into the crop water production model for simulation. Simultaneously, the real-time irrigation decision results are obtained by combining the optimal first target soil moisture threshold group. The irrigation decision results are shown in Table 4 below (taking the first 100 iterations as an example).

[0126]

[0127]

[0128]

[0129] Real-time meteorological data was incorporated into the model simulation process, making the simulation results more closely reflect the actual environmental conditions of the current growing season. This improves the accuracy of model predictions and makes the optimized soil moisture threshold combinations more practical and feasible.

[0130] Figure 1 This is a flowchart illustrating a method for determining a crop irrigation strategy in one embodiment. It should be understood that, although... Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order in which these steps are executed, and they can be performed in other orders. Furthermore, Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0131] Figure 3 The diagram schematically illustrates a structural block diagram of an apparatus for determining a crop irrigation strategy according to an embodiment of this application. Figure 3 As shown in the figure, this application provides an apparatus for determining a crop irrigation strategy, which may include:

[0132] The memory is configured to store instructions; and

[0133] A processor configured to retrieve instructions from memory and, when executing the instructions, to implement the aforementioned apparatus and method for determining a crop irrigation strategy.

[0134] This application also provides a machine-readable storage medium storing instructions for causing a machine to perform the aforementioned apparatus and method for determining a crop irrigation strategy.

[0135] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor A01, a network interface A02, memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A04. The database stores data used to determine crop irrigation strategies. The network interface A02 communicates with external terminals via a network connection. When executed by the processor A01, the computer program B02 implements a method for determining crop irrigation strategies.

[0136] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0137] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0138] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0139] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0140] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0141] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0142] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0143] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0144] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0145] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for determining crop irrigation strategies, characterized in that, The method includes: An initial population is generated, which includes multiple first moisture threshold groups, each of which includes multiple moisture thresholds for different reproductive stages. Obtain historical planting data for the target planting area; Each first soil moisture threshold group and the historical planting data are input into the crop water production model to obtain the objective function of crop yield and irrigation amount and the function value corresponding to each first soil moisture threshold group. The function value includes the first crop yield and the first irrigation amount of the target planting area. The objective function is to maximize the crop yield and minimize the irrigation amount of the target planting area. Based on the function values ​​corresponding to all the first moisture threshold groups, a first target moisture threshold group corresponding to the target planting area is determined, and an irrigation strategy for the target planting area is determined based on the first target moisture threshold group and the irrigation amount corresponding to the first target moisture threshold group.

2. The method for determining crop irrigation strategies according to claim 1, characterized in that, Also includes: Determine the non-dominated relationships among the function values ​​corresponding to all first moisture threshold groups, so as to determine the second moisture threshold group corresponding to each non-dominated level in all first moisture threshold groups; For each non-dominated level, determine the crowding distance corresponding to each second moisture threshold group; Based on the non-dominant level and the crowding distance, a second target moisture threshold group is selected from all the first moisture threshold groups, so that the second target moisture threshold group is used as the parent moisture threshold group. A crossover mutation operation is performed on multiple moisture thresholds in the parent generation moisture threshold group to generate a child generation moisture threshold group corresponding to the parent generation moisture threshold group. The parent generation moisture threshold set and the offspring generation moisture threshold set are merged to update the initial population.

3. The method for determining crop irrigation strategies according to claim 1, characterized in that, The method further includes: Obtain the preset prior distribution of each model parameter corresponding to the crop water production model; Samples are taken from a preset prior distribution for each model parameter to generate a set of candidate parameters for each iteration based on the parameter values ​​of all model parameters obtained from the sampling. Obtain measured crop data for the target planting area at the historical time. The acceptance probability corresponding to each candidate parameter set is determined based on the measured crop data. For each iteration, it is determined whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set, so as to determine the first target parameter set corresponding to each iteration; The historical planting data is input into the crop water production model corresponding to each first target parameter set, so that the crop water production model corresponding to each first target parameter set outputs the simulated crop data corresponding to each first target parameter set. Based on the measured crop data and all simulated crop data, a second set of objective parameters corresponding to the crop water production model is selected from the entire first set of objective parameters.

