Corn straw returning and fertilizing method and system

Abstract

The invention discloses a corn straw returning-to-field fertilization method and system, and relates to the technical field of farming, and the method comprises the steps of basic data preprocessing, digital twinborn body construction, dynamic decomposition model embedding, optimization decision generation and fertilization prescription output, and construction of a high-fidelity virtual farmland environment based on multi-source data fusion and a spatial interpolation algorithm. The limitation of insufficient description of field heterogeneity in a traditional method is broken through; by embedding the straw soil microorganism dynamic decomposition model, quantitative prediction of the straw decomposition rate and the nutrient release amount is realized, and a scientific basis is provided for the fertilization opportunity and dosage; by combining state space construction and a strategy learning algorithm, nutrient supply and demand dynamic matching and fertilization strategy adaptive optimization are realized, the fertilizer utilization efficiency is remarkably improved, the environmental pollution risk is reduced, a precise fertilization prescription is generated and output to a display terminal, and the intelligent level and operation convenience of agricultural production are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the field of farming technology, in particular to a corn straw returning to field fertilization method and system. BACKGROUND

[0002] Traditional agricultural fertilization management relies on experience judgment or fixed mode formula, and it is difficult to realize the dynamic matching of nutrient release and crop demand under the condition of straw returning to field; the existing straw returning to field management lacks digital twin and intelligent decision support, and it is difficult to realize the "data-model-decision" closed loop optimization, which restricts the straw resource utilization efficiency and the sustainable development of farmland.

[0003] The prior art such as the invention application patent with the announcement number CN119090668B discloses a fertilization method and device based on corn straw returning to field, which comprises the following steps: receiving planting data sent by a corn straw returning to field planter, obtaining a plurality of historical soil samples based on a straw returning to field area, detecting the plurality of historical soil samples by using a pre-constructed soil data detection method to obtain historical soil data, obtaining target soil data based on a target crop, calculating a unit area straw amount based on the planting data, extracting a straw decomposition curve set from a pre-constructed database based on the unit area straw amount and a corn variety, obtaining unit mixed fertilizer amount data by using the straw decomposition curve set, and performing fertilization operation on the straw returning to field area based on the unit mixed fertilizer amount data to complete the fertilization based on the corn straw returning to field.

[0004] For the above-mentioned scheme, the present application found that the above-mentioned technology at least has the following technical problems: 1, the current lack of through weighted fusion and standardization, improve the data quality and consistency, cannot avoid the problem of data source error accumulation in traditional method, cannot ensure the reliability of subsequent model input, so as to improve the overall prediction accuracy. Cannot guarantee the authenticity and reference of the analysis result, cannot provide reliable basis for subsequent decision. Lack of based on standardized data set and hyperspectral data, through spatial interpolation algorithm to construct high-fidelity virtual farmland simulation environment; it is impossible to realize the accurate digital mapping of farmland environment, it is impossible to provide high-resolution simulation environment, it is impossible to enhance the authenticity and prediction ability of the model, and it is impossible to support comprehensive and objective analysis.

[0005] 2, the current lack of through the combination of mechanism model and data driven, cannot accurately predict the nutrient dynamics, cannot avoid the idealized assumption of traditional single model, cannot improve the accuracy and adaptability of the model, cannot ensure the scientificity of fertilization decision, cannot reduce the waste and pollution caused by blind fertilization; it is impossible to realize the closed loop optimization and real-time decision of complex agricultural process, it is impossible to adjust the fertilization strategy in time, so as to improve the fertilizer utilization rate and crop yield.

[0006] 3. Currently, there is a lack of optimization decision-based fertilization amount calculation, conversion of decisions into operable prescriptions, and provision of variable fertilization recommendations and optimal window periods, which cannot achieve precision and automation of fertilization, generate precise prescriptions, and output to display terminals, cannot guarantee the pertinence and balance of field operations, and cannot improve the utilization efficiency and sustainability of agricultural resources. SUMMARY

[0007] In view of the above technical deficiencies, the purpose of the present application is to provide a corn straw field returning fertilization method and system.

[0008] To solve the above technical problems, the present application adopts the following technical solutions: In a first aspect, the present application provides a corn straw field returning fertilization method, which comprises the following steps: Step one, basic data preprocessing: based on the pre-acquired basic data, data cleaning and weighted fusion analysis are performed to generate a standardized data set.

[0009] Step two, digital twin construction: based on the standardized data set, a pre-set geographic information system (GIS) map, and hyperspectral image data, a spatial interpolation algorithm is used to construct a digital twin of the field to form a high-fidelity virtual farmland simulation environment.

[0010] Step three, dynamic decomposition model embedding: a pre-set straw-soil microbial dynamic decomposition model is embedded in the digital twin of the field, and the basic data is input to simulate the straw decomposition rate and nutrient release, to output decomposition rate prediction and nutrient release amount prediction; Step four, optimization decision generation: based on the decomposition rate prediction and nutrient release amount prediction, a state space is constructed to integrate the nutrient status, and then the optimal strategy is learned to generate an optimized fertilization decision.

[0011] Step five, fertilization prescription output: based on the optimized fertilization decision, the fertilization amount is calculated and a precise fertilization prescription is generated, which is then output to a display terminal.

[0012] Preferably, the basic data includes soil data, straw data, and environmental data.

[0013] Preferably, based on the pre-acquired basic data, data cleaning and weighted fusion analysis are performed to generate a standardized data set, which includes: based on soil data, straw data, and environmental data, data cleaning operations are performed to obtain a cleaned data set, wherein the data cleaning operations include missing value processing, outlier removal, and data standardization; based on the pre-set weight coefficients of data source accuracy, weighted fusion analysis is performed on the cleaned data set to generate a standardized data set, wherein the weighted fusion analysis is based on the weighted average formula.

