Fertilization management system for soybean planting

Abstract

The invention provides a soybean planting fertilization management system, and relates to the technical field of management systems.Soil physicochemical properties, a meteorological microenvironment, a plant growth state and agricultural machinery operation data are collected in real time through a field sensor network and an unmanned aerial vehicle multispectral imaging technology, and a farmland dynamic portrait is constructed; the intelligent decision-making unit uses a deep learning model and a GLCM texture analysis technology to dynamically identify plant nutrient deficiency pathological characteristics, determines a nutrient deficiency type in combination with an element concentration threshold and triggers early warning; the scheme decision module dynamically optimizes the nitrogen-phosphorus-potassium ratio through a reinforcement learning algorithm, generates a field-level fertilization prescription map by using Kriging interpolation, recommends the optimal fertilization opportunity and dosage, and the precise execution module is linked with a variable fertilizer applicator to realize integrated precise application of water and fertilizer. The user interaction module supports remote monitoring and intervention of farmers through a GIS visual interface and a mobile terminal APP, can dynamically adapt to environmental changes and crop requirements, and significantly improves the nutrient utilization efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of management systems, and in particular, to a soybean planting and fertilization management system. Background Art

[0002] During the growth process of soybeans, the coupling relationship between the dynamic demand for nutrients such as nitrogen, phosphorus, and potassium and soil fertility and climatic conditions is complex. There is an urgent need to achieve precise and dynamic nutrient management through intelligent technologies.

[0003] For example, the patent with the publication number CN118350772A discloses an orchard management system based on digital twins. The deep learning camera carried by the inspection trolley can detect pests and diseases on fruit trees, and can judge whether the fruit trees lack fertilizer or other elements according to the recognition of leaf color. It provides convenience for precise fertilization in the orchard and more accurate elimination of pests and diseases. The monitoring device is inserted in the fruit row, and the underground part can monitor the moisture content, fertilizer content, and mineral elements in the soil, etc., and the results detected by the above-ground inspection device are transmitted to the database together. However, simply detecting the content of a certain element in the crop soil cannot fully reflect the current nutritional status of the crop, thus affecting the accuracy of precise fertilization for the crop.

[0004] Therefore, a soybean planting and fertilization management system is introduced. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the defect in the prior art that simply detecting the content of a certain element in the crop soil cannot fully reflect the current nutritional status of the crop, thus affecting the accuracy of precise fertilization for the crop. The present invention proposes a soybean planting and fertilization management system.

[0006] To solve the above technical problem, the technical solution adopted by the present invention is: a soybean planting and fertilization management system, which specifically includes:

[0007] Multi-source data acquisition unit: Based on sensors scattered in the field and drone imaging, it can obtain farmland environment, soybean crop growth, and agricultural machinery operation data in real time, providing multi-source data input for system decision-making;

[0008] Intelligent decision-making unit: Based on multi-source data analysis, it generates a precise management plan for soybean crop fertilization, realizing dynamic optimization management of soybean growth and fertilization, specifically including:

[0009] Intelligent diagnosis module: Based on a deep learning model, it detects the fertilizer-deficiency pathological characteristics of soybean plants in drone images, combines GLCM texture analysis to analyze the severity of nutritional imbalance of the plants, and at the same time uses element concentration thresholds to determine the nutrient deficiency types of soybean plants, outputs the pathological classification results and nutrient deficiency warnings, and triggers the system to generate targeted fertilization prescriptions;

[0010] Expert system module: Used to call historical farming data and knowledge graphs, and provide emergency management suggestions and output soil improvement plans for complex scenarios of concurrent drought / high temperature;

[0011] Scenario decision-making module: Used to dynamically adjust the nitrogen, phosphorus, and potassium element ratios of the fertilization plan based on the triggered nutrient deficiency warning according to the reinforcement learning algorithm, and generate a field-level fertilization prescription map in combination with Kriging method, and recommend the best fertilization time and dosage;

[0012] Precision execution module: Used to operate the variable fertilizer applicator to adjust the fertilizer discharge amount for drip irrigation or sprinkler irrigation fertilization and irrigation according to the fertilization prescription of the soybean plants generated by the scenario decision-making module, and convert the decision instructions into precise physical operations;

[0013] User interaction module: Used to provide a human-computer interaction interface for farmers to display targeted fertilization prescriptions, and support farmers to monitor and remotely operate the equipment for intervention.

