Fitting analysis method for matching of potash magnesium sulphate fertilizer nutrient release and crop demand

Abstract

The invention discloses a potash magnesium sulphate fertilizer nutrient release and crop demand matching fitting analysis method, and relates to the technical field of intelligent agriculture, and the method comprises the following steps: S1, collecting fertilizer multi-condition nutrient release and crop whole growth period nutrient demand data, and carrying out standardization pretreatment; s2, constructing a fertilizer nutrient release dynamic curve by using a nonlinear fitting method; s3, performing segmented fitting to construct a crop nutrient demand dynamic curve; s4, constructing a dynamic fitting analysis model integrating multiple modules; s5, calculating a curve matching degree and identifying time offset and intensity difference; s6, establishing evaluation system rating and generating fertilization optimization suggestions; and S7, verifying and iteratively optimizing the model in combination with actual crop data. According to the method, complete-cycle accurate matching analysis of fertilizer nutrient release and crop requirements is realized, data acquisition is comprehensive, curve fitting is accurate, analysis results are refined, suitability can be quantitatively evaluated, practical fertilization suggestions can be generated, the model can be continuously optimized, and scientific data support is provided for accurate fertilization.

Description

Technical Field

[0001] This invention relates to the field of smart agriculture technology, and in particular to a fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand. Background Technology

[0002] Potassium magnesium sulfate fertilizer, as a highly efficient multi-element compound potassium fertilizer, contains the three essential nutrients for crops: potassium, magnesium, and sulfur. Its fertilizer effect is mild and nutritionally balanced, supplementing the nutrients needed for crop growth, improving photosynthetic efficiency, yield, and quality, while also improving soil physicochemical properties and alleviating soil nutrient deficiencies and compaction. It is increasingly widely used in the large-scale cultivation of grains, fruits, vegetables, and cash crops, and is one of the core fertilizers in modern agricultural precision fertilization systems. Achieving precise matching between the nutrient release rhythm of potassium magnesium sulfate fertilizer and the nutrient demand rhythm of crops throughout their entire growth cycle is key to maximizing fertilizer efficiency, improving fertilizer utilization, and reducing agricultural production costs. It is also an important research direction in the fields of digital agriculture and smart planting, and a crucial measure to promote green agricultural development and reduce non-point source pollution.

[0003] Currently, the application of potassium magnesium sulfate fertilizer in agricultural production still largely relies on growers' experience and judgment, resulting in a common problem of mismatch between fertilizer application amount, timing, and method and the actual nutrient requirements of crops. In some scenarios, the fertilizer releases nutrients too quickly, easily causing seedling burn and nutrient loss through leaching; while in other scenarios, the release is too slow, leading to insufficient nutrient supply during key growth stages such as flowering and grain filling, directly affecting crop growth and yield. This experience-based fertilization method not only significantly reduces the utilization efficiency of potassium magnesium sulfate fertilizer and increases planting costs, but also easily leads to ecological problems such as soil physicochemical deterioration and water eutrophication due to nutrient loss, which contradicts the requirements of modern agriculture's precision and green development.

[0004] Existing analytical methods for matching fertilizer nutrients with crop requirements, while incorporating data fitting and simple algorithmic analysis, still suffer from numerous technical shortcomings. Most methods focus only on static matching analyses of single nutrient amounts or single crop growth stages, failing to fully capture the dynamic patterns of the synergistic release of potassium, magnesium, and sulfur in potassium sulfate fertilizers. They also neglect the stage-specific and differentiated characteristics of nutrient requirements throughout the crop's entire growth cycle, making it difficult to achieve precise matching across both time and intensity dimensions. Furthermore, existing methods do not adequately consider the impact of key application conditions such as soil type, soil temperature and humidity, and fertilization depth. The fitting analysis models exhibit weak scenario adaptability and generalization ability, and lack a comprehensive adaptability evaluation index system. Consequently, the analytical results are mostly qualitative, making it difficult to directly translate them into field-operable fertilization optimization strategies.

[0005] With the acceleration of agricultural digital transformation, precision fertilization technology based on big data and algorithm models has become a development trend. Utilizing digital means to construct dynamic fitting analysis models to achieve full-cycle, multi-dimensional matching analysis of potassium magnesium sulfate fertilizer nutrient release and crop needs has become an urgent need to address the shortcomings of traditional fertilization. However, current related technologies still suffer from problems such as single data collection dimensions, insufficient curve fitting accuracy, lack of comprehensive matching degree calculation, and the absence of model verification and iteration mechanisms, which restrict their practical application in agricultural production. Therefore, developing an analytical method that can achieve multi-dimensional data fusion, dynamic curve fitting, accurate matching degree calculation, and scientific adaptability assessment is of significant practical importance for promoting the scientific application of potassium magnesium sulfate fertilizer and constructing a modern agricultural precision fertilization system. Summary of the Invention

[0006] This invention proposes a fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand, in order to solve the problems mentioned in the prior art.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop requirements, comprising the following steps: S1 multi-dimensional data acquisition and preprocessing: collect nutrient release data of potassium magnesium sulfate fertilizer under different application conditions; simultaneously collect nutrient requirement data of the target crop throughout its entire growth period; and perform outlier removal, missing value completion, and standardization processing on the collected data. S2 nutrient release curve construction: Based on the preprocessed nutrient release data, a nonlinear fitting algorithm is used to construct the dynamic nutrient release curve of potassium magnesium sulfate fertilizer. The curve parameters are optimized by the least squares method, and the release curve equations and characteristic parameters of each element are output. S3 Crop Demand Curve Construction: Based on the nutrient demand data of the target crop and the duration of the crop growth stages, a dynamic curve of nutrient demand throughout the entire growth period is constructed. A segmented fitting method is used to fit curve segments according to the demand characteristics of different growth stages, and the demand curve equations and characteristic parameters of each element are output. The S4 dynamic fitting analysis model is constructed using nutrient release curves and crop demand curves as core inputs. The model integrates three major functional modules: time alignment, intensity matching, and deviation analysis. S5 matching degree calculation and difference identification: The matching degree between the nutrient release curve and the crop demand curve is calculated through a dynamic fitting analysis model; the time offset and intensity difference between release and demand are identified. S6 compatibility assessment and optimization suggestions are generated. Based on the matching degree results, time offset and intensity difference value, a compatibility assessment index system is established, and the compatibility is divided into four levels: excellent, good, average and poor according to the scoring results. S7 model validation and iterative optimization involves collecting crop growth index data and yield data from actual application scenarios, using these as validation indicators; based on the validation results, the gradient descent algorithm is used to iteratively optimize the model parameters, and the curve fitting algorithm and matching degree calculation logic are updated.

[0008] Furthermore, it also includes a precise calculation step for the overall matching degree, executed in S5, which quantifies the overall adaptability to the requirements through a formula. The specific formula is as follows: For overall matching degree, As a time node, This represents the total number of time points. for Nutrient release rate at any time for Real-time crop demand for Time-weighted coefficients.