[0014] Preferably, the constructing the field digital twin through a spatial interpolation algorithm to form a high-fidelity virtual farmland simulation environment comprises: based on the standardized data set, the preset geographic information system GIS map and the hyperspectral image data, performing data alignment and coordinate unification processing to generate an integrated spatial data set; based on the integrated spatial data set, performing interpolation calculation on missing or low-resolution areas through Kriging spatial interpolation to generate a high-resolution spatial distribution map; based on the high-resolution spatial distribution map and the preset geographic information system GIS map, generating a digital twin of the field through three-dimensional modeling technology, which includes real-time soil data, straw data and environmental data, to form a high-fidelity virtual farmland simulation environment.

[0015] Preferably, the embedding a preset straw-soil microbial dynamic decomposition model in the field digital twin and inputting basic data to simulate straw decomposition rate and nutrient release to output decomposition rate prediction and nutrient release amount prediction comprises: A1, initializing model parameters, based on the field digital twin, loading model parameters of the preset straw-soil microbial dynamic decomposition model, wherein the model parameters include microbial activity coefficient, straw carbon-nitrogen ratio threshold, temperature sensitivity coefficient and humidity influence factor.

[0016] A2, inputting basic data, extracting basic data parameters including soil temperature, soil humidity, soil pH value, straw carbon-nitrogen ratio and microbial biomass from the standardized data set, and inputting the basic data parameters into the dynamic decomposition model.

[0017] A3, simulating decomposition rate and nutrient release: based on the straw-soil microbial dynamic decomposition model, calculating straw decomposition rate through a decomposition rate algorithm, wherein the decomposition rate algorithm integrates temperature, humidity and microbial activity parameters; based on the decomposition rate prediction, calculating nutrient release amount prediction through a nutrient release algorithm, wherein the nutrient release algorithm is associated with straw nutrient content and decomposition rate.

[0018] A4, outputting decomposition rate prediction and nutrient release amount prediction.

[0019] Preferably, the calculating straw decomposition rate through a decomposition rate algorithm comprises: through a calculation formula the straw decomposition rate is obtained , wherein is a reference decomposition rate constant, which is preset based on straw type and soil properties; is a temperature influence function, is a humidity influence function, is a microbial activity function.

[0020] Preferably, the calculating nutrient release amount prediction through a nutrient release algorithm based on the decomposition rate prediction comprises: through a calculation formula the nutrient release amount of nitrogen the nutrient release amount of phosphorus the nutrient release amount of potassium , , and respectively represent the nutrient content of nitrogen, the nutrient content of phosphorus and the nutrient content of potassium of the straw, is expressed as the time period corresponding to the nutrient release.

[0021] Preferably, based on the decomposition rate prediction and the nutrient release amount prediction, a state space is constructed to integrate the nutrient status, and then the optimal strategy is learned to generate the optimized fertilization decision, including: B1, based on the decomposition rate prediction and the nutrient release amount prediction, a state space is constructed, wherein the state variables include real-time soil nutrient content and predicted nutrient release amount, to integrate the nutrient status.

[0022] B2, define an action space, wherein the action includes fertilization type selection, fertilization amount setting and fertilization time period determination as fertilization decision options; design a reward function, which is optimized based on nutrient utilization rate, economic benefit and environmental benefit to calculate the reward value of simulation deduction; use a reinforcement learning algorithm to perform simulation deduction in the field digital twin based on the state space, action space and reward function to learn the optimal fertilization strategy.

[0023] B3, according to the learned optimal fertilization strategy, generate the optimized fertilization decision, including the fertilizer formula, the fertilization amount and the fertilization time period.

[0024] Preferably, based on the optimized fertilization decision, the fertilization amount is calculated and the precise fertilization prescription is generated, and then output to the display terminal, including: based on the optimized fertilization decision, the target nutrient amount and the crop nutrient demand parameters are extracted, combined with the existing soil nutrient amount and the released nutrient amount of the straw, the fertilization amount is calculated to obtain the actual fertilization amount, and then based on the actual fertilization amount and the fertilizer database, the fertilizer formula is determined, and at the same time based on the crop growth model and the real-time environmental data, the fertilization time period is determined, finally the precise fertilization prescription is integrated and generated to output to the display terminal.

[0025] The application provides a corn straw returning to field fertilization method system in the second aspect, including: a basic data preprocessing module, based on the pre-acquired basic data, data cleaning and weighted fusion analysis are carried out to generate a standardized data set.

[0026] A digital twin construction module constructs a field digital twin based on the standardized data set, a preset geographic information system (GIS) map and hyperspectral image data through a spatial interpolation algorithm to form a high-fidelity virtual farmland simulation environment.

[0027] The dynamic decomposition model embedding module is configured to embed a preset straw soil microbial dynamic decomposition model in the plot digital twin, and input basic data to simulate straw decomposition rate and nutrient release, so as to output decomposition rate prediction and nutrient release amount prediction.

[0028] The optimization decision generation module is configured to construct a state space based on the decomposition rate prediction and the nutrient release amount prediction to integrate nutrient conditions, and then learn an optimal strategy to generate an optimized fertilization decision.

[0029] The fertilization prescription output module is configured to calculate a fertilization amount and generate a precise fertilization prescription based on the optimized fertilization decision, and then output the precise fertilization prescription to a display terminal.

[0030] The corn straw field fertilization method and system provided by the application has the following beneficial effects: 1. The corn straw field fertilization method and system provided by the application is based on multi-source data fusion and spatial interpolation algorithm, and a high-fidelity virtual farmland environment is constructed, which breaks through the limitation of traditional methods that cannot accurately depict the heterogeneity of a field. By embedding a straw soil microbial dynamic decomposition model, the decomposition rate of straw and the nutrient release amount are quantitatively predicted, which provides a scientific basis for the timing and amount of fertilization. In combination with state space construction and strategy learning algorithm, dynamic matching of nutrient supply and demand and adaptive optimization of fertilization strategy are realized, which significantly improves the fertilizer utilization efficiency, reduces the risk of environmental pollution, generates a precise fertilization prescription and outputs the precise fertilization prescription to a display terminal, and improves the intelligent level and operation convenience of agricultural production.