[0014] Furthermore, the multi-source data acquisition unit specifically includes:

[0015] Soil monitoring module: Used to continuously monitor soil physical and chemical properties including pH value, nitrogen, phosphorus, potassium content, trace elements, and soil humidity through multi-parameter sensors, generate a soil fertility distribution map after calibrating the collected data, mark the acidified / salinized areas, and provide basic data support for fertilizer ratio;

[0016] Meteorological monitoring module: Used to deploy a field meteorological station to collect microclimate data including temperature, humidity, rainfall, wind speed, and light intensity, predict drought / flood risk in combination with satellite remote sensing data, output drought index, accumulated temperature value, and weather warning for the next 24 hours, and trigger environmental risk warning response;

[0017] Plant monitoring module: Used to analyze the chlorophyll content of leaves, the distribution of pathological characteristics, and the plant height / stem diameter growth curve by using UAV multi-spectral imaging, identify the nutrient deficiency pathological types of plants through the ResNet-50 model, output the pathological heat map and the root nodule number statistical results, and provide the input of nutrient deficiency pathological characteristics for the intelligent diagnosis module;

[0018] Agricultural machinery data interface: Used to receive the real-time operation data of seeders and fertilizer applicators, generate an agricultural machinery operation coverage report in combination with GPS positioning information, synchronize to the application management system to record the operation trajectory of agricultural machinery, and ensure the precise execution of the variable fertilizer applicator and the water and fertilizer integration system.

[0019] Furthermore, the intelligent decision-making unit also includes a simulation engine module and a digital twin module;

[0020] Simulation Engine Module: It is used to simulate the photosynthesis-transpiration-nutrient absorption process based on the DSSAT model, predict the impacts of different fertilization schemes on yield and soil, identify key regulatory factors through parameter sensitivity analysis, and provide optimized parameters for the Precision Execution Module;

[0021] Digital Twin Module: It is used to construct a three-dimensional model of the field using drone oblique photography, map sensor data in real time, generate a virtual farmland visualization interface and mark abnormal areas for the User Interaction Module to retrieve remote diagnosis.

[0022] Furthermore, the Intelligent Diagnosis Module calculates and determines the nutritional status of soybeans based on the Metabolism-Environment Imbalance Index Model, and the formula is as follows:

[0023]

[0024] Where: S i ′ represents the change in soybean plant metabolic indicators including the daily increment of stem diameter, the grain filling rate, and the daily increase in plant height; ΔE represents the change in environmental factors including the daily temperature difference and the fluctuation of soil pH; a represents the stress resistance adjustment coefficient, with a value of 0.2 for soybeans; P j represents the soybean plant morbidity characteristic values including the apical bud necrosis rate and the proportion of leaf color change.

[0025] Furthermore, the calculation steps of the Intelligent Diagnosis Module for the nutritional status of soybeans based on the Metabolism-Environment Imbalance Index Model are as follows:

[0026] 1. ) Metabolic Dynamics:

[0027] Stem diameter growth rate: k1 = 0.04, D is the daily increment of stem diameter, and H is the plant height;

[0028] Grain filling rate: k2 = 0.03, N is the grain nitrogen content, and T is the average daily temperature;

[0029] 2. ) Environmental Impact:

[0030] Soil pH mutation: WHC is the soil water holding capacity;

[0031] Sudden change in temperature: RH is the relative humidity, and ΔT max represents the maximum temperature change value;

[0032] 3. ) Threshold Judgment:

[0033] When MEI > 0.7, it is severely imbalanced and requires emergency intervention;

[0034] When 0.4 ≤ MEI ≤ 0.7, it is moderately imbalanced and requires the supplementation of trace elements;

[0035] When MEI < 0.4, it is in a mild imbalance and preventive management is carried out.