[0009] Furthermore, it also includes a sensitivity analysis step for application conditions, which is performed after S1 and before S2. The influence of each application condition on the nutrient release curve is analyzed by controlling the variable method. Gradient change values ​​for soil type, soil moisture, soil temperature, fertilizer application amount, and fertilizer application depth are set. The rate of change of characteristic parameters of nutrient release curve when a single condition changes is calculated, and the key influencing factors are determined by ranking them according to the magnitude of the rate of change.

[0010] Furthermore, in the S1 multi-dimensional data acquisition and preprocessing, nutrient release data is monitored in real time by soil nutrient sensors, with the monitoring period covering the entire period from the application of potassium magnesium sulfate fertilizer to crop maturity; crop demand data is obtained through a combination of field experiments and literature surveys; outlier removal adopts the Grubbs criterion, missing value completion adopts the linear interpolation method, and standardization processing adopts the Z-score standardization method.

[0011] Furthermore, in the construction of the S2 nutrient release curve, the Logistic growth curve algorithm was selected for nonlinear fitting, and the curve equation was in the form of an S-shaped curve. During the optimization of curve parameters, feature parameters were extracted by calculating the inflection points and extreme points of the curve through differentiation.

[0012] Furthermore, in the construction of the S3 crop demand curve, the piecewise fitting uses appropriate fitting algorithms for different growth stages: linear fitting algorithm for the seedling and maturity stages, quadratic polynomial fitting algorithm for the jointing and flowering stages, and exponential fitting algorithm for the grain-filling stage; each curve segment is processed through smooth transition, and the characteristic parameters are obtained by solving curve integrals and extrema.

[0013] Furthermore, in the construction of the S4 dynamic fitting analysis model, the time alignment module adopts a dynamic time warping algorithm to synchronize the time sequence of two curves by stretching or compressing the time axis; the intensity matching module uses a combination of correlation analysis and amplitude ratio analysis to calculate the Pearson correlation coefficient and amplitude ratio coefficient between the curves; and the deviation analysis module calculates the time offset and intensity difference value for each time period using the sliding window method.

[0014] Furthermore, in the S5 matching degree calculation and difference identification, the local matching degree calculation is performed separately for each reproductive stage, and the matching degree score of each element in each stage is output; a positive value indicates that the nutrient release peak is later than the demand peak, and a negative value indicates that the nutrient release peak is earlier than the demand peak; the intensity difference value calculation includes absolute difference value and relative difference value. The absolute difference value is the difference between the release amount and the demand amount, and the relative difference value is the ratio of the absolute difference value to the demand amount.

[0015] Furthermore, in the S6 adaptability assessment and optimization suggestion generation, the adaptability assessment index system uses the analytic hierarchy process (AHP) to determine the weight of each index. When generating optimization suggestions, for cases where the absolute value of the time offset is ≥3 days, if the time offset is positive, the fertilization time is adjusted to be earlier or fast-release potassium magnesium sulfate fertilizer is selected; if the time offset is negative, the fertilization time is adjusted to be later or slow-release potassium magnesium sulfate fertilizer is selected. For cases where the absolute value of the relative difference in intensity is ≥20%, the amount of fertilizer is adjusted. For cases where soil conditions are sensitive, the fertilization depth or soil improvement measures are optimized.

[0016] Furthermore, in the S7 model validation and iterative optimization, the evaluation of validation indicators adopts the comparative experiment method, setting up an optimization group and a control group. The optimization group adopts the application strategy generated by the model, while the control group adopts the conventional application strategy. The model effect is verified by comparing the differences in crop growth indicators and yield data between the two groups. During model iterative optimization, parameters are updated once every 10 sets of new experimental data are collected.

[0017] Compared with existing technologies, the beneficial effects of this invention are: The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand in this invention relies on multi-dimensional data collection, dynamic curve construction, and intelligent model fitting to achieve full-cycle, multi-dimensional, and accurate analysis of fertilizer nutrient release and crop demand. It comprehensively makes up for the technical shortcomings of existing analysis methods and provides digital and quantitative data analysis basis for the scientific application and strategy optimization of potassium magnesium sulfate fertilizer. It has significant advantages in improving the scientific nature of fertilization, maximizing fertilizer efficiency, and reducing planting costs, which is in line with the development trend of digital agriculture and smart planting.

[0018] First, multi-dimensional and comprehensive data collection and standardized preprocessing were carried out, covering the release data of potassium, magnesium, and sulfur in potassium sulfate and magnesium fertilizers, nutrient requirements at each stage of crop growth, and key application conditions such as soil type, temperature, and humidity. Combined with professional methods, outlier removal, missing value completion, and dimensional unification were completed, ensuring the accuracy and completeness of the basic data for analysis. This solved the problems of single data dimensions and poor data quality in existing methods, laying a reliable data foundation for subsequent curve construction and model fitting.

[0019] Secondly, dynamic curves are constructed using appropriate fitting algorithms to address the different characteristics of nutrient release and crop demand. For nutrient release, a nonlinear algorithm is used to accurately capture its slow-fast-gradual phased release pattern. For crop demand, a segmented fitting method is used to match the differences in demand at different growth stages. At the same time, the curve feature parameters are accurately extracted, which can truly reflect the temporal changes in release and demand, and greatly improve the accuracy and fit of curve fitting.

[0020] Furthermore, the dynamic fitting analysis model integrates three core functions: time alignment, intensity matching, and deviation analysis. It realizes the time synchronization of release and demand curves and the calculation of multi-dimensional matching degree. It can not only obtain the overall matching degree, but also realize the local matching degree analysis of elements and reproductive stages. At the same time, it accurately identifies the time offset and intensity difference between the two, making the matching analysis results more comprehensive and refined, and solving the core problems of static matching and one-sided analysis in existing methods.

[0021] Meanwhile, this invention constructs a comprehensive suitability assessment index system, scientifically determines the weight of each index using the analytic hierarchy process, achieves quantitative suitability level classification, and can accurately analyze the causes of differences in cases of poor suitability, generating specific and actionable fertilization optimization suggestions such as adjusting fertilizer application amount, optimizing fertilization time, and improving fertilization methods. The analysis results are directly transformed into field fertilization strategies, greatly improving the practicality and guidance of the method, and providing a clear direction for optimizing the application of potassium magnesium sulfate fertilizer in different soil environments and different crop varieties.

[0022] Finally, the design of the model validation and iterative optimization mechanism, combined with actual crop growth and yield data from field planting, verifies the model's effectiveness. The model parameters are continuously optimized using the gradient descent algorithm, constantly improving the model's generalization ability and prediction accuracy, thus adapting to the analytical needs of different planting scenarios. This method relies entirely on digital algorithms and models, which not only improves the utilization efficiency of potassium magnesium sulfate fertilizer and reduces planting costs, but also reduces nutrient loss and agricultural non-point source pollution, possessing significant economic and ecological value and contributing to the construction of a modern precision fertilization system. Attached Figure Description

[0023] Figure 1 The present invention proposes Figure 1A schematic diagram illustrating the matching of nutrient release from potassium magnesium sulfate fertilizer with crop requirements; Figure 2 A bar chart showing the matching degree between nutrient release and demand of potassium magnesium sulfate fertilizer at different crop growth stages; Figure 3 Line graphs showing the peak release time of potassium magnesium sulfate fertilizer under different application conditions; Figure 4 Line graph showing the crop yield increase rate under different fitness levels. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0026] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. The invention will now be described in further detail with reference to the accompanying drawings.

