Water and fertilizer application amount decision-making method based on cooperation of yield and water and fertilizer utilization efficiency

Abstract

The invention discloses a water and fertilizer application amount decision-making method based on cooperation of yield and water and fertilizer utilization efficiency, and relates to the technical field of agricultural intelligent decision-making. The method comprises the following steps: each layer in a multi-layer farmland crop water and fertilizer real-time decision module corresponds to a growth period of crops; the input data of the irrigation consumption decision-making module comprises a crop growth variable, a meteorological variable, an irrigation consumption predicted value of the last growth period and an ammonia fertilization amount predicted value of the last growth period, and the output data of the irrigation consumption decision-making module is an irrigation consumption recommended value of the current growth period; the input data of the ammonia consumption decision-making module further comprises an irrigation water consumption recommended value in the current growth period, and the output data of the ammonia consumption decision-making module is an ammonia fertilization amount recommended value in the current growth period; and the irrigation water consumption and the ammonia fertilization amount of the crops in the current growth and development period are predicted through the multi-layer farmland crop water and fertilizer real-time decision-making module. The method can dynamically adjust the water and fertilizer application amount.

Description

Technical Field

[0001] This invention relates to the field of agricultural intelligent decision-making technology, and in particular to a method for determining the amount of water and fertilizer applied in a way that coordinates yield and water and fertilizer use efficiency. Background Technology

[0002] Field water and fertilizer management is crucial for ensuring crop yield levels and balancing farmland ecological environment goals. Effective field management guidelines are urgently needed to address the increasing abiotic stresses caused by climate change, population growth, and rising food prices. This study aims to calibrate and validate crop models using measured climate, soil, and crop growth data, and to utilize field management practice scenario analysis to identify water and fertilizer application strategies suitable for local farmland crops.

[0003] However, existing agricultural crop water and fertilizer decision-making technologies are usually focused on optimizing the single objective of irrigation or fertilization. They complete the task of optimizing the supply strategy under a limited number of meteorological conditions and water and fertilizer application scenarios, and cannot be adjusted in a timely manner according to the actual situation. This hinders the identification of the optimal water and fertilizer pattern for crop response to agricultural inputs. Summary of the Invention

[0004] Therefore, it is necessary to provide a water and fertilizer application rate decision-making method that coordinates yield and water and fertilizer use efficiency to address the aforementioned technical problems. This method improves the accuracy of water and fertilizer application rate determination.

[0005] The present invention adopts the following technical solution: This invention provides a water and fertilizer application rate decision-making method that coordinates yield and water and fertilizer use efficiency, comprising: This system establishes a multi-layered real-time decision-making module for crop water and fertilizer application. Each layer in this module corresponds to a specific growth stage of the crop. Each layer includes an irrigation dosage decision module and an ammonium fertilizer dosage decision module. The input data for the irrigation dosage decision module includes crop growth variables, meteorological variables, predicted irrigation water usage for the previous growth stage, and predicted ammonium fertilizer application for the previous growth stage. The output data for the irrigation dosage decision module is the recommended irrigation water usage for the current growth stage. The input data for the ammonium fertilizer dosage decision module includes crop growth variables, meteorological variables, predicted irrigation water usage for the previous growth stage, predicted ammonium fertilizer application for the previous growth stage, and recommended irrigation water usage for the current growth stage. The output data for the ammonium fertilizer dosage decision module is the recommended ammonium fertilizer application for the current growth stage. The multi-layered real-time decision-making module for farmland crop water and fertilizer predicts the amount of irrigation water and nitrogen fertilizer applied during the current growth and development stage of crops.

[0006] Preferably, the construction process of the multi-layered real-time decision-making module for crop water and fertilizer management in farmland includes: Obtain the modeling dataset; the modeling dataset includes crop growth variables, meteorological data, and predicted values ​​of irrigation water consumption and nitrogen fertilizer application for each growth stage of the crop; Using the modeling dataset, the initial multi-layered real-time decision-making module for water and fertilizer in farmland crops was trained to obtain the multi-layered real-time decision-making module for water and fertilizer in farmland crops.

