Organic fertilizer application method and system for dynamic management of farmland nutrients

Abstract

The invention relates to the technical field of farmland nutrient management, and discloses an organic fertilizer application method and system for dynamic management of farmland nutrients, and the method comprises the steps: obtaining and standardizing original data of a farmland environment, extracting features to generate farmland feature vectors, and completing nutrient demand recognition and fertilization type analysis; integrating the data to generate an enhanced farmland matrix, performing multi-dimensional nutrient analysis, calculating application necessity and an environmental influence value, and generating decision data after verification; constructing an enhanced fertilizer feature space based on an organic fertilizer library and the decision data, completing fertilizer matching, proportioning and application path optimization, and outputting pre-inspection report data; and collecting application data to carry out anomaly detection and adaptive learning. The system comprises a processor and a memory, and instructions stored in the memory can implement the method. According to the invention, through multi-dimensional data integration and intelligent decision making, the precision and scientificity of farmland nutrient management are improved, and the system is suitable for precise fertilization requirements of modern agriculture.

Description

Technical Field

[0001] The present invention relates to the technical field of farmland nutrient management, and in particular to an organic fertilizer application method and system for dynamic management of farmland nutrients. Background Art

[0002] In agricultural production, farmland nutrient management is a key link in ensuring healthy crop growth and improving yield and quality. Nutrient monitoring, in particular, for medium- and low-yield farmland is particularly important for improving the physical and chemical properties of farmland and increasing crop yields. Traditional organic fertilizer application methods are often based on experience or simple soil testing, lacking precise response to dynamic changes in the farmland environment, leading to a prominent imbalance in nutrient supply and demand. On the one hand, excessive fertilization can cause environmental problems such as soil compaction and groundwater contamination, increasing agricultural production costs; on the other hand, insufficient fertilization can restrict crop growth, affecting yield and quality. With the development of smart agriculture, how to use information technology to achieve dynamic and precise management of farmland nutrients has become a research hotspot in the current agricultural field.

[0003] In existing technologies, farmland fertilization management systems mostly remain at the static data processing level, making it difficult to comprehensively consider multi-dimensional dynamic factors such as soil characteristics, crop growth stages, and climatic conditions. For example, when extracting soil characteristics, traditional methods can only perform simple nutrient content detection and are unable to construct a complete farmland feature vector to fully reflect the farmland conditions; in the fertilization decision-making process, there is a lack of in-depth mining of historical data and multiple verification mechanisms, resulting in insufficient scientificity and reliability of the decision-making; in terms of fertilizer matching and application path optimization, it is difficult to achieve the organic integration of dynamic features and associated features, and it is impossible to meet the personalized needs of different farmland scenarios. In addition, existing systems generally lack the ability to monitor and dynamically optimize the fertilization process in real time, making it difficult to cope with sudden changes in the field environment, resulting in poor adaptability of fertilization plans.

[0004] With the advancement of sensor technology, big data analytics, and machine learning algorithms, the field of farmland nutrient management urgently needs an organic fertilizer application method and system that can integrate dynamic data from multiple sources and achieve intelligent decision-making throughout the entire process. This approach must address core issues inherent in traditional technologies, such as extensive data processing, a single decision-making model, and insufficient dynamic adaptability. By building a refined farmland characteristic analysis system, a multi-dimensional nutrient analysis model, and an intelligent fertilization decision-making process, it aims to achieve precise, efficient, and sustainable organic fertilizer application, meeting the multiple demands of modern agriculture for resource conservation, environmental friendliness, and high yield and quality. Summary of the Invention

[0005] The purpose of the present invention is to provide an organic fertilizer application method and system for dynamic management of farmland nutrients to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: an organic fertilizer application method for dynamic management of farmland nutrients, the method comprising: S1. Obtain raw farmland environmental data and standardize it to obtain standardized farmland data. Based on the standardized farmland data, extract soil characteristics, calculate crop growth characteristics and climate characteristics, and generate farmland feature vectors. Based on the farmland feature vectors, identify nutrient requirements and analyze fertilization types to obtain preliminary analysis results. S2. Integrate the preliminary analysis result data and the standardized farmland data to generate an enhanced farmland matrix; perform multi-dimensional nutrient analysis based on the enhanced farmland matrix to obtain nutrient analysis result data; analyze the nutrient analysis result data to calculate the necessity score and environmental impact value of organic fertilizer application to form fertilization decision data; perform multiple verifications on the fertilization decision data to generate verified decision data; S3. Based on the pre-stored original information of the organic fertilizer library and the verified decision data, an enhanced fertilizer feature space is constructed; based on the enhanced fertilizer feature space, fertilizer matching and ratio optimization are performed to generate an optimized fertilizer selection plan; the application path of the optimized fertilizer selection plan is optimized to form an optimized application plan; based on the optimized application plan, a multi-dimensional pre-inspection is performed before application, and finally the pre-inspection report data is output.

[0007] Preferably, step S1 is further: S11. Obtaining raw farmland environmental data including soil samples, timestamps, plot identifiers, and crop identifiers; converting the soil samples in the raw farmland environmental data into a unified coding format, removing outliers and redundant information, and obtaining processed input data; performing data length normalization on the processed input data to generate standardized farmland data; S12. Based on the standardized farmland data, soil parameter values are calculated, key nutrient sets are extracted, crop types are identified, and basic characteristic data are generated; based on the basic characteristic data, a preconfigured growth analysis model is used to calculate the probability distribution of nutrient requirements and growth vectors of the crops to obtain growth characteristic data; historical records of the farmland are obtained, and based on the historical records, plot status information is extracted; the plot status information is combined with the growth characteristic data to form a farmland characteristic vector; S13. Based on the farmland feature vector, use the preset feature-demand mapping rules to identify the nutrient demand type, calculate the demand priority score, and generate demand feature data; analyze the demand feature data, estimate the fertilizer resource demand value and the time window demand value, and obtain the resource demand data; based on the demand feature data and the resource demand data, identify the dependency relationship between the demands and construct a dependency graph; based on the nutrient demand type, priority score, resource demand data and the dependency graph, generate preliminary analysis result data.

[0008] Preferably, step S2 is further: S21. Converting the preliminary analysis result data into a feature matrix, converting the standardized farmland data into a vector representation, and combining the feature matrix and the vector representation to form an initial farmland matrix; extracting relevant historical farmland records and calculating historical information weights; fusing the historical information weights with the initial farmland matrix to generate an enhanced farmland matrix; S22. Analyze the enhanced farmland matrix to identify the primary nutrient target; decompose the primary nutrient target into a set of sub-targets, construct a target dependency graph, and obtain target structure data; calculate the resource requirement vector and target priority matrix for each sub-target to generate target resource data; analyze fertilizer characteristic requirements based on the primary nutrient target and historical farmland records, and calculate fertilizer importance weights; based on the fertilizer importance weights, integrate the target structure data and target resource data into nutrient analysis result data; S23. Based on the nutrient analysis result data, calculate the necessity score of organic fertilizer application, evaluate the application risk value, and obtain application assessment data; based on the application assessment data, determine the application timing and generate an application priority list; based on the application priority list, formulate an adjustment strategy to form application strategy data; integrate the application assessment data and the application strategy data to generate an application path map; based on the application path map, calculate the confidence score and ultimately form fertilization decision data; S24. Perform internal consistency verification on the fertilization decision data to generate consistency verification data; based on the consistency verification data, verify resource availability, check technical constraints and seasonal restrictions, and obtain feasibility assessment data; based on the consistency verification data and feasibility assessment data, calculate the verification score, mark risk points, and generate optimization suggestions; based on the optimization suggestions, form verified decision data.

[0009] Preferably, step S3 is further: S31. Based on the pre-stored original information of the organic fertilizer library and the verified decision data, the functional feature vector, attribute index vector, and resource demand vector of each fertilizer are extracted to generate static feature data; the historical effect matrix, average application time vector, and resource consumption distribution of the fertilizer are calculated to form dynamic feature data; a fertilizer dependency graph is constructed, the fertilizer compatibility matrix and the fertilizer combination effect tensor are calculated to obtain associated feature data; the static feature data, dynamic feature data, and associated feature data are integrated into an enhanced fertilizer feature space; S32. Calculating functional suitability based on the enhanced fertilizer feature space; performing attribute constraint filtering based on the functional suitability to generate an initial set of candidate fertilizers; obtaining contextual features of the current farmland and calculating a contextual relevance score; adjusting the weights of the candidate fertilizers based on the contextual relevance score, and reordering the initial set of candidate fertilizers to obtain an optimized set of candidate fertilizers; constructing a set of feasible fertilizer combinations and calculating a combination synergy score; selecting an optimal combination scheme from the optimized set of candidate fertilizers based on the combination synergy score to form an optimized fertilizer selection scheme; S33. Based on the optimized fertilizer selection plan, construct an application dependency graph, calculate the critical path, and generate a parallel application plan; based on the parallel application plan, form application sequence data; based on the application sequence data, construct a resource allocation matrix and optimize the application timing; based on the optimized application timing, construct a cache strategy and obtain resource optimization data; based on the resource optimization data, construct a failure handling strategy, alternative plans, and a monitoring point set to generate fault tolerance mechanism data; integrate the application sequence data, resource optimization data, and fault tolerance mechanism data into the optimized application plan; S34. Perform an online status check on the fertilizer in the optimized application plan, verify resource adequacy, test interface response, and generate availability verification data; based on the availability verification data, perform permission check, risk assessment, and compliance verification to form security assessment data; based on the security assessment data, estimate response time, predict resource consumption, calculate success probability, and obtain performance prediction data; integrate the availability verification data, safety assessment data, and performance prediction data into pre-inspection report data.

