A wheat growth regulation and yield prediction integrated method based on artificial intelligence

Abstract

The application discloses a wheat growth regulation and yield prediction integrated method based on artificial intelligence, obtains wheat growth data and environmental factor data through a sensor network, and carries out standardization after pretreatment; microelement proportioning adjustment directions are determined by utilizing machine learning modeling analysis to identify the nutrient requirement change characteristics of the wheat growth stage; personalized nutrient supply schemes are generated in combination with dynamic regulation strategies, the regulation schemes are optimized through a resource utilization efficiency evaluation model, and a field management system is updated, and finally, wheat yield prediction results are obtained in combination with a yield prediction model. The method provided by the application realizes accurate nutrient management of the whole wheat growth process, improves resource utilization efficiency, and provides effective technical support for improving wheat yield and quality.

Description

Technical Field

[0001] This invention belongs to the field of agricultural technology, and in particular relates to an integrated method for wheat growth regulation and yield prediction based on artificial intelligence. Background Technology

[0002] As a vital global food crop, wheat's growth regulation and yield forecasting are directly related to food security and sustainable agricultural development, possessing extremely critical strategic significance. Scientific management of the wheat growth process and improving the accuracy of yield forecasting are not only core requirements for agricultural modernization but also important pathways to addressing climate change and resource scarcity.

[0003] However, current methods for wheat growth management and yield prediction still have significant shortcomings. Many traditional schemes rely on empirical judgment or fixed patterns, making it difficult to adapt to complex environmental changes in different regions and growth stages, and lacking in-depth analysis of the dynamic interaction between crops and the environment. When faced with variable climatic conditions and soil differences, these methods often lead to low resource utilization efficiency and even regulatory errors, affecting the final yield.

[0004] To address the aforementioned issues, there is an urgent need to propose an integrated method for wheat growth regulation and yield prediction based on artificial intelligence. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides an integrated method for wheat growth regulation and yield prediction based on artificial intelligence.

[0006] This invention proposes an integrated method for wheat growth regulation and yield prediction based on artificial intelligence, comprising the following steps:

[0007] Based on sensor networks, data on wheat growth stage differences and environmental factor-sensitive data are obtained from the field to obtain an initial dataset. After preprocessing, a standardized dataset is obtained.

[0008] For a standardized dataset, machine learning methods were used to model and analyze the correlation between differences in wheat growth stages and sensitivity to environmental factors, obtain the response patterns of each stage to changes in climate conditions and differences in soil properties, and obtain the sensitivity analysis results.

[0009] Based on the results of sensitivity analysis and plant status monitoring data, we can identify the characteristics of changes in the nutritional requirements of wheat at the current growth stage and determine the direction for adjusting the micronutrient ratio.

[0010] Based on the adjustment direction of the micronutrient ratio and combined with dynamic control strategies, personalized nutrient supply plans for different plots and growth stages are generated to obtain fertilization ratio parameters.

[0011] Based on the fertilization ratio parameters and the resource utilization efficiency assessment model, the applicability of the current scheme under changes in climate conditions and differences in soil characteristics is analyzed. If the resource utilization rate is lower than the preset threshold, the ratio parameters are readjusted and the optimized control scheme is determined.

[0012] By optimizing the control scheme, updating the dynamic control strategy database in the field management system, and applying the adjusted micronutrient ratios and control measures to actual production in a timely manner, wheat yield prediction results can be obtained by combining the wheat yield prediction model.

[0013] Optionally, the step of acquiring wheat growth stage difference data and environmental factor sensitive data from the field based on sensor networks to obtain an initial dataset, followed by preprocessing to obtain a standardized dataset, includes:

[0014] By using a sensor network to monitor different plots of land in real time, data on differences in wheat growth stages and sensitive data on environmental factors are obtained and stored as an initial dataset.

[0015] Based on the initial dataset, data cleaning, data normalization, and linear interpolation are performed to obtain a standardized dataset.

[0016] Optionally, for the standardized dataset, machine learning methods are used to model and analyze the correlation between differences in wheat growth stages and sensitivity to environmental factors, obtaining the response patterns of each stage to changes in climate conditions and differences in soil properties, and obtaining sensitivity analysis results, including:

[0017] For a standardized dataset, the support vector machine method was used to preliminarily model the correlation between wheat growth stages and sensitivity to environmental factors, analyze the characteristic performance of different stages, and obtain preliminary correlation model results.

[0018] Based on the preliminary correlation model results, corresponding stage-specific feature data were extracted for climate condition changes and soil property differences. The feature data were divided into several subsets using a data stratification processing method to determine the stratified feature subsets.

[0019] For the stratified feature subsets, clustering method is used to group the differences in wheat growth stages for analysis, obtain the correspondence between the features of each group and the sensitivity to environmental factors, and obtain the feature mapping results after grouping.

[0020] The grouped feature mapping results are filtered to obtain the core feature set.

[0021] For the core feature set after screening, the soil property differences and stage characteristics are jointly analyzed by data integration tools, and multi-dimensional information is integrated to determine the integrated comprehensive feature dataset.

[0022] Based on the integrated feature dataset, the linear regression method was used to quantitatively analyze the response pattern of wheat growth stage differences and sensitivity to environmental factors, and the final sensitivity analysis results were obtained.

[0023] Optionally, the step of identifying the characteristics of changes in wheat's nutritional requirements at the current growth stage and determining the direction for adjusting the micronutrient ratio by combining the results of sensitivity analysis with plant status monitoring data includes:

[0024] Based on the sensitivity analysis results, plant status monitoring data were integrated to analyze the characteristics of changes in nutrient requirements of wheat at the current growth stage and to obtain the specific patterns of demand changes.

[0025] Based on the specific patterns of demand changes, data processing tools are used to classify and stratify the characteristics of change, and to determine the degree of deviation of nutritional requirements under different categories.

[0026] If the deviation of nutrient requirements exceeds the preset threshold, a preliminary judgment result on the cause of the deviation will be obtained based on the plant status data at the current stage.

[0027] Based on the preliminary judgment results, the support vector machine method was used to further classify and analyze the causes of deviation and identify the core factors leading to the deviation.

[0028] Based on the classification results of the core factors, a preliminary plan for adjusting the proportion of trace elements is generated using data mapping tools to obtain a reference basis for the direction of adjustment;

[0029] If the reference basis for adjusting the direction conflicts with the data of historical control mechanisms, a second verification will be conducted through data comparison to determine the final direction of the allocation adjustment.