[0007] Preferably, obtaining the modeling dataset specifically includes: The genetic parameters of the crop model were calibrated based on the growth indicators and final yield of the crop at each growth stage as determined by field trials in the study area. Historical meteorological observation data of the study area were obtained. The irrigation water volume and nitrogen fertilizer application volume corresponding to different growth stages, as well as meteorological variables corresponding to historical meteorological observation data, were set to drive the calibrated crop model to generate crop growth variables and crop yields corresponding to different growth stages. The initial target yield model was trained using crop yield and irrigation water and nitrogen fertilizer application rates corresponding to different growth stages to obtain the target yield model. The input data for the target yield model were the irrigation water and nitrogen fertilizer application rates corresponding to different growth stages, and the output data was the crop yield. With crop yield and water and fertilizer utilization rate as the top preset percentages as the target, the parameters of the target yield model are optimized by particle swarm optimization algorithm to obtain the predicted values ​​of irrigation water consumption and nitrogen fertilizer application for each growth stage. The predicted values ​​of irrigation water consumption and nitrogen fertilizer application for each growth stage, along with the corresponding crop growth variables and meteorological variables, are determined as the modeling dataset.

[0008] Preferably, the genetic parameters of the crop model are calibrated based on the growth indicators and final yield of the crop at each growth stage as determined by field trials in the study area. Specifically, this includes: The crop model is run based on the initial values ​​of the crop model's genetic parameters to obtain the predicted values ​​of growth indicators and yield for each growth and development stage in the current cycle. Calculate the difference between the predicted growth index values ​​for each growth and development stage in the current cycle and the growth index values ​​for each growth and development stage to obtain the prediction error of the growth index for each growth and development stage. Calculate the difference between the predicted output value and the actual output value for the current round to obtain the output prediction error; Based on the prediction errors of growth indicators and yield, the genetic parameters are iteratively updated until both the prediction errors of growth indicators and yield are minimized.

[0009] Preferably, the formula corresponding to the target output model is: ; in, The crop yield output by the target yield model. For the target output model, W 1 , W 2 , W 3 , W 4 These are the irrigation water amounts for the first, second, third, and fourth growth and development stages, respectively. N 1 , N 2 , N 3 , N 4 The nitrogen fertilization amounts are for the first, second, third, and fourth growth stages, respectively.

[0010] Preferably, with crop yield and water and fertilizer utilization rate ranked as the top preset percentages as the target, the parameters of the target yield model are optimized using a particle swarm optimization algorithm to obtain the predicted values ​​of irrigation water consumption and nitrogen fertilizer application for each growth stage, specifically including: Initialize the parameters of the particle swarm optimization algorithm; the parameters include population size, inertia weight, learning factor, maximum number of iterations, and search space range; The position and velocity of each particle are randomly initialized, and the current position of each particle is set as the individual's historical best position, while the position of the best particle in the swarm is set as the global best position; the position of the particle is the predicted value of irrigation water consumption and nitrogen fertilizer application corresponding to different growth and development stages. Calculate the fitness function value for each particle; the fitness function is constructed based on crop yield and water and fertilizer use efficiency. Compare the current fitness value with the individual's historical best position. If the current fitness value is better, update the individual's historical best position using the current fitness value. Compare the current fitness value with the global historical best position. If the current fitness value is better, then update the global best position using the current fitness value. The velocity and position of each particle are iteratively updated according to the velocity update formula and the position update formula until the termination condition is reached; the termination condition is that the crop yield and water and fertilizer utilization rate are ranked first by a preset percentage. The global optimal position corresponding to the termination condition is determined as the predicted value of irrigation application rate and nitrogen fertilizer application rate.