[0031] 2. The application improves the data quality and consistency by weighted fusion and standardization, and avoids the problem of data source error accumulation in traditional methods. This ensures the reliability of the subsequent model input, thereby improving the overall prediction accuracy. The authenticity and reference value of the analysis results are ensured, and a reliable basis is provided for subsequent decision-making. Based on the standardized data set and hyperspectral data, a high-fidelity virtual farmland simulation environment is constructed by a spatial interpolation algorithm; accurate digital mapping of the farmland environment is realized, a high-resolution simulation environment is provided, and the authenticity and prediction ability of the model are enhanced. Comprehensive and objective analysis is supported, which is helpful for subsequent model embedding and optimization, and reduces the cost of field tests.

[0032] 3. The application accurately predicts nutrient dynamics by combining mechanism models and data-driven methods, avoiding the idealized assumptions of traditional single models. The accuracy and adaptability of the model are improved, the scientificity of the fertilization decision is ensured, and the waste and pollution caused by blind fertilization are reduced; the closed-loop optimization and real-time decision-making of complex agricultural processes are realized, the intelligence and efficiency of the system are improved, the fertilization strategy is adjusted in a timely manner, and thus the fertilizer utilization rate and crop yield are improved, and the environmental load is reduced.

[0033] 4、The application calculates the fertilization amount based on optimized decision-making, converts the decision-making into an operable prescription, provides variable fertilization recommendation and an optimal window period, realizes the precision and automation of fertilization, generates a precise prescription and outputs it to a display terminal, ensures the pertinence and balance of field operation, reduces human intervention, and improves the utilization efficiency and sustainability of agricultural resources. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0035] Figure 1 The present application is a method implementation step flow diagram.

[0036] Figure 2 The present application is a system structure connection diagram. DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0038] Please refer to Figure 1 As shown in the first aspect, the present application provides a corn straw returning to field fertilization method, comprising: step one, basic data preprocessing: based on the pre-acquired basic data, data cleaning and weighted fusion analysis are performed to generate a standardized data set.

[0039] In one specific example, the basic data includes soil data, straw data and environmental data.

[0040] It should be noted that the soil data includes soil humidity, soil pH value, soil organic matter content, soil nitrogen content, soil phosphorus content, soil potassium content, soil carbon content, soil porosity and soil temperature, etc.; the straw data includes straw biomass, straw carbon-nitrogen ratio, straw moisture content and straw decomposition rate, etc.; the environmental data includes air temperature, air humidity, light intensity, rainfall, wind speed and carbon dioxide concentration, etc.

[0041] Further, the soil humidity sensor and the soil temperature sensor collect field data in real time; soil sampling analysis determines the pH value, organic matter content, soil nitrogen content, soil phosphorus content, soil potassium content, and soil carbon content; the soil porosity tester measures the soil porosity parameters; the field sampling weighing method determines the straw biomass; the elemental analyzer determines the straw carbon-nitrogen ratio; the moisture tester measures the straw water content; the in-situ decomposition experiment determines the straw decomposition rate; these data are obtained by the following methods: the meteorological station automatically monitors the air temperature, humidity, rainfall, and wind speed; the photosynthetically active radiation sensor measures the light intensity; the carbon dioxide concentration sensor monitors the atmospheric content; all environmental data are transmitted to the data center in real time through the Internet of Things equipment.

[0042] It should be noted that the data acquisition adopts a multi-source heterogeneous collection method, including real-time monitoring by sensors, laboratory test analysis, and field measurement; and all collection equipment is calibrated regularly to ensure data accuracy and consistency, and the collection frequency is dynamically adjusted according to the data type and growth season.

[0043] In one specific example, based on the pre-acquired basic data, data cleaning and weighted fusion analysis are performed to generate a standardized data set, including: based on soil data, straw data, and environmental data, performing a data cleaning operation to obtain a cleaned data set, wherein the data cleaning operation includes missing value processing, outlier removal, and data standardization; based on the preset weight coefficient of the data source accuracy, performing weighted fusion analysis on the cleaned data set to generate a standardized data set, wherein the weighted fusion analysis is performed based on the weighted average formula.

[0044] It should be noted that in the data cleaning operation, the missing value processing adopts the mean interpolation method, and the specific way is: the interpolated value is equal to the sum of the values of each valid data point divided by the total number of valid data points; the physical meaning is to fill in the missing values by averaging to ensure data integrity; the outlier removal is based on the 3σ principle, and the specific way is: when the absolute deviation of the data point value from the mean value exceeds three times the standard deviation, the data point value is regarded as an outlier and is deleted; the physical meaning is to identify and remove outliers that deviate significantly to improve data quality; the data standardization adopts the Min-Max standardization method, and the specific way is: the standardized value is equal to the original value minus the minimum value of the data set divided by the difference between the maximum value and the minimum value; the physical meaning is to scale the data to the interval, eliminating the dimension effect.

[0045] Further, The principle is expressed as the Lai Da criterion, which is to first assume that a set of detection data only contains random errors, to calculate and process the standard deviation, to determine an interval according to a certain probability, to consider that any error exceeding the interval is not a random error but a gross error, and to eliminate the data containing the error.

[0046] It should be noted that the weighted fusion analysis is based on a weighted average formula, and the calculation method is that the fusion value in the standardized data set is equal to the sum of the data values of each data source after cleaning multiplied by the corresponding preset weight coefficient, wherein the weight coefficient is preset based on the data source accuracy, and the sum of all weight coefficients is 1, and the physical meaning is to reflect the data source reliability through the weight to improve the fusion accuracy.

[0047] Further, based on historical data collection records, the accuracy indicators of each data source are counted, including data collection error rate, data integrity and sampling frequency; according to the accuracy indicators, the initial weight coefficients of each data source are calculated through a weight distribution formula, and the weight distribution formula is: data collection error rate multiplied by 1 minus the weight factor corresponding to the data collection error rate, data integrity multiplied by the weight factor corresponding to the data integrity, and sampling frequency multiplied by the weight factor corresponding to the sampling frequency; based on the real-time data quality monitoring result, the initial weight coefficients are dynamically adjusted to obtain the final preset weight coefficients.

[0048] It should be noted that the weight factor corresponding to the data collection error rate, the weight factor corresponding to the data integrity, and the weight factor corresponding to the sampling frequency are obtained by the factor analysis method, which first condenses the information of the collection error rate, the data integrity and the sampling frequency, then obtains the variance explained rate after rotation, and obtains the weight by subtracting the cumulative variance explained rate.