[0030] Optionally, the step of adjusting the direction based on the micronutrient ratio, combined with a dynamic control strategy, to generate personalized nutrient supply plans for different plots and growth stages, yields fertilization ratio parameters, including:

[0031] Based on the adjustment direction of trace element ratios and combined with dynamic control strategies, personalized nutrient supply plans are generated for different plots and growth stages. Based on the differences in personalized nutrient supply needs of different plots and growth stages, data integration tools are used to perform hierarchical analysis of trace element demand distribution, and a personalized plan framework for different plots at the current growth stage is obtained.

[0032] For the personalized solution framework, a dynamic control strategy is integrated to compare the real-time collected environmental fluctuation data. If the environmental fluctuation data exceeds the preset threshold, the micronutrient ratio is adjusted to determine the preliminary fertilization ratio parameters.

[0033] By using the control measures generation tool, key adjustment points can be extracted from the initial fertilizer ratio parameters, and specific implementation plans for targeted control measures can be obtained by combining timely instructions and needs.

[0034] Based on the specific implementation plan, the fertilizer ratio parameters are refined and adjusted using data mapping to obtain the final nutrient supply plan applicable to different plots.

[0035] For the final nutrition supply plan, timely adjustment instructions are generated through instruction distribution to determine the execution time and allocation method of ratio parameters for each region.

[0036] Optionally, the step of updating the dynamic control strategy database in the field management system through the optimized control scheme, and applying the adjusted micronutrient ratios and control measures to actual production in a timely manner, includes:

[0037] By using the optimized control scheme and pre-established mapping tools, the adjustment of trace element ratios is linked to the production execution process, resulting in a preliminary synchronous update framework.

[0038] Based on the preliminary synchronous update framework, the ratio adjustment data is matched with the control measures. If the matching result shows that the data deviation exceeds the preset threshold, the correction process is triggered to determine the adjusted ratio parameters.

[0039] By using the adjusted proportioning parameters and data distribution tools, the control measures are synchronized to the production execution unit to obtain corresponding execution data feedback.

[0040] Based on the feedback of execution data, the dynamic strategy of the production execution process is tracked in real time. If the execution data is found to be inconsistent with the strategy record, the data calibration process is initiated to obtain the calibrated execution status.

[0041] By comparing and updating the dynamic policy records with the management database using the calibrated execution status and information integration tools, the final database update result is determined.

[0042] Based on the database update results, a log recording module is used to store the entire process of synchronous updates into the field management database to obtain complete strategy execution data.

[0043] Optionally, after obtaining the wheat yield prediction results by combining the wheat yield prediction model, the following may also be included:

[0044] Obtain comparative data between the predicted results and the actual output, and determine the magnitude and distribution characteristics of the deviation value through the deviation value calculation method;

[0045] Based on the distribution characteristics of the deviation values, the impact of dynamic control strategies on the final output is analyzed. A regression analysis model is used to quantify the influencing factors and obtain key indicators of the degree of impact.

[0046] For key indicators of impact, adjust the parameter settings of sensitivity analysis. If the key indicators exceed the preset threshold range, optimize the model parameters of the dynamic control strategy and obtain the adjusted parameter configuration.

[0047] Based on the adjusted parameter configuration, optimize the collection frequency of plant status monitoring. If the matching degree between the collection frequency and the sensitive data of environmental factors is lower than the predetermined standard, determine a new collection frequency scheme by automatically adjusting the collection interval.

[0048] By adopting a new data acquisition frequency scheme and updating the input conditions for data analysis, combined with the updated execution records and environmental data, a regression analysis model is used to recalculate the prediction results, resulting in optimized production prediction data.

[0049] The present invention also proposes a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0050] The present invention also proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.

[0051] The present invention also proposes a computer program product, including a computer program that, when executed by a processor, implements the steps of the method.

[0052] Compared with the prior art, the present invention has the following advantages and technical effects:

[0053] Precision Nutrition Management Based on Dynamic Regulation Strategy: This invention combines real-time sensor data with machine learning technology to generate personalized nutrition supply plans, achieving precise regulation. Compared to traditional fixed methods, this approach can dynamically adjust fertilization plans according to the specific needs of wheat at different growth stages and in different plots, avoiding resource waste and nutrient deficiencies or excesses, significantly improving resource utilization efficiency, promoting healthy wheat growth, and increasing yield and quality.

[0054] Comprehensive Decision Support Based on Multimodal Data Fusion: This invention integrates wheat growth data, environmental factors (temperature, humidity, soil properties, etc.), and yield prediction models to form a closed-loop optimization system. This multimodal data fusion approach can comprehensively analyze the factors influencing wheat growth, providing richer and more accurate information support for decision-making, effectively improving the accuracy of wheat yield prediction and the overall management level of agricultural production.

[0055] Dynamic optimization based on resource utilization efficiency assessment: This invention introduces a resource utilization efficiency assessment model and combines it with a dynamic threshold adjustment mechanism to monitor the implementation effect of fertilization programs in real time. When the resource utilization rate falls below a preset threshold, the system automatically adjusts the fertilization ratio parameters to optimize the fertilization program. This dynamic assessment and adjustment mechanism differs from traditional static fertilization strategies, better adapting to complex environmental changes and ensuring that the fertilization program is always in an optimal state, further improving resource utilization efficiency and reducing environmental pollution risks.

[0056] Based on continuous improvement of the entire intelligent decision-making chain: This invention establishes an entire intelligent decision-making method, from data acquisition to dynamic modeling, real-time control, and closed-loop optimization. Field data is collected in real time through sensor networks, and machine learning techniques are used for modeling and analysis to generate personalized control plans. The execution results are then fed back in real time to optimize model parameters. This entire intelligent decision-making method enables intelligent and refined management of agricultural production, automatically adjusting the optimization direction according to changes in the production process, and promoting the transformation of agricultural production methods towards high efficiency, intelligence, and sustainability.