[0011] This invention provides a water and fertilizer application rate decision-making device that coordinates yield and water and fertilizer use efficiency, comprising: The acquisition module is used to acquire data from the multi-layered real-time water and fertilizer decision-making module for farmland crops. Each layer in the multi-layered real-time water and fertilizer decision-making module corresponds to a growth stage of the crop. Each layer in the multi-layered real-time water and fertilizer decision-making module includes an irrigation amount decision-making module and an ammonia amount decision-making module. The input data of the irrigation amount decision-making module includes crop growth variables, meteorological variables, the predicted irrigation water amount and the predicted ammonia fertilizer amount for the previous growth stage, and the output data of the irrigation amount decision-making module is the recommended irrigation water amount for the current growth stage. The input data of the ammonia amount decision-making module includes crop growth variables, meteorological variables, the predicted irrigation water amount and the predicted ammonia fertilizer amount for the previous growth stage, and the recommended irrigation water amount for the current growth stage, and the output data of the ammonia amount decision-making module is the recommended ammonia fertilizer amount for the current growth stage. The prediction module is used to predict the amount of irrigation water and nitrogen fertilizer applied during the current growth and development stage of crops through a multi-layered real-time decision-making module for farmland crop water and fertilizer.

[0012] The present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for determining the amount of water and fertilizer applied in coordination with yield and water and fertilizer utilization efficiency.

[0013] The present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-mentioned method for determining the amount of water and fertilizer application that coordinates yield and water and fertilizer utilization efficiency.

[0014] The above-mentioned at least one technical solution adopted in this invention can achieve the following beneficial effects: This system establishes a multi-layered real-time decision-making module for crop water and fertilizer application. Each layer in this module corresponds to a specific growth stage of the crop. Each layer includes an irrigation dosage decision module and an ammonium fertilizer dosage decision module. The input data for the irrigation dosage decision module includes crop growth variables, meteorological variables, predicted irrigation water usage for the previous growth stage, and predicted ammonium fertilizer application for the previous growth stage. The output data for the irrigation dosage decision module is the recommended irrigation water usage for the current growth stage. The input data for the ammonium fertilizer dosage decision module includes crop growth variables, meteorological variables, predicted irrigation water usage for the previous growth stage, predicted ammonium fertilizer application for the previous growth stage, and recommended irrigation water usage for the current growth stage. The output data for the ammonium fertilizer dosage decision module is the recommended ammonium fertilizer application for the current growth stage. The system then predicts the irrigation water usage and ammonium fertilizer application for the crop during its current growth and development stage. By employing a hierarchical decision-making architecture—for example, with each layer corresponding to a specific crop growth stage—this method integrates growth variables, meteorological data, and historical water and fertilizer forecasts in real time to dynamically adjust irrigation and nitrogen application rates. This approach enables dynamic adjustment of water and fertilizer application rates. Attached Figure Description

[0015] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0016] Figure 1 A schematic diagram of a water and fertilizer application rate decision-making method that coordinates yield and water and fertilizer utilization efficiency provided by the present invention. Figure 2 The water and fertilizer application rate decision-making flowchart provided by this invention; Figure 3 A schematic diagram of the real-time decision-making model for water and fertilizer management of farmland crops provided by the present invention; Figure 4 A schematic diagram of a water and fertilizer application rate decision device that coordinates yield and water and fertilizer utilization efficiency, provided by the present invention. Figure 5 A schematic diagram of a computer device for a water and fertilizer application rate decision-making method that achieves synergy between yield and water and fertilizer utilization efficiency, provided by the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0018] Devices such as desktop computers, servers, and laptops are capable of executing the present invention. For ease of explanation, the following description will focus on servers as the executing entity.

[0019] The technical solutions provided by the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0020] Figure 1 This is a schematic diagram of a water and fertilizer application rate decision-making method that coordinates yield and water and fertilizer use efficiency in this invention, specifically including the following steps: S101: Obtain the multi-layered real-time decision-making module for farmland crop water and fertilizer; each layer in the multi-layered real-time decision-making module for farmland crop water and fertilizer corresponds to a growth stage of the crop; each layer in the multi-layered real-time decision-making module for farmland crop water and fertilizer includes an irrigation amount decision-making module and an ammonia amount decision-making module; the input data of the irrigation amount decision-making module includes crop growth variables, meteorological variables, the predicted irrigation water amount for the previous growth stage, and the predicted ammonia fertilizer amount for the previous growth stage, and the output data of the irrigation amount decision-making module is the recommended irrigation water amount for the current growth stage; the input data of the ammonia amount decision-making module includes crop growth variables, meteorological variables, the predicted irrigation water amount for the previous growth stage, the predicted ammonia fertilizer amount for the previous growth stage, and the recommended irrigation water amount for the current growth stage, and the output data of the ammonia amount decision-making module is the recommended ammonia fertilizer amount for the current growth stage.