[0027] Reference Figures 1 to 4 A fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop requirements includes the following steps: S1 multi-dimensional data acquisition and preprocessing collects nutrient release data of potassium magnesium sulfate fertilizer under different application conditions, including the release amount and rate of potassium, magnesium, and sulfur. Application conditions cover soil type, soil moisture, soil temperature, fertilizer application amount, and fertilizer application depth. Simultaneously, it collects nutrient requirement data of the target crop throughout its entire growth period, including the demand for potassium, magnesium, and sulfur at each growth stage and the demand rate. The growth stages are divided into sowing period, seedling stage, jointing stage, flowering stage, grain filling stage, and maturity stage. The collected data undergoes outlier removal, missing value completion, and standardization to unify data units and time scales, forming a standardized dataset.

[0028] S2 Nutrient Release Curve Construction: Based on preprocessed nutrient release data, a nonlinear fitting algorithm is used to construct the dynamic nutrient release curve of potassium magnesium sulfate fertilizer. The curve is plotted with time on the x-axis and nutrient release amount or release rate on the y-axis. The curve parameters are optimized using the least squares method to ensure that the curve can accurately reflect the temporal pattern of nutrient release under different application conditions. The release curve equations and characteristic parameters of each element are output, including maximum release amount, peak time, and release period.

[0029] The S3 crop demand curve construction method constructs a dynamic curve of nutrient demand throughout the entire growth period based on the nutrient demand data of the target crop and the duration of the crop growth stages. The curve is plotted with time on the horizontal axis and nutrient demand amount or demand rate on the vertical axis. A segmented fitting method is used to fit curve segments according to the demand characteristics of different growth stages, and outputs the demand curve equation and characteristic parameters of each element. The characteristic parameters include the maximum demand amount, peak demand time, and peak demand period.

[0030] The S4 dynamic fitting analysis model is constructed using the nutrient release curve and crop demand curve as core inputs. The model integrates three major functional modules: time alignment, intensity matching, and deviation analysis. The time alignment module achieves temporal synchronization of the two curves through time axis calibration. The intensity matching module calculates the amplitude fit between the curves. The deviation analysis module identifies the differences between release and demand in the time and intensity dimensions.

[0031] S5 matching degree calculation and difference identification calculates the matching degree between the nutrient release curve and the crop demand curve through a dynamic fitting analysis model. The matching degree calculation covers the overall matching degree and the local matching degree of each element and each growth stage. It identifies the time offset and intensity difference value between release and demand. The time offset refers to the time difference between the peak of nutrient release and the peak of crop demand, and the intensity difference value refers to the absolute difference or relative difference between the nutrient release amount and the crop demand amount at each time period.

[0032] The S6 compatibility assessment and optimization suggestion generation system establishes a compatibility assessment index system based on matching results, time offset, and intensity difference values. The indexes include overall compatibility score, element compatibility score, and growth stage compatibility score. The compatibility is divided into four levels: excellent, good, average, and poor, according to the score results. For cases with poor compatibility, the causes of the difference are analyzed, and specific strategy suggestions for adjusting fertilizer application rate, optimizing fertilizer application time, and improving fertilizer application method are generated in combination with application conditions and crop requirements. This provides data support for optimizing potassium magnesium sulfate fertilizer application schemes.

[0033] S7 model validation and iterative optimization involves collecting crop growth index data and yield data under actual application scenarios, including plant height, leaf area, biomass, seed setting rate, and yield, which are used as validation indicators to evaluate the predictive accuracy and reliability of the dynamic fitting analysis model. Based on the validation results, the gradient descent algorithm is used to iteratively optimize the model parameters, update the curve fitting algorithm and matching degree calculation logic, and improve the model's generalization ability under different crop varieties and soil conditions.

[0034] This invention also includes a step for accurately calculating the overall matching degree, which is executed in S5. The formula quantifies the overall suitability with the requirements, and the specific formula is as follows: This represents the overall match score, ranging from 0 to 1. A higher value indicates a higher match score. As a time node, This represents the total number of time points. for Nutrient release rate at any time for Real-time crop demand for The time-weighted coefficient is assigned a value based on the importance of the crop growth stage. The weighted coefficients for the grain-filling and flowering stages are higher than those for other stages. This calculation enables a comprehensive matching degree assessment of multiple time periods and multiple factors.

[0035] This invention also includes a sensitivity analysis step for application conditions, performed after S1 and before S2. The influence of each application condition on the nutrient release curve is analyzed using the controlled variable method. Gradient change values ​​for soil type, soil moisture, soil temperature, fertilizer application amount, and fertilizer application depth are set. The rate of change of characteristic parameters of the nutrient release curve when a single condition changes are calculated. Key influencing factors are determined by ranking them according to the magnitude of the rate of change. These key influencing factors are the focus of adjustment for application strategy optimization, thereby improving the pertinence and operability of the optimization suggestions.

[0036] In this invention, during the S1 multi-dimensional data acquisition and preprocessing, nutrient release data is monitored in real time by a soil nutrient sensor at a frequency of 24 hours / time, covering the entire period from the application of potassium magnesium sulfate fertilizer to crop maturity. Crop demand data is obtained through a combination of field trials and literature review. Three parallel field trials are set up, and the average value of the data is taken as the final demand data. Outlier removal adopts the Grubbs criterion, missing value completion adopts the linear interpolation method, and standardization processing adopts the Z-score standardization method.

[0037] In this invention, the nonlinear fitting algorithm used in the construction of the S2 nutrient release curve is the Logistic growth curve algorithm, and the curve equation is an S-shaped curve, which can accurately fit the three stages of slow, fast and flat nutrient release. In the process of curve parameter optimization, the upper limit of the number of iterations is set to 1000 times and the convergence threshold is 0.001 to ensure the accuracy of curve fitting. Feature parameters are extracted by calculating the inflection point and extreme point of the curve by differentiation.

[0038] In this invention, the construction of the S3 crop demand curve involves segmented fitting, with appropriate fitting algorithms selected for different growth stages. Linear fitting algorithms are used for the seedling and maturity stages, quadratic polynomial fitting algorithms are used for the jointing and flowering stages, and exponential fitting algorithms are used for the grain-filling stage. Each curve segment undergoes smooth transition processing to ensure the continuity of the entire demand curve, and the characteristic parameters are obtained through curve integration and extreme value solving.

