Accurate crop management method and system based on crop growth model

Abstract

The invention provides a crop accurate management method and system based on a crop growth model. The method comprises the following steps: collecting environment data of a crop growth period by integrating a remote sensing technology, a field intelligent sensor and an unmanned aerial vehicle, and carrying out space-time alignment to form a comprehensive environment data set; and dynamically adjusting parameters of the crop growth model by using a Bayesian rule, evaluating the influence of environmental uncertainty on crops in combination with Monte Carlo simulation, and providing yield prediction with a confidence interval. Based on the prediction, a particle swarm optimization algorithm is applied to formulate management schemes such as irrigation, fertilization, pest control and harvesting opportunities, and meanwhile, a multi-objective optimization strategy is adopted to balance economic benefits and environmental protection. After the scheme is implemented, the crop growth response is monitored in real time so as to reinforce learning and automatically adjust management measures, and precise agricultural management is realized. According to the technical scheme provided by the embodiment of the invention, the crop management precision and efficiency are improved.

Description

Technical Field

[0001] The present application relates to the field of precision agriculture technology, and in particular to a crop precision management method and system based on a crop growth model. Background Art

[0002] In modern agriculture, precision agricultural management requires real-time monitoring and analysis of the crop growth environment to optimize operations such as irrigation, fertilization, and pest control. With the advancement of technology, agricultural producers need to be able to process large amounts of spatiotemporal data from different sources (such as remote sensing, field sensors, and drones) and transform these data into valuable decision support information, thereby achieving precise management of crop growth and efficient use of resources.

[0003] At present, some agricultural management systems have adopted intelligent sensor networks and remote sensing technology to collect environmental data and apply statistical models or machine learning algorithms to predict crop yields. However, these systems usually lack the ability to align multi-source heterogeneous data in time and space, making it difficult to generate continuous time series data sets for more accurate crop growth modeling. In addition, crop growth models in existing solutions are often based on fixed parameters and cannot dynamically adapt to changing environmental conditions.

[0004] However, the existing crop management methods have the following main shortcomings: first, they fail to fully consider the uncertainty of environmental factors and their impact on crop growth; second, they lack an effective mechanism to quantify this uncertainty and its impact on yield prediction results; third, most existing solutions fail to combine the dual goals of economic benefits and environmental protection for optimization, which may lead to resource waste or environmental pollution problems in practical applications. Therefore, there is an urgent need for an accurate management method that can integrate multiple data sources, dynamically adjust crop growth model parameters, and balance economic and environmental goals at the same time. Summary of the invention

[0005] The embodiments of the present application provide a crop precision management method and system based on a crop growth model, so as to solve the problems of insufficient precision and low efficiency of crop management in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a crop precision management method based on a crop growth model, comprising:

[0007] Through remote sensing technology and intelligent sensors and drones deployed in the field, the first environmental data during the crop growth cycle and the second environmental data of the farmland are collected, and the first environmental data and the second environmental data are aligned in time and space to ensure the correspondence of environmental data from different sources at the same time point and the same geographical area, so as to generate a comprehensive environmental data set with a time series;

[0008] According to the comprehensive environmental data set, the model parameters of the pre-trained crop growth model are dynamically adjusted by the Bayesian update rule, and on the basis of the adjusted crop growth model, multiple crop growth scenarios are simulated by the Monte Carlo simulation method, and the influence of the uncertainty of environmental factors on crop production is quantified, so as to generate a crop yield prediction result including a confidence interval; wherein the crop growth model is trained according to the crop type, variety characteristics and historical crop growth records;

[0009] Based on the crop yield prediction results and the degree of influence of uncertainty of environmental factors on crop production, a particle swarm optimization algorithm is applied to generate a crop management plan for the operation of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time. At the same time, a multi-objective optimization strategy is used to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal;

[0010] The adjusted crop management plan is implemented, and the actual growth response information of the crops is obtained in real time and used as the input of the reinforcement learning algorithm. The crop management plan is automatically adjusted under a specific reward mechanism by training the intelligent agent to achieve precise management of crop growth.

[0011] Optionally, dynamically adjusting the model parameters of the pre-trained crop growth model by Bayesian update rule according to the comprehensive environmental data set includes:

[0012] Using the environmental data in the comprehensive environmental data set as input to a pre-trained crop growth model, so that the crop growth model outputs predicted crop growth conditions based on the input and in combination with crop types, variety characteristics and historical growth records;

[0013] Determining actual crop growth conditions from the comprehensive environmental data set, and comparing the predicted crop growth conditions with the actual crop growth conditions to obtain a growth condition difference result;

[0014] By using the Bayesian updating method, the probability distribution of the model parameters of the crop growth model is adjusted based on the growth condition difference result, so that the crop growth condition predicted by the crop growth model is consistent with the actual crop growth condition.

[0015] Optionally, based on the adjusted crop growth model, multiple crop growth scenarios are simulated by Monte Carlo simulation method, and the influence of uncertainty of environmental factors on crop production is quantified to generate crop yield prediction results including confidence intervals, including:

[0016] A plurality of samples are randomly drawn from the probability distribution of the adjusted model parameters, each sample representing a different set of environmental data combinations;

[0017] Combining each set of environmental data as a crop growth scenario through a Monte Carlo simulation method, simulating the crop growth conditions of the crop under the crop generation scenario, and calculating a predicted value of the crop yield based on the crop growth conditions;

[0018] By quantifying the influence of the uncertainty of the obtained environmental factors on the crop growth, adjusting the predicted value of the crop yield, and statistically analyzing the distribution of crop yields under all crop growth scenarios, the statistical characteristics of the crop yield are determined;

[0019] Based on the statistical characteristics, a crop yield prediction result including a confidence interval is generated, wherein the confidence interval represents a numerical range within which the crop yield falls with a certain probability under the condition of uncertainty of given environmental factors.

[0020] Optionally, adjusting the crop yield by quantifying the influence of the uncertainty of the environmental factors obtained on the crop growth comprises:

[0021] For each crop growth scenario, calculating key indicators of crop growth under the crop growth scenario, and evaluating the variation range and volatility of the key indicators to obtain evaluation results;

[0022] Based on the evaluation results, evaluating the degree of influence of different environmental factors on key indicators of crop growth, so as to identify key environmental factors that have a significant impact on the key indicators of crop growth;

[0023] According to the key environmental factors, the predicted value of crop yield under each of the crop growth scenarios is adjusted, wherein the process of adjusting the predicted value of crop yield under each of the crop growth scenarios comprises: if the degree of influence of any environmental factor on crop growth meets a preset condition, appropriately increasing or decreasing the predicted value of crop yield in the corresponding crop growth scenario to quantify the degree of influence of the uncertainty of the environmental factor on crop growth;

[0024] The statistical distribution of crop yields under all crop growth scenarios and the determination of statistical characteristics of crop yields include:

[0025] Collecting the forecast values ​​of crop yields adjusted under all crop growth scenarios, constructing crop yield distribution, and analyzing the statistical characteristics of the crop yield distribution, including mean, standard deviation, skewness and kurtosis, to determine the variation range and central tendency of crop yields and obtain statistical analysis results;

[0026] Based on the statistical analysis results, key statistical characteristics of crop yield are determined to describe the central position and dispersion of crop yield through key statistical characteristics, thereby reflecting the distribution of crop yield.

[0027] Optionally, determining the actual crop growth conditions from the comprehensive environmental data set comprises:

[0028] Extracting observation data reflecting the actual growth status of crops from the environmental data of the comprehensive environmental data set as actual crop growth conditions;

[0029] The method further comprises: adjusting the probability distribution of the model parameters of the crop growth model based on the growth difference result by a Bayesian updating method, so that the crop growth condition predicted by the crop growth model is consistent with the actual crop growth condition.

[0030] The posterior probability of each model parameter of the crop growth model under the current probability distribution is calculated, and the probability distribution of the model parameters is updated in combination with the prior knowledge and the growth condition difference results, so that the crop growth condition predicted by the crop growth model is closer to the actual crop growth condition.

[0031] Optionally, based on the crop yield prediction result and the degree of influence of uncertainty of environmental factors on crop production, a particle swarm optimization algorithm is applied to generate a crop management plan for the operation of irrigation amount, fertilization ratio, pest control measures and harvesting time. At the same time, a multi-objective optimization strategy is adopted to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal, including:

[0032] According to the crop yield prediction results and the degree of influence of uncertainty of environmental factors on crop growth, the particle swarm optimization algorithm is applied to optimize the operating parameters of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time to generate a preliminary crop management plan;

[0033] Using the preliminary crop management plan, defining multiple optimization objectives, converting the multiple optimization objectives into mathematical expressions, and building a multi-objective optimization model in combination with actual constraints, the multiple optimization objectives at least include maximizing crop yield, minimizing input costs, and minimizing environmental pollution;

[0034] According to the multi-objective optimization model, a multi-objective optimization algorithm is used to solve the optimization problem to obtain a set of non-inferior solutions, each solution representing a feasible crop management solution;

[0035] Based on the decision maker's preferences or priorities in a specific situation, one or more feasible crop management solutions are selected from the non-inferior solution set, and with the optimization goal of balancing economic benefits and environmental protection, one or more feasible crop management solutions are adjusted to determine the final crop management solution.