[0049] It should be noted that the factor analysis method is a known technology, which is a multivariate statistical analysis method of reducing a number of variables with complex relationships to a few comprehensive factors from the dependent relationship of internal correlation of variables; information condensation is expressed as calculating the median; variance explained rate is the amount of information of factor extraction, and variance explained rate = eigenvalue / total number of analysis items; the variance explained rate after rotation represents the variance explained rate of the factor after maximum variance rotation.

[0050] Step two, digital twin construction: based on the standardized data set, the preset geographic information system GIS map and the hyperspectral image data, a plot digital twin is constructed through a spatial interpolation algorithm to form a high-fidelity virtual farmland simulation environment.

[0051] It should be noted that hyperspectral image data is a kind of spectral image data containing hundreds of narrow bands, which can capture the detailed reflection characteristics of ground objects at different wavelengths, and is used for fine classification and analysis, such as vegetation health monitoring and soil property identification; hyperspectral image data is obtained through remote sensing platforms, such as satellite sensors or unmanned aerial vehicles equipped with hyperspectral cameras, to provide high-resolution spectral information.

[0052] In one specific example, the construction of the digital twin of the plot by the spatial interpolation algorithm to form a high-fidelity virtual farmland simulation environment includes: based on the standardized data set, the preset geographic information system GIS map and the hyperspectral image data, performing data alignment and coordinate unification processing to generate an integrated spatial data set; based on the integrated spatial data set, performing interpolation calculation on missing or low-resolution areas by Kriging spatial interpolation method to generate a high-resolution spatial distribution map; based on the high-resolution spatial distribution map and the preset geographic information system GIS map, generating the digital twin of the plot by three-dimensional modeling technology, which includes real-time soil data, straw data and environmental data, to form a high-fidelity virtual farmland simulation environment.

[0053] It should be noted that the data alignment processing is to match the soil parameters, straw parameters and environmental parameters in the standardized data set with the spatial coordinates of the GIS map and the pixel coordinates of the hyperspectral image, to ensure that all data are in the same coordinate system; the coordinate unification processing adopts the WGS84 coordinate system, and different sources of data are converted into unified coordinates by a coordinate conversion formula: wherein represents the original coordinates, represents the converted unified coordinates, represents the coordinate conversion function, and represents a linear conversion based on affine transformation; the physical meaning is to eliminate data space deviation through a unified coordinate system to ensure the accuracy of subsequent interpolation.

[0054] Further, the specific implementation of the coordinate conversion function includes translation, rotation and scaling operations, and least squares fitting is performed based on control points to minimize conversion errors, thereby completing the conversion of coordinates and generating the integrated spatial data set.

[0055] It should be noted that the WGS-84 coordinate system (World Geodetic System 1984 Coordinate System) is an international adopted geocentric coordinate system.

[0056] It should be noted that the Kriging spatial interpolation method is a statistical optimal unbiased estimation method; the interpolation result of the prediction point is obtained by the calculation formula wherein represents the prediction point, denotes the known points, denotes the known points the weight coefficient of the prediction point, is the known point the measured value, denotes the number corresponding to the known point, , denotes the total number of known points.

[0057] Further, the physical meaning of the Kriging spatial interpolation method is to predict the unknown point value by weighted averaging the known point value, while considering the spatial autocorrelation to improve the interpolation accuracy.

[0058] It should be noted that the interpolation result, such as the soil nutrient value, represents the simulated value of the point in the digital twin; the measured value is extracted from the standardized dataset; the weight coefficient of the known point to the prediction point is obtained by factor analysis method, and satisfies to ensure the unbiasedness.

[0059] It should be noted that the three-dimensional modeling technology adopts a voxel-based rendering method to convert the interpolated spatial distribution map into a three-dimensional mesh model, where each voxel contains parameter values (such as humidity and temperature, etc.), and the farmland environment is simulated through a real-time rendering engine; the physical meaning is to create a virtual copy consistent with the real plot, supporting dynamic simulation and interaction.

[0060] Further, the high-fidelity virtual farmland simulation environment includes a visualization interface that allows users to view soil profiles at different depths and real-time environmental changes, and is implemented based on OpenGL or similar graphics libraries.

[0061] The present application improves data quality and consistency through weighted fusion and standardization, avoiding the problem of cumulative data source errors in traditional methods. This ensures the reliability of subsequent model input, thereby improving overall prediction accuracy. It ensures the authenticity and reference value of the analysis results, providing a reliable basis for subsequent decision-making. Based on the standardized dataset and hyperspectral data, a high-fidelity virtual farmland simulation environment is constructed through a spatial interpolation algorithm; accurate digital mapping of the farmland environment is achieved, providing a high-resolution simulation environment and enhancing the authenticity and prediction ability of the model. It supports comprehensive and objective analysis, which helps subsequent model embedding and optimization, reducing the cost of field trials.

[0062] Step three, dynamic decomposition model embedding: embed the preset straw soil microorganism dynamic decomposition model in the plot digital twin, and input the basic data to simulate the straw decomposition rate and nutrient release, to output the decomposition rate prediction and nutrient release amount prediction.

[0063] In one embodiment, the preset straw-soil microbial dynamic decomposition model is embedded in the plot digital twin, and the base data is input to simulate the straw decomposition rate and nutrient release, to output the decomposition rate prediction and nutrient release prediction, including: A1, model parameter initialization is performed, and the model parameters of the preset straw-soil microbial dynamic decomposition model are loaded based on the plot digital twin, wherein the model parameters include a microbial activity coefficient, a straw carbon-nitrogen ratio threshold, a temperature sensitivity coefficient, and a humidity influence factor.

[0064] A2, base data input is performed, and base data parameters including soil temperature, soil humidity, soil pH, straw carbon-nitrogen ratio, and microbial biomass are extracted from the standardized data set, and the base data parameters are input to the dynamic decomposition model.