[0039] In this invention, in the construction of the S4 dynamic fitting analysis model, the time alignment module adopts a dynamic time warping algorithm to achieve temporal synchronization of two curves by stretching or compressing the time axis, with the synchronization error controlled within 24 hours; the intensity matching module adopts a combination of correlation analysis and amplitude ratio analysis to calculate the Pearson correlation coefficient and amplitude ratio coefficient between the curves; the deviation analysis module calculates the time offset and intensity difference value for each time period using the sliding window method.

[0040] In this invention, in the S5 matching degree calculation and difference identification, the local matching degree calculation is performed separately for each reproductive stage, and the matching degree score of each element in each stage is output; the time offset calculation is accurate to the day, with a positive value indicating that the nutrient release peak is later than the demand peak, and a negative value indicating that the nutrient release peak is earlier than the demand peak; the intensity difference value calculation includes absolute difference value and relative difference value, with the absolute difference value being the difference between the release amount and the demand amount, and the relative difference value being the ratio of the absolute difference value to the demand amount.

[0041] In this invention, during the S6 compatibility assessment and optimization suggestion generation, the compatibility assessment index system uses the analytic hierarchy process (AHP) to determine the weight of each index. The overall compatibility score accounts for 40% of the weight, the element compatibility score accounts for 30%, and the growth stage compatibility score accounts for 30%. When generating optimization suggestions, for cases where the absolute value of the time offset is ≥3 days, if the time offset is positive, the fertilization time is adjusted to be earlier or fast-release potassium magnesium sulfate fertilizer is selected; if the time offset is negative, the fertilization time is adjusted to be later or slow-release potassium magnesium sulfate fertilizer is selected. For cases where the absolute value of the relative difference in intensity is ≥20%, the fertilization amount is adjusted. For cases where soil conditions are sensitive, the fertilization depth or soil improvement measures are optimized.

[0042] In this invention, during the S7 model validation and iterative optimization, the evaluation of validation indicators adopts a comparative experiment method. An optimization group and a control group are set up. The optimization group adopts the application strategy generated by the model, while the control group adopts the conventional application strategy. The model effect is verified by comparing the differences in crop growth indicators and yield data between the two groups. During model iterative optimization, parameters are updated once every 10 sets of new experimental data are collected to continuously improve the model's adaptability and prediction accuracy in different scenarios.

[0043] The following two examples further illustrate the specific implementation of this system: Example 1: Application of winter wheat cultivation in alluvial soil in the North China Plain This embodiment is applied to the field planting scenario of winter wheat in the alluvial soil region of the North China Plain. The soil moisture in this region is 18%-22%, the annual temperature of the cultivated layer is 10-28℃, and the entire growth period of winter wheat is about 230 days, which is divided into six stages: sowing period, seedling period, overwintering period, jointing period, grain filling period, and maturity period. The goal of this method is to achieve precise matching of potassium, magnesium and sulfur nutrient release of potassium sulfate fertilizer with the nutrient requirements of winter wheat at each growth stage, thereby optimizing the fertilization strategy.

[0044] S1 multi-dimensional data acquisition and preprocessing involved deploying soil nutrient sensors and environmental monitoring sensors in the experimental field to monitor the release amount and rate of potassium, magnesium, and sulfur from potassium magnesium sulfate fertilizer at an application rate of 20 kg / mu and a fertilization depth of 15 cm in real time. Monitoring was conducted 24 hours a day, covering the period from fertilization to the maturity of winter wheat. Simultaneously, nutrient requirements data for each growth stage of winter wheat were obtained through field trials and literature review. Three parallel field trials were set up to determine the demand and rate of potassium, magnesium, and sulfur at each stage. Outliers were removed from the collected data using the Grubbs criterion, missing values ​​were filled using linear interpolation, and Z-score standardization was used to unify the data units and time scale, forming a standardized dataset.

[0045] S2 nutrient release curve construction: Based on preprocessed nutrient release data, a nonlinear dynamic release curve is constructed using the Logistic growth curve algorithm. With growth time as the x-axis and nutrient release rate as the y-axis, the curve parameters are optimized using the least squares method. The upper limit of the number of iterations is set to 1000, and the convergence threshold is 0.001. The three-stage law of slow-fast-gradual nutrient release is accurately fitted, and the release curve equations of potassium, magnesium and sulfur are output. Feature parameters such as maximum release amount, peak time and release cycle are extracted.

[0046] The S3 crop demand curve construction method uses a segmented fitting approach to construct dynamic demand curves based on winter wheat nutrient demand data and the duration of each growth stage. Linear fitting algorithms are used for the seedling and overwintering stages, quadratic polynomial fitting algorithms are used for the jointing stage, exponential fitting algorithms are used for the grain-filling stage, and linear fitting algorithms are used for the maturity stage. Each curve segment is smoothed to ensure continuity. The equations of demand curves for each element are output, and characteristic parameters such as maximum demand, peak demand time, and demand peak cycle are extracted.

[0047] The S4 dynamic fitting analysis model is constructed using the nutrient release curves of potassium, magnesium, and sulfur in potassium magnesium sulfate fertilizer and the nutrient requirements of crops for potassium, magnesium, and sulfur throughout their entire growth period as core inputs. The model integrates time alignment, intensity matching, and deviation analysis modules. The time alignment module uses a dynamic time warping algorithm to stretch and compress the time axis, achieving time synchronization of the two curves with a synchronization error controlled within 24 hours. The intensity matching module combines correlation analysis and amplitude proportion analysis to calculate the Pearson correlation coefficient and amplitude proportion coefficient. The deviation analysis module uses a sliding window method to calculate the time offset and intensity difference value for each time period.

[0048] S5 matching degree calculation and difference identification: The model calculates the overall matching degree and the local matching degree for each element and each growth stage. The time offset is calculated to the day. Positive values ​​indicate that the peak release of each element is later than the peak demand, and negative values ​​indicate that it is earlier. The intensity difference values ​​are calculated as absolute and relative differences. The results show that under traditional fertilization, the peak release of potassium sulfate magnesium fertilizer is 7 days later than the peak demand during the grain filling stage of winter wheat, and the relative difference in potassium intensity during the grain filling stage reaches 35%.

[0049] The S6 compatibility assessment and optimization suggestions were generated using the analytic hierarchy process (AHP) to determine the weights of the assessment indicators: overall compatibility score (40%), element compatibility score (30%), and growth stage compatibility score (30%). Compatibility was categorized into four levels: excellent, good, average, and poor, with the traditional fertilization strategy classifying it as average. Regarding the time deviation and intensity differences, the analysis indicated that the causes were late fertilization time and insufficient fertilizer application. Optimization suggestions were generated: advance the fertilization time by 7 days, adjust the fertilizer application rate to 25 kg / mu, and maintain the fertilization depth at 15 cm.

[0050] The S7 model was validated and iteratively optimized. An optimization group and a control group were set up. The optimization group used the fertilization strategy generated by this method, while the control group used the traditional fertilization strategy. At harvest, the plant height, leaf area, and yield of winter wheat in both groups were measured and used as validation indicators to evaluate the model's performance. The results showed that the optimization group outperformed the control group in all indicators. The gradient descent algorithm was used to optimize the model parameters based on the validation results. Parameter updates were performed every 10 sets of new experimental data to improve the model's generalization ability in winter wheat planting scenarios in alluvial soil regions.