[0036] Optionally, according to the crop yield prediction result and the degree of influence of the uncertainty of environmental factors on crop growth, a particle swarm optimization algorithm is applied to optimize the irrigation amount, fertilization ratio, pest and disease control measures and harvesting timing to generate a preliminary crop management plan, including:

[0037] Using the iterative search process of the particle swarm optimization algorithm, multiple groups of operating parameter combinations are simulated, wherein the operating parameter combinations include one or a combination of any several operations of adjusting the irrigation amount, adjusting the fertilization ratio, adjusting the pest control measures, and adjusting the harvesting time;

[0038] According to the crop yield prediction results and the degree of influence of the uncertainty of environmental factors on crop growth, the value range of each operating parameter in each group of operating parameter combinations is dynamically adjusted to ensure that the generated preliminary crop management plan can reflect the actual needs and environmental conditions of crop growth to the greatest extent;

[0039] After multiple iterations, the optimal operating parameter combination is selected from the adjusted multiple operating parameter combinations as the preliminary crop management plan.

[0040] In a second aspect, the embodiment of the present application provides a crop precision management system based on a crop growth model, including:

[0041] A collection module is used to collect first environmental data in the crop growth cycle and second environmental data of the farmland through remote sensing technology and intelligent sensors and drones deployed in the field, and to align the first environmental data with the second environmental data in time and space to ensure the correspondence of environmental data from different sources at the same time point and in the same geographical area, so as to generate a comprehensive environmental data set with a time series;

[0042] A generation module, used to dynamically adjust the model parameters of a pre-trained crop growth model according to the comprehensive environmental data set by using the Bayesian update rule, and simulate multiple crop growth scenarios by using the Monte Carlo simulation method based on the adjusted crop growth model, and quantify the influence of the uncertainty of environmental factors on crop generation, and generate a crop yield prediction result including a confidence interval; wherein the crop growth model is trained according to crop types, variety characteristics and historical crop growth records;

[0043] The generation module is also used to generate a crop management plan based on the crop yield prediction result and the degree of influence of the uncertainty of environmental factors on crop generation, using a particle swarm optimization algorithm, targeting the operation of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time, and at the same time, using a multi-objective optimization strategy to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal;

[0044] The adjustment module is used to implement the adjusted crop management plan, obtain the actual growth response information of the crop in real time and use it as the input of the reinforcement learning algorithm, and automatically adjust the crop management plan under a specific reward mechanism by training the intelligent agent to achieve precise management of crop growth.

[0045] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a precise crop management method based on a crop growth model as described in the first aspect above.

[0046] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a crop precision management method based on a crop growth model as described in the first aspect.

[0047] In an embodiment of the present application, first environmental data within the crop growth cycle and second environmental data of the farmland are collected through remote sensing technology and intelligent sensors and drones deployed in the field, and the first environmental data and the second environmental data are aligned in time and space to ensure the correspondence of environmental data from different sources at the same time point and the same geographical area, so as to generate a comprehensive environmental data set with a time series; according to the comprehensive environmental data set, the model parameters of the pre-trained crop growth model are dynamically adjusted through the Bayesian update rule, and on the basis of the adjusted crop growth model, multiple crop growth scenarios are simulated through the Monte Carlo simulation method, and the degree of influence of the uncertainty of environmental factors on crop generation is quantified to generate a confidence interval. Crop yield prediction results; wherein the crop growth model is trained according to crop types, variety characteristics and historical crop growth records; based on the crop yield prediction results and the degree of influence of uncertainty of environmental factors on crop generation, a particle swarm optimization algorithm is applied to generate a crop management plan for the operation of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time, and at the same time, a multi-objective optimization strategy is adopted to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal; the adjusted crop management plan is implemented, the actual growth response information of the crop is obtained in real time and used as the input of the reinforcement learning algorithm, and the crop management plan is automatically adjusted by training the intelligent agent under a specific reward mechanism to achieve precise management of crop growth.

[0048] The technical solution of this application has the following beneficial effects:

[0049] This application makes agricultural management more precise by driving decisions with real-time data. Quantify the impact of environmental factors, enhance the accuracy of yield prediction, and ensure stable crop yields. Optimize water and fertilizer use, reduce waste, and improve resource utilization. Multi-objective optimization takes environmental factors into account to reduce the impact of agricultural production on the environment. Optimizing management plans helps increase farmers' income and improve the economic returns of agricultural production. Use reinforcement learning to continuously improve management strategies to adapt to different growing seasons and climate changes.

[0050] Furthermore, the embodiment of the present application also outputs predicted crop growth conditions by using the environmental data in the comprehensive environmental data set as the input of the pre-trained crop growth model, combined with specific crop types, variety characteristics and historical growth records. Then, by comparing the predicted and actual crop growth conditions, the parameter probability distribution of the crop growth model is adjusted using the Bayesian update rule to ensure that the prediction results are closer to the actual situation. In addition, based on the crop yield prediction results and the impact of environmental uncertainty on crop growth, the particle swarm optimization algorithm is used to optimize the operating parameters such as irrigation amount, fertilization ratio, pest and disease control measures and harvest time, generate a preliminary management plan, and by constructing a multi-objective optimization model (including maximizing yield, minimizing cost and environmental pollution), a non-inferior solution set is obtained. Finally, the optimal crop management plan is selected and adjusted according to the decision maker's preferences or situational priorities to balance economic benefits and environmental protection.

[0051] Through the above method, not only the accuracy of crop growth prediction is improved, but also the adaptability of the model is enhanced by dynamically adjusting the model parameters. At the same time, the crop management plan formulated through the multi-objective optimization strategy not only ensures the economic benefits of agricultural production, but also takes into account the needs of environmental protection, and realizes efficient utilization of resources and sustainable development. The real-time adjustment of the management plan further improves the flexibility and response speed of crop management, helps to stabilize crop yields and reduce the negative impact of agricultural production on the environment.

[0052] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 A flow chart of a crop precision management method based on a crop growth model provided by the present application is shown;

[0055] Figure 2 A schematic diagram of the structure of a crop precision management system based on a crop growth model provided by the present application is shown;

[0056] Figure 3 A schematic diagram of the structure of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0057] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0058] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0059] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0060] Figure 1 A flowchart of a crop precision management method based on a crop growth model is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0061] 101. Using remote sensing technology and intelligent sensors and drones deployed in the field, collect first environmental data during the crop growth cycle and second environmental data of the farmland, and align the first environmental data and the second environmental data in time and space to ensure the correspondence of environmental data from different sources at the same time point and in the same geographical area, so as to generate a comprehensive environmental data set with a time series;

[0062] In precision agriculture, the collection of primary and secondary environmental data is accomplished by combining remote sensing technology, field smart sensor networks, and drones. The primary environmental data includes macro-scale information from satellites or high-altitude platforms, such as soil moisture, vegetation index, etc., while the secondary environmental data involves more detailed field conditions, such as temperature, humidity, light intensity, and local meteorological conditions. These data are used to construct a background framework for crop growth, and the spatiotemporal alignment algorithm is used to ensure the precise matching of all data points in time and geographic location, thereby generating a comprehensive environmental data set with time series characteristics. This step is crucial for subsequent model training and prediction because it provides accurate basic information.

[0063] First, smart sensors and drones deployed in the field regularly collect environmental parameters during the crop growth cycle and transmit these data to a central processing system. Then, a specially designed spatiotemporal alignment algorithm is used, which can identify and correct data inconsistencies caused by timestamp differences or geographic location offsets of different devices, and finally output a high-precision time series dataset that integrates multiple sources. This dataset provides a detailed and continuous record of environmental changes for subsequent steps.

[0064] Take wheat planting as an example. After sowing in spring, farmers use drones equipped with GPS modules to inspect the fields once a week. At the same time, multiple temperature and humidity sensors are arranged in the fields to monitor local climate conditions in real time. All data is uploaded to the cloud server, and after time and space alignment processing, a wheat growth environment database covering the entire growth period is formed. This database not only contains daily temperature, precipitation and other information, but also records the changing trend of soil moisture content, providing a solid foundation for subsequent model adjustments.