[0057] This invention achieves precise nutritional management and optimized regulation of wheat throughout its entire growth process through innovative methods such as dynamic control strategies, multimodal data fusion, resource utilization efficiency assessment, and whole-chain intelligent decision-making. This significantly improves resource utilization efficiency, increases wheat yield and quality, and provides an efficient and intelligent technical solution for the development of agricultural modernization. Attached Figure Description

[0058] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0059] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0060] Figure 2 This is a schematic diagram of the process for adjusting fertilizer ratio parameters according to an embodiment of the present invention. Detailed Implementation

[0061] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0062] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0063] Example 1

[0064] This embodiment provides an integrated method for wheat growth regulation and yield prediction based on artificial intelligence, including the following steps:

[0065] Based on sensor networks, data on wheat growth stage differences and environmental factor-sensitive data are obtained from the field to obtain an initial dataset. After preprocessing, a standardized dataset is obtained.

[0066] For a standardized dataset, machine learning methods were used to model and analyze the correlation between differences in wheat growth stages and sensitivity to environmental factors, obtain the response patterns of each stage to changes in climate conditions and differences in soil properties, and obtain the sensitivity analysis results.

[0067] Based on the results of sensitivity analysis and plant status monitoring data, we can identify the characteristics of changes in the nutritional requirements of wheat at the current growth stage and determine the direction for adjusting the micronutrient ratio.

[0068] Based on the adjustment direction of the micronutrient ratio and combined with dynamic control strategies, personalized nutrient supply plans for different plots and growth stages are generated to obtain fertilization ratio parameters.

[0069] Based on the fertilization ratio parameters and the resource utilization efficiency assessment model, the applicability of the current scheme under changes in climate conditions and differences in soil characteristics is analyzed. If the resource utilization rate is lower than the preset threshold, the ratio parameters are readjusted and the optimized control scheme is determined.

[0070] By optimizing the control scheme, updating the dynamic control strategy database in the field management system, and applying the adjusted micronutrient ratios and control measures to actual production in a timely manner, wheat yield prediction results can be obtained by combining the wheat yield prediction model.

[0071] As a specific implementation method, such as Figure 1 As shown, it includes the following steps:

[0072] Step S101: Data on differences in wheat growth stages and sensitive data on environmental factors are acquired from the field through a sensor network. Real-time data collection is performed on the climate conditions and soil characteristics of different plots, and the data is stored as an initial dataset. Based on the initial dataset, data preprocessing methods are used to clean and standardize the collected data on differences in growth stages and sensitive data on environmental factors, outliers are removed and missing values ​​are interpolated to fill in the gaps, and the processed standardized dataset is determined.

[0073] Real-time monitoring of different plots in the field using a sensor network acquires data on wheat growth stage differences and environmentally sensitive data, which are stored as an initial dataset. Based on this initial dataset, data cleaning methods are used to preliminarily process the collected data, including outlier removal using preset threshold ranges to obtain a cleaned preliminary dataset. For this cleaned preliminary dataset, standardization methods are used to normalize the data, unifying the units of measurement from different sources to determine a standardized intermediate dataset. If missing values ​​exist in the intermediate dataset, interpolation methods are used to fill in the missing values, combining data from adjacent time points through linear interpolation to obtain a complete dataset.

[0074] In feasible real-time monitoring of different plots in a field using a sensor network, a specific scenario can be envisioned: A wheat planting base is divided into 10 plots, each equipped with temperature and humidity sensors, soil moisture sensors, and light sensors to collect data in real time. Monitoring reveals that the wheat in plot A is in the jointing stage, while plot B is still in the tillering stage, showing a significant difference in growth stage. Simultaneously, environmental data, such as the average daily temperature of 18 degrees Celsius in plot A and 15 degrees Celsius in plot B, demonstrate the influence of environmental factors. This data is stored as an initial dataset, laying the foundation for subsequent analysis.

[0075] During the data cleaning phase, for outlier removal, a reasonable threshold for temperature data can be set between -5 degrees Celsius and 40 degrees Celsius. If the temperature data collected for a certain plot at a certain time point is 50 degrees Celsius, which significantly exceeds the reasonable range, it will be marked as an outlier and removed. The cleaned preliminary dataset is more reliable, avoiding interference from extreme values ​​in subsequent analysis and helping to improve data quality.

[0076] When standardizing the initial dataset after cleaning, the min-max normalization method can be used to unify data of different dimensions, such as temperature and humidity, into the range of 0 to 1.

[0077] For example, if the minimum temperature data is 10 degrees and the maximum is 30 degrees, a temperature value of 20 degrees at a certain time point will be normalized to 0.5. This process eliminates the dimensional differences between data from different sources, facilitating subsequent modeling and analysis and enhancing the comparability of the data.

[0078] If missing values ​​exist in the standardized intermediate dataset, they can be filled using linear interpolation. For example, if humidity data for a certain plot of land is missing for a particular day, but the humidity value the previous day was 0.6 and the next day was 0.8, then the missing value can be estimated to be approximately 0.7. This method combines data from adjacent time points to reasonably fill in the gaps, ensuring the continuity of the dataset and providing complete data support for subsequent analysis, effectively avoiding analytical bias caused by missing values.

[0079] Specifically, the implementation of each of the above technical steps is closely aligned with the operational needs of wheat growth monitoring. From data collection and cleaning to standardization and missing value imputation, each step aims to improve data quality and ensure the accuracy of the analysis results.

[0080] For example, outlier removal avoids interference from invalid data in judging differences in growth stages, while standardization provides a unified standard for the correlation analysis between environmental factors and growth stages. Ultimately, the complete dataset can support precision agriculture decisions, such as adjusting irrigation or fertilization strategies for different plots, thereby improving wheat yield and resource utilization efficiency.

[0081] Specifically, the application of methods such as linear interpolation also demonstrates the flexibility of the technology. In cases where data for certain plots is significantly missing, more complex interpolation adjustments can be made by combining historical data trends to ensure the rationality of data completion. This approach not only improves data completeness but also provides a solid foundation for subsequent predictive model construction, contributing to the scientific management of agricultural production.

[0082] Step S102: For the standardized dataset, machine learning methods are applied to model and analyze the correlation between differences in wheat growth stages and sensitivity to environmental factors, obtain the response patterns of each stage to changes in climate conditions and differences in soil properties, and obtain the sensitivity analysis results.