[0051] Table 1 Comparison of Fertilization Optimization Effects in Winter Wheat Planting Scenarios Evaluation indicators Traditional fertilization strategies This method optimizes fertilization strategies. Potassium magnesium sulfate fertilizer utilization rate 38% 65% Overall matching degree between nutrient release and demand 0.52 0.89 winter wheat yield per mu 1100 jin 1320 jin Nutrient leaching loss rate 28% 10% Fertilizer cost reduction rate 0 12% Table 1 clearly demonstrates the optimization effect of this method in winter wheat planting. Traditional fertilization strategies suffer from low fertilizer utilization, poor matching between nutrient release and demand, and insufficient supply of key nutrients during the grain-filling stage, leading to low yields and severe nutrient loss. This method, through precise curve fitting and model analysis, identifies the time deviation and intensity difference between release and demand, and optimizes fertilization timing and amount accordingly. This significantly improves fertilizer utilization and overall matching, effectively reduces nutrient loss, and achieves a significant increase in winter wheat yield while lowering fertilization costs. This fully demonstrates the practicality and effectiveness of this method in field grain crop planting scenarios and provides a scientific basis for precision fertilization of winter wheat.

[0052] Example 2: Application of Spring Corn Planting in Northeast Black Soil This embodiment is applied to the field planting scenario of spring maize in the black soil region of Northeast China. The soil moisture in this region is 20%-25%, the temperature of the tillage layer is 15-30℃, and the entire growth period of spring maize is about 120 days, which is divided into six stages: sowing period, seedling period, jointing period, large trumpet stage, grain filling period, and maturity period. The black soil has high fertility but strong water and fertilizer retention capacity. It is necessary to accurately match the nutrient release rhythm of potassium sulfate magnesium fertilizer to avoid nutrient accumulation or insufficient supply.

[0053] S1 multi-dimensional data acquisition and preprocessing involved deploying sensors in the experimental field to monitor potassium, magnesium, and sulfur release data of potassium magnesium sulfate fertilizer at an application rate of 25 kg / mu and a fertilization depth of 20 cm, with a monitoring frequency of 24 hours / time, covering the entire growth period. Nutrient requirements at each growth stage of spring maize were determined through three parallel field trials, and the dataset was refined based on literature review. Outlier removal, missing value completion, and standardization were performed using the Grubbs criterion, linear interpolation, and Z-score standardization methods, with a unified time scale of days and a unit of measurement of mg / kg.

[0054] S2 nutrient release curve construction: The Logistic growth curve algorithm was used to construct dynamic curves for potassium, magnesium, and sulfur nutrient release. With time as the x-axis and nutrient release amount as the y-axis, the curve parameters were optimized using the least squares method with 1000 iterations and a convergence threshold of 0.001. The nutrient release pattern of fertilizer under black soil environment was accurately fitted, and the release curve equations of each element were output. Characteristic parameters such as maximum release amount, peak time, and release cycle were extracted. It was found that the peak time of nutrient release under black soil environment was 3 days later than that under alluvial soil environment.

[0055] The S3 crop demand curve is constructed by segmenting and fitting data based on the nutrient requirements of spring maize and the duration of its growth stages. A linear fitting algorithm is used for the seedling stage, a quadratic polynomial fitting algorithm is used for the jointing and tasseling stages, an exponential fitting algorithm is used for the grain-filling stage, and a linear fitting algorithm is used for the maturity stage. Each curve segment is smoothed to ensure continuity. The demand curve equation is output, key feature parameters are extracted, and the tasseling and grain-filling stages of spring maize are identified as critical periods for nutrient demand.

[0056] The S4 dynamic fitting analysis model is constructed by integrating three major functional modules. The time alignment module uses a dynamic time warping algorithm to synchronize the release and demand curves, with the synchronization error controlled within 24 hours. The intensity matching module calculates the Pearson correlation coefficient and amplitude ratio coefficient between the curves to quantify the degree of intensity fit. The deviation analysis module identifies time offset and intensity difference on a time-by-time basis using the sliding window method, with the window size set to 5 days.

[0057] S5 matching degree calculation and difference identification: The overall matching degree and the local matching degree of each element and growth stage were calculated. The results showed that the overall matching degree under traditional fertilization was 0.48. The magnesium release during the large trumpet stage was 40% lower than the demand. The nutrient release peak was 5 days earlier than the demand peak during the grain filling stage, with a time offset of -5 days. There was an excessive release of sulfur during the maturity stage, with a relative difference value of 28%.

[0058] The S6 compatibility assessment and optimization suggestions were generated using the analytic hierarchy process (AHP) to determine the weights of the assessment indicators: overall compatibility 40%, element compatibility 30%, and growth stage compatibility 30%. The traditional fertilization strategy was deemed to have poor compatibility. The analysis revealed that the discrepancies were caused by excessive fertilization depth leading to slow nutrient release in the early stages and unreasonable fertilizer distribution. The optimization suggestions were: adjust the fertilization depth to 15cm, maintain a total fertilizer application rate of 25kg / mu, applied in two applications: 15kg / mu at sowing and 10kg / mu as a top dressing at the large bell-shaped tassel stage.

[0059] The S7 model was validated and iteratively optimized. Comparative experiments were conducted between an optimization group and a control group. The optimization group adopted the fertilization strategy proposed in this method, while the control group adopted a traditional one-time fertilization strategy. At harvest, the biomass, seed setting rate, and yield per acre of spring maize in both groups were measured to verify the accuracy of the model analysis results. Based on the validation data, the gradient descent algorithm was used to iteratively optimize the model parameters and update the curve fitting and matching degree calculation logic. Each iteration of the model was completed every 10 sets of new spring maize planting data from the black soil region, improving the model's adaptability to this scenario.

[0060] Table 2 Comparison of Fertilization Optimization Effects in Spring Maize Planting Scenarios Evaluation indicators Traditional fertilization strategies This method optimizes fertilization strategies. Potassium magnesium sulfate fertilizer utilization rate 35% 68% Overall matching degree between nutrient release and demand 0.48 0.91 Spring corn yield per mu 1200 jin 1560 jin Nutrient accumulation rate 22% 5% Crop seed setting rate 85% 96% Table 2 shows the data highlighting the optimization value of this method in the black soil planting scenario for spring maize. Under the traditional one-time deep fertilization strategy, the matching degree between fertilizer nutrient release and the needs of spring maize is low, the supply of magnesium is insufficient during the critical period, nutrient release ends prematurely during the grain-filling period, and sulfur is excessive during the maturity period. This not only results in low fertilizer utilization but also leads to nutrient accumulation, affecting the crop's seed setting rate. This method, through precise analysis of the difference between release and demand, optimizes the fertilization depth and method, and adopts a multi-stage fertilization strategy to achieve a high degree of matching between nutrient release and crop needs. This significantly improves fertilizer utilization and seed setting rate, effectively reduces nutrient accumulation, and significantly increases the yield per acre of spring maize. This proves that this method is suitable for the water and fertilizer retention characteristics of black soil and provides scientific strategic support for precision fertilization of spring maize.