[0065] 102. According to the comprehensive environmental data set, dynamically adjust the model parameters of the pre-trained crop growth model by using the Bayesian update rule, and simulate multiple crop growth scenarios by using the Monte Carlo simulation method based on the adjusted crop growth model, and quantify the influence of the uncertainty of environmental factors on crop production, and generate a crop yield prediction result including a confidence interval; wherein the crop growth model is trained according to crop types, variety characteristics and historical crop growth records;

[0066] Based on the generated comprehensive environmental dataset, this step uses the Bayesian update rule to adjust the parameters of the pre-trained crop growth model. The crop growth model is a mathematical expression that reflects how a specific crop type (such as corn, soybean) and its variety characteristics (such as drought resistance, cold resistance) respond to environmental changes under historical conditions. By continuously updating the probability distribution of model parameters, the crop growth laws under the current environment can be better captured, thereby improving the prediction accuracy. On this basis, the Monte Carlo simulation method further simulates various possible growth scenarios, evaluates the quantitative impact of environmental uncertainty, and gives yield prediction results with confidence intervals.

[0067] The comprehensive environmental data set is input into the pre-trained crop growth model, which generates preliminary predictions based on crop types and historical growth records. Then, by comparing the differences between actual observations and model predictions, the Bayesian update rule is used to adjust the model parameters to make the predictions more realistic. After that, with the help of Monte Carlo simulation, a series of possible future growth scenarios are simulated, taking into account the randomness and uncertainty of environmental factors, and finally generating yield forecasts with confidence intervals to help farmers understand potential risks and opportunities.

[0068] Continuing with the wheat planting example above, after obtaining a complete environmental data set, the researchers applied the Bayesian update rule to fine-tune the wheat growth model optimized for local soil and climate conditions. Subsequently, they used the Monte Carlo simulation method to explore the fluctuation range of wheat yields under different weather patterns (such as drought and floods), providing farmers with reliable estimates of expected harvests in the coming months, including the most likely yield range and possible results in extreme cases. This prediction is critical for planning irrigation plans and marketing strategies.

[0069] 103. Based on the crop yield prediction results and the degree of influence of uncertainty of environmental factors on crop production, a particle swarm optimization algorithm is applied to generate a crop management plan for the operation of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time. At the same time, a multi-objective optimization strategy is used to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal;

[0070] Based on the crop yield prediction results and environmental uncertainty analysis obtained above, this step uses the particle swarm optimization algorithm to develop a set of refined crop management plans. This plan specifically covers the amount of irrigation, fertilization ratio, pest and disease control measures, and the choice of harvest time. Multi-objective optimization strategy plays an important role in this process. It aims to find the best practice combination that can maximize economic benefits and minimize environmental pollution. In this way, not only the efficiency of crop production is improved, but also the realization of sustainable development goals is promoted.

[0071] After obtaining crop growth predictions and environmental uncertainty impact assessments, the particle swarm optimization algorithm begins to work to find the optimal operating parameter settings. These parameters include but are not limited to irrigation frequency, fertilizer application, pesticide spraying times, etc. In order to balance economic benefits and environmental protection, a multi-objective optimization model is introduced, in which several optimization goals are set (such as maximizing yield, minimizing costs, and reducing carbon emissions). Finally, by solving the multi-objective optimization problem, a non-inferior solution set is obtained, and the most suitable management plan is selected according to the specific situation.

[0072] Back to the wheat planting example, after obtaining a detailed forecast of future yields, agronomists used the particle swarm optimization algorithm to calculate the best irrigation and fertilization plan to ensure that excessive water use and fertilizer waste are avoided while meeting the needs of wheat growth. In addition, considering the importance of environmental protection, the team also paid special attention to how to minimize the risk of water pollution caused by nitrogen fertilizer loss. Through multi-objective optimization, a solution that is both economically beneficial and conducive to ecological and environmental protection was found, providing farmers with a set of scientific and reasonable management guidelines.

[0073] 104. Implement the adjusted crop management plan, obtain the actual growth response information of the crops in real time and use it as input to the reinforcement learning algorithm, and automatically adjust the crop management plan under a specific reward mechanism by training the intelligent agent to achieve precise management of crop growth.

[0074] The importance of real-time monitoring of crop growth responses was emphasized, and adaptive adjustments to crop management plans were achieved through reinforcement learning algorithms. Reinforcement learning is a form of machine learning that allows an agent (such as a computer program) to learn optimal behavior strategies through trial and error. In this scenario, the agent is rewarded or punished based on the actual growth performance of the crop (such as leaf color and root development), and gradually optimizes its decision-making rules until it finds the best way to promote healthy crop growth.

[0075] Once the initial crop management plan is determined, it is implemented and the status of the crops is continuously monitored throughout the growing season. Various indicators of crop growth are collected as feedback information and input into the reinforcement learning algorithm. The algorithm automatically adjusts management measures, such as increasing or decreasing irrigation water in a timely manner, according to the set reward mechanism (such as giving positive incentives when the ideal growth state is achieved). As the system iterates and learns, the management plan gradually approaches the ideal state, thereby achieving precise control of crop growth.

[0076] During the wheat growing season, farmers check the growth of crops in the field every day and upload relevant data to the cloud system through mobile phone applications. The intelligent agent in the system instantly evaluates the effectiveness of current management measures based on the uploaded information and makes suggestions for improvement. For example, if it is found that the wheat leaves in a certain area are yellowing, it may be caused by nitrogen deficiency. The intelligent agent will recommend an appropriate increase in nitrogen fertilizer application in that area. Over time, the intelligent agent has learned to flexibly adjust the management plan according to the needs of different stages, ensuring that each wheat plant can receive the most appropriate care, and ultimately achieving higher yield and quality standards.

[0077] In order to solve the problem of deviation between model prediction and actual situation, in some embodiments, the step 102 dynamically adjusts the model parameters of the pre-trained crop growth model according to the comprehensive environmental data set by using the Bayesian update rule, including:

[0078] The environmental data in the comprehensive environmental data set is used as the input of a pre-trained crop growth model, so that the crop growth model outputs predicted crop growth conditions based on the input and in combination with crop types, variety characteristics and historical growth records; the actual crop growth conditions are determined from the comprehensive environmental data set, and the predicted crop growth conditions are compared with the actual crop growth conditions to obtain growth condition difference results; and the probability distribution of the model parameters of the crop growth model is adjusted based on the growth condition difference results through a Bayesian update method, so that the crop growth conditions predicted by the crop growth model are consistent with the actual crop growth conditions.

[0079] In this embodiment, in order to further improve the accuracy and adaptability of the crop growth model, the Bayesian update method is used in some embodiments to dynamically adjust the model parameters. The key to this step is to use the environmental data in the comprehensive environmental data set as input, so that the crop growth model can output the predicted crop growth conditions in combination with specific crop types, variety characteristics and historical growth records. Then, the actual crop growth conditions are determined from the comprehensive environmental data set, and the predicted results are compared with the actual conditions to obtain the growth condition difference results. Through this comparison, the gap between the model prediction and reality can be identified. Based on these differences, the probability distribution of the model parameters is adjusted using the Bayesian update rule to ensure that the model prediction is closer to the actual growth conditions.

[0080] In the embodiment of the present application, the process first requires inputting a comprehensive environmental data set (including multi-source information such as temperature, humidity, and light intensity) into a pre-trained crop growth model. The model predicts the growth status of crops under current environmental conditions based on the specific type and characteristics of the crops, as well as historical data accumulated in the past. Next, real crop growth data such as plant height, leaf area index, biomass, etc. are obtained through field measurements or high-resolution remote sensing image analysis for comparison with the model predictions. Finally, the Bayesian statistical method is used to adjust the probability distribution of the model's internal parameters based on the difference between the two, thereby correcting the model to better reflect crop growth patterns in the real world.

[0081] Here is a specific example:

[0082] In the case of a wheat-growing area, researchers deployed a variety of sensor networks to monitor key environmental factors such as soil moisture, temperature, and rainfall, and regularly used drones to take photos of the fields to assess vegetation cover. All collected data was integrated into a comprehensive environmental dataset and input into a crop growth model designed specifically for local wheat varieties. The model preliminarily predicted the development trend of wheat throughout the growth cycle. However, at some stages, such as the jointing stage, the wheat height predicted by the model was significantly different from the actual observation. Therefore, the team members applied the Bayesian update rule to recalibrate the model parameters based on the actual measured wheat growth data. After several iterations, the model successfully improved its prediction accuracy for future growth stages, especially showing stronger adaptability in response to extreme weather events caused by climate change. This improvement allows farmers to optimize irrigation plans and fertilization strategies based on more accurate forecast information, ultimately promoting a steady increase in wheat yields.

[0083] In this way, not only the possible prediction bias problem of the initial model is solved, but also the model's ability to respond to new environmental changes is enhanced, providing strong support for precision agriculture.