[0083] For a standardized dataset, a support vector machine (SVM) method was used to initially model the correlation between wheat growth stages and sensitivity to environmental factors, analyzing the characteristic performance of different stages and obtaining preliminary correlation model results. Based on the preliminary correlation model results, corresponding stage-specific feature data were extracted for climate condition changes and soil characteristic differences. Using a data stratification method, the feature data was divided into multiple subsets to determine the stratified feature subsets. For the stratified feature subsets, a clustering method was used to group and analyze the differences in wheat growth stages, obtaining the correspondence between each group of features and sensitivity to environmental factors, resulting in grouped feature mapping results. Based on the grouped feature mapping results, if the correlation between a certain group of features and climate condition changes was lower than a preset threshold, the data in that group was further filtered to remove irrelevant features, obtaining the core feature set after filtering. For the core feature set after filtering, a data integration tool was used to jointly analyze soil characteristic differences and stage-specific features, fusing multi-dimensional information to determine the integrated comprehensive feature dataset. Based on the integrated comprehensive feature dataset, a linear regression method was used to quantitatively analyze the response pattern of wheat growth stage differences and sensitivity to environmental factors, obtaining the final sensitivity analysis results.

[0084] Feasible data were collected from multiple plots within a wheat-growing region, including data on growth stages and environmental factors such as temperature, humidity, and soil nutrient content. Using support vector machines, the characteristics of wheat from the tillering to heading stages were analyzed. The results showed that temperature significantly affected the growth rate during the jointing stage, while humidity had a more pronounced effect on the tillering stage. This preliminary modeling helped identify the correlation between key environmental factors and growth stages, laying the foundation for subsequent analysis.

[0085] When performing data stratification on the results of the preliminary correlation model, the feature data can be divided into subsets such as high temperature and high humidity, and low temperature and low humidity, based on climatic conditions. For example, a plot of land with an average daily temperature of 20 degrees Celsius and humidity of 70% would be classified into the high temperature and high humidity subset, while another plot with a temperature of 12 degrees Celsius and humidity of 40% would be classified into the low temperature and low humidity subset. This stratification method facilitates differentiated analysis of growth characteristics under different environmental conditions.

[0086] When using clustering methods to group and analyze the stratified feature subsets, the differences in wheat growth stages can be divided into three groups using clustering algorithms. Assume the first group of plots is mostly in the jointing stage, characterized by high temperatures; the second group is mostly in the tillering stage, characterized by low humidity; and the third group is in a mixed stage. This grouping clarifies the correspondence between the characteristics of each group and environmental factors, providing a clear classification basis for further analysis.

[0087] If the correlation between a set of features and changes in climate conditions is lower than a preset threshold, such as a correlation coefficient less than 0.3, then that set of data undergoes a secondary screening. For example, if a set of data includes wind speed, but analysis reveals that its impact on the growth stage is negligible, then it is removed, retaining core features such as temperature and humidity. This screening ensures the relevance and accuracy of subsequent analyses.

[0088] When conducting joint analysis on the selected core feature set, data integration tools can be used to combine soil characteristics such as pH and organic matter content with growth stage characteristics. For example, suppose a plot of soil has a pH of 6.5 and a high organic matter content, which is associated with rapid growth during the jointing stage. By fusing multi-dimensional information, a comprehensive feature dataset is formed, supporting a more comprehensive analysis.

[0089] When using linear regression for quantitative analysis, we can analyze the specific impact of environmental factors on the differences in wheat growth stages using a comprehensive feature dataset. For example, the analysis might find that for every 1-degree Celsius increase in temperature, the jointing stage is advanced by approximately 2 days, while a 10% increase in humidity extends the tillering stage by approximately 1 day. These quantitative results provide an intuitive basis for understanding the response patterns of environmental factors and help in developing targeted agricultural management measures.

[0090] Step S103: Based on the sensitivity analysis results and plant status monitoring data, identify the characteristics of changes in the nutritional requirements of wheat at the current growth stage, determine if the current requirements deviate from the preset threshold, trigger the nutritional regulation mechanism, and determine the direction of micronutrient ratio adjustment.

[0091] Based on the sensitivity analysis results, plant status monitoring data is integrated to analyze the nutrient requirement changes of wheat at the current growth stage and obtain specific patterns of requirement changes. According to these patterns, data processing tools are used to stratify and classify the change characteristics, determining the degree of deviation in nutrient requirements under different categories. If the deviation exceeds a preset threshold, the information integration module extracts plant status data for the current stage to obtain a preliminary judgment of the cause of the deviation. Based on this preliminary judgment, a support vector machine method is used to further classify and analyze the causes of the deviation, identifying the core factors leading to the deviation. Based on the classification results of the core factors, a preliminary plan for adjusting the micronutrient ratio is generated using a data mapping tool to obtain a reference basis for the adjustment direction. If the reference basis for the adjustment direction conflicts with historical control mechanism data, a secondary verification is performed using a data comparison module to determine the final adjustment direction.

[0092] Feasible sensitivity analysis results typically reflect the response patterns of wheat to environmental factors at different growth stages, while plant status monitoring data provides real-time growth indicators such as leaf color, plant height, and leaf area. Combining the two allows for a more comprehensive analysis of the changing characteristics of wheat's nutrient requirements at the current stage. For example, in a wheat-growing region, sensitivity analysis might show a high nitrogen requirement during the jointing stage, while monitoring data reveals yellowing leaves and lower plant height in some plots, potentially indicating insufficient nitrogen supply. Data fusion can provide a preliminary indication of the direction of deviation in nutrient requirements.

[0093] When stratifying and classifying specific patterns of demand changes, data processing tools can be used to categorize the changing characteristics into three types: high demand, medium demand, and low demand. For example, a plot of land might require 50 kg of nitrogen per hectare during the jointing stage, falling into the high demand category, while another plot requires only 20 kg per hectare, classifying it as low demand. This classification method facilitates the identification of the degree of nutrient deviation between different plots. If the actual supply to a high-demand plot is only 30 kg per hectare, exceeding a preset threshold of 10 kg, further analysis is required.

[0094] When extracting plant condition data and initially determining the cause of deviations, leaf nutrient content and soil nutrient data can be integrated. For example, if the leaf nitrogen content in a certain plot is 2 percentage points lower than normal, and the soil nitrogen content is also low, the initial assessment is insufficient soil nutrient supply. This initial assessment provides direction for subsequent analysis.

[0095] When using the support vector machine (SVM) method to classify and analyze the causes of deviations, they can be categorized into three main types: soil factors, climate factors, and management factors. If the analysis reveals that the nitrogen deviation in a certain plot is primarily related to soil nutrient loss, rather than irrigation or fertilization management issues, then soil factors are identified as the core factors. This classification helps to accurately pinpoint the root cause of the problem.