[0036] Furthermore, the intelligent diagnosis module determines whether the soybean plants are lacking in fertilizer based on element association, and the calculation formula is:

[0037]

[0038] Where: C i represents the effective concentration of element i; represents the optimal concentration range of element i; W ki represents the functional association strength of element k with respect to i; d ki represents the metabolic distance of element k with respect to i; b represents the environmental interference coefficient, and b = 0.3 during drought.

[0039] Furthermore, the calculation steps for the intelligent diagnosis module to determine whether the soybean plants are lacking in fertilizer based on element association are as follows:

[0040] 1. ) Node energy:

[0041] 2. ) Edge weight calculation: T is the soil temperature;

[0042] 3. ) Threshold for nutrient deficiency determination:

[0043] Nitrogen deficiency: END(N) > 0.65; Phosphorus deficiency: END(P) > 0.6; Potassium deficiency: END(K) > 0.55; Calcium deficiency: END(Ca) > 0.7; Magnesium deficiency: END(Mg) > 0.5; Boron deficiency: END(B) > 0.8; Molybdenum deficiency: END(Mo) > 0.75; Iron deficiency: END(Fe) > 0.85.

[0044] Furthermore, the user interaction module specifically includes:

[0045] Web management module: Used to overlay the fertilization area and the soil fertility grade layer based on the GIS map engine, dynamically display the yield trend and the fertilizer efficiency curve, support farmers to set target yield parameters, trigger the system to recalculate the fertilization prescription, and output the monthly fertilization cost statistics and the farming operation report;

[0046] Mobile APP module: Used to bind the farmer's mobile APP, send real-time alerts through the message push service, support remote control of the start and stop of the variable rate fertilizer applicator, and reverse transmit the data manually intervened by the farmer to the intelligent decision-making unit to correct the model parameters and generate an operation confirmation log;

[0047] Blockchain traceability module: Used to record the fertilization operations of farmers on soybeans in real time online, and at the same time generate NFT traceability tags for downstream enterprises to query.

[0048] Furthermore, the system also includes a feedback optimization unit, which is used to provide optimization suggestions for the fertilization plan through closed-loop feedback according to the growth situation of soybeans after each fertilization according to the soybean fertilization prescription generated by the system, so as to continuously optimize the system performance. Specifically, it includes:

[0049] Growth monitoring feedback module: It is used to use Kalman filter to fuse the measured data of the drone and the model prediction value, recalibrate the photosynthetic parameters of the DSSAT model every 7 days, and output the model correction coefficient and the fertilization amount adjustment suggestion for the next cycle;

[0050] Environmental risk warning module: It is used to predict the rainfall probability in the next 15 days based on the LSTM model, preset at least 10 emergency scenario plan libraries, and trigger the precise execution module to suspend fertilization or start drainage operations when an environmental risk warning is issued;

[0051] Farmer behavior analysis module: It is used to monitor the fertilization management of farmers in real time, generate farmer behavior reports and optimization suggestions, and cooperate with the expert system module to generate an adapted management plan.

[0052] Furthermore, the system also includes a system management unit, which is used to monitor and manage the operation of the system in real time to ensure the safe and stable operation of the system. Specifically, it includes:

[0053] Permission management module: It is used to dynamically allocate role permissions using the RBAC model, protect data transmission with SSL / TLS encryption, record operation audit logs, and limit the function access scope of the user interaction module;

[0054] Device monitoring module: It is used to monitor the online status of sensors or agricultural machinery through heartbeat detection, determine the cause of device abnormalities based on a rule-based expert system, trigger the downgraded mode of the precise execution module, and output a device health report and a standby device activation instruction;

[0055] Data security module: It is used to encrypt the transmission and storage of data generated by soybean planting and fertilization activities, strengthen the confidentiality of data transmission and storage, and reduce the probability of farmland data leakage.