[0084] In order to solve the problem of the impact of environmental factor uncertainty on crop yield prediction, in some embodiments, the method of simulating multiple crop growth scenarios by Monte Carlo simulation method based on the adjusted crop growth model in step 102, and quantifying the impact of environmental factor uncertainty on crop production, generates a crop yield prediction result including a confidence interval, including:

[0085] A plurality of samples are randomly selected from the probability distribution of the adjusted model parameters, each sample representing a different combination of environmental data; each group of environmental data is combined as a crop growth scenario through the Monte Carlo simulation method, and the crop growth conditions under the crop generation scenario are simulated, and the predicted value of the crop yield is calculated based on the crop growth conditions; the predicted value of the crop yield is adjusted by quantifying the degree of influence of the uncertainty of the obtained environmental factors on crop growth, and the crop yield distribution under all crop growth scenarios is statistically analyzed to determine the statistical characteristics of the crop yield; based on the statistical characteristics, a crop yield prediction result including a confidence interval is generated, wherein the confidence interval represents a numerical range within which the crop yield falls with a certain probability under the condition of given environmental factor uncertainty.

[0086] In this embodiment, in order to further improve the reliability and accuracy of crop yield prediction, a Monte Carlo simulation method is used in some embodiments to evaluate the impact of environmental factor uncertainty on crop growth. Specifically, multiple samples are randomly selected from the probability distribution of the adjusted model parameters, and each sample represents a set of different environmental data combinations (such as temperature, humidity, light, etc.). These samples are used to simulate a variety of possible crop growth scenarios to calculate crop yield prediction values ​​under different scenarios. In this way, not only a single prediction value can be obtained, but also the yield distribution under all crop growth scenarios can be statistically analyzed to determine the statistical characteristics of crop yield. Finally, based on these statistical characteristics, a crop yield prediction result containing a confidence interval is generated, where the confidence interval represents the numerical range within which the crop yield falls with a certain probability under the condition of given environmental factor uncertainty.

[0087] In this embodiment, the scheme first uses the probability distribution of model parameters adjusted by the Bayesian update rule to randomly extract multiple groups of samples, each of which corresponds to a hypothetical combination of environmental conditions. Then, the Monte Carlo simulation method is used to take each group of environmental conditions as input to simulate the development of crops throughout the growth cycle, and the corresponding crop yield prediction value is calculated based on the simulation results. Next, by quantifying the impact of uncertainty in environmental factors (such as precipitation patterns and temperature fluctuations) on crop growth, the predicted value of crop yield is adjusted to ensure that the prediction results can reflect the variability in reality. Finally, the crop yield distribution under all simulated scenarios is statistically analyzed, and its statistical characteristics such as mean value and standard deviation are analyzed, and then the confidence interval is constructed to provide a more comprehensive yield prediction framework to help decision makers understand potential risks and uncertainties.

[0088] Here is a specific example:

[0089] In an application case of a corn-growing area, the research team conducted a large-scale Monte Carlo simulation experiment after completing the model parameter adjustment. They randomly selected 10,000 samples from the adjusted model parameter probability distribution, each of which represented a specific combination of environmental conditions, such as different precipitation, temperature variation, and sunshine duration. For each set of samples, the researchers simulated the growth of corn throughout the growing season, including key stages such as germination, heading, and filling, and calculated the expected yield under each scenario. By comparing the differences between different scenarios, the impact of extreme weather events such as drought or floods on corn yields was quantified. After multiple iterative simulations, the team obtained a large number of data points on corn yields and conducted statistical analysis on them to determine the average yield, yield range, and 95% confidence interval. This not only provides farmers with more reliable yield forecasts, but also helps them better plan management measures such as irrigation and fertilization to cope with possible climate change risks in the future.

[0090] In this way, Monte Carlo simulation not only addresses the challenges brought about by the uncertainty of environmental factors, but also enhances the transparency and credibility of the prediction results, enabling agricultural producers to make more informed decisions in the face of a complex and changing natural environment.

[0091] In order to solve the problem of the influence of environmental factor uncertainty on crop growth and yield prediction, in some embodiments, the step 102 of adjusting the crop yield by quantifying the influence of the uncertainty of the environmental factor on crop growth includes:

[0092] For each crop growth scenario, the key indicators of crop growth under the crop growth scenario are calculated, and the variation range and volatility of the key indicators are evaluated to obtain an evaluation result; based on the evaluation result, the influence of different environmental factors on the key indicators of crop growth is evaluated to identify the key environmental factors that have a significant influence on the key indicators of crop growth; according to the key environmental factors, the predicted value of crop yield under each crop growth scenario is adjusted, wherein the process of adjusting the predicted value of crop yield under each crop growth scenario includes: if the influence degree of any environmental factor on crop growth meets the preset conditions, appropriately increase or decrease the predicted value of crop yield in the corresponding crop growth scenario; The predicted value of crop yield is reduced to quantify the influence of uncertainty of environmental factors on crop growth; the statistical crop yield distribution under all crop growth scenarios is determined to determine the statistical characteristics of crop yield, including: collecting the adjusted predicted values ​​of crop yield under all crop growth scenarios, constructing crop yield distribution, and analyzing the statistical characteristics of the crop yield distribution, including mean, standard deviation, skewness and kurtosis, to determine the variation range and central trend of crop yield, and obtain statistical analysis results; based on the statistical analysis results, the key statistical characteristics of crop yield are determined to describe the central position and dispersion of crop yield through key statistical characteristics, and reflect the distribution of crop yield.

[0093] In this embodiment, in order to further improve the accuracy and reliability of crop yield prediction, a method for quantifying environmental uncertainty is introduced in some embodiments. The key to this step is to calculate the key indicators of crop growth (such as plant height, leaf area index LAI, biomass, etc.) under each crop growth scenario, and evaluate the range and volatility of these key indicators. Based on this evaluation result, it is possible to identify which environmental factors (such as temperature, precipitation, light intensity) have a significant impact on crop growth, and adjust the predicted value of crop yield according to these key environmental factors. Specifically, if the impact of a certain environmental factor on crop growth meets the preset conditions (for example, drought or flood risk exceeding a certain threshold), the predicted value of crop yield is appropriately increased or decreased in the corresponding crop growth scenario to reflect the real impact of environmental uncertainty. In addition, by statistically analyzing the adjusted yield prediction values ​​under all crop growth scenarios, the distribution characteristics of crop yield, including mean, standard deviation, skewness and kurtosis, are determined, thereby describing the central position and discrete degree of crop yield.

[0094] In the embodiment of the present application, the scheme first calculates the key growth indicators for each crop growth scenario, and evaluates its range of variation and volatility to obtain preliminary evaluation results. Then, based on these evaluation results, the degree of influence of different environmental factors on key growth indicators is analyzed, and the key environmental factors that have a significant impact on crop growth are identified. Next, the crop yield prediction value under each crop growth scenario is adjusted according to the identified key environmental factors. If the impact of any environmental factor on crop growth meets the preset conditions, the prediction value is increased or decreased accordingly to more accurately quantify the impact of environmental uncertainty on crop growth. Finally, all adjusted crop yield prediction values ​​are collected, the crop yield distribution is constructed, and the key statistical characteristics of crop yield are determined by statistical analysis (such as calculating the mean, standard deviation, skewness and kurtosis) to describe the central tendency and range of variation of crop yield and provide more comprehensive prediction information.

[0095] Here is a specific example:

[0096] In an application case in a rice-growing area, the research team completed the model parameter adjustment and conducted a detailed Monte Carlo simulation experiment. They extracted 5,000 samples from the adjusted model parameter probability distribution, each of which represented a specific combination of environmental conditions. For each sample, the researchers simulated the growth of rice throughout the growing season and calculated key growth indicators such as plant height, leaf area index (LAI), biomass, etc. By comparing the differences between different scenarios, the degree of influence of environmental factors such as temperature fluctuations and precipitation changes on rice growth was quantified. In particular, they found that extreme temperatures (above 35°C or below 15°C) and excessive precipitation (monthly rainfall exceeding 200 mm) had a significant negative impact on rice growth.

[0097] Based on these assessment results, the team identified temperature and precipitation as key environmental factors. Subsequently, they adjusted the predicted rice yield in each crop growth scenario according to the degree of impact of these key environmental factors. For example, when extreme high temperatures occur in a scenario, the predicted yield is appropriately reduced; when precipitation is moderate, the predicted yield remains unchanged or increases slightly. Finally, the team collected all adjusted crop yield forecasts, constructed a complete crop yield distribution, and performed statistical analysis on it to obtain key statistical characteristics such as mean yield, standard deviation, skewness, and kurtosis. These statistical characteristics not only describe the central trend and range of rice yield, but also reflect the distribution of yield under different environmental conditions, providing farmers with more scientific and reasonable decision-making support.

[0098] In this way, the method not only solves the challenges brought by the uncertainty of environmental factors, but also enhances the transparency and credibility of the prediction results, enabling agricultural producers to make more informed decisions in the face of a complex and changing natural environment. At the same time, this method helps to optimize agricultural management strategies, ensure the efficient use of resources, and promote sustainable agricultural development.