[0065] A3, decomposition rate and nutrient release simulation is performed: based on the straw-soil microbial dynamic decomposition model, a straw decomposition rate is calculated by a decomposition rate algorithm, wherein the decomposition rate algorithm integrates temperature, humidity, and microbial activity parameters; based on the decomposition rate prediction, a nutrient release prediction is calculated by a nutrient release algorithm, wherein the nutrient release algorithm is associated with straw nutrient content and decomposition rate.

[0066] A4, the decomposition rate prediction and the nutrient release prediction are output.

[0067] In one embodiment, the straw decomposition rate is calculated by the decomposition rate algorithm, including: the straw decomposition rate is calculated by the formula wherein is a reference decomposition rate constant, which is preset based on straw type and soil properties; is a temperature influence function, is a humidity influence function, is a microbial activity function.

[0068] ​It should be noted that the straw decomposition rate is expressed as the percentage decrease in straw mass per unit time; the reference decomposition rate constant is pre-set based on straw type and soil properties, including: extracting the carbon-to-nitrogen ratio of corn straw in historical records, analyzing the straw decomposition rate data of corn straw under various soil properties; various soil properties such as soil pH and organic matter content; soil properties are divided into multiple categories or ranges by soil pH and organic matter content; for example, soil properties are divided into acidic, neutral and alkaline according to pH, and low, medium and high according to organic matter content. The average value of the historical decomposition rate is calculated as the reference decomposition rate constant by combining the carbon-to-nitrogen ratio of corn straw and the soil properties; for example, the carbon-to-nitrogen ratio of corn straw is about 25 and the neutral soil pH is 6.5-7.5, and the organic matter content is 3%, the reference decomposition rate constant may be pre-set to 0.05% / day. The accuracy of the reference decomposition rate constant is verified by controlling experiments, for example, measuring the decomposition rate under specific conditions and comparing it with the pre-set value, and adjusting the reference decomposition rate constant to ensure reliability.

[0069] It should be noted that the temperature influence function is dimensionless, reflecting the modulation of temperature on the decomposition rate, including: the temperature influence function is expressed as an activation energy parameter, is expressed as a reference temperature, is the input soil temperature, is expressed as a natural constant.

[0070] Further, the activation energy parameter is pre-set based on the metabolic characteristics of microorganisms, the specific process being as follows: extract metabolic rate data of each microorganism type (such as bacteria and fungi, etc.) at a specific temperature in the database; calculate the activation energy parameter based on the metabolic rate data of each microorganism type at a specific temperature by linear regression; finally, pre-set the classification parameter (such as the activation energy parameter of mesophilic aerobic bacteria is pre-set to 50 kJ / mol) according to the microorganism classification (such as aerobic and anaerobic bacteria, etc.), and dynamically adjust the parameter value based on real-time microbial biomass data.

[0071] It should be noted that the humidity influence function is dimensionless, reflecting the modulation of humidity on the decomposition rate, including: the humidity influence function is expressed as the input soil humidity, is expressed as the reference value of soil humidity, pre-set based on soil type.

[0072] Further, the reference value of soil humidity is pre-set based on soil type, expressed as the optimal humidity corresponding to the soil type; the physical meaning of the humidity influence function is to simulate the nonlinear effect of humidity on microbial activity, and excessive or insufficient humidity will inhibit decomposition.

[0073] ​​It should be noted that the microbial activity function is dimensionless, reflecting the modulation of microbial biomass on the decomposition rate, including: the microbial activity function , is expressed as the input microbial biomass, is expressed as the maximum value of the microbial biomass; Further, the maximum value of the microbial biomass is preset based on the soil health status, indicating the allowable value of the microbial biomass of the soil in a healthy state; the physical meaning of the microbial activity function is that the number of microorganisms directly proportional to the decomposition rate.

[0074] In one specific example, the decomposition rate prediction is used to predict the nutrient release amount by a nutrient release algorithm, including: by calculating the formula the nutrient release amount of nitrogen , the nutrient release amount of phosphorus , and the nutrient release amount of potassium , , and respectively represent the nutrient content of nitrogen, the nutrient content of phosphorus and the nutrient content of potassium of the straw, is expressed as the corresponding time period of nutrient release.

[0075] It should be noted that the nutrient release amount represents the amount of nutrients released in the corresponding time period; the nutrient content of nitrogen, the nutrient content of phosphorus and the nutrient content of potassium of the straw are extracted from the straw data; the physical meaning is that the nutrient release amount is directly proportional to the decomposition rate and the nutrient content, simulating the cumulative release process over time.

[0076] Step four, optimization decision generation: based on the decomposition rate prediction and the nutrient release amount prediction, the state space is constructed to integrate the nutrient status, and then the optimal strategy is learned to generate the optimized fertilization decision; In one specific example, based on the decomposition rate prediction and the nutrient release amount prediction, the state space is constructed to integrate the nutrient status, and then the optimal strategy is learned to generate the optimized fertilization decision, including: B1, based on the decomposition rate prediction and the nutrient release amount prediction, the state space is constructed, wherein the state variables include real-time soil nutrient content and predicted nutrient release amount, to integrate the nutrient status.

[0077] It should be noted that the construction of the state space, the real-time soil nutrient content and the predicted nutrient release amount in the state variable include soil nitrogen content, soil phosphorus content, soil potassium content, predicted nitrogen release amount, predicted phosphorus release amount and predicted potassium release amount, etc. These variables are updated in real time based on the decomposition rate prediction and the nutrient release amount prediction; the physical meaning is to fully characterize the nutrient dynamics of farmland by integrating predicted and real-time data, and to provide environmental state input for reinforcement learning.

[0078] Further, the update frequency of the state variables is synchronized with the data collection, ensuring that the state space reflects the latest nutrient status and is displayed in real-time through the digital twin visualization module.

[0079] B2, defining an action space, wherein the actions include fertilizer type selection, fertilizer amount setting, and fertilizer time period determination as fertilizer decision options; designing a reward function that optimizes nutrient use efficiency, economic benefits, and environmental benefits to calculate the reward value of the simulation deduction; using a reinforcement learning algorithm to perform simulation deduction in the field digital twin based on the state space, action space, and reward function to learn the optimal fertilization strategy.