[0056] Compared with the prior art, the beneficial effects of the present invention include: through multi-source data fusion and intelligent decision-making mechanisms, dynamic optimization management of the entire fertilization process is achieved. The system can analyze the farmland environment and the growth status of soybean plants in real time, automatically generate adaptive fertilization plans, accurately control the fertilizer ratio and application timing, effectively improve nutrient utilization efficiency and reduce resource waste. Its intelligent diagnosis and emergency response module can quickly identify disease and extreme environment risks, cooperate with variable rate fertilizer applicators to achieve precise intervention, and enhance the stress resistance of crops; at the same time, the user interaction module supports remote monitoring and manual correction, forming a data-driven closed-loop optimization system, taking into account production efficiency and operation convenience, and providing a standardized and sustainable technical solution for large-scale planting. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The disclosure of the present invention will be described with reference to the accompanying drawings. It should be understood that the drawings are only for illustrative purposes and are not intended to limit the scope of protection of the present invention. In the drawings, the same reference numerals are used to refer to the same components. Among them:

[0058] Figure 1 Schematically shows an overall architecture diagram of a soybean planting fertilization management system according to an embodiment of the present invention;

[0059] Figure 2 Schematically shows an architecture diagram of a user interaction module of a soybean planting fertilization management system according to an embodiment of the present invention;

[0060] Figure 3 Schematically shows an architecture diagram of a system management unit of a soybean planting fertilization management system according to an embodiment of the present invention;

[0061] Figure 4 Schematically shows a schematic diagram of a method for monitoring soybean fertilization status and disease symptoms of a soybean planting fertilization management system according to an embodiment of the present invention;

[0062] Figure 5 Schematically shows a comparison diagram of the results of a fertilization method of a soybean planting fertilization management system according to an embodiment of the present invention with a traditional method. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] It is easy to understand that according to the technical solution of the present invention, without changing the essence of the present invention, those of ordinary skill in the art can propose various structural ways and implementation ways that can be mutually replaced. Therefore, the following detailed embodiments and the accompanying drawings are only exemplary descriptions of the technical solutions of the present invention, and should not be regarded as the whole of the present invention or as a limitation or restriction on the technical solutions of the present invention.

[0064] According to an embodiment of the present invention in combination with Figures 1 - 5Shown. A soybean planting fertilization management system, which specifically includes:

[0065] Multi-source data acquisition unit: Based on sensors and UAV imaging dispersed in the field, it can obtain farmland environment, soybean crop growth and agricultural machinery operation data in real time, providing multi-source data input for system decision-making; specifically including:

[0066] Soil monitoring module: It is used to monitor the physical and chemical properties of the soil in real time through multi-parameter sensors (pH probe, conductivity sensor, spectral analysis module), including pH value, nitrogen, phosphorus and potassium content, trace elements (boron / molybdenum / zinc) and soil humidity, etc. After the collected data is calibrated, it generates a soil fertility distribution map, marks acidified / salinized areas, and provides basic data support for fertilizer ratio; and

[0067] Meteorological monitoring module: It is used to deploy a field meteorological station to collect microclimate data such as temperature, humidity, rainfall, wind speed, and light intensity, and combine satellite remote sensing data (such as MODIS cloud map) to predict drought / flood risk, output drought index, accumulated temperature cumulative value and weather warning for the next 24 hours, and trigger the linkage response of the environmental risk warning module; and

[0068] Plant monitoring module: It is used to analyze the chlorophyll content of leaves, the distribution of disease spots and the growth curves of plant height / stem diameter by using UAV multi-spectral imaging (visible light + near-infrared band), identify disease types such as leaf spot disease and root rot disease through the ResNet-50 model, and output disease heat maps and root nodule number statistics results, providing disease characteristics input for the intelligent diagnosis module; and

[0069] Agricultural machinery data interface: Receive the real-time operation data of seeders and fertilizer applicators (such as seeding depth, fertilization amount, row spacing), generate an agricultural machinery operation coverage report in combination with GPS positioning information, and synchronize to the application management system to record the operation trajectory of agricultural machinery, ensuring the accurate execution of variable rate fertilizer applicators and water and fertilizer integration systems.