[0099] In order to solve the problem of deviation between the crop growth model prediction and the actual growth situation, in some embodiments, the determining the actual crop growth situation from the comprehensive environmental data set in step 102 includes:

[0100] Observation data reflecting the actual growth status of crops are extracted from the environmental data of the comprehensive environmental data set as the actual crop growth situation; the probability distribution of the model parameters of the crop growth model is adjusted based on the growth situation difference result through the Bayesian updating method, so that the crop growth situation predicted by the crop growth model is consistent with the actual crop growth situation, including: calculating the posterior probability of each model parameter of the crop growth model under the current probability distribution, combining prior knowledge and the growth situation difference result, and updating the probability distribution of the model parameters, so that the crop growth situation predicted by the crop growth model is closer to the actual crop growth situation.

[0101] In this embodiment, in order to further improve the accuracy of the crop growth model, some embodiments use observation data reflecting the actual growth status of crops extracted from the comprehensive environmental data set as the actual crop growth situation. These observation data may include but are not limited to soil moisture, temperature, light intensity measured by field sensors, and vegetation index (such as NDVI), plant height and other information provided by drones or satellite remote sensing images. Through the integration of these multi-source data, the actual growth status of crops can be more comprehensively reflected. Subsequently, the Bayesian update method is used to adjust the probability distribution of model parameters of the crop growth model based on the growth difference results to ensure that the model prediction is closer to the actual situation.

[0102] Specifically, the Bayesian update method is implemented by calculating the posterior probability of each model parameter under the current probability distribution. This step combines prior knowledge (i.e., understanding of the initial estimates of the model parameters) with newly obtained growth difference results (i.e., the gap between the model prediction value and the actual observation value) to dynamically update the probability distribution of the model parameters. The purpose of this is to enable the model to continuously learn and adapt to new data, thereby improving its ability to predict future crop growth.

[0103] In an embodiment of the present application, the scheme first screens out observation data that reflects the actual growth status of crops from a comprehensive environmental data set, such as real-time environmental conditions obtained by field intelligent sensors and high-resolution images taken by drones. Then, these observation data are compared with the prediction results of the crop growth model to identify the differences between the two. Next, the Bayesian statistical method is used to calculate the posterior probability of each model parameter based on these differences, and the probability distribution of the model parameters is updated in combination with prior knowledge. This process enables the model to gradually approach the actual crop growth conditions in each iteration, improving the prediction accuracy and adaptability of the model.

[0104] Here is a specific example:

[0105] In an application case of apple orchard management, the research team deployed various types of sensors to monitor microclimate conditions in the orchard, including soil moisture, air temperature, and relative humidity. At the same time, they regularly used drones equipped with multispectral cameras to take photos of the orchard to assess the health of the fruit trees and the development of the fruit. All collected data were integrated into a comprehensive environmental dataset, from which key observational data reflecting the actual growth status of the apple trees were extracted, such as leaf moisture content, fruit size and color changes, etc.

[0106] The researchers compared these observations with the predictions of a pre-trained apple tree growth model and found several significant differences, especially during the fruit expansion period, when the model predicted fruit size that was significantly smaller than the actual observed value. To correct this deviation, the team applied the Bayesian update method to calculate the posterior probability of each parameter in the model (such as water absorption efficiency, photosynthesis rate, etc.) under the current probability distribution. They combined existing prior knowledge (such as historical best practices and literature reports) with the newly obtained growth difference results to update the probability distribution of the model parameters.

[0107] After several iterations, the model successfully improved its prediction accuracy for future fruit development stages, especially showing stronger adaptability when dealing with different weather patterns. Ultimately, farmers can optimize irrigation plans and fertilization strategies based on more accurate forecast information, ensuring that each apple tree receives the most appropriate care, thereby promoting a steady increase in fruit yield and quality.

[0108] In this way, not only the prediction bias problem that may exist in the initial model is solved, but also the model's ability to respond to new environmental changes is enhanced, providing strong support for precision agriculture. This approach helps to optimize agricultural production management, ensure the efficient use of resources, and promote sustainable agricultural development.

[0109] In order to solve the problem of balancing economic benefits and environmental protection in crop management schemes, in some embodiments, the step 103, based on the crop yield prediction result and the degree of influence of uncertainty of environmental factors on crop production, applies a particle swarm optimization algorithm to generate a crop management scheme for the operation of irrigation amount, fertilization ratio, pest control measures and harvesting time, and at the same time, adopts a multi-objective optimization strategy to adjust the crop management scheme with balancing economic benefits and environmental protection as the optimization goal, including:

[0110] According to the crop yield prediction results and the degree of influence of the uncertainty of environmental factors on crop growth, a particle swarm optimization algorithm is applied to optimize the operating parameters of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time to generate a preliminary crop management plan; using the preliminary crop management plan, multiple optimization objectives are defined, and the multiple optimization objectives are converted into mathematical expressions, and a multi-objective optimization model is constructed in combination with actual constraints, and the multiple optimization objectives at least include maximizing crop yield, minimizing input costs, and minimizing environmental pollution; according to the multi-objective optimization model, a multi-objective optimization algorithm is used to solve the optimization problem to obtain a non-inferior solution set, each solution represents a feasible crop management plan; based on the decision maker's preference or priority in a specific situation, one or more feasible crop management plans are selected from the non-inferior solution set, and with balancing economic benefits and environmental protection as the optimization goal, one or more feasible crop management plans are adjusted to determine the final crop management plan.

[0111] In this embodiment, in order to further improve the scientificity and sustainability of the crop management scheme, a particle swarm optimization (PSO) algorithm and a multi-objective optimization strategy are introduced in some embodiments. Specifically, according to the crop yield prediction results and the degree of influence of environmental factor uncertainty on crop growth, the particle swarm optimization algorithm is first applied to optimize the operation parameters such as irrigation amount, fertilization ratio, pest and disease control measures and harvesting time to generate a preliminary crop management scheme. These operation parameters include but are not limited to the amount of water per irrigation, the amount of fertilizer applied, the frequency of pesticide spraying and the optimal harvesting time point. Then, multiple optimization objectives are defined, and these objectives are converted into mathematical expressions. Combined with actual constraints (such as water resource restrictions, soil carrying capacity, etc.), a multi-objective optimization model is constructed. The model aims to achieve multiple objectives such as maximizing crop yield, minimizing input costs and minimizing environmental pollution. Finally, by solving the multi-objective optimization problem, a set of non-inferior solutions is obtained, each of which represents a feasible crop management scheme. Decision makers can select the most appropriate solution from this set of non-inferior solutions according to their own preferences or priorities in a specific situation, and make appropriate adjustments to it to finally determine the optimal crop management scheme.

[0112] In the embodiment of the present application, the scheme first uses the particle swarm optimization algorithm to optimize the key operating parameters based on the crop yield prediction results and the impact assessment of the uncertainty of environmental factors to generate a preliminary crop management plan. Then, multiple optimization objectives are defined and converted into mathematical expressions, and a multi-objective optimization model is constructed in combination with actual constraints. Next, a multi-objective optimization algorithm is used to solve the optimization problem to obtain a non-inferior solution set, that is, a series of feasible crop management plans. Finally, based on the decision maker's preferences or the needs of a specific situation, one or more solutions are selected from the non-inferior solution set and fine-tuned to ensure that the final solution meets both economic efficiency requirements and environmental protection goals.

[0113] Here is a specific example:

[0114] In a vineyard management application case, the research team hopes to develop a management plan that can both increase grape yields and reduce environmental impacts. They first optimized the irrigation amount, fertilization ratio, pest and disease control measures, and harvest timing based on the crop yield prediction results and the impact of environmental factors uncertainty, and generated a preliminary crop management plan. For example, the optimized irrigation plan takes into account the changes in water demand at different growth stages, and the fertilization ratio is precisely adjusted according to the soil nutrient status.

[0115] Next, the team defined three main optimization objectives: maximizing grape yield, minimizing input costs (such as the cost of water, fertilizer, and pesticides), and minimizing environmental pollution (such as fertilizer loss and pesticide residues). These objectives were converted into specific mathematical expressions, and combined with practical constraints (such as the availability of local water resources and soil carrying capacity), a multi-objective optimization model was constructed. By solving this model, the researchers obtained a set of non-inferior solutions, each of which represents a possible management plan covering different irrigation strategies, fertilization methods, and pest and disease control measures.

[0116] Faced with this set of non-inferior solutions, the vineyard managers selected two solutions with the greatest potential based on their own economic capabilities and environmental awareness, and conducted detailed evaluations. One solution focused on high yield and efficiency, while the other focused more on eco-friendliness. After comprehensive consideration, a compromise solution was finally determined, which not only ensured a higher grape yield, but also significantly reduced the use of chemical inputs and reduced the impact on the surrounding environment. In addition, the solution also particularly emphasized the importance of timely harvesting to ensure that the fruit quality is in the best state, thereby improving market competitiveness.

[0117] In this way, the method not only solves the problem of resource waste or environmental damage that may be caused by traditional single-objective optimization, but also enhances the flexibility and adaptability of crop management programs, helping agricultural producers to actively fulfill their social responsibilities while pursuing economic benefits and promoting the sustainable development of agriculture.