[0021] In an exemplary embodiment, the construction process of the multi-layered farmland crop water and fertilizer real-time decision-making module includes: acquiring a modeling dataset; the modeling dataset includes crop growth variables, meteorological data, and predicted values ​​of irrigation water consumption and nitrogen fertilizer application for each growth stage of the crop; using the modeling dataset, training the initial multi-layered farmland crop water and fertilizer real-time decision-making module to obtain the multi-layered farmland crop water and fertilizer real-time decision-making module.

[0022] In an exemplary embodiment, obtaining the modeling dataset specifically includes: calibrating the genetic parameters of the crop model based on the growth indicators and final yield of crops at each growth stage as measured by field trials in the study area; obtaining historical meteorological observation data of the study area, and driving the calibrated crop model by setting irrigation water volume and nitrogen fertilizer application corresponding to different growth stages, as well as meteorological variables corresponding to the historical meteorological observation data, to generate crop growth variables and crop yields corresponding to different growth stages; training the initial target yield model with crop yield and irrigation water volume and nitrogen fertilizer application corresponding to different growth stages to obtain the target yield model; the input data of the target yield model is the irrigation water volume and nitrogen fertilizer application corresponding to different growth stages, and the output data of the target yield model is the crop yield; optimizing the parameters of the target yield model using a particle swarm optimization algorithm with the target being the top preset percentage of crop yield and water and fertilizer utilization rate, to obtain the predicted values ​​of irrigation water volume and nitrogen fertilizer application corresponding to each growth stage; and determining the predicted values ​​of irrigation water volume and nitrogen fertilizer application corresponding to each growth stage, along with the corresponding crop growth variables and meteorological variables, as the modeling dataset.

[0023] Specifically, the predicted values ​​of irrigation water volume and nitrogen fertilizer application for each growth stage are input into the crop model to determine the crop growth variables for each growth stage. Among the water-ammonia application scenarios corresponding to different growth stages and their corresponding irrigation water volume and nitrogen fertilizer application, the water-ammonia application scenario that is closest to the predicted values ​​of irrigation water volume and nitrogen fertilizer application is identified, and the crop growth variables for each growth stage corresponding to this water-ammonia application scenario are determined as the crop growth variables for each growth stage.

[0024] Specifically, historical meteorological observation data from national meteorological stations is obtained based on the latitude and longitude coordinates of the farmland. Simultaneously, for specific crop types, different application rate gradients are set at key growth stages according to the upper and lower limits of local irrigation and fertilization rates to create water and fertilizer application scenarios. For example, taking winter rapeseed as an example, for each key growth stage, four gradients are set for irrigation application rates: 0, 40, 60, and 80 mm, and four gradients for nitrogen fertilizer application rates: 0, 30, 60, and 90 kg / ha. Therefore, these correspond to 256 irrigation scenarios and 256 nitrogen application scenarios, totaling 65,536 (256*256) water and nitrogen application scenarios. Thus, 65,536 variables and yield data for winter rapeseed growth processes can be generated annually. Based on this, the DSSAT crop model is driven by historical meteorological observation data to form datasets of winter rapeseed growth variables, meteorological variables, and yield.

[0025] In an exemplary embodiment, the genetic parameters of the crop model are calibrated based on the growth indicators and final yield of the crop at each growth stage as determined by field trials in the study area. Specifically, this includes: running the crop model based on the initial values ​​of the crop model's genetic parameters to obtain the predicted values ​​of growth indicators and yield for each growth and development stage in the current cycle; calculating the difference between the predicted values ​​of growth indicators and the growth indicators for each growth and development stage in the current cycle to obtain the prediction error of growth indicators for each growth and development stage; calculating the difference between the predicted yield and the actual yield in the current cycle to obtain the prediction error of yield; and iteratively updating the genetic parameters based on the prediction errors of growth indicators and yield until both the prediction errors of growth indicators and yield are minimized.