[0070] Intelligent decision-making unit: Generate a precise management plan for soybean crop fertilization based on multi-source data analysis, and realize the dynamic optimization management of soybean growth and fertilization, specifically including:

[0071] Intelligent diagnosis module: It is used to detect the disease spot characteristics (such as shape, texture, color) in UAV images through the YOLOv7 deep learning model algorithm, quantify the severity of plant nutrient imbalance in combination with GLCM texture analysis (contrast, entropy, energy), and at the same time use element concentration thresholds such as boron / molybdenum to determine the nutrient deficiency types of soybean plants, output the disease state classification results (such as the probability of soybean apical bud necrosis is 85%) and nutrient deficiency warnings (such as boron deficiency risk level), triggering the plan decision-making module to generate targeted fertilization prescriptions;

[0072] Expert System Module: Used to call historical farming data and knowledge graphs (including more than 3,000 rules), provide emergency management suggestions for complex scenarios such as concurrent drought / high temperature (such as "spray tebuconazole immediately"), process uncertain conditions (such as sudden changes in soil pH) through fuzzy logic, and output soil improvement plans (such as calculation of lime application rate); and

[0073] Scenario Decision Module: Used to dynamically adjust the nitrogen, phosphorus, and potassium ratios of fertilization plans based on reinforcement learning algorithms, generate field-level fertilization prescription maps in combination with Kriging interpolation method, recommend the best fertilization time and dosage (such as increasing potassium fertilizer by 15% during the flowering period), and simulate the impacts of different plans on yield and soil through the digital twin module, output yield increase potential assessment (such as a 15% yield increase) and the change trend of soil organic matter;

[0074] Precision Execution Module: Used to operate the variable rate fertilizer applicator to adjust the fertilizer discharge amount for drip or sprinkler fertilization and irrigation according to the fertilization prescription of soybean plants generated by the scenario decision module, and convert decision instructions into precise physical operations;

[0075] User Interaction Module: Used to provide a human-computer interaction interface for farmers to display targeted fertilization prescriptions, support farmers to monitor and remotely operate devices for intervention, specifically including:

[0076] Web Management Module: Used to overlay the fertilization area and soil fertility grade layers based on the GIS map engine, dynamically display the yield trend and fertilizer efficiency curve (ECharts visualization), support farmers to set target yield parameters (such as 300 kg per mu), trigger the system to recalculate the fertilization prescription, output monthly fertilization cost statistics and farming operation reports, deeply integrate geospatial data with agricultural production elements through visualization technology, enable farmers to intuitively grasp the spatio-temporal correlation characteristics of farmland fertility distribution and crop growth, dynamically adjust management strategies, and significantly improve decision-making efficiency and accuracy; and

[0077] Mobile APP Module: Used to bind the farmer's mobile APP, send real-time alerts (such as "Potassium supplementation is required in the northwest region") through the message push service (Firebase Cloud Messaging), support remote control of the start and stop of the variable rate fertilizer applicator (send instructions through the MQTT protocol), and the data manually intervened by farmers is transmitted back to the intelligent decision-making unit to correct the model parameters, generate operation confirmation logs. Through the instant messaging ability and lightweight interaction design of the mobile terminal, break the spatial limitations of traditional agricultural management, enable farmers to participate in farmland management anytime and anywhere, form a two-way feedback mechanism of "cloud decision-making - terminal execution - manual verification", and effectively improve the fault tolerance of the system and the participation of farmers;