[0118] In order to solve the problem of optimizing operating parameters in the crop management plan, in some embodiments, the particle swarm optimization algorithm is applied in step 103 to optimize the irrigation amount, fertilization ratio, pest control measures and harvesting time according to the crop yield prediction result and the influence of the uncertainty of environmental factors on crop growth, so as to generate a preliminary crop management plan, including:

[0119] By using the iterative search process of the particle swarm optimization algorithm, multiple groups of operating parameter combinations are simulated, and the operating parameter combinations include one or a combination of any several operations of adjusting the irrigation amount, adjusting the fertilization ratio, adjusting the pest and disease control measures, and adjusting the harvesting time; according to the crop yield prediction results and the influence of the uncertainty of environmental factors on crop growth, the value range of each operating parameter in each group of operating parameter combinations is dynamically adjusted to ensure that the generated preliminary crop management plan can reflect the actual needs and environmental conditions of crop growth to the greatest extent; after multiple iterations, the optimal operating parameter combination is selected from the adjusted multiple operating parameter combinations as the preliminary crop management plan.

[0120] In this embodiment, in order to further improve the scientificity and adaptability of the crop management scheme, a particle swarm optimization (PSO) algorithm is used in some embodiments to optimize key operating parameters. Specifically, the particle swarm optimization algorithm is an iterative search method based on swarm intelligence, which simulates the behavior patterns of bird flocks or fish schools to find the optimal solution in the solution space. In this scenario, the algorithm is used to simulate multiple groups of operating parameter combinations, which may include one or a combination of any several operations of adjusting the irrigation amount, adjusting the fertilization ratio, adjusting the pest control measures, and adjusting the harvesting time. The value range in each group of operating parameter combinations will be dynamically adjusted according to the crop yield prediction results and the degree of influence of environmental factors uncertainty on crop growth, so as to ensure that the generated preliminary crop management plan can reflect the actual needs and environmental conditions of crop growth to the greatest extent. Through multiple iterative searches, the best one is finally selected from multiple adjusted operating parameter combinations as the preliminary crop management plan.

[0121] In the embodiment of the present application, the scheme first uses the iterative search process of the particle swarm optimization algorithm to simulate multiple groups of operating parameter combinations. These combinations cover operations such as different levels of irrigation, fertilization ratios, pest control measures, and harvest timing. Then, based on the crop yield prediction results and the impact of environmental factor uncertainty on crop growth, the value range in each group of operating parameter combinations is dynamically adjusted to ensure that they meet the current crop growth needs and environmental conditions. After multiple iterations, the algorithm selects the best group from all adjusted operating parameter combinations as a preliminary crop management plan. This process not only improves the selection accuracy of operating parameters, but also enhances the adaptability of management plans to environmental changes.

[0122] Here is a specific example:

[0123] In an application case of a rice-growing area, the research team hopes to develop a management plan that can both increase yields and reduce resource waste. They first used the particle swarm optimization algorithm to simulate a variety of operating parameter combinations, covering different irrigation frequencies, fertilizer application rates, pest and disease control strategies, and optimal harvesting time points. For example, an operating parameter combination may include irrigation once a day, application of 200 kg of nitrogen fertilizer per hectare, spraying pesticides once a week, and harvesting on the 120th day.

[0124] Next, the team dynamically adjusted the range of each operating parameter based on the crop yield forecast results and the impact of environmental uncertainty on rice growth. For example, if the forecast results show that there will be more precipitation in the future, the number of irrigations will be appropriately reduced; if the soil nutrient test shows that the nitrogen content is high, the amount of nitrogen fertilizer applied will be reduced. At the same time, considering the need for environmental protection, the team set a strict upper limit on the use of pesticides and explored the possibility of biological control.

[0125] After multiple iterations of optimization, the researchers selected the best combination of operating parameters from the many adjusted combinations as the initial crop management plan. This plan not only ensures the healthy development of rice throughout the entire growth cycle, but also significantly reduces the consumption of water resources and fertilizers, and reduces the risk of environmental pollution. In addition, the team also paid special attention to the importance of timely harvesting to ensure that the quality of rice is in the best state, thereby improving market competitiveness.

[0126] In this way, the method not only solves the problem of resource waste or unstable yield caused by traditional fixed parameter management, but also enhances the flexibility and adaptability of crop management solutions, helping agricultural producers make more informed decisions in the face of complex and changing natural environments. This method is applicable to rice cultivation and other crop fields, and helps achieve refined management and sustainable development of agricultural production.

[0127] Through the above methods, the research team not only developed a more scientific and reasonable crop management plan, but also found a balance between resource utilization efficiency and environmental protection. The application of particle swarm optimization algorithm makes the selection of operating parameters more accurate and flexible, while the dynamic adjustment mechanism enhances the management plan's ability to adapt to environmental changes. This method is not only suitable for rice cultivation, but can also be widely used in other crop planting fields, helping farmers better plan and execute agricultural production activities, and promoting the development of agriculture towards a green, efficient and modern direction.

[0128] This application considers that in modern agriculture, crop growth models are widely used to predict crop yields and optimize agricultural management decisions. However, the impact of environmental uncertainty (such as climate change, soil condition fluctuations, etc.) on crop growth is difficult to fully predict, which poses a challenge to crop yield prediction. Traditional statistical methods and simple simulation techniques often fail to fully consider these complexities and uncertainties.

[0129] In order to improve the accuracy of crop yield prediction, researchers have developed a complex model based on the Bayesian update rule and Monte Carlo simulation. This method can not only dynamically adjust the parameters of the crop growth model to adapt to new observations, but also generate multiple crop growth scenarios through Monte Carlo simulation and quantify the impact of environmental uncertainty, thereby generating crop yield prediction results with confidence intervals. This method provides a more reliable and flexible decision support tool for agricultural production, so a new optional solution is proposed, which includes:

[0130] Based on the adjusted crop growth model, multiple crop growth scenarios are simulated by Monte Carlo simulation method, and the impact of environmental uncertainty on crop production is quantified to generate crop yield prediction results with confidence intervals, including:

[0131] Monte Carlo method is used to obtain the probability distribution of the adjusted model parameters. Randomly draw multiple samples from , each sample represents a different combination of environmental data, and combines importance sampling to improve the representativeness of sampling, where the sampling formula is defined as: ; in, is the recommended distribution, It is weights of group samples; Combine each set of environmental data As a crop growth scenario, the Monte Carlo simulation method is used to simulate the crop growth situation under the crop growth scenario, and the predicted value of crop yield is calculated based on the crop growth situation. , where the formula for the predicted value of crop yield is defined as: ; in, is the environmental data in the comprehensive environmental dataset, It is The model parameters of the group are randomly drawn, is a crop growth model. is a random error term, where , is the error variance, is the variance of the environmental data; By quantifying the impact of the uncertainty of environmental factors on crop growth, the predicted value of crop yield is adjusted , the adjustment formula is defined as: ; in, is the adjusted crop yield forecast, is the adjustment factor, It is The effect of changes in environmental factors on crop yields. It is The weight of the combination of group environment data, is the time decay factor, It is a time delay; Among them, in calculating the weight When the exponential decay factor is introduced to better reflect the long-term impact of changes in environmental factors and consider nonlinear effects, the exponential decay factor is expressed as: ; Weight After considering the exponential decay factor and nonlinear effects, it is defined as: ; in is a parameter that controls the speed of weight decay. It is The difference measure between the combination of group environment data and the actual data, is the nonlinear influence coefficient; Statistically calculate the crop yield distribution under all crop growth scenarios and determine the statistical characteristics of crop yield, where the statistical characteristics include at least: weighted average and weighted standard deviation , calculated by the following formula: ; In addition, considering the nonlinear effect, the higher-order moments are introduced: ; Based on the statistical characteristics and high-order moments, a crop yield prediction result including a confidence interval is generated, wherein the confidence interval represents the numerical range within which the crop yield falls with a certain probability under the condition of uncertainty of given environmental factors, and the confidence interval calculation formula is defined as: ; in, is the critical value of the standard normal distribution corresponding to the desired confidence level (e.g., Confidence level ). Confidence interval It means that under the condition of given environmental factors uncertainty, crop yield has a certain probability (such as ) falls within this interval.

[0132] The following is a detailed explanation of each parameter: : Proposed distribution. This is a distribution chosen to improve sampling efficiency, usually an approximate distribution that is easy to sample. It is obtained based on prior knowledge or preliminary model fitting.

[0133] : Given the observed data (e.g. crop yields) and environmental data (such as temperature, humidity, etc.), the model parameters The posterior probability distribution of . It is a core concept in Bayesian statistics and is used to describe the probability distribution of model parameters after considering the observed data.

[0134] : No. The weight of each sample is used to adjust the importance of each sample to ensure that the samples drawn from the proposed distribution can represent the actual probability distribution. The true distribution is calculated using the Bayesian formula and the suggested distribution Get the ratio of .