[0010] Preferably, step S12 is further: S121. Read soil segments from the standardized farmland data, calculate parameter values, nutrient values, and moisture values for each soil segment, and generate soil statistical data; based on the soil statistical data, use analysis tools to segment the soil segments, obtain nutrient statistics, and generate nutrient characteristic data; combine the soil statistical data and nutrient characteristic data to construct complete basic characteristic data; S122. Obtain nutrient information from the basic characteristic data. Based on the nutrient information, calculate the environmental association strength of each nutrient using the bidirectional association network of the model to generate nutrient-level growth association data. Based on the nutrient-level growth association data, construct a growth similarity matrix, calculate key growth units, and generate growth unit data. Map the growth unit data to a predefined demand space, calculate the demand probability distribution, and obtain growth characteristic data. S123. Obtain the operation sequence in the pre-stored farmland historical records, construct a time series feature vector, and generate historical farmland data; obtain and analyze the current plot status, including the plot usage time, operation rounds, and environmental continuity, to form plot status data; perform feature fusion on the growth feature data, historical farmland data, and plot status data, and output the final farmland feature vector.

[0011] Preferably, step S22 is further: S221. Read nutrient description information in the enhanced farmland matrix, construct an environmental dependency tree based on the nutrient description information using a model, extract core action nodes, and generate action sequence data; identify key nutrient targets based on the action sequence data, calculate the strength of logical relationships between targets, and form target association data; combine the action sequence data and the target association data to output nutrient target data; S222. Based on the nutrient target data, pattern matching is performed using a predefined model target decomposition template library to identify decomposable sub-target units and generate an initial sub-target set; the execution conditions and completion criteria of each sub-target in the initial sub-target set are analyzed, a sub-target constraint relationship diagram is constructed, and target constraint data is obtained; based on the target constraint data, the initial sub-target set is optimized and reorganized to output sub-target sequence data; S223. Based on the sub-goal sequence data, extract the input-output dependency relationship of each sub-goal, construct a data flow graph, and generate data dependency data; based on the data dependency data, analyze the execution order constraints, identify parallel execution opportunities, construct a target execution network, and generate execution dependency data; integrate the data dependency data and execution dependency data to construct a complete target dependency graph; S224. Obtain historical execution records, and based on the sub-goal sequence data and historical execution records, calculate the processing complexity and resource consumption characteristics of each sub-goal to generate resource characteristic data; based on the resource characteristic data, analyze the time sensitivity and priority factors of the sub-goals, construct a target scheduling weight matrix, and form scheduling characteristic data; combine the resource characteristic data, scheduling characteristic data and the target dependency graph to output the final nutrient analysis result data.

[0012] Preferably, step S23 is further: S231, reading resource demand information from the nutrient analysis result data, calculating a resource utilization threshold based on the resource demand information and pre-stored historical application records, and generating resource assessment data; analyzing the target completion time requirement based on the resource assessment data, calculating a time pressure coefficient in combination with the current load state of the system, and forming time assessment data; calculating an organic fertilizer application necessity score matrix based on the resource assessment data and the time assessment data, and outputting application necessity data; S232. Based on the application necessity data, extract characteristic patterns of historical application failure cases, construct risk characteristic vectors, and generate risk pattern data; based on the risk pattern data, analyze the similarity between the current target and pre-stored historical high-risk scenarios, calculate multi-dimensional risk coefficients, and form risk assessment data; based on the risk pattern data and risk assessment data, construct a risk-benefit assessment matrix and output application risk data; S233. Based on the application necessity data and the application risk data, construct an organic fertilizer application time series network, calculate the optimal application time window, and generate application time series data; based on the application time series data, analyze the priority dependency relationship between fertilizers, establish an application priority queue, and generate priority data; combine the application time series data and the priority data to construct an application execution plan and output application strategy data; S234. Based on the application strategy data and pre-stored historical failure processing records, a fault processing decision tree is constructed to generate fault recovery data; based on the fault recovery data, a multi-level adjustment plan is constructed, including alternative fertilizer chains and simplified strategies, to form adjustment strategy data; the application strategy data, fault recovery data, and adjustment strategy data are integrated, the strategy reliability score is calculated, a complete application path map is constructed, and finally the fertilization decision data is output.

[0013] Preferably, step S32 is further: S321. Based on the functional feature vectors and decision requirements in the enhanced fertilizer feature space, the functional applicability score of each fertilizer is calculated using the model to generate functional applicability data. Based on the functional applicability data, the attribute indicator vector is used to perform constraint filtering to select a set of fertilizers that meet the attribute requirements, thereby forming attribute filtering data. Combining the functional applicability data and the attribute filtering data, a preliminary fertilizer list and its scoring matrix are constructed, and initial candidate data are output. S322. Obtain contextual features of the current farmland, including seasonal window, resource status, and target priority, and generate contextual feature data. Based on the contextual feature data, analyze the fertilizer application effects under similar historical scenarios, construct a scenario correlation matrix, and generate scenario matching data. Calculate a context adjustment coefficient based on the contextual feature data and scenario matching data, and output context scoring data. S323. Based on the initial candidate data and contextual scoring data, a dynamic weight algorithm is used to adjust the fertilizer score to generate adjusted weight data; based on the historical success rate and stability index of the fertilizer, the fertilizer credibility score is updated to form credibility data; based on the adjusted weight data and credibility data, the preliminary fertilizer list is reordered to output optimized candidate data; S324. Based on the fertilizer combination feature tensors in the optimization candidate data, construct a feasible fertilizer combination scheme set and generate combination scheme data; based on the combination scheme data, calculate the synergistic effect scores of different combination schemes, including functional complementarity and attribute gain, to form synergistic evaluation data; based on the synergistic evaluation data, analyze the complexity and risk factors of the combination scheme data, construct a comprehensive evaluation matrix, and obtain scheme evaluation data; S325. Based on the combination scheme data, collaborative evaluation data and scheme evaluation data, a multi-objective optimization algorithm is used to calculate the comprehensive score of each combination scheme and generate optimization scoring data; based on the optimization scoring data, the optimal combination scheme is selected, and a detailed fertilizer application sequence is constructed to form application sequence data; the optimization scoring data and application sequence data are integrated, and finally the optimized fertilizer selection scheme is output.

[0014] Preferably, the method further comprises: S4. Collect the execution data flow and historical monitoring data during the application process of organic fertilizer to generate a comprehensive monitoring data packet; based on the comprehensive monitoring data packet, perform anomaly detection and early warning analysis, and output the anomaly analysis result data; based on the anomaly analysis result data and the comprehensive monitoring data packet, generate a dynamic optimization strategy to form an optimization strategy set data; based on the optimization strategy set data and the pre-stored historical optimization effect data, perform adaptive learning, and finally output the optimization update data packet.

[0015] Preferably, the present invention also includes an organic fertilizer application system for dynamic management of farmland nutrients, the system comprising: at least one processor; and, a memory communicatively coupled to at least one of the processors; The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned organic fertilizer application method for dynamic management of farmland nutrients.

[0016] Compared with the prior art, the present invention has the following beneficial effects: At the data processing level, raw farmland environmental data is acquired and standardized, extracting soil, crop growth, and climate characteristics to generate a comprehensive farmland feature vector. This process not only achieves unified coding of soil samples and removes outliers, but also calculates the probability distribution of crop nutrient requirements through growth analysis models. Combined with historical records, a complete farmland feature vector is constructed, ensuring data integrity and accuracy, providing a solid foundation for subsequent nutrient requirement identification and fertilization type analysis.

[0017] During the decision analysis phase, an enhanced farmland matrix is generated by integrating preliminary analysis results with standardized farmland data, enabling multi-dimensional nutrient analysis. This process incorporates historical information weights to achieve data fusion, decompose nutrient targets, construct a target dependency graph, calculate resource demand vectors and a target priority matrix, and simultaneously assess the necessity and environmental impact of organic fertilizer application. A multi-factor validation mechanism ensures the scientific and reliable nature of the decision. This multi-dimensional analytical framework comprehensively considers the dynamic needs and potential risks of farmland, avoiding decision-making biases caused by a single factor.