[0026] Specifically, crop growth indicators and final yield measured in field trials are used as benchmarks to calibrate the genetic parameters of the crop model to minimize the difference between the model's simulated values ​​and actual observed values. The model calibration process involves running the crop model using these coefficients in 10,000 development and growth iterations of generalized likelihood uncertainty estimation. This iterative process includes running the model with the genetic coefficients involved, comparing the output with the measured values ​​to adjust the coefficients until a satisfactory fit is achieved.

[0027] For example, taking the calibration of the Decision Support System for Agrotechnology Transfer (DSSAT) crop model to simulate the growth process of winter rapeseed as an example, the DSSAT crop model is calibrated using various growth indicators (leaf area index, aboveground biomass, leaf nitrogen content, and plant height) and final yield of winter rapeseed during key growth stages (overwintering stage, budding stage, flowering stage, and pod-forming stage) measured in field trials over many years as benchmarks. This ensures that the simulated values ​​output by the model are closer to the measured values. The Mean Absolute Percentage Error (MAPE) can be used to evaluate the error metric because it quantifies the relative magnitude of the error between the measured and simulated values.

[0028] In an exemplary embodiment, the formula corresponding to the target output model is formula (1): (1); in, The crop yield output by the target yield model. For the target output model, W 1 , W 2 , W 3 , W 4These are the irrigation water amounts for the first, second, third, and fourth growth and development stages, respectively. N 1 , N 2 , N 3 , N 4 The nitrogen fertilization amounts are for the first, second, third, and fourth growth stages, respectively.

[0029] In an exemplary embodiment, with crop yield and water and fertilizer utilization rate ranking as the top preset percentages, the parameters of the target yield model are optimized using a particle swarm optimization (PSO) algorithm to obtain predicted irrigation water consumption and nitrogen fertilizer application rates for each growth stage. Specifically, this includes: initializing the parameters of the PSO algorithm; these parameters include population size, inertia weight, learning factor, maximum number of iterations, and search space range; randomly initializing the position and velocity of each particle, setting the current position of each particle as its historical best position, and setting the position of the best particle in the swarm as its global best position; the particle positions represent the predicted irrigation water consumption and nitrogen fertilizer application rates for different growth and development stages; and calculating the predicted values ​​for each growth stage. The fitness function value of a particle is constructed based on crop yield and water and fertilizer use efficiency. The current fitness value is compared with the individual's historical best position; if the current fitness value is better, the individual's historical best position is updated using the current fitness value. The current fitness value is also compared with the global historical best position; if the current fitness value is better, the global best position is updated using the current fitness value. The velocity and position of each particle are iteratively updated according to the velocity update formula and the position update formula until a termination condition is reached. The termination condition is a preset percentage of crop yield and water and fertilizer use efficiency ranking at the top. The global best position corresponding to reaching the termination condition is determined as the predicted irrigation application rate and the predicted nitrogen fertilizer application rate.

[0030] Specifically, the preset percentage is set according to the specific engineering practice; for example, the preset percentage is 10%.

[0031] Specifically, with yield and water and fertilizer use efficiency as optimization objectives, parameter optimization methods are used to screen suitable water and fertilizer application plans. For example, a target yield model established based on winter rapeseed growth process variables and yield datasets is optimized with the goal of ranking the crop yield and water and fertilizer use efficiency in the top 10%. Particle Swarm Optimization (PSO) algorithm is used to optimize the input variable values ​​of the target yield model, namely the amount of irrigation nitrogen fertilizer applied at each growth stage (e.g., the numerical range is set to 0~80 mm and 0~90 kg / ha, respectively), in order to determine the appropriate water and nitrogen application plan.