[0135] : No. The crop yield forecast under a crop growth scenario. This is the crop yield forecast based on a specific combination of environmental conditions and model parameters.

[0136] : Crop growth model function. This is a function that describes environmental factors. How the function affects crop yield depends on the model used (e.g. linear regression, machine learning model, etc.). The model trained with historical data is usually fitted using statistical or machine learning methods.

[0137] :Environmental data in the comprehensive environmental data set. Including temperature, humidity, precipitation, etc. Historical and real-time observation data acquisition.

[0138] : No. Set of randomly drawn model parameters. Monte Carlo sampling as above.

[0139] : Crop growth model. This is a function that describes how environmental factors affect crop yield. Models trained on historical data (e.g. linear regression, machine learning models).

[0140] : Random error term, assuming . This is to simulate unpredictable random fluctuations. The acquisition method is random generation.

[0141] : Error variance. Represents the uncertainty of the model prediction error. It is estimated through the model validation process.

[0142] : Variance of environmental data. Indicates the variability of the environmental data itself. Statistical analysis of historical records of environmental data.

[0143] : The adjusted The crop yield forecast under the crop growth scenario is adjusted for changes in environmental factors and aims to more accurately reflect the actual crop yield.

[0144] :Adjustment coefficient. Used to control the impact of environmental factors on crop yield. Obtained based on experience or through sensitivity analysis.

[0145] : No. A measure of the effect of changes in a set of environmental factors on crop yields. It is obtained by comparing current environmental conditions with historical averages or other benchmark conditions.

[0146] : No. The effect of the change in a group of environmental factors on crop yield is quantified. Changes in environmental factors relative to baseline conditions and their impact on crop yields under different scenarios.

[0147] : Time attenuation factor. Used to reflect the weakening of the influence of environmental factors over time. Set according to the changing rules of environmental factors.

[0148] : Time delay. Indicates the time interval from the occurrence of environmental changes to the manifestation of their effects. Determined based on the crop growth cycle and environmental change characteristics.

[0149] : Exponential decay factor. Used to adjust the time decay effect of environmental changes.

[0150] : Parameter that controls the speed of weight decay. Used to adjust the relative importance between different combinations of environmental data. Determined through experiments or historical data analysis.

[0151] : No. A measure of the difference between a set of environmental data and actual data. It is obtained by comparing the difference between the simulated environmental data and the actual observed data.

[0152] :Non-linear influence coefficient. Used to consider the non-linear effect of changes in environmental factors. Determined by model fitting or sensitivity analysis.

[0153] : The number of Monte Carlo simulations or samples. Represents the probability distribution of the adjusted model parameters The total number of samples randomly selected from .

[0154] : Weighted average. Indicates the average level of crop yields under all simulation scenarios. Obtained by weighted summation.

[0155] : Weighted standard deviation. Indicates the variability of crop yield. Obtained by adding the weighted variance to the error variance.

[0156] : High-order moment. Used to describe the kurtosis of the distribution, reflecting the asymmetry and tail behavior of the data. Obtained by dividing the fourth variance by the square of the second variance.

[0157] : The critical value of the standard normal distribution corresponding to the desired confidence level. For example, Confidence level .

[0158] : Error variance, which represents the variability of the model prediction errors. It is estimated by the model validation process, usually through cross-validation or residual analysis.

[0159] : Weighted variance, which is the sum of the squares of the differences between the adjusted forecast value and the weighted mean, reflects the degree of dispersion of the forecast results. It is obtained by weighted summing the differences between each forecast value and the mean.

[0160] Here is a specific example: Crop type: Wheat Location: North China Plain, China Time frame: Spring planting to fall harvest in 2024 (assuming planting date is March 1 and harvest date is September 30) Comprehensive environmental data set: including historical and real-time data of multiple environmental factors such as temperature, humidity, precipitation, light intensity, soil fertility, etc. The specific data are as follows: Average temperature: ; Humidity: 60%; Precipitation: 500mm; Light intensity: 6 hours per day on average; Soil fertility: medium (NPK ratio is 1:1:1); The following are the parameter settings: : Number of Monte Carlo simulations; : adjustment coefficient; : time decay factor; : Parameters that control the speed of weight decay; : nonlinear influence coefficient; : error variance; : Variance of environmental data; : corresponds to The confidence level of the standard normal distribution critical value; From the probability distribution of the adjusted model parameters Randomly select 1000 samples from .

[0161] Using the importance sampling formula: ; For each sample Run simulations to calculate crop yield forecasts Assume that a simplified linear regression model is used as of the form, and known .

[0162] Calculation formula: ; Assumptions It is a linear regression model trained based on historical data. Its specific form is: ; in is the temperature, is humidity, is the amount of precipitation, is the light intensity, It's soil fertility.

[0163] Calculating adjusted crop yield forecasts : ; in It is The effect of changes in environmental factors on crop yields. It is The weight of the combination of group environment data, is the time decay factor, It's a time delay.

[0164] Introducing an exponential decay factor , and consider the nonlinear effects: ; Calculate weighted average and weighted standard deviation : ; Introducing higher order moments : ; Based on the statistical characteristics and high-order moments, the crop yield prediction results including confidence intervals are generated: ; Assume the following simulation results:

[0165] Substitute specific values ​​for detailed calculations: Calculating weights : for : ; Similarly, calculate all .

[0166] Adjusted forecast value

[0167] for : ; Similarly, calculate all .

[0168] Weighted average : ; Weighted standard deviation : ; Higher-order moments : ; Confidence interval : ; Assume that the final calculated result is: Weighted average Tons / hectare; Weighted standard deviation Tons / hectare; Higher-order moments ; The confidence interval for: ;

[0169] Given the uncertainty of environmental factors, the average wheat yield is expected to be 810 tons / hectare, and the 95% confidence interval is [785.8, 834.2] tons / hectare. The width of the confidence interval reflects the degree of impact of environmental uncertainty on crop yield. A wider confidence interval indicates that the uncertainty is high and environmental management measures need to be further optimized. By introducing complex formulas and Monte Carlo simulation, the model improves the accuracy and reliability of crop yield prediction, providing a scientific basis and decision-making support for agricultural production.

[0170] Based on the results of the confidence interval, more sophisticated irrigation, fertilization, and pest and disease control measures can be taken to reduce the impact of environmental uncertainty on crop yields. Environmental data should be continuously collected and updated, and model parameters should be adjusted regularly to ensure the accuracy of the prediction. The model should be applied to other crops and regions to verify its universality and effectiveness, and promote agricultural modernization and technological progress.

[0171] This complete implementation demonstrates how to apply the complex formula to predict crop yield in the wheat planting field, ensuring the flexibility and accuracy of the model while providing reliable decision support for agricultural production.

[0172] Figure 2 A structural diagram of a crop precision management system based on a crop growth model is provided for an embodiment of the present application. Figure 2 As shown, the system includes:

[0173] The acquisition module 21 is used to collect the first environmental data in the crop growth cycle and the second environmental data of the farmland through remote sensing technology and intelligent sensors and drones deployed in the field, and to align the first environmental data with the second environmental data in time and space to ensure the correspondence of environmental data from different sources at the same time point and in the same geographical area, so as to generate a comprehensive environmental data set with a time series;

[0174] A generation module 22 is used to dynamically adjust the model parameters of the pre-trained crop growth model according to the comprehensive environmental data set through the Bayesian update rule, and simulate multiple crop growth scenarios through the Monte Carlo simulation method based on the adjusted crop growth model, and quantify the influence of the uncertainty of environmental factors on crop growth, and generate a crop yield prediction result including a confidence interval; wherein the crop growth model is trained according to crop types, variety characteristics and historical crop growth records;

[0175] The generation module is also used to generate a crop management plan based on the crop yield prediction result and the degree of influence of the uncertainty of environmental factors on crop generation, using a particle swarm optimization algorithm, targeting the operation of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time, and at the same time, using a multi-objective optimization strategy to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal;

[0176] The adjustment module 23 is used to implement the adjusted crop management plan, obtain the actual growth response information of the crop in real time and use it as the input of the reinforcement learning algorithm, and automatically adjust the crop management plan under a specific reward mechanism by training the intelligent agent to achieve accurate management of crop growth.

[0177] Figure 2 The crop precision management device based on the crop growth model can be executed Figure 1 The implementation principle and technical effect of the crop precision management method based on the crop growth model described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the crop precision management device based on the crop growth model in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.

[0178] In one possible design, Figure 2 The crop precision management device based on the crop growth model of the illustrated embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0179] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0180] The processing component 32 is used for the above Figure 1 The embodiment provides a crop precision management method based on a crop growth model.

[0181] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0182] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0183] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0184] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0185] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0186] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0187] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is a method for precise crop management based on a crop growth model.