[0018] For fertilizer matching and application optimization, an enhanced fertilizer feature space, comprising static, dynamic, and associated features, is constructed based on pre-stored raw information from the organic fertilizer library and verified decision data. Functional suitability calculations, attribute constraint filtering, and contextual relevance analysis enable precise fertilizer matching and combination optimization. Furthermore, an application dependency graph and fault-tolerance mechanism are constructed to optimize application timing and resource allocation, ensuring efficient execution and risk mitigation of fertilization plans. This intelligent matching and optimization process significantly improves fertilizer utilization, reduces resource waste, and reduces environmental risks.

[0019] In terms of system scalability, the system collects execution data streams and historical monitoring data during the fertilization process to enable anomaly detection, dynamic optimization, and adaptive learning. This functionality enables the system to adjust fertilization strategies in real time based on changes in the actual field environment, continuously optimize decision-making models, and enhance the system's adaptability and long-term performance, providing a technical path for continuous improvement in the dynamic management of farmland nutrients.

[0020] Through a data-driven intelligent decision-making process, the present invention realizes automated management of the entire process from farmland characteristic analysis, nutrient demand identification to fertilizer matching and application optimization, effectively solving the problems of extensive data processing, insufficient scientific decision-making and poor dynamic adaptability in traditional methods, and provides an efficient and reliable technical solution for precision fertilization in modern agriculture, with significant economic value and environmental benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a working principle diagram of the organic fertilizer application method for dynamic management of farmland nutrients according to the present invention; Figure 2 Flowchart for the generation and preliminary analysis of farmland feature vectors; Figure 3 Flowchart generated for enhanced field matrix analysis and fertilization decision making. DETAILED DESCRIPTION

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0023] See also Figure 1-Figure 3 The present invention relates to an organic fertilizer application method for dynamic management of farmland nutrients, and the specific implementation steps are as follows: Step S1: First, raw farmland environmental data, including soil samples, timestamps, plot identifiers, and crop identifiers, is acquired. The soil samples are converted to a unified coding format, and outliers and redundant information are removed to obtain processed input data. This data is then length-normalized to generate standardized farmland data. Based on the standardized farmland data, soil parameter values are calculated, key nutrient sets are extracted, and crop types are identified to generate basic feature data. A preconfigured growth analysis model is used to calculate the probability distribution of crop nutrient requirements and growth vectors to generate growth feature data. Farmland historical records are obtained, and plot status information (such as plot usage time, operation cycles, and environmental continuity) is extracted. This information is combined with the growth feature data to form a farmland feature vector. Furthermore, based on the farmland feature vector, pre-set feature-requirement mapping rules are used to identify nutrient requirement types, calculate requirement priority scores, and generate requirement feature data. The requirement feature data is analyzed to estimate fertilizer resource requirements and time window requirements to generate resource requirement data. Dependencies between requirements are identified and a dependency graph is constructed. Finally, the nutrient requirement types, priority scores, resource requirement data, and dependency graph are combined to generate preliminary analysis results data.

[0024] Step S2: The preliminary analysis results are converted into a feature matrix, and the standardized farmland data are converted into a vector representation. These are combined to form an initial farmland matrix. Relevant historical farmland records are extracted and historical information weights are calculated. This information is then integrated with the initial farmland matrix to generate an enhanced farmland matrix. The enhanced farmland matrix is analyzed to identify the primary nutrient target, decompose it into a set of sub-targets, and construct a target dependency graph to obtain target structure data. The resource requirement vector and target priority matrix are calculated for each sub-target to generate target resource data. Based on the primary nutrient target and historical farmland records, fertilizer feature requirements are analyzed and fertilizer importance weights are calculated. The target structure data and target resource data are integrated into the nutrient analysis result data. Furthermore, based on the nutrient analysis result data, the necessity score of organic fertilizer application is calculated, the application risk value is assessed, the application timing is determined, and an application priority list is generated. Adjustment strategies are formulated and integrated into fertilization decision data. The fertilization decision data is verified for internal consistency, resource availability, technical constraints, and seasonal restrictions. Verification scores are calculated, risk points are identified, and optimization recommendations are generated, ultimately forming the verified decision data.

[0025] Step S3: Based on the pre-stored raw information of the organic fertilizer library and the verified decision data, the functional feature vector, attribute index vector, and resource requirement vector of each fertilizer are extracted to generate static feature data. The historical effect matrix, average application time vector, and resource consumption distribution of the fertilizer are calculated to form dynamic feature data. A fertilizer dependency graph is constructed, and the fertilizer compatibility matrix and combination effect tensor are calculated to obtain associated feature data, which are integrated to form an enhanced fertilizer feature space. Based on the enhanced fertilizer feature space, the functional suitability is calculated and attribute constraint filtering is performed to generate an initial set of candidate fertilizers. The contextual features of the current farmland (such as seasonal window, resource status, target priority, etc.) are obtained, the contextual relevance score is calculated, the candidate fertilizer weights are adjusted, and the reordering is performed to obtain an optimized set of candidate fertilizers. A set of feasible fertilizer combinations is constructed, the combination synergy score is calculated, and the optimal combination scheme is selected to form an optimized fertilizer selection scheme. Based on the optimized fertilizer selection plan, an application dependency graph is constructed, the critical path is calculated, and a parallel application plan is generated to form application sequence data; a resource allocation matrix is constructed to optimize the application timing, and the cache strategy, failure handling strategy, alternative plan and monitoring point set are combined to integrate into an optimized application plan; the optimized application plan is subjected to online status check, resource adequacy verification, interface response test, authority check, risk assessment, compliance verification, response time estimation, resource consumption prediction and success probability calculation, and integrated to generate pre-inspection report data.

[0026] The present invention will be further described below in conjunction with Examples 1 to 5: Example

[0027] This embodiment further defines the specific implementation of step S1. When acquiring raw farmland environmental data, multi-dimensional information, including soil samples, timestamps, plot identifiers, and crop identifiers, must be fully collected. Soil samples cover soil samples at different depths (e.g., 0-20 cm tillage layer, 20-40 cm subsoil layer). Timestamps are accurate to the minute to record the sampling time. Plot identifiers include the plot number and geographic coordinates (longitude and latitude). Crop identifiers include crop type (e.g., wheat, corn), variety name, and growth cycle stage (e.g., seedling stage, tillering stage). Crop type identification during basic feature data generation requires further refinement to the variety and growth stage. For example, image recognition technology can be used to analyze features such as seedling leaf morphology and color (input is RGB images captured by a farmland camera, classified using a pre-trained ResNet model). This verifies the accuracy of the original crop identifiers and supplements the variety information (e.g., "Jimai 22"). In addition, although the input data processing is mainly based on soil samples, the growth cycle stage information (such as "tillering period") in the crop identification needs to be integrated when generating standardized farmland data, and then associated with soil parameters to form a feature matrix in the spatiotemporal dimension.

[0028] When converting soil samples from the original farmland environment data into a unified coding format, standardized coding is performed for the physical and chemical property detection parameters: physical property parameters: soil bulk density (unit: g / cm³), porosity (%), moisture content (%), are encoded in floating-point format; chemical property parameters: pH value (dimensionless), organic matter content (g / kg), total nitrogen (g / kg), available phosphorus (mg / kg), and available potassium (mg / kg) are uniformly stored in the key-value pair format of "parameter name_value" (such as "pH_7.2" and "organic matter_18.5").

[0029] When outliers are removed through statistical methods, the basis for determining outliers is the industry standard threshold value of the above-mentioned detection parameters. For example, the reasonable range of organic matter content is set to ≤100g / kg (according to NY / T 1121.6-2006). Values outside the range are regarded as outliers and eliminated. Redundant information processing is carried out by identifying repeated sampling records of the same plot at the same time point (with timestamp and plot identifier as unique keys) and retaining the first detection value to obtain the processed input data.

[0030] When performing data length normalization on the processed input data, a fixed-length sliding window technique is used to account for differences in crop types and sampling periods. For example, for annual crops, such as wheat, which has a 120-day growth cycle, the entire growth cycle is divided into sampling windows every 10 days. Within each window, data such as soil parameters and crop growth indicators are uniformly populated into a 12-dimensional vector. Insufficient data is supplemented using linear interpolation to generate standardized farmland data.

[0031] When calculating soil parameter values based on standardized farmland data, each soil sample must undergo physical and chemical property analysis. Soil parameters specifically include eight indicators: physical properties (bulk density, porosity, and water content) and chemical properties (pH, organic matter, and nitrogen, phosphorus, and potassium content). These indicators are populated into vectors using a fixed-length sliding window technique. For crop growth indicators (such as plant height and leaf area index), the data source is either crop identification information included in the raw data (such as manually entered growth monitoring data) or real-time data collected by IoT sensors. During standardization, these data are aligned with the soil parameter timestamps and merged to form a multidimensional vector. For example, for wheat with a growth cycle of 120 days, it is divided into 12 sampling windows, and the vector dimension of each window is "8 soil parameters + 3 growth indicators" (a total of 11 dimensions). The insufficient part is supplemented by linear interpolation; the chemical property parameters include pH value (glass electrode method), organic matter content (potassium dichromate oxidation method), total nitrogen content (Kjeldahl method), available phosphorus content (sodium bicarbonate extraction-molybdenum antimony colorimetry), available potassium content (ammonium acetate extraction-flame photometry), etc. The average value and standard deviation of each parameter are calculated as the statistical characteristic value of the soil.