[0061] Example 3: Application of Grape Cultivation in Sandy Loam Soil of the Old Course of the Yellow River This embodiment is applied to the open-field grape planting scenario in the sandy loam soil area of ​​the old course of the Yellow River. The soil in this area has good aeration but poor fertilizer retention, with soil moisture of 15%-18% and tillage temperature of 12-32℃. Grapes are perennial vines, and their annual growth period is divided into six stages: budding stage, new shoot growth stage, flowering stage, fruit expansion stage, coloring stage, and ripening stage. The requirements for potassium, magnesium, and sulfur nutrients vary significantly. The fruit expansion stage and coloring stage are the critical periods for demand, and it is necessary to accurately match the release of potassium sulfate and magnesium fertilizer to reduce nutrient leaching and loss.

[0062] S1 multi-dimensional data acquisition and preprocessing involved deploying sensors in vineyards to monitor potassium, magnesium, and sulfur release data under conditions of 30 kg / mu (approximately 20 kg / acre) and 18 cm (approximately 18 cm) of potassium magnesium sulfate fertilizer. Monitoring was conducted 24 hours per session over a one-year growth period. Nutrient requirements at various growth stages of grapes were determined through three parallel field trials, and the dataset was refined using professional literature on grape cultivation. Outliers were removed using the Grubbs criterion, missing values ​​were filled using linear interpolation, and Z-score standardization was used for data preprocessing, resulting in a standardized dataset with consistent time scale and dimensions.

[0063] S2 Nutrient Release Curve Construction: Addressing the poor fertilizer retention of sandy loam soil, a Logistic growth curve algorithm was used to construct a dynamic nutrient release curve. With growth time as the x-axis and nutrient release rate as the y-axis, the curve parameters were optimized using the least squares method, iterating 1000 times with a convergence threshold of 0.001. The curve fitted the pattern of rapid and gradual release of fertilizer nutrients in sandy loam soil, outputting the release curve equations for each element and extracting characteristic parameters such as maximum release amount, peak time, and release cycle.

[0064] The S3 crop demand curve was constructed by segmenting and fitting the nutrient requirements of grapes at each growth stage according to their characteristics and duration. A linear fitting algorithm was used for the budding and new shoot growth stages, a quadratic polynomial fitting algorithm was used for the flowering stage, an exponential fitting algorithm was used for the fruit expansion and coloring stages, and a linear fitting algorithm was used for the ripening stage. Each curve segment was smoothed to ensure overall continuity. The demand curve equation was output, key feature parameters were extracted, and potassium during the fruit expansion stage and magnesium during the coloring stage were identified as the key elements of demand.

[0065] The S4 dynamic fitting analysis model is constructed. The time alignment module uses a dynamic time warping algorithm to synchronize the release and demand curves, with the synchronization error controlled within 24 hours. The intensity matching module quantifies the degree of fit through correlation analysis and amplitude ratio analysis. The deviation analysis module uses a 7-day sliding window to calculate the time offset and intensity difference value for each time period, accurately identifying the matching problems caused by the rapid release and easy loss of nutrients in sandy loam soil.

[0066] S5 matching degree calculation and difference identification: Calculate the overall matching degree and the local matching degree of each element and growth stage. Under traditional fertilization, the overall matching degree is 0.50. The potassium release during the fruit expansion period is 38% lower than the demand. The nutrient release peak is 8 days earlier than the demand peak during the coloring period, with a time offset of -8 days. The sulfur release during the new shoot growth period is excessive, with a relative difference value of 30%. In addition, due to the poor fertilizer retention of sandy loam soil, the nutrient supply is insufficient in the later stage.

[0067] The S6 compatibility assessment and optimization suggestions were generated. The analytic hierarchy process (AHP) was used to determine the weights of the assessment indicators: overall compatibility 40%, element compatibility 30%, growth stage compatibility 30%, and the compatibility of traditional fertilization strategies was rated as average. The analysis revealed that the differences were caused by insufficient fertilization amount, a single fertilization method, and poor fertilizer retention in sandy loam soil leading to nutrient loss in the later stages. The optimization suggestions were: adjust the fertilization amount to 35 kg / mu, using a combination of furrow application and foliar spraying; apply 30 kg / mu of basal fertilizer; and apply 5 kg / mu of diluted potassium magnesium sulfate fertilizer solution as a foliar spray during the fruit expansion period, maintaining a fertilization depth of 18 cm.

[0068] The S7 model was validated and iteratively optimized. Comparative experiments were conducted with an optimization group and a control group. The optimization group adopted the fertilization strategy proposed in this method, while the control group adopted the traditional single-furrow application strategy. At grape maturity, the single fruit weight, sugar content, and yield per acre were measured in both groups as validation indicators to evaluate the model's effectiveness. Based on the validation data, the gradient descent algorithm was used to optimize the model parameters, and the curve fitting algorithm and matching degree calculation logic were updated. A model iteration was completed every 10 sets of grape planting data in sandy loam soil areas to improve the model's generalization ability in fruit tree planting scenarios.

[0069] Table 3 Comparison of Fertilization Optimization Effects in Grape Cultivation Scenarios Evaluation indicators Traditional fertilization strategies This method optimizes fertilization strategies. Potassium magnesium sulfate fertilizer utilization rate 32% 66% Overall matching degree between nutrient release and demand 0.50 0.88 Grape yield per mu 3000 jin 3900 jin Average sugar content of fruit 15.2°Bx 18.5°Bx Nutrient leaching loss rate 32% 11% Table 3 shows the data, which clearly demonstrates the optimization effect of this method in grape cultivation. Under the traditional single furrow application strategy, the poor fertilizer retention of sandy loam soil leads to excessive release of potassium magnesium sulfate fertilizer, resulting in excessive sulfur in the early stage and insufficient supply of key nutrients during the fruit expansion and coloring stages. This results in low fertilizer utilization, severe nutrient leaching and loss, and also affects the quality and yield of grapes. This method, through precise analysis of the difference between release and demand, optimizes the amount and method of fertilizer application based on the characteristics of sandy loam soil. It adopts a combined strategy of furrow application and foliar spraying to compensate for the poor soil fertilizer retention, achieving a precise match between nutrient release and grape demand. This significantly improves fertilizer utilization, reduces nutrient loss, and significantly increases the sugar content and yield per acre of grapes. This proves that this method is suitable for the nutrient requirements of fruit trees and the characteristics of sandy loam soil, providing guidance for the scientific application of potassium magnesium sulfate fertilizer in fruit tree cultivation.