[0032] In one exemplary embodiment, the present invention provides as follows Figure 2 The flowchart shown below illustrates the decision-making process for water and fertilizer application rates. Figure 2 As shown, Step 1: Calibrate the crop model parameters using data collected from field trials. Step 2: Based on years of climate observation and pre-defined water and fertilizer application scenarios, generate a growth process and yield dataset using the crop model calibrated in Step 1. Step 3: Establish a target yield model based on the generated crop growth process and yield dataset. Step 4: Based on the established target yield model and parameter optimization methods, determine a suitable water and fertilizer application strategy with crop yield and water and fertilizer utilization rate as objectives. Step 5: Use the suitable water and fertilizer application strategy and the crop model to generate crop growth process and yield data to establish an integrated water and fertilizer decision-making model.

[0033] The calibrated crop model generates modeling data to train a real-time water and fertilizer decision-making model for farmland crops. Taking the establishment of a real-time water and nitrogen decision-making model for winter rapeseed as an example, in the irrigation decision module, the input variables are set as leaf area index, leaf nitrogen concentration, aboveground biomass, canopy height, air, precipitation, reference crop evapotranspiration, and irrigation and nitrogen fertilizer application rates in the previous growth period. The output variable is the recommended irrigation application rate for the current growth period. In the nitrogen fertilizer decision module, the input variables are based on the irrigation decision module, with the addition of the recommended irrigation application rate for the current growth period. The output variable is the recommended nitrogen fertilizer application rate for the current growth period. The winter rapeseed water and nitrogen application plan selected in step four drives the DSSAT calibrated in step one to generate a dataset of winter rapeseed growth process and yield for training the irrigation and fertilization decision modules. The modeling techniques used include random forest, partial least squares, support vector machine, adaptive reinforcement learning, gradient boosting regression tree algorithm, Gaussian process regression, and distributed gradient reinforcement learning, among other machine learning algorithms.

[0034] In one exemplary embodiment, the present invention provides as follows Figure 3 The diagram shown is a real-time decision-making model for water and fertilizer management of farmland crops. Figure 3 As shown, each layer corresponds to a water and fertilizer decision-making module for each growth stage. This module includes two decision-making modules: irrigation and fertilization. For the irrigation decision-making module, the input variables include crop growth indicators, meteorological indicators, and the irrigation and fertilizer application rates of the previous growth stage; the output variable is the recommended irrigation and fertilizer application rates for the current growth stage. For the fertilizer decision-making module, the input variables include crop growth indicators, meteorological indicators, the irrigation and fertilizer application rates of the previous growth stage, and the recommended irrigation application rate for the current growth stage; the output variable is the recommended fertilizer application rate for the current growth stage. This design is to fully consider the differences in the impact of water and fertilizer on crop growth at different growth stages and the influence of historical water and fertilizer application information on the decision-making results.

[0035] When applying the water and fertilizer application rate decision method that synergizes yield and water and fertilizer use efficiency provided by this invention, it is not necessary to consider... Figure 1The steps shown are executed in sequence. The specific execution order of each step can be determined as needed, and this invention does not impose any restrictions on it.

[0036] The above describes a water and fertilizer application rate decision-making method that coordinates yield and water and fertilizer use efficiency, provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding water and fertilizer application rate decision-making device that coordinates yield and water and fertilizer use efficiency, such as... Figure 4 As shown.

[0037] Figure 4 A schematic diagram of a water and fertilizer application rate decision device that coordinates yield and water and fertilizer use efficiency provided by the present invention includes: The acquisition module 401 is used to acquire the multi-layer real-time decision-making module for farmland crop water and fertilizer. Each layer in the multi-layer real-time decision-making module for farmland crop water and fertilizer corresponds to a growth stage of the crop. Each layer in the multi-layer real-time decision-making module for farmland crop water and fertilizer includes an irrigation amount decision-making module and an ammonia amount decision-making module. The input data of the irrigation amount decision-making module includes crop growth variables, meteorological variables, the predicted irrigation water amount for the previous growth stage, and the predicted ammonia fertilizer amount for the previous growth stage. The output data of the irrigation amount decision-making module is the recommended irrigation water amount for the current growth stage. The input data of the ammonia amount decision-making module includes crop growth variables, meteorological variables, the predicted irrigation water amount for the previous growth stage, the predicted ammonia fertilizer amount for the previous growth stage, and the recommended irrigation water amount for the current growth stage. The output data of the ammonia amount decision-making module is the recommended ammonia fertilizer amount for the current growth stage.