[0032] To extract key nutrient sets, principal component analysis (PCA) is used to reduce the dimensionality of soil chemical parameters. Nutrient factors corresponding to principal components with cumulative contributions exceeding 85% are selected as key nutrients. For example, for wheat fields, nitrogen, phosphorus, and potassium are identified as the key nutrient set after analysis, and basic feature data containing the content and ratio of each nutrient is generated. Crop type identification is achieved through image recognition technology. High-definition cameras installed in the fields capture images of crop leaves at the seedling stage. These images are input into a pre-trained convolutional neural network (CNN) model for classification, and the output is a crop type label (such as "winter wheat").

[0033] When calculating crop growth characteristics using a preconfigured growth analysis model based on basic characteristic data, the growth characteristic data calculated by the preconfigured growth analysis model (such as the LSTM-GARCH hybrid model) includes two parts: the nutrient requirement probability distribution and the growth vector: Nutrient requirement probability distribution: Calculate the demand probability of key nutrients such as nitrogen, phosphorus, and potassium based on the crop growth period (e.g., jointing stage) (e.g., nitrogen fertilizer demand probability of 75%); Growth vector: This includes three indicators: plant height growth rate (cm / day), leaf area expansion rate (cm² / day), and biomass accumulation rate (g / day). It is generated by extracting the time series features of historical growth data through the model.

[0034] To obtain historical farmland records, it is necessary to collect fertilization records (fertilizer type, application rate, and application time), irrigation records (water volume and timing), crop yield data (yield per unit area, quality indicators), and pest and disease occurrence data from the past 3-5 years. When extracting plot status information, plot use time is determined by calculating the number of consecutive years of cultivation, and operation rounds are calculated by counting the number of crop crops planted in a year (e.g., one crop per year, two crops per year). Environmental continuity analysis assesses climate stability by calculating the coefficient of variation of annual average temperature and annual precipitation, generating plot status data with time series characteristics.

[0035] When combining plot status information with growth characteristic data to form a farmland feature vector, a multidimensional vector is constructed using feature concatenation. For example, the vector dimensions include soil nutrient content (3 dimensions), growth indicators (2 dimensions), plot usage time (1 dimension), operation cycles (1 dimension), and environmental continuity indicators (2 dimensions), for a total of 9 dimensions. Each dimension is normalized (scaled to the range [0, 1]) to eliminate dimensionality effects.

[0036] When identifying nutrient needs based on farmland feature vectors, pre-set feature-requirement mapping rules are constructed using a decision tree algorithm. The root node of the decision tree is the soil nitrogen content, and the branching condition is whether it falls below a critical value (e.g., 30 mg / kg). If so, the requirement is determined to be nitrogen deficiency; otherwise, the decision proceeds to the next node (soil phosphorus content). After identifying the nutrient requirement type (e.g., nitrogen and phosphorus deficiency, potassium excess) using this rule, the analytic hierarchy process (AHP) is used to calculate the requirement priority score. Nutrient requirements are weighted according to the urgency of the crop growth stage (e.g., the nitrogen requirement at the heading stage is weighted 0.5) to generate demand characteristic data.

[0037] When analyzing demand characteristic data to estimate fertilizer resource requirements, the fertilizer application rate is calculated based on crop nutrient absorption patterns (e.g., 3kg of nitrogen, 1.2kg of phosphorus, and 2.5kg of potassium are required to produce 100kg of wheat) and soil nutrient abundance and deficiency indicators, combined with target yield. The time window demand value is determined based on the crop growth calendar. For example, the wheat jointing period is the optimal time window for nitrogen fertilizer application, which lasts approximately 10 days. When identifying dependencies between requirements, the order of nutrient application is determined through causal analysis (e.g., applying phosphorus fertilizer first to improve the soil phosphorus base, followed by nitrogen fertilizer to promote growth). A directed acyclic graph (DAG) is used to construct a dependency graph, where nodes represent nutrient requirement types and edges represent dependency relationships (e.g., "phosphorus fertilizer application" → "nitrogen fertilizer application").

[0038] When the preliminary analysis result data is finally generated, the nutrient requirement type, priority score, resource requirement value (such as urea application rate of 15 kg / mu and superphosphate application rate of 20 kg / mu), time window (such as the 3rd to 12th day of the jointing period) and dependency graph are integrated and stored in a structured data format (such as JSON). It contains fields such as "demand type", "priority score", "fertilizer type", "application amount", "time window start date", "dependency", etc., providing a basis for subsequent data integration and fertilization decisions.

[0039] Data quality control mechanisms are established at every stage of data processing. For example, soil samples are collected using a three-point sampling method (at the diagonal vertices and midpoints of the plots) to ensure representativeness. During data cleaning, a log is kept of outlier removal for easy traceability. Model predictions are cross-validated with historical data to ensure the reliability of the analysis. The entire S1 process strictly adheres to the logical chain of data standardization, feature engineering, and requirements analysis, achieving systematic processing from raw data to preliminary analysis results, providing a scientific basis for dynamic nutrient management in farmland. Example

[0040] This embodiment further defines the specific implementation of step S2. During data integration and fertilization decision generation, the preliminary analysis results generated in step S1 must be structured and integrated with the standardized farmland data. This preliminary analysis data integrates demand characteristic data, resource demand data (e.g., fertilizer application rate), and a dependency graph, and is converted into a feature matrix. The rows of the matrix represent different farmland plots or sampling periods. Normalization is performed using spatiotemporal dimensions: if the row index is a plot, the columns contain the characteristics of the plot for each sampling period (e.g., soil nitrogen content and demand priority at different time points). Data is aligned using timestamps. If the row index is a sampling period, the columns contain the characteristics of all plots for that period, with missing values interpolated from neighboring plots. For example, for the demand priority data for three plots (A, B, and C) over four sampling periods (T1-T4), a 3×4 matrix is constructed, with rows corresponding to plots and columns corresponding to periods. The values represent the priority scores for each plot in the corresponding period, ensuring a structured mapping of data across different dimensions. The column dimensions of the feature matrix include "demand type," "priority score," and "resource demand value." At the same time, the soil parameters, crop types and other information in the standardized farmland data are converted into vector representations. Each vector corresponds to the multidimensional data of a plot or time point, such as a 10-dimensional vector containing indicators such as soil pH, organic matter content, and crop growth stage.

[0041] After combining the feature matrix and vector representation to form the initial farmland matrix, historical farmland records need to be introduced to enhance the temporal relevance of the data. Historical information includes fertilization records (fertilizer type, application amount, and time) from the past 3-5 years, soil nutrient monitoring data (monthly nitrogen, phosphorus, and potassium levels), crop yield data (yield per unit area), and climate data (annual average temperature and precipitation). The entropy weighting method is used to calculate the weight of historical information. The specific steps are as follows: Feature normalization: normalize historical soil nutrients, yield and other data to [0,1]; Entropy calculation: Calculate information entropy for each feature dimension to reflect the degree of data dispersion; Weight assignment: Weight = 1 - entropy. A higher weight indicates a greater impact of the historical feature on the current decision. During matrix fusion, principal component analysis (PCA) is used to reduce the dimensionality of historical information to the same dimension as the initial farmland matrix. For example, if the initial matrix is 10-dimensional, the historical information is reduced to a 10-dimensional weight vector. Weighted fusion is achieved through matrix dot multiplication (e.g., enhanced farmland matrix = initial matrix × historical weight diagonal matrix) to ensure dimensionality matching.

[0042] The entropy weighting method is used to calculate the weight of historical information, which reflects the reference value of historical records for current decision-making (for example, recent records are weighted higher than more distant records). The historical information weights are then weighted and integrated with the initial farmland matrix, for example, by matrix multiplication to assign weights. This generates an enhanced farmland matrix that incorporates historical trend characteristics and more comprehensively reflects the dynamic changes in farmland nutrient demand.

[0043] When performing multidimensional nutrient analysis on an enhanced farmland matrix, the primary nutrient objective is first identified through cluster analysis or topic modeling. For example, for an enhanced farmland matrix in a particular corn-growing region, clustering results revealed that 80% of the plots were nitrogen deficient, leading to the identification of "increasing soil nitrogen content" as the primary nutrient objective. Subsequently, the primary nutrient objective is broken down into a set of sub-objectives, such as "selecting the appropriate nitrogen fertilizer type," "determining the optimal application rate," and "planning the fertilization time window." Causal analysis is then used to construct a goal dependency graph, where nodes represent sub-objectives and edges represent the logical relationships between sub-objectives (e.g., "determining the application rate" depends on the results of "soil nitrogen content testing").