[0096] When generating micronutrient ratio adjustment plans using data mapping tools, the nitrogen, phosphorus, and potassium ratios can be adjusted based on core factors. For example, assuming a plot of land has insufficient nitrogen in its soil, the initial plan suggests increasing the nitrogen fertilizer ratio from 20% to 30%, while keeping the phosphorus and potassium ratio unchanged. This adjustment direction provides a reference for subsequent implementation.

[0097] If the adjustment direction conflicts with historical control mechanism data, a secondary verification can be conducted. For example, if historical data indicates that the nitrogen fertilizer ratio should not exceed 25%, while the initial plan is 30%, then by comparing the current soil conditions with historical effects, the final ratio can be determined to be 27%. This verification ensures the rationality of the adjustment plan while taking into account both historical experience and current needs.

[0098] Step S104: Adjust the direction according to the micronutrient ratio, combine with dynamic control strategy to generate personalized nutrient supply plan for different plots and growth stages, obtain timely control measures instructions, and obtain specific fertilization ratio parameters.

[0099] like Figure 2 As shown, based on the adjustment direction of micronutrient ratios and combined with dynamic control strategies, personalized nutrient supply plans are generated for different plots and growth stages. According to the differences in personalized nutrient supply needs across different plots and growth stages, a data integration tool is used to perform a stratified analysis of micronutrient demand distribution, resulting in a personalized plan framework for different plots at the current growth stage. For this personalized plan framework, a dynamic control strategy is integrated, comparing real-time collected environmental fluctuation data. If the environmental fluctuation data exceeds a preset threshold, the micronutrient ratio is adjusted to determine the initial fertilization ratio parameters. A control measure generation tool extracts key adjustment points from the initial fertilization ratio parameters and, combined with timely instruction requirements, obtains specific implementation plans for targeted control measures. Based on the specific implementation plans, the fertilization ratio parameters are refined to obtain the final nutrient supply plan applicable to different plots. For the final nutrient supply plan, a timely control instruction is generated through an instruction distribution module, determining the execution time and ratio parameter allocation method for each region.

[0100] The core of an feasible, personalized solution lies in identifying the unique nutrient requirements of wheat at different growth stages and customizing the design based on soil characteristics and environmental conditions. For example, if one plot of soil has low potassium content during the jointing stage, while another plot shows phosphorus deficiency, data integration tools can be used to perform a stratified analysis of micronutrient requirements, classifying them into high, medium, and low levels, thus constructing a preliminary framework for a personalized solution.

[0101] By integrating dynamic control strategies into the personalized solution framework, real-time comparison of environmental fluctuation data can be achieved. For example, if soil moisture suddenly increases after irrigation, exceeding a preset threshold, it may affect nitrogen absorption efficiency. In this case, the system will adjust the micronutrient ratio based on the fluctuation data, initially determining to reduce the nitrogen fertilizer proportion from 25% to 20% to avoid the risk of root damage from excessive application. This dynamic adjustment method can respond promptly to environmental changes.

[0102] When extracting key adjustment points using the regulatory measures generation tool, specific implementation plans can be generated by combining real-time command requirements. For example, suppose a plot of land needs increased zinc supply during the heading stage. Initial mixing parameters indicate a zinc fertilizer ratio of 5%. However, by analyzing historical data and current needs, the system generates a specific plan to apply the zinc fertilizer in two stages, each time at a ratio of 2.5%, to improve absorption efficiency. This staged application method better matches the wheat's growth rhythm.

[0103] When refining the fertilization ratio parameters, the plan can be further optimized based on the specific characteristics of the plot. For example, assuming a plot has highly acidic soil, the initial plan might have a phosphate fertilizer ratio of 15%. However, considering the phosphorus fixation effect of acidic soil, the system would adjust the ratio to 18% and recommend adding a small amount of alkaline regulator. This refined adjustment can improve the availability of nutrients.

[0104] When generating timely adjustment instructions through the instruction distribution module for the final nutrient supply plan, the execution time and allocation method of ratio parameters for each region can be clearly defined. For example, in a certain planting area, the system instruction specifies that the first plot should be fertilized with nitrogen fertilizer at 8:00 AM at a ratio of 22%, while the second plot should be fertilized at 3:00 PM at a ratio of 18%, to adapt to the different light and temperature conditions of the plots. This precise allocation of time and parameters can optimize resource utilization efficiency.

[0105] From an overall perspective, the close integration of the above-mentioned links, from demand analysis to dynamic adjustment and final instruction distribution, forms a complete closed-loop management system. This approach not only specifically meets the nutritional needs of wheat at different growth stages but also reduces resource waste and improves the scientific nature of planting management through real-time regulation.

[0106] Step S105: Based on the fertilization ratio parameters and the resource utilization efficiency assessment model, analyze the applicability of the current scheme under changes in climate conditions and differences in soil characteristics. If the resource utilization rate is lower than the preset threshold, readjust the ratio parameters and determine the optimized control scheme.

[0107] For fertilization ratio parameters, a pre-established resource utilization efficiency assessment model is used to analyze the adaptability of the current fertilization ratio, determining whether the resource utilization rate is lower than a preset threshold, and thus determining the applicability assessment result of the current plan. If the applicability assessment result shows that the resource utilization rate is lower than the preset threshold, the fertilization ratio is recalculated using a parameter adjustment tool. Combined with the analysis of differences in climate conditions and soil characteristics, the adjusted ratio parameters are obtained. For the adjusted ratio parameters, a data mapping module is used to compare them with relevant indicators of condition changes and efficiency assessment to obtain an optimized control plan framework. Based on the optimized control plan framework, an instruction generation tool is used to integrate the utilization rate threshold and the results of the adaptability analysis to determine the specific execution parameters and time allocation method. For the specific execution parameters and time allocation method, an information distribution module is used to distribute the control plan to the execution units of the corresponding plots to obtain the final plan deployment status. The status monitoring module continuously tracks the plan deployment status. If a significant change in climate conditions or soil characteristics is detected, the parameter adjustment process is triggered to recalculate the fertilization ratio, resulting in an updated control plan.

[0108] This step is feasible, and its core lies in comprehensively assessing the compatibility of fertilizer application ratios with soil, climate, and other conditions to determine whether resources are being used effectively. For example, if wheat in a certain plot is in the jointing stage and nitrogen fertilizer accounts for 30% of the fertilizer application, model analysis reveals that the resource utilization rate is only 60%, lower than the preset threshold of 75%. This indicates that the current plan carries a risk of waste and needs adjustment.