[0038] The prediction module 402 is used to predict the irrigation water and nitrogen fertilizer application amount during the current growth and development stage of crops through the multi-layer farmland crop water and fertilizer real-time decision module.

[0039] Specific limitations regarding the water and fertilizer application rate decision-making device that coordinates yield and water and fertilizer use efficiency can be found in the limitations of the water and fertilizer application rate decision-making method that coordinates yield and water and fertilizer use efficiency, and will not be repeated here. Each module in the aforementioned water and fertilizer application rate decision-making device that coordinates yield and water and fertilizer use efficiency can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0040] The present invention also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described... Figure 1 The proposed method provides a synergistic approach to determining water and fertilizer application rates that integrates yield and water and fertilizer use efficiency.

[0041] The present invention also provides Figure 5The schematic diagram of the computer device shown is as follows: Figure 5 As shown, at the hardware level, this computer device includes a processor, internal bus, network interface, memory, and non-volatile memory, and may also include other hardware required for business operations. The processor reads the corresponding computer program from the non-volatile memory into memory and then executes it to achieve the above. Figure 1 The proposed method provides a synergistic approach to determining water and fertilizer application rates that integrates yield and water and fertilizer use efficiency.

[0042] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0043] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this invention.

Claims

1. A method for determining the amount of water and fertilizer applied in a way that coordinates yield and water and fertilizer use efficiency, characterized in that, include: A multi-layered real-time decision-making module for water and fertilizer application in farmland crops is established. Each layer in the multi-layered real-time decision-making module corresponds to a growth stage of the crop. Each layer in the multi-layered real-time decision-making module for water and fertilizer application includes an irrigation amount decision-making module and an ammonia amount decision-making module. The input data of the irrigation amount decision-making module includes crop growth variables, meteorological variables, predicted irrigation water amount for the previous growth stage, and predicted ammonia fertilizer application amount for the previous growth stage. The output data of the irrigation amount decision-making module is the recommended irrigation water amount for the current growth stage. The input data for the ammonia application rate decision module are crop growth variables, meteorological variables, predicted irrigation water volume for the previous growth period, predicted ammonia fertilizer application rate for the previous growth period, and recommended irrigation water volume for the current growth period. The output data for the ammonia application rate decision module is the recommended ammonia fertilizer application rate for the current growth period. The multi-layered real-time decision-making module for farmland crop water and fertilizer predicts the amount of irrigation water and nitrogen fertilizer applied during the current growth and development stage of crops.

2. The method as described in claim 1, characterized in that, The construction process of the multi-layered real-time decision-making module for farmland crop water and fertilizer includes: Obtain the modeling dataset; the modeling dataset includes crop growth variables, meteorological data, and predicted values ​​of irrigation water consumption and nitrogen fertilizer application for each growth stage of the crop; Using the modeling dataset, the initial multi-layered real-time decision-making module for water and fertilizer in farmland crops was trained to obtain the multi-layered real-time decision-making module for water and fertilizer in farmland crops.

3. The method as described in claim 2, characterized in that, Obtain the modeling dataset, specifically including: The genetic parameters of the crop model were calibrated based on the growth indicators and final yield of the crop at each growth stage as determined by field trials in the study area. Historical meteorological observation data of the study area were obtained. The irrigation water volume and nitrogen fertilizer application volume corresponding to different growth stages, as well as meteorological variables corresponding to historical meteorological observation data, were set to drive the calibrated crop model to generate crop growth variables and crop yields corresponding to different growth stages. The initial target yield model is trained using the crop yield and the irrigation water and nitrogen fertilizer application rates corresponding to different growth stages to obtain the target yield model. The input data of the target yield model are the irrigation water and nitrogen fertilizer application rates corresponding to different growth stages, and the output data of the target yield model is the crop yield. With crop yield and water and fertilizer utilization rate as the top preset percentages as the target, the parameters of the target yield model are optimized by particle swarm optimization algorithm to obtain the predicted values ​​of irrigation water consumption and nitrogen fertilizer application for each growth stage. The predicted values ​​of irrigation water consumption and nitrogen fertilizer application for each growth stage, along with the corresponding crop growth variables and meteorological variables, are determined as the modeling dataset.