[0044] When calculating the resource requirement vector for each sub-goal, the type and quantity of resources required to achieve the sub-goal must be clearly defined. For example, the resource requirement vector for "Selecting the Appropriate Type of Nitrogen Fertilizer" includes a list of available fertilizer types, their supply cycles, and a cost budget. The goal priority matrix ranks sub-goals based on their impact on crop yield, with priority coefficients determined using expert scoring or the Analytic Hierarchy Process (AHP). (For example, the priority coefficient for "Fertilization Time Window Planning" is 0.4, higher than the 0.3 for "Fertilizer Cost Control"). Based on key nutrient targets and historical farmland records, fertilizer characteristic requirements (such as slow-release properties and compatibility with soil pH) are analyzed. Fertilizer importance weights are calculated using the Delphi method. For example, decomposed organic fertilizers are assigned a weight of 0.6 due to their long-term effectiveness, while chemical fertilizers are assigned a weight of 0.4. Ultimately, the goal structure data, goal resource data, and fertilizer importance weights are integrated to form nutrient analysis results that include a nutrient target hierarchy, resource requirements, and fertilizer compatibility.

[0045] During the fertilization decision data generation phase, the necessity score for organic fertilizer application is calculated based on the nutrient analysis results. The specific method is: combining resource demand information (such as the difference between the current soil nitrogen content and the target value) and historical application records, the resource utilization threshold is calculated using a linear weighted model. For example, when the soil nitrogen gap exceeds 20 mg / kg, the utilization threshold triggers the calculation of the necessity score. At the same time, the target completion time requirements (such as fertilization must be completed before the crop jointing period) and the current system load status (such as the available time of fertilization equipment) are analyzed to calculate the time pressure coefficient. The time pressure coefficient is calculated using the formula:

[0046] Among them, "remaining time" is the number of days from the current time to the jointing stage, and "target completion time threshold" is the fertilization window period specified by industry standards (such as 7 days before the jointing stage of wheat).

[0047] Resource utilization thresholds are calculated using the quantiles of historical usage records:

[0048] The administration necessity score matrix elements are calculated as:

[0049] in, 、 is the weight coefficient ( , ), and optimization is performed by inferring the historical decision-making results. The rows of this matrix correspond to different fertilizer types, the columns correspond to plots, and the values reflect the priority of applying different fertilizers to each plot.

[0050] Application risk assessment is based on the characteristic patterns of historical application failures, such as crop burns caused by improper fertilization timing. Risk characteristics such as soil moisture, temperature, and fertilizer type are extracted to construct a risk feature vector. A cosine similarity algorithm is used to analyze the match between current targets and historical high-risk scenarios. Multi-dimensional risk factors (such as climate risk and operational risk) are calculated, and a risk-benefit assessment matrix is constructed to output application risk data for each fertilizer plan.

[0051] When determining the timing of application, a time series network for organic fertilizer application is constructed based on application necessity data and risk data, and the optimal time window is calculated using the critical path method. For example, the soil moisture of a plot of land must reach 60%-70% before fertilizer can be applied. Combined with weather forecast data, the next 5-7 days are determined to be the optimal application period. When establishing an application priority queue, the priority dependencies between fertilizers are analyzed (for example, base fertilizer must be applied before sowing, and topdressing must be applied after the seedling stage), and a priority scheduling algorithm is used to generate a queue in which base fertilizer application has a higher priority than topdressing. Combining time series data and priority queues, an application execution plan is constructed that includes specific operational steps, responsible persons, and time nodes, such as "purchase organic fertilizer on the first day, deep tillage on the third day, and spreading fertilizer on the fifth day."

[0052] When building a fault handling mechanism, a decision tree algorithm is used to generate a fault handling process based on application strategy data and historical failure records. For example, if the fertilization equipment fails, the first-level response is to activate the backup equipment, and the second-level response is to adjust the fertilization method (such as manual spreading instead of mechanical application). Multi-level adjustment plans include alternative fertilizer chains (such as switching to biogas residue fertilizer when the main organic fertilizer is out of stock) and simplified strategies (such as reducing the number of fertilizations but ensuring that the total application amount remains unchanged). Integrate application strategy data, fault recovery data, and adjustment strategy data, and calculate the strategy reliability score through reliability theory. For example, the score ranges from 0 to 100 points based on factors such as the availability of backup equipment and the stability of the supply of alternative fertilizers. Finally, a complete application path map is constructed that includes risk points, response measures, and execution processes.

[0053] In the multiple verification phase, the fertilization decision data is first verified for internal consistency to check the logical compatibility between sub-goals. For example, whether there is a conflict between "applying high-phosphorus fertilizer" and "adjusting the soil pH to alkaline" is automatically verified through the rule engine to generate consistency verification data. Then, the resource availability is verified, including fertilizer inventory, equipment operating status, manpower allocation, etc. For example, it is confirmed that the selected organic fertilizer inventory is sufficient and the fertilization machinery is fault-free, and feasibility assessment data is generated. Based on the consistency and feasibility data, the verification score is calculated (such as a percentage score), high-risk points are marked (such as the backup fertilizer supplier is far away), and optimization suggestions are generated (such as signing an emergency supply agreement with the backup supplier in advance). Finally, verified decision data is formed to ensure the scientific nature and feasibility of the fertilization plan.

[0054] The implementation process of the entire step S2 runs through the logic of data integration, target decomposition, risk assessment and verification optimization. Through historical data-driven weight calculation, multi-dimensional target decomposition model and dynamically adjusted fault handling mechanism, a systematic deduction from data to decision-making is achieved, providing a reliable basis for the precise application of organic fertilizer in farmland.

[0055] Example 3: This embodiment further defines the specific implementation method of step S3. When constructing the enhanced fertilizer feature space, it is necessary to integrate the multi-dimensional features of the original information of the organic fertilizer library and the decision data after verification. The static feature data of each fertilizer is extracted, including the functional feature vector (such as the ability to provide nitrogen, the effect of improving soil aggregate structure, etc.), the attribute index vector (physical and chemical indicators such as pH value, water content, organic matter content), and the resource demand vector (production cycle, transportation cost, storage conditions, etc.). For example, the functional feature vector of decomposed sheep manure can be expressed as [high organic matter, medium nitrogen content, acidity], and the attribute index vector includes data such as pH value 5.5 and water content 30%.

[0056] Dynamic characteristic data is calculated based on historical application records, including the fertilizer's historical effect matrix (such as crop yield increases and soil nutrient changes in different plots of land in different years), the average application time vector (the average number of days from procurement to completion of application), and resource consumption distribution (the proportion of labor, machinery, and energy consumed). Taking commercial organic fertilizer as an example, its historical effect matrix may record an average yield increase of 8%-12% in wheat fields over the past three years, an average application time vector of 5 days, and a mechanical contribution of 60% to resource consumption.

[0057] Correlating feature data is achieved by constructing a fertilizer dependency graph, where nodes represent fertilizer types and edges represent compatibility (e.g., sheep manure and superphosphate can be mixed, while wood ash and ammonium nitrogen fertilizer cannot be mixed). The fertilizer compatibility matrix is calculated by combining expert knowledge with historical data on the effects of mixed applications. For example, the compatibility score for sheep manure and superphosphate is 90 out of 100, while the compatibility score for wood ash and ammonium sulfate is 20. The fertilizer combination effect tensor quantifies the synergistic effects of combined applications. For example, combining sheep manure with cake fertilizer can increase the rate of soil organic matter accumulation by 30%.

[0058] In the fertilizer matching and ratio optimization stage, the functional suitability is first calculated based on the enhanced fertilizer feature space. The cosine similarity algorithm is used. The specific steps are as follows: Feature vectorization: Convert the current farmland context features (seasonal window, resource status, target priority) into a multi-dimensional vector. For example: Seasonal window: one-hot encoding (spring = 100, summer = 010, autumn = 001, winter = 000); Resource status: The adequacy of organic fertilizer inventory is quantified as a value in the interval [0,1] (e.g., sufficient = 1, moderate = 0.5, insufficient = 0); Target priority: Urgent / High / Medium / Low are encoded as one-hot vectors such as [1,0,0,0], [0,1,0,0], etc.

[0059] After merging, an 8-dimensional context feature vector is formed, such as [1,0,0,1,0,1,0,0] The current farmland nutrient requirement characteristic vector (e.g., nitrogen deficiency, need for increased organic matter) is matched with the fertilizer functional characteristic vector to generate a functional suitability score. For example, if the demand characteristics of a plot are [nitrogen gap 20mg / kg, organic matter target increase 1%], the functional suitability score of decomposed sheep manure is 85 points, while that of chemical fertilizer urea is 70 points.