[0080] It should be noted that the definition of the action space includes fertilizer type selection, fertilizer amount setting, and fertilizer time period determination expressed in growth stages, etc.; the fertilizer type selection is, for example, nitrogen fertilizer, phosphorus fertilizer, potassium fertilizer, or compound fertilizer, etc.; the fertilizer amount setting is in units of kilograms per hectare; and the fertilizer time period determination is expressed in growth stages; the physical meaning of defining the action space is to convert the fertilization operation into continuous decision options for reinforcement learning algorithm exploration and optimization.

[0081] Further, the design of the action space is based on agronomic knowledge and historical data to ensure decision feasibility and practicality, and prevent over-fertilization through constraint conditions.

[0082] It should be noted that the reward function is expressed as wherein is expressed as the reward value of taking action in state , is expressed as a nutrient use efficiency function, is expressed as an economic benefit function, is expressed as an environmental benefit function, , and are respectively expressed as the weight factor corresponding to the nutrient use efficiency function, the weight factor corresponding to the economic benefit function, and the weight factor corresponding to the environmental benefit function; the physical meaning of the reward function is to guide the reinforcement learning algorithm to optimize the decision direction towards high nutrient use efficiency, high economic benefit, and low environmental impact through weighted combination of multi-objective indicators.

[0083] Further, , , , The weight factor corresponding to the nutrient utilization rate function, the weight factor corresponding to the economic benefit function and the weight factor corresponding to the environmental benefit function are obtained by a factor analysis method. First, the information of the nutrient utilization rate function, the economic benefit function and the environmental benefit function is condensed to obtain the weight by the cumulative variance explanation rate; the calculation method of the nutrient utilization rate function is the ratio of the nutrient absorption amount of the crop to the fertilization nutrient amount, the calculation method of the economic benefit function is the value of the crop yield minus the fertilization cost, and the calculation method of the environmental benefit function is the negative environmental pollution index (such as the amount of nitrogen and phosphorus loss).

[0084] It should be noted that the reward value of the simulation deduction is calculated, wherein the Q-learing algorithm is used for simulation deduction, and the update formula is: , wherein represents a state-action value function, represents a learning rate, represents a discount factor, represents a next state; the physical meaning is to update the value of by iteration, learn the long-term expected reward of taking an action in a given state, and finally converge to obtain an optimal strategy.

[0085] Further, the simulation deduction is performed in the digital twin of the field, the high-fidelity environmental simulation state transition and reward calculation are utilized, the learning rate and the discount factor are preset based on historical data, such as the learning rate being equal to 0.1 and the discount factor being equal to 0.9, and the algorithm stability is ensured through multiple simulations.

[0086] B3, according to the learned optimal fertilization strategy, an optimized fertilization decision is generated, including a fertilizer formula, a fertilization amount and a fertilization time period.

[0087] It should be noted that based on the optimal strategy learned by the Q-learning algorithm, the action combination with the highest reward value is selected, and the output is a fertilizer formula, a fertilization amount and a fertilization time period, such as the ratio of nitrogen, phosphorus and potassium, the specific numerical value of the fertilization amount, and the specific date of the fertilization time period; precise fertilization is realized, the optimized fertilization decision is output by the control system, guiding the on-site fertilization operation, and can be dynamically adjusted according to real-time data.

[0088] The present application accurately predicts the nutrient dynamics by combining mechanism models and data-driven methods, avoiding the idealized assumptions of traditional single models. The accuracy and adaptability of the model are improved, ensuring the scientificity of the fertilization decision, reducing the waste and pollution caused by blind fertilization; the closed-loop optimization and real-time decision of the complex agronomic process are realized, the intelligence and efficiency of the system are improved, the fertilization strategy is adjusted in time, thereby improving the fertilizer utilization rate and crop yield, and reducing the environmental load.

[0089] Step five, fertilizer prescription output: based on the optimized fertilization decision, calculate the fertilization amount and generate the precise fertilization prescription, and then output to the display terminal.

[0090] In one specific example, the fertilization amount is calculated based on the optimized fertilization decision, and the precise fertilization prescription is generated and output to the display terminal, including: based on the optimized fertilization decision, the target nutrient amount and crop nutrient demand parameters are extracted, combined with the existing soil nutrient amount and straw released nutrient amount, the fertilization amount is calculated to obtain the actual fertilization amount, and then based on the actual fertilization amount and the fertilizer database, the fertilizer formula is determined, and based on the crop growth model and real-time environmental data, the fertilization time period is determined, and finally the precise fertilization prescription is integrated to output to the display terminal.

[0091] It should be noted that the actual fertilization amount is calculated based on the actual fertilization amount, and the specific process is as follows: the actual fertilization amount is equal to the crop nutrient demand parameter minus the soil existing nutrient amount minus the straw released nutrient amount, and the result is divided by the fertilizer utilization rate; the unit of actual fertilization amount is kilogram per hectare, indicating the mass of fertilizer needed to be applied, and its physical meaning is the fertilization amount calculated based on the principle of nutrient balance to ensure the growth demand of crops; the unit of crop nutrient demand parameter is kilogram per hectare, indicating the total nutrient amount required by crops during the entire growth period, which is obtained from a pre-set database based on crop type and expected yield; the unit of soil existing nutrient amount is kilogram per hectare, indicating the current available nutrient amount in the soil, which is extracted from the standardized data set; the straw released nutrient amount represents the nutrient amount released from straw decomposition, with a unit of kilogram per hectare, which is obtained from nutrient release prediction; the fertilizer utilization rate is a dimensionless value, which is pre-set based on soil properties and fertilizer type, with a value range of 0 to 1, and its physical meaning is to reflect the actual absorption efficiency of fertilizer in the soil.

[0092] Further, the crop nutrient demand parameter is obtained by querying the crop nutrient demand table, which is constructed based on historical experimental data, including nitrogen, phosphorus and potassium demand values of different crop types (such as corn and wheat, etc.) at different yield levels; the soil existing nutrient amount is determined by soil sampling analysis, including nitrogen, phosphorus and potassium content, and is stored in the standardized data set; the straw released nutrient amount is based on the nutrient release prediction output by the straw soil microbial dynamic decomposition model, simulating the release dynamics of nutrients during straw decomposition; the fertilizer utilization rate is calibrated by field test, considering the influence of soil pH value, organic matter content and fertilizer type (such as urea and compound fertilizer, etc.).