4. The method for determining crop irrigation strategies according to claim 3, characterized in that, The step of sampling from a preset prior distribution for each model parameter to generate a candidate parameter set for each iteration based on the parameter values ​​of all sampled model parameters includes: Samples are taken from a preset prior distribution for each model parameter to generate an initial parameter set based on the parameter values ​​of all model parameters obtained from the sampling. The candidate parameter set is generated for each iteration based on the preset distribution and the initial parameter set.

5. The method for determining crop irrigation strategies according to claim 4, characterized in that, The step of generating a candidate parameter set corresponding to each iteration based on the preset distribution and the initial parameter set includes: Determine the covariance matrix of the preset distribution; A set of candidate parameters corresponding to each iteration is generated based on the covariance matrix and the initial parameter set.

6. The method for determining crop irrigation strategies according to claim 4, characterized in that, For each iteration, determining whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set, to determine the first target parameter set corresponding to each iteration includes: Generate a random number corresponding to each iteration, wherein the random number follows a uniform distribution; For each iteration, if the acceptance probability corresponding to the candidate parameter set is greater than or equal to the random number, it is determined to perform an update operation based on the candidate parameter set, so as to determine the candidate parameter set as the first target parameter set corresponding to this iteration; For each iteration, if the acceptance probability corresponding to the candidate parameter set is less than the random number, the update operation based on the candidate parameter set is prohibited, and the first target parameter set corresponding to the previous iteration is determined as the first target parameter set corresponding to this iteration.

7. The method for determining crop irrigation strategies according to claim 4, characterized in that, The step of determining the acceptance probability for each candidate parameter set based on the measured crop data includes: For each iteration, a first likelihood function corresponding to the candidate parameter set and a second likelihood function corresponding to the first target parameter set of the previous iteration are determined based on the measured crop data. For each set of candidate parameters, the acceptance probability corresponding to the set of candidate parameters is determined based on the first likelihood value and the second likelihood value.

8. The method for determining crop irrigation strategies according to claim 7, characterized in that, For each iteration, it is determined whether to perform an update operation based on the acceptance probability corresponding to the candidate parameter set, so as to determine the first target parameter set corresponding to each iteration. The acceptance probability corresponding to the candidate parameter set is calculated according to the following formula (1): Where t refers to the t-th iteration, θ t-1 It refers to the set of the first objective parameters corresponding to the (t-1)th iteration, θ * It refers to the set of candidate parameters corresponding to the t-th iteration, and α refers to the set of candidate parameters θ corresponding to the t-th iteration. * The corresponding acceptance probability, D, refers to the measured crop data, P(D|θ) * ) refers to the set of candidate parameters θ corresponding to the t-th iteration. * The corresponding first likelihood function, P(D|θ) t-1 ) refers to the first objective parameter set θ corresponding to the (t-1)th iteration. t-1 The corresponding second likelihood function, P(θ) * ) refers to the set of candidate parameters θ corresponding to the t-th iteration. * The corresponding prior probability distribution, P(θ) t-1 ) refers to the first objective parameter set θ corresponding to the (t-1)th iteration. t-1 The corresponding prior probability distribution.

9. The method for determining crop irrigation strategies according to claim 1, characterized in that, The step of determining the first target moisture threshold group corresponding to the target planting area based on the function values ​​corresponding to all the first moisture threshold groups includes: Determine the sales revenue and irrigation costs of the irrigation strategy corresponding to each first soil moisture threshold group; Sort all first soil moisture threshold groups according to sales revenue; The profit growth corresponding to each first soil moisture threshold group after sorting is determined based on irrigation costs; The first moisture threshold group that meets the preset conditions for profit growth is determined as the first target moisture threshold group corresponding to the target planting area.

10. An apparatus for determining a crop irrigation strategy, characterized in that, include: The memory is configured to store instructions; A processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the method for determining a crop irrigation strategy according to any one of claims 1 to 9.

11. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to perform the method for determining a crop irrigation strategy according to any one of claims 1 to 9.

Images