[0109] When resource utilization rates fall below a threshold, recalculating the fertilizer mix using parameter adjustment tools can incorporate differentiated analysis based on specific climate and soil characteristics. For example, if the current soil is acidic and recent rainfall has led to nitrogen loss, the nitrogen fertilizer ratio can be adjusted from 30% to 25%, with a small amount of slow-release fertilizer added to improve absorption efficiency. This adjustment method is better suited to local conditions.

[0110] When comparing the adjusted ratio parameters with changes in conditions, attention should be paid to the impact of indicators such as soil moisture and temperature on nutrient absorption. For example, if the adjusted ratio parameters show a 20% phosphate fertilizer content, but data mapping reveals that current soil moisture is high, increasing the risk of phosphorus fixation, it is recommended to slightly adjust the phosphate fertilizer content to 22% and apply it in two applications to reduce losses.

[0111] When determining execution parameters and time allocation using the instruction generation tool for the optimized control scheme framework, a refined design can be made according to the specific needs of each plot. For example, if a plot needs increased potassium fertilizer supply during the heading stage, the system generates an instruction to apply potassium fertilizer at 7:00 AM, accounting for 15% of the total fertilizer application, to fully utilize the lower morning temperatures and higher absorption efficiency.

[0112] When the control plan is issued to the execution unit, the accurate transmission of instructions can be ensured. Suppose that for two adjacent plots, the system issues different fertilization times and ratios. For the first plot, 20% nitrogen fertilizer is applied at 8:00 AM, and for the second plot, 18% nitrogen fertilizer is applied at 2:00 PM. This differentiated arrangement can adapt to changes in the local microenvironment.

[0113] Furthermore, continuous monitoring of the deployment status of the plan can promptly identify the impact of environmental changes. For example, if a plot of land is suddenly hit by strong winds after fertilization, potentially leading to uneven fertilizer distribution, the system will trigger a parameter adjustment process, suggesting an additional small topdressing, at 5% of the original amount, to compensate for the losses.

[0114] Furthermore, when triggering parameter adjustment procedures and recalculating fertilization ratios, a comprehensive judgment can be made by combining historical data and the current environment. For example, if soil nutrient testing in a certain plot shows a persistent zinc deficiency, the updated plan recommends increasing the zinc fertilizer ratio from 3% to 5% and suggests applying it concurrently with irrigation to improve effectiveness. This dynamic adjustment can better match crop needs.

[0115] Step S106: Update the dynamic control strategy database in the field management system using the optimized control scheme, apply the adjusted micronutrient ratios and control measures to actual production in a timely manner, and obtain the updated execution records.

[0116] By optimizing the control scheme and utilizing a pre-established mapping tool, the adjustment of trace element ratios is linked to the production execution process, resulting in a preliminary synchronous update framework. Based on this framework, an information comparison module matches the ratio adjustment data with the control measures. If the matching result shows a data deviation exceeding a preset threshold, a correction process is triggered to determine the adjusted ratio parameters. Using the adjusted ratio parameters, a data distribution tool synchronizes the control measures to the production execution unit, obtaining corresponding execution data feedback. Based on this feedback, a status monitoring module tracks the dynamic strategies in the production execution process in real time. If inconsistencies are detected between the execution data and the strategy records, a data calibration process is initiated to obtain the calibrated execution status. Using the calibrated execution status, an information integration tool compares and updates the dynamic strategy records with the management database, determining the final database update result. Based on the database update result, a log recording module stores the entire synchronous update process data into the field management database, obtaining a complete strategy execution archive.

[0117] In practice, when using pre-established mapping tools to link micronutrient ratio adjustments with production execution, one can begin by understanding the basic principles of these tools. The core of a mapping tool lies in matching micronutrient ratio parameters with actual field production needs through data correlation, forming a dynamic, synchronously updated framework. For example, if a plot of land needs to adjust the zinc ratio during the mid-growth stage of the crop, the mapping tool will adjust the zinc fertilizer ratio from 2% to 4% based on soil testing data and crop requirements, and correlate this with the fertilization time and method, ensuring that the adjusted parameters directly guide production execution.

[0118] The process of matching the adjusted nitrogen fertilizer ratio with the control measures can be further illustrated through a specific scenario. Suppose the adjusted nitrogen fertilizer ratio is 25%, but comparison reveals that the actual nitrogen fertilizer application rate in the control measures is too high, with a data deviation reaching 8%, exceeding the preset 5% threshold. This triggers a correction process. The correction process will consider soil moisture and recent rainfall, recommending a reduction in the nitrogen fertilizer ratio to 23%, and suggesting application in multiple stages to reduce the risk of runoff.

[0119] When using data distribution tools to synchronize regulatory measures to production execution units, the accuracy of instruction issuance should be emphasized. For example, if the phosphate fertilizer application rate for a specific plot is adjusted to 18%, the data distribution tool will simultaneously send this parameter and the application time (6:00 AM) to the field equipment and collect feedback data after execution, such as whether the actual application amount and time are consistent. This method ensures the effective transmission of instructions.

[0120] Real-time tracking of dynamic strategies in the production execution process can be demonstrated through specific case studies. For example, suppose a plot of land experiences fertilizer loss due to sudden rainfall after fertilization. If inconsistencies are detected between the execution data and the strategy records, a data calibration process is immediately initiated, recommending an additional 3% of the planned fertilizer application and adjusting the application time to adapt to the weather change.

[0121] When using information integration tools to compare and update dynamic strategy records with the management database, it's crucial to ensure data consistency. For example, if the adjusted potash fertilizer ratio is 15%, the integration tool will compare this data with historical records in the database. If discrepancies are found, the database will be updated to ensure the accuracy of subsequent management decisions.

[0122] The process of synchronously updating and storing the entire process data in the field management database can be illustrated through specific operations. Suppose that for a certain plot, the entire process data from ratio adjustment to implementation feedback, including changes in trace element ratios, application time, and environmental conditions, is archived by the log recording module, forming a complete strategy execution file for easy subsequent traceability and analysis. This method provides reliable data support for field management.