[0060] During the attribute constraint filtering phase, fertilizers are selected based on conditions such as soil pH and crop type. For example, when the soil pH is 8.0 (slightly alkaline), acidic fertilizers with a pH below 5.0 are excluded, while neutral or slightly alkaline fertilizers (such as wood ash with a pH of 10.0 and compost with a pH of 7.5) are retained to generate an initial set of candidate fertilizers. The contextual features of the current farmland (such as the spring planting season, sufficient organic fertilizer inventory, and urgent priority) are obtained. By matching historical similar scenarios (e.g., fertilizer usage records for similar plots in the spring of the past three years), a weighted cosine similarity algorithm is used to calculate a contextual relevance score. First, the current farmland's seasonal window, resource status, and other features are converted into multidimensional vectors. Similarity is calculated with the historical scenario vectors, and a time decay factor and historical effect weights are introduced. The resulting relevance score is used to adjust the weights of the candidate fertilizers. For example, basal fertilizer is a high priority during the spring planting season, and the historical usage rate of decomposed sheep manure as basal fertilizer is 70%, which can increase its contextual relevance score by 15%.

[0061] When constructing a feasible set of fertilizer combinations, combinations with compatibility scores below a threshold (e.g., below 60 points) are excluded to generate a list of options for mixed application. When calculating the synergy score for a combination, functional complementarity (e.g., nitrogen fertilizer and organic fertilizer combined to meet both short-term and long-term crop nutrient needs) and attribute gain (e.g., mixed application can adjust soil pH) are considered. A hierarchical analysis method and a fuzzy comprehensive evaluation model are used: functional complementarity accounts for 60%, using a judgment matrix to calculate scores for indicators such as nutrient release cycle matching; attribute gain accounts for 40%, quantifying the degree to which mixed application improves soil pH, organic matter, and other factors. The synergy score is ultimately derived by weighted integration. For example, the combined synergy score of sheep manure and urea is 88 points, significantly higher than the score for either single application.

[0062] During the application path optimization phase, an application dependency graph is constructed based on the optimized fertilizer selection plan to clarify the fertilization sequence and parallelization opportunities. For example, base fertilizer application should be completed before sowing, while topdressing can be performed simultaneously with irrigation after the seedling stage. The shortest application cycle is calculated using the critical path method. When constructing the resource allocation matrix, human and mechanical resources are allocated according to time nodes, such as mechanical fertilizer transportation on day 1, manual base fertilizer application on day 2, and sowing on day 3, to ensure maximum resource utilization.

[0063] Caching strategies are developed based on fertilizer supply cycles and application sequences. Fertilizers with long procurement cycles (such as customized organic fertilizers) are stockpiled in advance, maintaining a safety stock level. Fault-tolerance mechanisms include failure handling strategies (such as activating backup equipment in the event of a mechanical failure), backup options (such as switching to an equivalent fertilizer when the primary fertilizer is out of stock), and a collection of monitoring points (such as real-time fertilizer application rate monitoring sensors and soil moisture meters) to ensure the stability of the application process.

[0064] Multi-dimensional pre-application inspections include: online status checks (verifying fertilizer inventory and equipment operating status), resource adequacy verification (calculating whether the total application amount exceeds inventory), and interface response testing (delays in receiving commands from the fertilizer equipment control system). Safety assessments include permission checks (operator qualification review), risk assessments (the impact of sudden weather changes on fertilization effectiveness), and compliance verification (compliance with environmental standards). Performance predictions include response time estimates, resource consumption forecasts (such as fuel consumption and man-hours), and success probability calculations (based on the historical success rates of similar solutions). For example, if pre-inspection reveals insufficient inventory of a particular fertilizer, this triggers a switch to an alternative solution and regenerates the pre-inspection report data.

[0065] The entire step S3 achieves accurate mapping from fertilizer feature space to specific application plans through multi-dimensional feature modeling, dynamic weight adjustment and path optimization, and combines the pre-inspection mechanism to ensure the feasibility and reliability of the plan, providing technical support for the scientific application of organic fertilizers in farmland.

[0066] Embodiment 4: This embodiment further defines the specific implementation of step S12. In the process of generating farmland feature vectors, it is necessary to perform multi-dimensional analysis and feature fusion on the standardized farmland data. Soil segments in the standardized farmland data are read, and physical parameter values (such as bulk density and porosity), chemical nutrient values (such as nitrogen, phosphorus, and potassium content), and moisture values are calculated for each segment. Soil statistical data are generated through descriptive statistics, including the mean, standard deviation, and distribution range of each parameter. For example, soil statistical data for the 0-20 cm soil layer of a certain plot of land show that the mean organic matter content is 15 g / kg and the standard deviation is 2.3 g / kg, showing a normal distribution characteristic.

[0067] Based on soil statistical data, a sliding window algorithm or density clustering algorithm is used to segment soil segments and classify them into high-nutrient, medium-nutrient, and low-nutrient zones. For example, a plot can be divided into three nutrient zones using K-means clustering, with soil nutrient values within each zone varying by less than 10%. Nutrient statistics for each zone are then obtained (e.g., the mean nitrogen content in the high-nutrient zone is 120 mg / kg, the medium-nutrient zone is 80 mg / kg, and the low-nutrient zone is 50 mg / kg), generating nutrient signature data. Combining soil statistical data with nutrient signature data, a complete set of basic soil signature data, encompassing the physical structure, chemical composition, and spatial distribution of the soil, can be constructed. For example, this data can be stored in a three-dimensional matrix, with rows representing sampling points, columns representing parameter types, and the third dimension representing different soil depths.

[0068] After obtaining nutrient information from the basic characteristic data, a bidirectional association network model is used to calculate the environmental correlation strength of each nutrient. This model inputs include environmental factors such as soil nutrient values, temperature and humidity, and daylight duration, and outputs a correlation coefficient matrix between nutrients and environmental factors. For example, the correlation coefficient between nitrogen fertilizer demand and temperature is 0.6, and the correlation coefficient with precipitation is -0.3, indicating that rising temperatures may increase nitrogen fertilizer demand, while increased precipitation may reduce nitrogen fertilizer effectiveness. Based on nutrient-level growth correlation data, a growth similarity matrix is constructed. A matrix decomposition algorithm (such as singular value decomposition) is used to extract key growth units. For example, a combination of nitrogen, phosphorus, and potassium with a ratio of 2:1:1 is identified as a key unit affecting seedling growth, generating growth unit data. This growth unit data is then mapped to a predefined demand space (e.g., nutrient demand intervals based on the crop's growth period) and the demand probability distribution is calculated. For example, at the jointing stage, the probability of demand for the key unit nitrogen, phosphorus, and potassium is 75%, generating growth characteristic data.

[0069] This approach uses pre-stored farmland operation sequences (such as the timing and type of fertilization, irrigation, and tillage) to construct time series feature vectors through time series analysis. For example, the operation sequences over the past three years show that base fertilizer was applied seven days before sowing in spring, and topdressing was performed during the summer seedling stage, forming a periodic time series feature. The generated historical farmland data includes information such as operation type, time interval, and cumulative number of operations. For example, a plot of land had organic fertilizer applied four times over the past three years, with an average interval of 120 days.

[0070] When analyzing the current status of a plot, the plot's duration is determined by the number of consecutive years of cropping, such as five consecutive years of wheat cultivation. The number of cropping cycles within the current year is counted as a cycle, such as 1 in a single-crop system. Environmental continuity is assessed by calculating the coefficient of variation of climate data (average annual temperature and precipitation) over the past three years. A smaller coefficient of variation indicates a more stable environment. For example, a coefficient of variation of 5% for average annual temperature and 8% for precipitation indicates high environmental continuity, which is then used to generate the plot's status data.

[0071] When fusing growth characteristic data, historical farmland data, and plot status data, a multi-layer perceptron (MLP) model performs nonlinear transformations. The input layer contains growth characteristics (10 dimensions), historical time series features (5 dimensions), and plot status features (3 dimensions), for a total of 18 dimensions. The hidden layer extracts high-order features using the ReLU activation function, and the output layer generates a 12-dimensional farmland feature vector. For example, in the fused vector, features related to fertilization frequency have a weight of 0.3, while features related to the spatial distribution of soil nutrients have a weight of 0.4, highlighting key influencing factors.

[0072] During the feature fusion process, a data quality control mechanism is established. Chemical analysis of soil samples utilizes standard laboratory methods to ensure data accuracy; historical operation records are stored using blockchain technology to prevent data tampering; and model parameters are optimized through cross-validation to ensure the representativeness of feature vectors. Step S12, through soil feature extraction, growth feature calculation, historical data integration, and multi-source feature fusion, generates a feature vector that comprehensively reflects the current state and historical trends of the farmland, providing a critical basis for subsequent nutrient demand identification and fertilization decisions.

[0073] Example 5: This embodiment further limits the specific implementation method of step S22. When performing multi-dimensional nutrient analysis on the enhanced farmland matrix, the nutrient description information in the matrix is first read, such as the target value of soil nutrient content, the description of nutrient requirements during the crop growth stage, etc. Based on this information, a natural language processing model is used to construct an environmental dependency tree. The root node of the tree is the core nutrient target (such as "increase soil nitrogen content"), and the child nodes are environmental factors that affect the target (such as temperature, precipitation, soil pH value) and operation nodes (such as "apply nitrogen fertilizer" and "adjust soil moisture"). By recursively traversing the tree structure, the core action nodes (such as "select nitrogen fertilizer type" and "determine fertilizer application amount") are extracted to generate action sequence data, which reflects the order of key steps to achieve the nutrient target.