[0070] Example 4: Application of Tomato Facility Cultivation in Soil Scenario This embodiment applies to tomato cultivation in a greenhouse-grown soil area. The soil moisture content is 22%-26%, the temperature of the tillage layer is 18-35℃, and the environment is relatively closed. The entire growth period of tomatoes is about 110 days, which is divided into six stages: sowing, seedling, flowering and fruit setting, fruit expansion, color change, and maturity. Tomatoes under greenhouse cultivation have a large fertilizer requirement, and the nutrient requirements of each growth stage are closely linked. It is necessary to accurately match the release of potassium sulfate magnesium fertilizer to avoid fertilizer damage or nutrient supply interruption.

[0071] S1 multi-dimensional data acquisition and preprocessing involved deploying sensors in a greenhouse to monitor potassium, magnesium, and sulfur release data under conditions of 20 kg / mu (approximately 10 kg / acre) and 12 cm (approximately 12 cm) of potassium magnesium sulfate fertilizer. Monitoring was conducted 24 hours a day, covering the entire tomato growth cycle. Nutrient requirements at each growth stage of tomatoes were determined through three parallel field trials, and the dataset was refined by combining data from literature on greenhouse tomato cultivation. Outlier removal, missing value completion, and standardization were performed using the Grubbs criterion, linear interpolation, and Z-score standardization methods to unify the time scale and dimensions, resulting in a standardized dataset.

[0072] S2 Nutrient Release Curve Construction: Taking advantage of the good fertilizer and water retention properties of soil in facility cultivation, the Logistic growth curve algorithm was used to construct a dynamic nutrient release curve. With time as the x-axis and nutrient release amount as the y-axis, the curve parameters were optimized using the least squares method, with 1000 iterations and a convergence threshold of 0.001. This accurately fits the fertilizer nutrient release pattern under facility conditions, outputs the release curve equations for each element, and extracts characteristic parameters such as maximum release amount, peak time, and release cycle.

[0073] The S3 crop demand curve was constructed by segmenting and fitting data on nutrient requirements and duration at each growth stage of tomatoes. A linear fitting algorithm was used for the seedling stage, a quadratic polynomial fitting algorithm was used for the flowering and fruit setting stage, an exponential fitting algorithm was used for the fruit expansion and color change stages, and a linear fitting algorithm was used for the maturity stage. Each curve segment was smoothed to ensure continuity. The demand curve equation was output, key feature parameters were extracted, and the fruit expansion stage was determined to be the peak period for potassium, magnesium, and sulfur nutrient requirements of tomatoes.

[0074] The S4 dynamic fitting analysis model is constructed. The time alignment module uses a dynamic time warping algorithm to synchronize the release and demand curves, with the synchronization error controlled within 24 hours. The intensity matching module calculates the Pearson correlation coefficient and amplitude ratio coefficient to quantify the degree of intensity fit. The deviation analysis module uses a 5-day sliding window to calculate the time offset and intensity difference value for each time period, accurately identifying nutrient matching problems under facility cultivation.

[0075] S5 matching degree calculation and difference identification calculates the overall matching degree and the local matching degree of each element and growth stage. Under traditional fertilization, the overall matching degree is 0.49. The magnesium release during the fruit expansion period is 42% lower than the demand. The nutrient release peak is 6 days later than the demand peak during the fruit expansion period, with a time offset of 6 days. Excessive fertilization during the seedling stage leads to a relative difference of 33% in potassium, which can easily cause seedling burn.

[0076] The S6 compatibility assessment and optimization suggestions were generated. The analytic hierarchy process (AHP) was used to determine the weights of the assessment indicators: overall compatibility 40%, element compatibility 30%, and growth stage compatibility 30%. The traditional fertilization strategy was deemed to have poor compatibility. The analysis revealed that the discrepancies were caused by insufficient fertilization depth, excessive fertilization during the seedling stage, and insufficient nutrient supply during the fruit expansion stage. The optimization suggestions were: adjust the fertilization depth to 15cm, with a total fertilization amount of 22kg / mu, and adopt phased fertilization: 6kg / mu during the seedling stage and 16kg / mu during the flowering and fruit setting stage, to avoid fertilizer damage during the seedling stage and ensure nutrient supply during the fruit expansion stage.

[0077] The S7 model was validated and iteratively optimized. Comparative experiments were conducted between an optimization group and a control group. The optimization group adopted the fertilization strategy described in this method, while the control group used a traditional one-time shallow fertilization strategy. Fruit set rate, single fruit weight, and yield per acre were measured in both groups at the tomato maturity stage to verify the accuracy of the model's analysis results. Based on the validation data, the gradient descent algorithm was used to iteratively optimize the model parameters and update the curve fitting and matching degree calculation logic. A model iteration was completed every 10 sets of greenhouse tomato planting data to improve the model's adaptability and prediction accuracy in greenhouse fruit and vegetable cultivation scenarios.

[0078] Table 4 Comparison of Fertilization Optimization Effects in Tomato Facility Cultivation Scenarios Evaluation indicators Traditional fertilization strategies This method optimizes fertilization strategies. Potassium magnesium sulfate fertilizer utilization rate 34% 67% Overall matching degree between nutrient release and demand 0.49 0.90 Tomato yield per acre 4500 jin 5850 jin Crop fruit setting rate 82% 95% Incidence of fertilizer damage in seedlings 18% 2% Table 4 shows that the data fully demonstrates the optimization value of this method in the tomato greenhouse cultivation scenario. Under the traditional one-time shallow fertilization strategy, the good soil fertility retention in greenhouse cultivation leads to a high incidence of fertilizer damage during the seedling stage, insufficient supply of key nutrients during the fruit expansion stage, low matching degree between nutrient release and demand, poor fertilizer utilization, and both tomato fruit setting rate and yield are affected. This method combines the environmental characteristics of greenhouse cultivation with the nutrient requirements of tomatoes, accurately analyzes the time deviation and intensity difference between release and demand, optimizes fertilization depth and staged fertilization amount, effectively avoids fertilizer damage during the seedling stage, ensures sufficient nutrient supply during the fruit expansion stage, significantly improves fertilizer utilization and fruit setting rate, and significantly increases tomato yield per acre. This proves that this method is suitable for the closed environment of greenhouse cultivation and the nutrient requirements of fruit and vegetable crops, providing a scientific and effective analytical method and strategy support for precision fertilization in greenhouse fruit and vegetable cultivation.

[0079] Reference Figure 2 This graph illustrates the differences in the matching between potassium magnesium sulfate fertilizer nutrient release and crop requirements at different growth stages. The matching degree is low during the seedling and grain-filling stages, reflecting a mismatch between the fertilizer nutrient release rate and the crop's demand rate during these stages. The matching degree during the grain-filling stage, a critical period for crop nutrient demand, is only 0.48, easily leading to insufficient nutrient supply and affecting yield. The matching degree is moderate during the jointing stage, and high during the flowering and maturity stages, indicating a better match between the nutrient release rhythm and demand during these stages. This graph can quickly identify growth stages with weak matching degrees, providing a visual basis for targeted adjustments to fertilization strategies. For example, to address the low matching degree during the grain-filling stage, fertilization timing can be optimized or slow-release fertilizers can be selected to improve the precision of nutrient supply during critical growth stages.