[0123] Step S107: Based on the execution record and combined with the yield prediction accuracy-related model, analyze the impact of the dynamic control strategy on the final yield, obtain the deviation value between the predicted result and the actual yield, and obtain an evaluation report on the data analysis accuracy. Based on the evaluation report on the data analysis accuracy, adjust the model parameters of sensitivity analysis and dynamic control strategy, optimize the collection frequency of plant status monitoring and environmental factor sensitive data, and determine the improved analysis framework.

[0124] Historical data is retrieved from execution records, and information related to yield prediction and dynamic regulation is categorized and organized. A structured storage method is used to store the categorized data hierarchically, resulting in a pre-processed dataset. Based on this dataset, yield prediction results are calculated using an accuracy model, and a comparison between the predicted and actual yields is obtained. The magnitude and distribution characteristics of the deviation values ​​are determined using a deviation value calculation method. Based on the distribution characteristics of the deviation values, the impact of the dynamic regulation strategy on the final yield is analyzed. A regression analysis model is used to quantify the influencing factors, obtaining key indicators of the impact. For these key indicators, the parameter settings for sensitivity analysis are adjusted. If the key indicators exceed the preset threshold range, the model parameters of the dynamic regulation strategy are optimized, resulting in an adjusted parameter configuration. Based on the adjusted parameter configuration, the collection frequency of plant status monitoring is optimized. If the matching degree between the collection frequency and the sensitive data of environmental factors is lower than the predetermined standard, a new collection frequency scheme is determined by automatically adjusting the collection interval. Using the new collection frequency scheme, the input conditions for data analysis are updated. Combined with the updated execution records and environmental data, the prediction results are recalculated using a regression analysis model, resulting in optimized yield prediction data. Based on the optimized production forecast data, the improvement effect of the dynamic control strategy is verified. If the deviation value is still higher than the preset threshold range, the parameters and sampling frequency of the sensitivity analysis are adjusted cyclically to determine the final analysis framework configuration.

[0125] Feasible practice is to first categorize and organize the historical data in the execution records. For example, suppose the yield forecasting data involved in field management includes planting records from the past three years, recording the micronutrient ratios, irrigation frequency, and final yield for different crops each year. This data can be categorized by crop type and control measures, forming a structured storage method. For instance, rice and wheat data could be stored in layers for easier subsequent analysis. This classification method helps to quickly extract the historical control effects of specific crops, providing a clear data foundation for subsequent forecasts.

[0126] For example, the calculation of yield forecasts can be analyzed using a precision model. Suppose that historical data predicts this year's crop yield to be 5000 kg per hectare, while the actual yield is 4800 kg. The deviation calculation method reveals a 4% deviation. Further analysis of the deviation distribution characteristics shows that the deviation is mainly concentrated in inappropriate control measures during certain specific months, such as insufficient irrigation. This analytical approach helps identify weaknesses in dynamic control strategies, providing a basis for subsequent optimization.

[0127] When analyzing the impact of dynamic control strategies on yield, regression analysis models can be used to quantify key indicators. Suppose that model analysis reveals that changes in the content of a certain trace element have the greatest impact on yield, with an impact weight of 30%. Based on this, this element can be used as a key indicator, and the parameter settings for sensitivity analysis can be adjusted. If this indicator exceeds a preset threshold, such as a content deviation exceeding 5%, the model parameters of the control strategy need to be optimized, for example, by increasing the application rate of this element or adjusting the application timing.

[0128] When optimizing the sampling frequency for plant status monitoring, adjustments can be made based on the matching degree of sensitive data related to environmental factors. For example, if the current sampling frequency is once a day, but it is found that changes in temperature and humidity have a significant impact on plant status, and the daily frequency cannot capture short-term abnormal fluctuations, then if the matching degree is lower than a predetermined standard, such as below 80%, the sampling interval can be automatically adjusted to once every 6 hours, forming a new sampling frequency scheme. This adjustment can more promptly reflect the impact of environmental changes on plants.

[0129] After updating the input conditions for data analysis, the prediction results can be recalculated by combining the new execution records and environmental data. Assuming that more detailed temperature and humidity data are obtained through the adjusted collection frequency, and the yield is re-predicted using a regression analysis model, the deviation between the predicted and actual yields decreased from 4% to 2%. This indicates that the optimized data input conditions significantly improve the accuracy of the predictions.

[0130] When verifying the effectiveness of the dynamic control strategy, if the deviation value still exceeds the preset threshold, such as 1.5%, it is necessary to iteratively adjust the sensitivity analysis parameters and the sampling frequency. Assuming that through multiple iterations, the sampling frequency is adjusted to once every 4 hours, and the trace element ratio parameters are optimized, the deviation value can eventually be controlled within 1%. This iterative adjustment method can gradually approach the optimal configuration, forming the final analytical framework. The advantage of this method is that it allows for continuous improvement of the strategy, ensuring that production forecasting and control measures are more closely aligned with actual production needs.

[0131] Example 2

[0132] This embodiment also discloses a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in Embodiment 1.

[0133] Example 3

[0134] This embodiment also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method described in Embodiment 1.

[0135] Example 4

[0136] This embodiment also discloses a computer program product, including a computer program that, when executed by a processor, implements the steps of the method described in Embodiment 1.