[0074] Based on action sequence data, key nutrient targets are identified through semantic similarity calculations. For example, "increasing nitrogen content" in the action sequence is matched with "nitrogen supplementation during the wheat jointing period" in the historical target database to identify it as the current primary target. An association rule algorithm is used to calculate the strength of the logical relationship between targets. For example, if "increasing nitrogen content" and "promoting tillering" have a support level of 80% and a confidence level of 90%, target association data is generated. The action sequence data and target association data are combined to output nutrient target data containing the target hierarchy and logical relationships. For example, the target name, sub-target list, and association strength matrix are stored in JSON format.

[0075] Based on nutrient target data, a predefined target decomposition template library is used for pattern matching. The template library contains target decomposition patterns for different crop types and growth stages. For example, the "Balanced Nutrient Supply" template for corn seedlings can be decomposed into sub-target units such as "Base Fertilizer Application," "Seedling Topdressing," and "Micronutrient Fertilizer Supplementation." Template matching identifies decomposable sub-targets and generates an initial set of sub-targets. For example, "Increasing Soil Nitrogen Content" can be decomposed into sub-targets such as "Soil Nitrogen Content Testing," "Nitrogen Fertilizer Type Selection," and "Fertilization Timing." Each sub-target's execution conditions (e.g., test results must be completed 24 hours before fertilization) and completion criteria (e.g., nitrogen content must reach 120 mg / kg) are analyzed. A sub-target constraint diagram is constructed. For example, "Nitrogen Fertilizer Type Selection" must be executed only after "Soil Nitrogen Content Testing" is completed and the results meet the criteria. This generates target constraint data. Based on this constraint data, the initial set of sub-targets is optimized and reorganized. For example, the sub-target execution order is adjusted to "Test → Select → Determine Time," outputting logically coherent sub-target sequence data.

[0076] Based on the sub-goal sequence data, extract the input-output dependencies of each sub-goal. For example, the output of "Soil Nitrogen Content Testing" is a test report, which serves as the input for "Nitrogen Fertilizer Type Selection." Construct a data flow graph and generate data dependency data. Analyze execution order constraints and identify sub-goals that can be executed in parallel. For example, "Fertilizer Equipment Debugging" and "Soil Testing" have no dependencies and can be performed simultaneously. Construct a target execution network and generate execution dependency data. Integrate the data dependency data and execution dependency data to construct a complete target dependency graph, where nodes represent sub-goals and directed edges represent data flow or execution order relationships.

[0077] Historical execution records are obtained, including data on the actual processing time and resource consumption (e.g., man-hours, fertilizer costs) for each sub-goal. Based on the sub-goal sequence data and historical records, the processing complexity of each sub-goal is calculated. For example, "Nitrogen Fertilizer Selection" receives a complexity score of 7 (out of 10) due to the need to compare data from multiple suppliers. Resource consumption characteristics include unit cost (e.g., organic fertilizer costs 2,000 yuan per ton) and time consumption (e.g., testing requires 48 hours), generating resource characteristic data. Sub-goals are analyzed for their time sensitivity (e.g., "Fertilization Time Determination" requires completion before the rainy season, which is highly sensitive) and priority factors (e.g., a weight of 0.6 on yield impact). A target scheduling weight matrix is constructed, with the matrix elements representing the sub-goal priority coefficients (e.g., 0.4, 0.5, 0.6), generating scheduling characteristic data. The resource characteristic data and scheduling characteristic data are combined with the target dependency graph to output the final nutrient analysis results, such as a multidimensional dataset containing sub-goal priorities, resource requirements, and the execution network.

[0078] In the entire step, only the construction of the target scheduling weight matrix is involved, and its mathematical expression is:

[0079] in, represents the target scheduling weight matrix, is the number of sub-goals, Dimensions of factors influencing priority (e.g., time sensitivity, resource consumption, and output impact); Indicates the The sub-goal is The weight coefficient under each influencing factor has a value range of , determined through the Analytic Hierarchy Process or expert scoring. This matrix is used to quantify the priority of sub-goals, guide resource allocation and timing arrangements for subsequent fertilization decisions, and ensure the scientific and efficient nature of the nutrient management plan.

[0080] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0081] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for applying organic fertilizer for dynamic management of farmland nutrients, characterized in that: The steps include: S1. Obtain raw farmland environmental data and standardize it to obtain standardized farmland data. Based on the standardized farmland data, extract soil characteristics, calculate crop growth characteristics and climate characteristics, and generate farmland feature vectors. Based on the farmland feature vectors, identify nutrient requirements and analyze fertilization types to obtain preliminary analysis results. S2, integrating the preliminary analysis result data and the standardized farmland data to generate an enhanced farmland matrix; Based on the enhanced farmland matrix, multi-dimensional nutrient analysis is performed to obtain nutrient analysis result data; the nutrient analysis result data is analyzed to calculate the necessity score and environmental impact value of organic fertilizer application to form fertilization decision data; the fertilization decision data is multi-verified to generate verified decision data; S3, constructing an enhanced fertilizer feature space based on the pre-stored original information of the organic fertilizer library and the verified decision data; Based on the enhanced fertilizer feature space, fertilizer matching and ratio optimization are performed to generate an optimized fertilizer selection plan; the application path of the optimized fertilizer selection plan is optimized to form an optimized application plan; based on the optimized application plan, multi-dimensional pre-application inspection is performed, and finally the pre-inspection report data is output.

2. The organic fertilizer application method for dynamic management of farmland nutrients according to claim 1, characterized in that: Step S1 is further as follows: S11. Obtaining raw farmland environmental data including soil samples, timestamps, plot identifiers, and crop identifiers; converting the soil samples in the raw farmland environmental data into a unified coding format, removing outliers and redundant information, and obtaining processed input data; performing data length normalization on the processed input data to generate standardized farmland data; S12. Based on the standardized farmland data, soil parameter values are calculated, key nutrient sets are extracted, crop types are identified, and basic characteristic data are generated; based on the basic characteristic data, a preconfigured growth analysis model is used to calculate the probability distribution of nutrient requirements and growth vectors of the crops to obtain growth characteristic data; historical records of the farmland are obtained, and based on the historical records, plot status information is extracted; the plot status information is combined with the growth characteristic data to form a farmland characteristic vector; S13. Based on the farmland feature vector, use the preset feature-demand mapping rules to identify the nutrient demand type, calculate the demand priority score, and generate demand feature data; analyze the demand feature data, estimate the fertilizer resource demand value and the time window demand value, and obtain the resource demand data; based on the demand feature data and the resource demand data, identify the dependency relationship between the demands and construct a dependency graph; based on the nutrient demand type, priority score, resource demand data and the dependency graph, generate preliminary analysis result data.

3. The organic fertilizer application method for dynamic management of farmland nutrients according to claim 2, characterized in that: Step S2 is further as follows: S21. Converting the preliminary analysis result data into a feature matrix, converting the standardized farmland data into a vector representation, and combining the feature matrix and the vector representation to form an initial farmland matrix; extracting relevant historical farmland records and calculating historical information weights; fusing the historical information weights with the initial farmland matrix to generate an enhanced farmland matrix; S22. Analyze the enhanced farmland matrix to identify the main nutrient target; decompose the main nutrient target into a set of sub-targets, construct a target dependency graph, and obtain target structure data; Calculate the resource requirement vector and target priority matrix of each sub-goal to generate target resource data; Analyze fertilizer characteristic requirements and calculate fertilizer importance weights based on primary nutrient targets and historical farmland records; Based on the fertilizer importance weight, the target structure data and target resource data are integrated into the nutrient analysis result data; S23. Based on the nutrient analysis result data, calculate the necessity score of organic fertilizer application, evaluate the application risk value, and obtain application assessment data; based on the application assessment data, determine the application timing and generate an application priority list; based on the application priority list, formulate an adjustment strategy to form application strategy data; integrate the application assessment data and the application strategy data to generate an application path map; based on the application path map, calculate the confidence score and ultimately form fertilization decision data; S24, performing internal consistency verification on the fertilization decision data to generate consistency verification data; Based on consistency verification data, verify resource availability, check technical constraints and seasonal restrictions, and obtain feasibility assessment data. Based on consistency verification data and feasibility assessment data, calculate verification scores, mark risk points, and generate optimization suggestions. Based on the optimization suggestions, verified decision data is formed.