[0080] Reference Figure 3This graph clearly shows the influence of different application conditions on the peak release time of potassium magnesium sulfate fertilizer. Soil moisture, fertilizer application rate, and peak release time are positively correlated. The longest peak release time is achieved at 25% moisture and 30 kg / mu fertilizer application rate, at 22 days and 25 days respectively. The shortest peak release time is only 15 days at 15% moisture. The peak release time is at a moderate level at a fertilization depth of 15 cm. The peak release time directly affects the timing of nutrient release and crop demand. This graph can help select suitable application conditions. For example, for crops with peak demand around 20 days, a fertilization depth of 15 cm can be selected to ensure a precise match between the peak nutrient release and the peak crop demand.

[0081] Reference Figure 4 This graph clearly demonstrates the positive correlation between nutrient release, demand matching, and crop yield increase. When the matching is poor, yield increases by only 5%; when the matching improves to good, the yield increase reaches 20%; and at excellent and superb levels, the increase further increases to 28% and 35%, respectively. This indicates that the higher the matching level, the more precisely the fertilizer nutrient release matches crop needs, allowing crops to absorb nutrients more efficiently, resulting in a more significant yield increase. This graph visually illustrates the application value of this method, clarifying the target direction for matching optimization for growers and promoting the implementation of precision fertilization.

[0082] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop requirements, characterized in that, Includes the following steps: S1 multi-dimensional data acquisition and preprocessing: collecting nutrient release data of potassium magnesium sulfate fertilizer under different application conditions; Nutrient requirement data of the target crop throughout its entire growth period are collected simultaneously; outlier removal, missing value completion, and standardization are performed on the collected data. S2 nutrient release curve construction: Based on the preprocessed nutrient release data, a nonlinear fitting algorithm is used to construct the dynamic nutrient release curve of potassium magnesium sulfate fertilizer. The curve parameters are optimized by the least squares method, and the release curve equations and characteristic parameters of each element are output. S3 Crop Demand Curve Construction: Based on the nutrient demand data of the target crop and the duration of the crop growth stages, a dynamic curve of nutrient demand throughout the entire growth period is constructed. A segmented fitting method is used to fit curve segments according to the demand characteristics of different growth stages, and the demand curve equations and characteristic parameters of each element are output. The S4 dynamic fitting analysis model is constructed using nutrient release curves and crop demand curves as core inputs. The model integrates three major functional modules: time alignment, intensity matching, and deviation analysis. S5 matching degree calculation and difference identification: The matching degree between the nutrient release curve and the crop demand curve is calculated through a dynamic fitting analysis model. Identify the time offset and intensity difference between release and demand; S6 compatibility assessment and optimization suggestions are generated. Based on the matching degree results, time offset and intensity difference value, a compatibility assessment index system is established, and the compatibility is divided into four levels: excellent, good, average and poor according to the scoring results. S7 model validation and iterative optimization involves collecting crop growth index data and yield data from actual application scenarios, using these as validation indicators; based on the validation results, the gradient descent algorithm is used to iteratively optimize the model parameters, and the curve fitting algorithm and matching degree calculation logic are updated.

2. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, It also includes a precise calculation step for the overall matching degree, executed in S5, which quantifies the overall adaptability to the requirements through a formula, the specific formula being: For overall matching degree, As a time node, This represents the total number of time points. for Nutrient release rate at any time for Real-time crop demand for Time-weighted coefficients.

3. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, It also includes a sensitivity analysis step for application conditions, which is performed after S1 and before S2. The influence of each application condition on the nutrient release curve is analyzed by controlling the variable method. The gradient change values ​​of soil type, soil moisture, soil temperature, fertilizer application amount and fertilizer depth are set. The change rate of characteristic parameters of nutrient release curve when a single condition changes is calculated. The key influencing factors are determined by sorting them according to the magnitude of the change rate.

4. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, In the S1 multi-dimensional data acquisition and preprocessing, nutrient release data is monitored in real time through soil nutrient sensors, and the monitoring period covers the entire period from the application of potassium magnesium sulfate fertilizer to crop maturity. Crop demand data were obtained through a combination of field trials and literature reviews; outlier removal was performed using the Grubbs criterion, missing value completion was performed using linear interpolation, and standardization was performed using the Z-score standardization method.

5. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, In the construction of the S2 nutrient release curve, the Logistic growth curve algorithm was selected for nonlinear fitting, and the curve equation was in the form of an S-shaped curve. During the optimization of curve parameters, feature parameters were extracted by calculating the inflection points and extreme points of the curve through differentiation.

6. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, In constructing the S3 crop demand curve, a suitable fitting algorithm is selected for different growth stages for segmented fitting. A linear fitting algorithm is used for the seedling and maturity stages, a quadratic polynomial fitting algorithm is used for the jointing and flowering stages, and an exponential fitting algorithm is used for the grain-filling stage. Each curve segment is processed through smoothing transition, and the characteristic parameters are obtained by solving curve integrals and extrema.

7. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, In the construction of the S4 dynamic fitting analysis model, the time alignment module adopts the dynamic time warping algorithm to synchronize the time sequence of two curves by stretching or compressing the time axis; the intensity matching module adopts a combination of correlation analysis and amplitude ratio analysis to calculate the Pearson correlation coefficient and amplitude ratio coefficient between the curves. The deviation analysis module calculates the time offset and intensity difference value for each time period using the sliding window method.

8. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, In S5 matching degree calculation and difference identification, local matching degree calculation is performed separately for each reproductive stage, and the matching degree score of each element in each stage is output; a positive value indicates that the peak of nutrient release is later than the peak of demand, and a negative value indicates that the peak of nutrient release is earlier than the peak of demand; the intensity difference value calculation includes absolute difference value and relative difference value. The absolute difference value is the difference between the amount released and the amount demanded, and the relative difference value is the ratio of the absolute difference value to the amount demanded.

9. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, In the S6 compatibility assessment and optimization suggestion generation process, the compatibility assessment index system uses the analytic hierarchy process (AHP) to determine the weight of each index. When generating optimization suggestions, for cases where the absolute value of the time offset is ≥3 days, if the time offset is positive, the fertilization time is adjusted to be earlier or fast-release potassium magnesium sulfate fertilizer is selected; if the time offset is negative, the fertilization time is adjusted to be later or slow-release potassium magnesium sulfate fertilizer is selected. For cases where the absolute value of the relative difference in intensity is ≥20%, the amount of fertilizer is adjusted. For cases where soil conditions are sensitive, the fertilization depth or soil improvement measures are optimized.

10. The fitting analysis method for matching the nutrient release of potassium magnesium sulfate fertilizer with crop demand according to claim 1, characterized in that, In the S7 model validation and iterative optimization, the evaluation of validation indicators adopts the comparative experiment method. An optimization group and a control group are set up. The optimization group adopts the application strategy generated by the model, while the control group adopts the conventional application strategy. The model effect is verified by comparing the differences in crop growth indicators and yield data between the two groups. During the model iterative optimization, the parameters are updated once every 10 sets of new experimental data are collected.

Images