[0093] It should be noted that the fertilizer formula is determined based on the actual fertilization amount and the nutrient ratio, by querying the pre-set fertilizer database to match the optimal fertilizer combination, wherein the fertilizer database contains the nutrient content (such as nitrogen, phosphorus and potassium percentage) and applicable conditions (such as soil type and pH range, etc.) of various fertilizers.

[0094] Further, the fertilizer database is constructed through laboratory analysis and field trials, including fertilizer name, nutrient composition, solubility and environmental adaptability, and the matching process adopts the minimum cost algorithm or nutrient balance algorithm to ensure economy and effectiveness.

[0095] It should be noted that the determination of the fertilization time period is based on the crop growth model and real-time environmental data, the peak of crop nutrient absorption is predicted through the growth model, and the appropriate fertilization time is selected in combination with the weather forecast to avoid rain washing or high temperature evaporation.

[0096] Further, the crop growth model adopts a physiological development stage model, and the real-time environmental data includes temperature, humidity and rainfall prediction, which are obtained in real time from the weather station and integrated into the decision through data fusion technology.

[0097] The application calculates the fertilization amount based on the optimized decision, converts the decision into an operable prescription, provides variable fertilization recommendation and the best window period, realizes the precision and automation of fertilization, generates a precise prescription and outputs it to a display terminal, ensures the pertinence and balance of field operation, reduces human intervention, and improves the utilization efficiency and sustainability of agricultural resources.

[0098] Please refer to Figure 2 The application provides a corn straw returning to field fertilization method system in the second aspect.

[0099] The corn straw returning to field fertilization method and system provided by the application break through the limitation of traditional methods on insufficient description of field heterogeneity based on multi-source data fusion and spatial interpolation algorithm, construct a high-fidelity virtual farmland environment, realize quantitative prediction of straw decomposition rate and nutrient release amount by embedding a straw soil microbial dynamic decomposition model, provide a scientific basis for fertilization timing and amount, realize dynamic matching of nutrient supply and demand and adaptive optimization of fertilization strategy by combining state space construction and strategy learning algorithm, significantly improve the fertilizer utilization efficiency, reduce the risk of environmental pollution, generate a precise fertilization prescription and output it to a display terminal, and improve the intelligent level and operation convenience of agricultural production.

[0100] The corn straw returning to field fertilization method system 100 can be installed in an electronic device. According to the realized function, the corn straw returning to field fertilization method system 100 can include a basic data preprocessing module 101, a digital twin construction module 102, a dynamic decomposition model embedding module 103, an optimized decision generation module 104 and a fertilization prescription output module 105. The modules of the application can also be called units, which refer to a series of computer program segments that can be executed by an electronic device processor and can complete a fixed function, which are stored in the memory of the electronic device.

[0101] In the embodiment, the functions of the modules / units are as follows: the basic data preprocessing module performs data cleaning and weighted fusion analysis based on the pre-acquired basic data to generate a standardized data set.

[0102] The digital twin construction module constructs a plot digital twin based on the standardized data set, a preset geographic information system (GIS) map and hyperspectral image data through a spatial interpolation algorithm to form a high-fidelity virtual farmland simulation environment.

[0103] The dynamic decomposition model embedding module is configured to embed a preset straw-soil microorganism dynamic decomposition model in the plot digital twin and input the basic data to simulate straw decomposition rate and nutrient release to output decomposition rate prediction and nutrient release amount prediction.

[0104] The optimization decision generation module is configured to construct a state space based on the decomposition rate prediction and nutrient release amount prediction to integrate nutrient conditions, learn an optimal strategy and generate an optimized fertilization decision.

[0105] The fertilization prescription output module is configured to calculate a fertilization amount and generate a precise fertilization prescription based on the optimized fertilization decision and output the precise fertilization prescription to a display terminal.

[0106] In the embodiments provided in the present application, it should be understood that the disclosed method and system can be implemented in other manners. For example, the system embodiments described above are merely illustrative. For example, the division of the modules is merely a logical function division. In actual implementation, another division manner can be used.

[0107] The modules illustrated as separated components can or can not be physically separated, and the components illustrated as modules can or can not be physical units, i.e., can be located in one place or distributed on a plurality of network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments.

[0108] In addition, each functional module in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of hardware plus software functional modules.

[0109] It is obvious for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application.

[0110] The embodiments of the present application can acquire and process related data based on artificial intelligence technology. The artificial intelligence is a theory, method, technology and application system for simulating, extending and expanding human intelligence by using a digital computer or a machine controlled by a digital computer, perceiving an environment, acquiring knowledge and using the knowledge to obtain optimal results.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A method for returning corn straw to fields for fertilization, characterized in that: include: Step 1: Basic data preprocessing: Based on the pre-acquired basic data, perform data cleaning and weighted fusion analysis to generate a standardized data set; Step 2: Digital twin construction: Based on the standardized dataset, the preset geographic information system (GIS) map, and the hyperspectral image data, a digital twin of the plot is constructed using a spatial interpolation algorithm to form a high-fidelity virtual farmland simulation environment. Step 3: Embed a dynamic decomposition model: Embed a preset straw soil microbial dynamic decomposition model into the plot digital twin, input basic data to simulate straw decomposition rate and nutrient release, and output decomposition rate prediction and nutrient release prediction; Step 4: Optimization decision generation: Based on the decomposition rate prediction and nutrient release prediction, a state space is constructed to integrate nutrient status, and then the optimal strategy is learned to generate an optimized fertilization decision; Step 5: Outputting a fertilization prescription: Based on the optimized fertilization decision, the amount of fertilizer to be applied is calculated and a precise fertilization prescription is generated, which is then output to a display terminal.

2. The method for returning corn straw to fields and applying fertilizer according to claim 1, wherein: The basic data includes soil data, straw data and environmental data.