[0188] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0189] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0190] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0191] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A crop precision management method based on a crop growth model, characterized in that: include: Through remote sensing technology and intelligent sensors and drones deployed in the field, the first environmental data during the crop growth cycle and the second environmental data of the farmland are collected, and the first environmental data and the second environmental data are aligned in time and space to ensure the correspondence of environmental data from different sources at the same time point and the same geographical area, so as to generate a comprehensive environmental data set with a time series; According to the comprehensive environmental data set, the model parameters of the pre-trained crop growth model are dynamically adjusted by the Bayesian update rule, and on the basis of the adjusted crop growth model, multiple crop growth scenarios are simulated by the Monte Carlo simulation method, and the influence of the uncertainty of environmental factors on crop production is quantified, so as to generate a crop yield prediction result including a confidence interval; wherein the crop growth model is trained according to the crop type, variety characteristics and historical crop growth records; Based on the crop yield prediction results and the degree of influence of uncertainty of environmental factors on crop production, a particle swarm optimization algorithm is applied to generate a crop management plan for the operation of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time. At the same time, a multi-objective optimization strategy is used to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal; The adjusted crop management plan is implemented, and the actual growth response information of the crops is obtained in real time and used as the input of the reinforcement learning algorithm. The crop management plan is automatically adjusted under a specific reward mechanism by training the intelligent agent to achieve precise management of crop growth.

2. The method according to claim 1, characterized in that The method of dynamically adjusting the model parameters of the pre-trained crop growth model according to the comprehensive environmental data set by using the Bayesian update rule includes: Using the environmental data in the comprehensive environmental data set as input to a pre-trained crop growth model, so that the crop growth model outputs predicted crop growth conditions based on the input and in combination with crop types, variety characteristics and historical growth records; Determining actual crop growth conditions from the comprehensive environmental data set, and comparing the predicted crop growth conditions with the actual crop growth conditions to obtain a growth condition difference result; By using the Bayesian updating method, the probability distribution of the model parameters of the crop growth model is adjusted based on the growth condition difference result, so that the crop growth condition predicted by the crop growth model is consistent with the actual crop growth condition.

3. The method according to claim 2, characterized in that Based on the adjusted crop growth model, multiple crop growth scenarios are simulated by Monte Carlo simulation method, and the influence of uncertainty of environmental factors on crop production is quantified to generate crop yield prediction results with confidence intervals, including: A plurality of samples are randomly drawn from the probability distribution of the adjusted model parameters, each sample representing a different set of environmental data combinations; Combining each set of environmental data as a crop growth scenario through a Monte Carlo simulation method, simulating the crop growth conditions of the crop under the crop generation scenario, and calculating a predicted value of the crop yield based on the crop growth conditions; By quantifying the influence of the uncertainty of the obtained environmental factors on the crop growth, adjusting the predicted value of the crop yield, and statistically analyzing the distribution of crop yields under all crop growth scenarios, the statistical characteristics of the crop yield are determined; Based on the statistical characteristics, a crop yield prediction result including a confidence interval is generated, wherein the confidence interval represents a numerical range within which the crop yield falls with a certain probability under the condition of uncertainty of given environmental factors.

4. The method according to claim 3, characterized in that The step of adjusting the crop yield by quantifying the influence of the uncertainty of the environmental factors obtained on the crop growth comprises: For each crop growth scenario, calculating key indicators of crop growth under the crop growth scenario, and evaluating the variation range and volatility of the key indicators to obtain evaluation results; Based on the evaluation results, evaluating the degree of influence of different environmental factors on key indicators of crop growth, so as to identify key environmental factors that have a significant impact on the key indicators of crop growth; According to the key environmental factors, the predicted value of crop yield under each of the crop growth scenarios is adjusted, wherein the process of adjusting the predicted value of crop yield under each of the crop growth scenarios comprises: if the degree of influence of any environmental factor on crop growth meets a preset condition, appropriately increasing or decreasing the predicted value of crop yield in the corresponding crop growth scenario to quantify the degree of influence of the uncertainty of the environmental factor on crop growth; The statistical distribution of crop yields under all crop growth scenarios and the determination of statistical characteristics of crop yields include: Collecting the forecast values ​​of crop yields adjusted under all crop growth scenarios, constructing crop yield distribution, and analyzing the statistical characteristics of the crop yield distribution, including mean, standard deviation, skewness and kurtosis, to determine the variation range and central tendency of crop yields and obtain statistical analysis results; Based on the statistical analysis results, key statistical characteristics of crop yield are determined to describe the central position and dispersion of crop yield through key statistical characteristics, thereby reflecting the distribution of crop yield.

5. The method according to claim 2, characterized in that: Determining the actual crop growth conditions from the comprehensive environmental data set comprises: Extracting observation data reflecting the actual growth status of crops from the environmental data of the comprehensive environmental data set as actual crop growth conditions; The method further comprises: adjusting the probability distribution of the model parameters of the crop growth model based on the growth difference result by a Bayesian updating method, so that the crop growth condition predicted by the crop growth model is consistent with the actual crop growth condition. The posterior probability of each model parameter of the crop growth model under the current probability distribution is calculated, and the probability distribution of the model parameters is updated in combination with the prior knowledge and the growth condition difference results, so that the crop growth condition predicted by the crop growth model is closer to the actual crop growth condition.

6. The method according to claim 1, characterized in that Based on the crop yield prediction results and the degree of influence of uncertainty of environmental factors on crop production, a particle swarm optimization algorithm is applied to generate a crop management plan for the operation of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time. At the same time, a multi-objective optimization strategy is adopted to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal, including: According to the crop yield prediction results and the degree of influence of uncertainty of environmental factors on crop growth, the particle swarm optimization algorithm is applied to optimize the operating parameters of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time to generate a preliminary crop management plan; Using the preliminary crop management plan, defining multiple optimization objectives, converting the multiple optimization objectives into mathematical expressions, and building a multi-objective optimization model in combination with actual constraints, the multiple optimization objectives at least include maximizing crop yield, minimizing input costs, and minimizing environmental pollution; According to the multi-objective optimization model, a multi-objective optimization algorithm is used to solve the optimization problem to obtain a set of non-inferior solutions, each solution representing a feasible crop management solution; Based on the decision maker's preferences or priorities in a specific situation, one or more feasible crop management solutions are selected from the non-inferior solution set, and with the optimization goal of balancing economic benefits and environmental protection, one or more feasible crop management solutions are adjusted to determine the final crop management solution.

7. The method according to claim 6, characterized in that According to the crop yield prediction results and the degree of influence of uncertainty of environmental factors on crop growth, the particle swarm optimization algorithm is applied to optimize the irrigation amount, fertilization ratio, pest and disease control measures and harvesting time operations to generate a preliminary crop management plan, including: Using the iterative search process of the particle swarm optimization algorithm, multiple groups of operating parameter combinations are simulated, wherein the operating parameter combinations include one or a combination of any several operations of adjusting the irrigation amount, adjusting the fertilization ratio, adjusting the pest control measures, and adjusting the harvesting time; According to the crop yield prediction results and the degree of influence of the uncertainty of environmental factors on crop growth, the value range of each operating parameter in each group of operating parameter combinations is dynamically adjusted to ensure that the generated preliminary crop management plan can reflect the actual needs and environmental conditions of crop growth to the greatest extent; After multiple iterations, the optimal operating parameter combination is selected from the adjusted multiple operating parameter combinations as the preliminary crop management plan.

8. A crop precision management system based on crop growth model, characterized in that: The following steps are involved: A collection module is used to collect first environmental data in the crop growth cycle and second environmental data of the farmland through remote sensing technology and intelligent sensors and drones deployed in the field, and to align the first environmental data with the second environmental data in time and space to ensure the correspondence of environmental data from different sources at the same time point and in the same geographical area, so as to generate a comprehensive environmental data set with a time series; A generation module, used to dynamically adjust the model parameters of a pre-trained crop growth model according to the comprehensive environmental data set by using the Bayesian update rule, and simulate multiple crop growth scenarios by using the Monte Carlo simulation method based on the adjusted crop growth model, and quantify the influence of the uncertainty of environmental factors on crop generation, and generate a crop yield prediction result including a confidence interval; wherein the crop growth model is trained according to crop types, variety characteristics and historical crop growth records; The generation module is also used to generate a crop management plan based on the crop yield prediction result and the degree of influence of the uncertainty of environmental factors on crop generation, using a particle swarm optimization algorithm, targeting the operation of irrigation amount, fertilization ratio, pest and disease control measures and harvesting time, and at the same time, using a multi-objective optimization strategy to adjust the crop management plan with a balance between economic benefits and environmental protection as the optimization goal; The adjustment module is used to implement the adjusted crop management plan, obtain the actual growth response information of the crop in real time and use it as the input of the reinforcement learning algorithm, and automatically adjust the crop management plan under a specific reward mechanism by training the intelligent agent to achieve precise management of crop growth.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a crop precision management method based on a crop growth model as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a crop precision management method based on a crop growth model as described in any one of claims 1 to 7 is implemented.

Images