[0078] Blockchain Traceability Module: It is used to record the fertilization operations carried out by farmers on soybeans in real time and online, and at the same time generate NFT traceability tags for downstream enterprises to query. Through the distributed ledger feature and cryptographic algorithms of blockchain technology, the data immutability is guaranteed, and a fully transparent traceability system from the field to the table is established. This not only meets consumers' information demands for the quality and safety of agricultural products, but also provides farmers with digital farming operation vouchers. At the same time, the uniqueness and verifiability of NFTs are used to enhance the brand premium ability of agricultural products and promote the value reconstruction of the agricultural industry chain.

[0079] The intelligent diagnosis module calculates and determines the nutritional status of soybeans based on the metabolism-environment imbalance index model. The formula is as follows:

[0080]

[0081] Where: S i ′ represents the change in soybean plant metabolic indicators including the daily increment of stem diameter, the grain filling rate, and the daily increase in plant height, reflecting the dynamic changes of crop metabolic activities and directly reflecting the plant growth status; ΔE represents the change in environmental factors including the daily temperature difference and soil pH fluctuation. The larger ΔE is, the higher the risk of metabolic imbalance. For example, a sudden increase in pH may cause phosphorus fixation and inhibit root absorption; a represents the stress resistance adjustment coefficient, with a value of 0.2 for soybeans. The larger the a value, the more significant the inhibitory effect of environmental stress on metabolism; P j represents the soybean plant morbidity characteristic values including the apical bud necrosis rate and the proportion of leaf color change.

[0082] The calculation steps of the intelligent diagnosis module for the nutritional status of soybeans based on the metabolism-environment imbalance index model are as follows:

[0083] 1. ) Metabolic dynamics:

[0084] Stem diameter growth rate: k1 = 0.04, D is the daily increment of stem diameter, H is the plant height, and S1′ reflects the mechanical strength and water absorption capacity of the plant;

[0085] Grain filling rate: k2 = 0.03, N is the grain nitrogen content, T is the average daily temperature, and S2′ reflects the transfer efficiency of photosynthetic products to grains;

[0086] 2. ) Environmental impact:

[0087] Soil pH mutation: WHC is the soil water holding capacity;

[0088] Sudden change in air temperature: RH is the relative humidity, ΔT max represents the maximum air temperature change value, and extreme temperature differences will damage the cell membrane stability;

[0089] 3. Threshold determination:

[0090] When MEI > 0.7, it indicates severe imbalance and urgent intervention is required.

[0091] When 0.4 ≤ MEI ≤ 0.7, it indicates moderate imbalance and trace element supplementation is required.

[0092] When MEI < 0.4, it indicates mild imbalance and preventive management is needed.

[0093] The intelligent diagnosis module determines whether the soybean plants lack fertilizer based on element association. The calculation formula is:

[0094]

[0095] Where: C i represents the effective concentration of element i; represents the optimal concentration range of element i; W ki represents the functional association strength of element k with respect to i; d ki represents the metabolic distance of element k with respect to i; b represents the environmental interference coefficient, and b = 0.3 during drought.

[0096] The calculation steps for the intelligent diagnosis module to determine whether the soybean plants lack fertilizer based on element association are as follows:

[0097] 1. Node energy:

[0098] 2. Edge weight calculation: T is the soil temperature;

[0099] 3. Nutrient deficiency determination threshold:

[0100] Nitrogen deficiency: END(N) > 0.65; Phosphorus deficiency: END(P) > 0.6; Potassium deficiency: END(K) > 0.55; Calcium deficiency: END(Ca) > 0.7; Magnesium deficiency: END(Mg) > 0.5; Boron deficiency: END(B) > 0.8; Molybdenum deficiency: END(Mo) > 0.75; Iron deficiency: END(Fe) > 0.85.