4. The organic fertilizer application method for dynamic management of farmland nutrients according to claim 3, characterized in that: Step S3 is further as follows: S31. Based on the pre-stored original information of the organic fertilizer library and the verified decision data, the functional feature vector, attribute index vector, and resource demand vector of each fertilizer are extracted to generate static feature data; the historical effect matrix, average application time vector, and resource consumption distribution of the fertilizer are calculated to form dynamic feature data; a fertilizer dependency graph is constructed, the fertilizer compatibility matrix and the fertilizer combination effect tensor are calculated to obtain associated feature data; the static feature data, dynamic feature data, and associated feature data are integrated into an enhanced fertilizer feature space; S32. Calculating functional suitability based on the enhanced fertilizer feature space; Based on functional suitability, attribute constraint filtering is performed to generate an initial set of candidate fertilizers; context features of the current farmland are obtained and context relevance scores are calculated; Based on the contextual relevance scores, the weights of candidate fertilizers are adjusted and the initial set of candidate fertilizers is re-ranked to obtain an optimized set of candidate fertilizers; a feasible fertilizer combination set is constructed and the combination synergy score is calculated; Based on the combination synergy score, the optimal combination scheme is selected from the optimized candidate fertilizer set to form an optimized fertilizer selection scheme; S33. Based on the optimized fertilizer selection plan, construct an application dependency graph, calculate the critical path, and generate a parallel application plan; Based on the parallel application plan, application sequence data is generated; based on the application sequence data, a resource allocation matrix is constructed to optimize the application sequence; based on the optimized application sequence, a cache strategy is constructed to obtain resource optimization data; Based on resource optimization data, failure handling strategies, alternative solutions, and monitoring point sets are constructed to generate fault-tolerant mechanism data; application sequence data, resource optimization data, and fault-tolerant mechanism data are integrated into an optimized application plan; S34. Perform an online status check on the fertilizer in the optimized application plan to verify resource adequacy, test interface response, and generate availability verification data; Based on the availability verification data, we conduct permission checks, risk assessments, and compliance verifications to generate security assessment data. Based on the security assessment data, we estimate response time, predict resource consumption, and calculate success probability to generate performance prediction data. Integrate usability verification data, safety assessment data, and performance prediction data into pre-inspection report data.

5. The organic fertilizer application method for dynamic management of farmland nutrients according to claim 4, characterized in that: Step S12 is further as follows: S121. Read soil segments from the standardized farmland data, calculate parameter values, nutrient values, and moisture values for each soil segment, and generate soil statistical data; based on the soil statistical data, use analysis tools to segment the soil segments, obtain nutrient statistics, and generate nutrient characteristic data; combine the soil statistical data and nutrient characteristic data to construct complete basic characteristic data; S122. Obtain nutrient information from the basic characteristic data. Based on the nutrient information, calculate the environmental association strength of each nutrient using the bidirectional association network of the model to generate nutrient-level growth association data. Based on the nutrient-level growth association data, construct a growth similarity matrix, calculate key growth units, and generate growth unit data. Map the growth unit data to a predefined demand space, calculate the demand probability distribution, and obtain growth characteristic data. S123. Obtain the operation sequence in the pre-stored farmland historical records, construct a time series feature vector, and generate historical farmland data; obtain and analyze the current plot status, including the plot usage time, operation rounds, and environmental continuity, to form plot status data; perform feature fusion on the growth feature data, historical farmland data, and plot status data, and output the final farmland feature vector.

6. The organic fertilizer application method for dynamic management of farmland nutrients according to claim 4, characterized in that: Step S22 is further as follows: S221, reading nutrient description information in the enhanced farmland matrix, building an environment dependency tree using a model based on the nutrient description information, extracting core action nodes, and generating action sequence data; Based on the action sequence data, key nutrient targets are identified, the strength of the logical relationship between targets is calculated, and target association data is generated; the action sequence data and target association data are combined to output nutrient target data; S222. Based on the nutrient target data, a predefined model target decomposition template library is used to perform pattern matching, identify decomposable sub-target units, and generate an initial sub-target set; the execution conditions and completion standards of each sub-target in the initial sub-target set are analyzed, a sub-target constraint relationship diagram is constructed, and target constraint data is obtained; Based on the target constraint data, the initial sub-target set is optimized and reorganized, and the sub-target sequence data is output; S223. Based on the sub-goal sequence data, extract the input-output dependency relationship of each sub-goal, construct a data flow graph, and generate data dependency data; Based on data dependency data, we analyze execution order constraints, identify parallel execution opportunities, build a target execution network, and generate execution dependency data. We also integrate data dependency data and execution dependency data to build a complete target dependency graph. S224. Obtain historical execution records, calculate the processing complexity and resource consumption characteristics of each sub-goal based on the sub-goal sequence data and historical execution records, and generate resource characteristic data; Based on resource characteristic data, the time sensitivity and priority factors of sub-goals are analyzed, and the target scheduling weight matrix is constructed to form scheduling characteristic data; the resource characteristic data, scheduling characteristic data and target dependency graph are combined to output the final nutrient analysis result data.

7. The organic fertilizer application method for dynamic management of farmland nutrients according to claim 4, characterized in that: Step S23 is further as follows: S231, reading resource demand information from the nutrient analysis result data, calculating a resource utilization threshold based on the resource demand information and pre-stored historical application records, and generating resource assessment data; Based on resource assessment data, analyze the target completion time requirements, combine the current system load status, calculate the time pressure coefficient, and form time assessment data; Based on resource assessment data and time assessment data, calculate the organic fertilizer application necessity score matrix and output application necessity data; S232. Based on the application necessity data, extract characteristic patterns of historical application failure cases, construct risk characteristic vectors, and generate risk pattern data; Based on risk pattern data, analyze the similarity between the current target and pre-stored historical high-risk scenarios, calculate multi-dimensional risk coefficients, and form risk assessment data; Based on risk model data and risk assessment data, a risk-benefit assessment matrix is constructed to output application risk data; S233. Based on the application necessity data and the application risk data, construct an organic fertilizer application time series network, calculate the optimal application time window, and generate application time series data; based on the application time series data, analyze the priority dependency relationship between fertilizers, establish an application priority queue, and generate priority data; combine the application time series data and the priority data to construct an application execution plan and output application strategy data; S234: Based on the application strategy data and pre-stored historical failure processing records, a fault processing decision tree is constructed to generate fault recovery data; Based on the fault recovery data, a multi-level adjustment plan is constructed, including alternative fertilizer chains and simplified strategies, to form adjustment strategy data; Integrate application strategy data, fault recovery data, and adjustment strategy data, calculate the strategy reliability score, build a complete application path map, and finally output fertilization decision data.

8. The organic fertilizer application method for dynamic management of farmland nutrients according to claim 4, characterized in that: Step S32 is further as follows: S321. Based on the functional feature vectors and decision requirements in the enhanced fertilizer feature space, the functional applicability score of each fertilizer is calculated using the model to generate functional applicability data; based on the functional applicability data, the attribute indicator vector is used to perform constraint filtering to select a set of fertilizers that meet the attribute requirements, thereby generating attribute filtering data; Combine functional applicability data and attribute filtering data to build a preliminary fertilizer list and its scoring matrix, and output initial candidate data; S322. Obtain contextual features of the current farmland, including season window, resource status, and target priority, and generate contextual feature data; Based on contextual feature data, the effects of fertilizer use in similar historical scenarios are analyzed, a scenario correlation matrix is constructed, and scenario matching data is generated; Based on the context feature data and scene matching data, the context adjustment coefficient is calculated and the context score data is output; S323. Based on the initial candidate data and contextual scoring data, a dynamic weight algorithm is used to adjust the fertilizer score to generate adjusted weight data; and based on the historical success rate and stability index of the fertilizer, the fertilizer credibility score is updated to form credibility data. Based on the adjusted weight data and credibility data, the preliminary fertilizer list is re-ranked and the optimized candidate data is output; S324. Based on the fertilizer combination feature tensor in the optimization candidate data, construct a feasible fertilizer combination solution set and generate combination solution data; Based on the combination scheme data, the synergistic effect scores of different combination schemes are calculated, including functional complementarity and attribute gain, to form synergistic evaluation data; Based on collaborative evaluation data, analyze the complexity and risk factors of the combined solution data, build a comprehensive evaluation matrix, and obtain solution evaluation data; S325. Based on the combination scheme data, collaborative evaluation data, and scheme evaluation data, a multi-objective optimization algorithm is used to calculate the comprehensive score of each combination scheme and generate optimized scoring data; Based on the optimization scoring data, the optimal combination scheme is selected, a detailed fertilizer application sequence is constructed, and application sequence data is formed; the optimization scoring data and application sequence data are integrated, and finally the optimized fertilizer selection scheme is output.

9. The organic fertilizer application method for dynamic management of farmland nutrients according to claim 8, characterized in that: Also includes: S4, collecting the execution data flow and historical monitoring data during the organic fertilizer application process to generate a comprehensive monitoring data package; Based on the comprehensive monitoring data package, perform anomaly detection and early warning analysis, and output anomaly analysis result data; Generate dynamic optimization strategies based on abnormal analysis result data and comprehensive monitoring data packets to form optimization strategy set data; Based on the optimization strategy set data and pre-stored historical optimization effect data, adaptive learning is performed and the optimization update data package is finally output.

10. An organic fertilizer application system for dynamic management of farmland nutrients, characterized in that: include: at least one processor; and, a memory communicatively coupled to at least one of the processors; Wherein, the memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the organic fertilizer application method for dynamic management of farmland nutrients as described in any one of claims 1-9.

Images