3. The method for returning corn straw to fields and applying fertilizer according to claim 1, wherein: The data cleaning and weighted fusion analysis based on the pre-acquired basic data are performed to generate a standardized data set, including: Based on soil data, straw data and environmental data, a data cleaning operation is performed to obtain a cleaned data set, wherein the data cleaning operation includes missing value processing, outlier removal and data standardization; based on the preset weight coefficient of the data source accuracy, a weighted fusion analysis is performed on the cleaned data set to generate a standardized data set, wherein the weighted fusion analysis is performed based on the weighted average formula.

4. The method for returning corn straw to fields and applying fertilizer according to claim 1, wherein: The digital twin of the plot is constructed by using a spatial interpolation algorithm to form a high-fidelity virtual farmland simulation environment, including: Based on the standardized data set, the preset geographic information system (GIS) map and the hyperspectral image data, data alignment and coordinate unification processing are performed to generate an integrated spatial data set; based on the integrated spatial data set, interpolation calculations are performed on missing or low-resolution areas using the Kriging spatial interpolation method to generate a high-resolution spatial distribution map; based on the high-resolution spatial distribution map and the preset geographic information system (GIS) map, a digital twin of the plot is generated using three-dimensional modeling technology, which includes real-time soil data, straw data and environmental data, to form a high-fidelity virtual farmland simulation environment.

5. The method for returning corn straw to fields and applying fertilizer according to claim 1, characterized in that: The preset straw soil microbial dynamic decomposition model is embedded in the plot digital twin, and basic data is input to simulate the straw decomposition rate and nutrient release to output the decomposition rate prediction and nutrient release amount prediction, including: A1. Initialize model parameters. Based on the digital twin of the plot, load the preset model parameters of the straw soil microbial dynamic decomposition model. The model parameters include the microbial activity coefficient, straw carbon-nitrogen ratio threshold, temperature sensitivity coefficient, and humidity influencing factor. A2. Basic data input is performed, and basic data parameters are extracted from the standardized data set, including soil temperature, soil moisture, soil pH value, straw carbon-nitrogen ratio, and microbial biomass, and the basic data parameters are input into the dynamic decomposition model; A3. Simulate decomposition rate and nutrient release: Based on the dynamic decomposition model of straw-soil microorganisms, calculate the straw decomposition rate using a decomposition rate algorithm that integrates temperature, humidity, and microbial activity parameters. Based on the decomposition rate prediction, calculate the nutrient release amount using a nutrient release algorithm that links straw nutrient content and decomposition rate. A4. Output decomposition rate prediction and nutrient release prediction.

6. The method for returning corn straw to fields and applying fertilizer according to claim 5, characterized in that: The method of calculating the straw decomposition rate by using a decomposition rate algorithm comprises: By calculating the formula Straw decomposition rate ,in Expressed as a baseline decomposition rate constant, preset based on straw type and soil properties; Expressed as a temperature influence function, Expressed as humidity influence function, Expressed as a function of microbial activity.

7. The method for returning corn straw to fields and applying fertilizer according to claim 5, characterized in that: The method of calculating the nutrient release amount prediction based on the decomposition rate prediction by using a nutrient release algorithm includes: By calculating the formula Determine the nitrogen nutrient release , phosphorus nutrient release and potassium nutrient release , 、 and They are expressed as the nitrogen nutrient content, phosphorus nutrient content and potassium nutrient content of straw, Expressed as the time period corresponding to nutrient release.

8. The method for returning corn straw to fields and applying fertilizer according to claim 1, characterized in that: The state space is constructed based on the decomposition rate prediction and the nutrient release amount prediction to integrate the nutrient status, and then the optimal strategy is learned to generate an optimized fertilization decision, including: B1. Based on the decomposition rate prediction and nutrient release prediction, construct a state space, wherein the state variables include real-time soil nutrient content and predicted nutrient release to integrate nutrient status; B2. Define an action space, where actions include selecting a fertilization type, setting a fertilization amount, and determining a fertilization time period as fertilization decision options; design a reward function that is optimized based on nutrient utilization, economic benefits, and environmental benefits to calculate a reward value for the simulation; and use a reinforcement learning algorithm to conduct simulations in the digital twin of the plot based on the state space, action space, and reward function to learn the optimal fertilization strategy. B3. Generate optimized fertilization decisions based on the learned optimal fertilization strategy, including fertilizer formula, fertilizer amount, and fertilization time period.

9. The method for returning corn straw to fields and applying fertilizer according to claim 1, characterized in that: The method of calculating the amount of fertilizer to be applied and generating a precise fertilizer prescription based on the optimized fertilization decision, and then outputting the result to a display terminal, includes: Based on the optimized fertilization decision, the target nutrient amount and crop nutrient requirement parameters are extracted, and the fertilization amount is calculated based on the existing nutrient amount in the soil and the nutrient amount released by the straw to obtain the actual fertilization amount. The fertilizer formula is then determined based on the actual fertilization amount and the fertilizer database. At the same time, the fertilization time period is determined based on the crop growth model and real-time environmental data. Finally, a precise fertilization prescription is generated and output to the display terminal.

10. A system for implementing the method for returning corn straw to fields and applying fertilizer according to any one of claims 1 to 9, characterized in that: include: The basic data preprocessing module performs data cleaning and weighted fusion analysis based on the pre-acquired basic data to generate a standardized data set; A digital twin construction module, based on the standardized data set, a preset geographic information system (GIS) map, and hyperspectral image data, constructs a digital twin of the plot through a spatial interpolation algorithm to form a high-fidelity virtual farmland simulation environment; The dynamic decomposition model embedding module is used to embed the preset straw soil microbial dynamic decomposition model in the plot digital twin, and input basic data to simulate the straw decomposition rate and nutrient release to output the decomposition rate prediction and nutrient release amount prediction; an optimization decision generation module, which constructs a state space based on the decomposition rate prediction and the nutrient release amount prediction to integrate the nutrient status, and then learns the optimal strategy to generate an optimized fertilization decision; The fertilization prescription output module calculates the fertilization amount and generates a precise fertilization prescription based on the optimized fertilization decision, and then outputs it to the display terminal.

Images