[0101] Traditional fertilization methods rely on manual judgment of the growth status of soybean plants to decide on their own what kind of fertilizers to apply to soybeans. This requires very experienced workers. In the prior art, the conventional technical means that are easily thought of are to generate a fertilization plan for crops through a combination of single detection of the pH value, trace elements, etc. of the soil where the crops are planted and digital twin technology, so as to determine the fertilization prescription for the crops. For example, the patent with the publication number CN118489387B discloses an intelligent management system for rice planting fertilization, which realizes the accurate monitoring of soil fertility at different growth stages of rice planting by intelligently analyzing whether the nitrogen content in the paddy soil at different growth stages of rice planting meets the standard requirements; scientifically measures the nitrogen fertilizer application rate at the monitoring points of the rice planting paddy field and independently and quantitatively performs fertilization operations on the monitoring points of the rice planting paddy field. However, simply detecting the content of a certain element in the crop soil cannot fully reflect the current nutritional status of the crop, thus affecting the accuracy of precise fertilization of the crop.

[0102] In this embodiment, first, based on multi-spectral imaging technology and sensor data, physiological indicators of soybean plants are collected, including metabolic dynamic parameters such as daily increment of stem diameter, grain filling rate, and chlorophyll content. At the same time, environmental factors such as soil pH value fluctuations and temperature and humidity changes are monitored. Through the comprehensive calculation of the metabolic-environment imbalance index model (MEI), this model weights and correlates metabolic indicators such as stem growth rate and grain nitrogen transport efficiency with environmental factor impacts (such as sudden pH changes and extreme temperature differences), and uses the stress resistance adjustment coefficient and environmental fluctuation factor to dynamically evaluate the degree of plant nutrient imbalance. When the MEI value exceeds the threshold, the system further activates the element correlation determination model (END), combines the available nutrient concentration in the soil with the optimal concentration range of the crop, and calculates the deficiency probability of each element through the functional correlation intensity matrix and metabolic distance correction. For example, the system determines whether the nitrogen deficiency risk exceeds the 0.65 threshold by comparing the ratio of the actual concentration of nitrogen, phosphorus, and potassium to the optimal concentration and combining parameters such as soil temperature and drought coefficient. After determining the type and degree of nutrient deficiency, the system calls the reinforcement learning algorithm to dynamically optimize the nitrogen, phosphorus, and potassium ratio, generates a field-level fertilization prescription map using Kriging interpolation method, and at the same time combines digital twin technology to simulate the impact of the fertilization plan on yield and soil organic matter. Finally, a precise management plan including fertilization time, dosage, and ratio is output, realizing the full-chain intelligence from physiological index collection to fertilization decision-making, being able to dynamically adapt to environmental changes and crop requirements, significantly improving nutrient use efficiency, achieving precise regulation and management of the nitrogen-deficient state of soybean plants, and avoiding the error rate of manual experience-based fertilization judgment.

[0103] As Figure 1 shown, the intelligent decision-making unit also includes a simulation engine module and a digital twin module;

[0104] Simulation Engine Module: It is used to simulate the photosynthesis-transpiration-nutrient absorption process based on the DSSAT model, predict the impacts of different fertilization schemes on yield and soil, identify key regulatory factors through parameter sensitivity analysis, and provide optimized parameters for the Precision Execution Module;

[0105] Digital Twin Module: It is used to construct a three-dimensional model of the field using drone oblique photography, map sensor data in real time, generate a virtual farm visualization interface and mark abnormal areas for the User Interaction Module to retrieve remote diagnosis.