[0137] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An integrated method for wheat growth regulation and yield prediction based on artificial intelligence, characterized in that, Includes the following steps: Based on sensor networks, data on wheat growth stage differences and environmental factor-sensitive data are obtained from the field to obtain an initial dataset. After preprocessing, a standardized dataset is obtained. For a standardized dataset, machine learning methods were used to model and analyze the correlation between differences in wheat growth stages and sensitivity to environmental factors, obtain the response patterns of each stage to changes in climate conditions and differences in soil properties, and obtain the sensitivity analysis results. Based on the results of sensitivity analysis and plant status monitoring data, we can identify the characteristics of changes in the nutritional requirements of wheat at the current growth stage and determine the direction for adjusting the micronutrient ratio. Based on the adjustment direction of the micronutrient ratio and combined with dynamic control strategies, personalized nutrient supply plans for different plots and growth stages are generated to obtain fertilization ratio parameters. Based on the fertilization ratio parameters and the resource utilization efficiency assessment model, the applicability of the current scheme under changes in climate conditions and differences in soil characteristics is analyzed. If the resource utilization rate is lower than the preset threshold, the ratio parameters are readjusted and the optimized control scheme is determined. By optimizing the control scheme, updating the dynamic control strategy database in the field management system, and applying the adjusted micronutrient ratios and control measures to actual production in a timely manner, wheat yield prediction results are obtained by combining the wheat yield prediction model. The process involves adjusting the direction based on the micronutrient ratio, combined with a dynamic control strategy, to generate personalized nutrient supply plans for different plots and growth stages, resulting in fertilization ratio parameters, including: Based on the adjustment direction of trace element ratios and combined with dynamic control strategies, personalized nutrient supply plans are generated for different plots and growth stages. Based on the differences in personalized nutrient supply needs of different plots and growth stages, data integration tools are used to perform hierarchical analysis of trace element demand distribution, and a personalized plan framework for different plots at the current growth stage is obtained. For the personalized solution framework, a dynamic control strategy is integrated to compare the real-time collected environmental fluctuation data. If the environmental fluctuation data exceeds the preset threshold, the micronutrient ratio is adjusted to determine the preliminary fertilization ratio parameters. By using the control measures generation tool, key adjustment points can be extracted from the initial fertilizer ratio parameters, and specific implementation plans for targeted control measures can be obtained by combining timely instructions and needs. Based on the specific implementation plan, the fertilizer ratio parameters are refined and adjusted using data mapping to obtain the final nutrient supply plan applicable to different plots. For the final nutrition supply plan, timely adjustment instructions are generated through instruction distribution to determine the execution time and allocation method of ratio parameters for each region; The optimized control scheme updates the dynamic control strategy database in the field management system, and the adjusted micronutrient ratios and control measures are applied to actual production in a timely manner, including: By using the optimized control scheme and pre-established mapping tools, the adjustment of trace element ratios is linked to the production execution process, resulting in a preliminary synchronous update framework. Based on the preliminary synchronous update framework, the ratio adjustment data is matched with the control measures. If the matching result shows that the data deviation exceeds the preset threshold, the correction process is triggered to determine the adjusted ratio parameters. By using the adjusted proportioning parameters and data distribution tools, the control measures are synchronized to the production execution unit to obtain corresponding execution data feedback. Based on the feedback of execution data, the dynamic strategy of the production execution process is tracked in real time. If the execution data is found to be inconsistent with the strategy record, the data calibration process is initiated to obtain the calibrated execution status. By comparing and updating the dynamic policy records with the management database using the calibrated execution status and information integration tools, the final database update result is determined. Based on the database update results, a log recording module is used to store the entire process of synchronous updates into the field management database to obtain complete strategy execution data; After obtaining the wheat yield prediction results by combining the wheat yield prediction model, the following steps are also included: Obtain comparative data between the predicted results and the actual output, and determine the magnitude and distribution characteristics of the deviation value through the deviation value calculation method; Based on the distribution characteristics of the deviation values, the impact of dynamic control strategies on the final output is analyzed. A regression analysis model is used to quantify the influencing factors and obtain key indicators of the degree of impact. For key indicators of impact, adjust the parameter settings of sensitivity analysis. If the key indicators exceed the preset threshold range, optimize the model parameters of the dynamic control strategy and obtain the adjusted parameter configuration. Based on the adjusted parameter configuration, optimize the collection frequency of plant status monitoring. If the matching degree between the collection frequency and the sensitive data of environmental factors is lower than the predetermined standard, determine a new collection frequency scheme by automatically adjusting the collection interval. By adopting a new data acquisition frequency scheme and updating the input conditions for data analysis, combined with the updated execution records and environmental data, a regression analysis model is used to recalculate the prediction results, resulting in optimized production prediction data.

2. The method according to claim 1, characterized in that, The method involves acquiring wheat growth stage variation data and environmental factor-sensitive data from the field using a sensor network to obtain an initial dataset. After preprocessing, a standardized dataset is obtained, including: By using a sensor network to monitor different plots of land in real time, data on differences in wheat growth stages and sensitive data on environmental factors are obtained and stored as an initial dataset. Based on the initial dataset, data cleaning, data normalization, and linear interpolation are performed to obtain a standardized dataset.

3. The method according to claim 1, characterized in that, For the standardized dataset, machine learning methods were used to model and analyze the correlation between differences in wheat growth stages and sensitivity to environmental factors. The response patterns at each stage to changes in climate conditions and soil properties were obtained, yielding sensitivity analysis results, including: For a standardized dataset, the support vector machine method was used to preliminarily model the correlation between wheat growth stages and sensitivity to environmental factors, analyze the characteristic performance of different stages, and obtain preliminary correlation model results. Based on the preliminary correlation model results, corresponding stage-specific feature data were extracted for climate condition changes and soil property differences. The feature data were divided into several subsets using a data stratification processing method to determine the stratified feature subsets. For the stratified feature subsets, clustering method is used to group the differences in wheat growth stages for analysis, obtain the correspondence between the features of each group and the sensitivity to environmental factors, and obtain the feature mapping results after grouping. The grouped feature mapping results are filtered to obtain the core feature set. For the core feature set after screening, the soil property differences and stage characteristics are jointly analyzed by data integration tools, and multi-dimensional information is integrated to determine the integrated comprehensive feature dataset. Based on the integrated feature dataset, the linear regression method was used to quantitatively analyze the response pattern of wheat growth stage differences and sensitivity to environmental factors, and the final sensitivity analysis results were obtained.

4. The method according to claim 1, characterized in that, The process involves using sensitivity analysis results, combined with plant status monitoring data, to identify the characteristics of changes in wheat's nutritional requirements at the current growth stage and determine the direction for adjusting the micronutrient ratio, including: Based on the sensitivity analysis results, plant status monitoring data were integrated to analyze the characteristics of changes in nutrient requirements of wheat at the current growth stage and to obtain the specific patterns of demand changes. Based on the specific patterns of demand changes, data processing tools are used to classify and stratify the characteristics of change, and to determine the degree of deviation of nutritional requirements under different categories. If the deviation of nutrient requirements exceeds the preset threshold, a preliminary judgment result on the cause of the deviation will be obtained based on the plant status data at the current stage. Based on the preliminary judgment results, the support vector machine method was used to further classify and analyze the causes of deviation and identify the core factors leading to the deviation. Based on the classification results of the core factors, a preliminary plan for adjusting the proportion of trace elements is generated using data mapping tools to obtain a reference basis for the direction of adjustment; If the reference basis for adjusting the direction conflicts with the data of historical control mechanisms, a second verification will be conducted through data comparison to determine the final direction of the allocation adjustment.

5. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1-4.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1-4.

7. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1-4.

Images