4. The method as described in claim 3, characterized in that, The genetic parameters of the crop model are calibrated based on the crop growth indicators and final yields measured in field trials in the study area at each growth stage. Specifically, this includes: The crop model is run based on the initial values ​​of the crop model's genetic parameters to obtain the predicted values ​​of growth indicators and yield for each growth and development stage in the current cycle. Calculate the difference between the predicted growth index values ​​for each growth and development stage in the current cycle and the growth index values ​​for each growth and development stage to obtain the prediction error of the growth index for each growth and development stage. Calculate the difference between the predicted output value and the actual output value for the current round to obtain the output prediction error; Based on the growth index prediction error and the yield prediction error, the genetic parameters are iteratively updated until both the growth index prediction error and the yield prediction error reach their minimum.

5. The method as described in claim 3, characterized in that, The formula corresponding to the target output model is: ； in, The crop yield output by the target yield model. For the target output model, W 1 , W 2 , W 3 , W 4 These are the irrigation water amounts for the first, second, third, and fourth growth and development stages, respectively. N 1 , N 2 , N 3 , N 4 The nitrogen fertilization amounts are for the first, second, third, and fourth growth stages, respectively.

6. The method as described in claim 3, characterized in that, The process involves prioritizing crop yield and water / fertilizer utilization rates as a predetermined percentage, then using a particle swarm optimization algorithm to optimize the parameters of the target yield model. This yields predicted irrigation water consumption and nitrogen fertilizer application rates for each growth stage. Specifically, this includes: Initialize the parameters of the particle swarm optimization algorithm; the parameters include population size, inertia weight, learning factor, maximum number of iterations, and search space range; The position and velocity of each particle are randomly initialized, and the current position of each particle is set as the individual's historical best position, while the position of the best particle in the swarm is set as the global best position; the position of the particle is the predicted value of irrigation water consumption and nitrogen fertilizer application corresponding to different growth and development stages. Calculate the fitness function value for each particle; the fitness function is constructed based on crop yield and water and fertilizer use efficiency. Compare the current fitness value with the individual's historical best position. If the current fitness value is better, update the individual's historical best position using the current fitness value. Compare the current fitness value with the global historical best position. If the current fitness value is better, then update the global best position using the current fitness value. The velocity and position of each particle are iteratively updated according to the velocity update formula and the position update formula until a termination condition is reached; the termination condition is the crop yield and water and fertilizer utilization rate ranked first by a preset percentage. The global optimal position corresponding to the termination condition is determined as the predicted value of irrigation application rate and nitrogen fertilizer application rate.

7. A water and fertilizer application rate decision-making device that coordinates yield and water and fertilizer use efficiency, characterized in that, include: The acquisition module is used to acquire the multi-layer real-time decision-making module for farmland crop water and fertilizer. Each layer in the multi-layer real-time decision-making module for farmland crop water and fertilizer corresponds to a growth stage of the crop. Each layer in the multi-layer real-time decision-making module for farmland crop water and fertilizer includes an irrigation amount decision-making module and an ammonia amount decision-making module. The input data of the irrigation amount decision-making module includes crop growth variables, meteorological variables, the predicted value of irrigation water amount for the previous growth stage, and the predicted value of ammonia fertilizer amount for the previous growth stage. The output data of the irrigation amount decision-making module is the recommended value of irrigation water amount for the current growth stage. The input data for the ammonia application rate decision module are crop growth variables, meteorological variables, predicted irrigation water volume for the previous growth period, predicted ammonia fertilizer application rate for the previous growth period, and recommended irrigation water volume for the current growth period. The output data for the ammonia application rate decision module is the recommended ammonia fertilizer application rate for the current growth period. The prediction module is used to predict the amount of irrigation water and nitrogen fertilizer applied during the current growth and development stage of crops through a multi-layered real-time decision-making module for farmland crop water and fertilizer.

8. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the method as described in any one of claims 1 to 6.

9. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as described in any one of claims 1 to 6.

Images