[0106] Specifically, the Simulation Engine Module relies on the authoritative agricultural model DSSAT to dynamically simulate the physiological processes of soybean photosynthesis, transpiration, and nutrient absorption. According to the current soil physical and chemical parameters, meteorological data, and plant growth status, it automatically deduces the impact laws of changes in nitrogen, phosphorus, and potassium ratios on dry matter accumulation and grain yield, and conducts parameter sensitivity analysis through the Monte Carlo method to identify key regulatory factors affecting output (such as critical nitrogen concentration dilution curve parameters), providing a dynamic optimization parameter set for the subsequent variable fertilizer applicator. At the same time, the Digital Twin Module uses the multi-angle oblique photography technology carried by the drone to construct a three-dimensional digital model of the field including terrain and crop canopy structure, and accesses the data streams of soil sensors, weather stations, and drones in real time through the 5G network, synchronously mapping the real state of the farmland in the virtual space, generating a dynamic visualization interface with a spatio-temporal coordinate system, and automatically marking temperature abnormal areas, nutrient deficiency hotspots, and potential disease threat areas through AI algorithms. Farmers can retrieve the digital twin scenario through the Web or mobile terminal to intuitively observe the comparison of crop growth under different fertilization schemes. The system background continuously corrects the prediction accuracy of the simulation engine based on the virtual-real interaction data, forming a spiral ascending decision-making link of "simulation verification - scheme optimization - field verification". This virtual-real fusion technology architecture not only realizes the visual preview and risk assessment of fertilization schemes, but also significantly shortens the decision-making cycle through real-time feedback in the digital space, improving the response timeliness of precision fertilization by more than 40%, providing strong support for dynamic decision-making in complex farmland scenarios.

[0107] As Figures 1 - 3 shown, the system also includes a Feedback Optimization Unit, which is used to provide optimization suggestions for the fertilization scheme through closed-loop feedback for the soybean growth situation after each fertilization according to the soybean fertilization prescription generated by the system, so as to continuously optimize the system performance, specifically including:

[0108] Growth Monitoring Feedback Module: It is used to fuse the actual measurement data of the drone and the model prediction value using Kalman filtering, recalibrate the photosynthetic parameters of the DSSAT model every 7 days, and output the model correction coefficient and the fertilization amount adjustment suggestion for the next cycle;

[0109] Environmental Risk Early Warning Module: It is used to predict the rainfall probability in the next 15 days based on the LSTM model, pre-set at least 10 emergency scenario plan libraries, and trigger the precise execution module to suspend fertilization or start drainage operations when an environmental risk warning is issued;

[0110] Farmer Behavior Analysis Module: It is used to monitor the fertilization management of farmers in real time, generate farmer behavior reports and optimization suggestions, and cooperate with the expert system module to generate adapted management plans.

[0111] The system also includes a system management unit, which is used to monitor and manage the operation of the system in real time to ensure the safe and stable operation of the system. Specifically, it includes:

[0112] Permission Management Module: It is used to dynamically allocate role permissions using the RBAC model, protect data transmission with SSL / TLS encryption, record operation audit logs, and limit the function access scope of the user interaction module;

[0113] Device Monitoring Module: It is used to monitor the online status of sensors or agricultural machinery through heartbeat detection, determine the cause of device anomalies based on a rule-based expert system, trigger the downgrade mode of the precise execution module, and output device health reports and standby device activation instructions;

[0114] Data Security Module: It is used to encrypt the transmission and storage of data generated by soybean planting and fertilization activities, strengthen the confidentiality of data transmission and storage, and reduce the probability of farmland data leakage.

[0115] Specifically, by constructing a dynamic closed-loop mechanism, the decision-making accuracy is continuously improved. The growth monitoring and feedback module uses the Kalman filter algorithm to fuse parameters such as the canopy spectrum and leaf water content measured by the drone with the predicted values of the DSSAT model, and Bayesian updates the photosynthetic parameters every 7 days to generate a calibration report containing the light energy utilization correction coefficient and the nitrogen, phosphorus, and potassium response sensitivity, providing a basis for dynamically adjusting the fertilization amount in the next cycle; the environmental risk warning module trains the LSTM neural network model based on historical meteorological data, extracts the spatio-temporal distribution characteristics of rainfall through the attention mechanism, predicts the probability of drought / flood occurrence 15 days in advance, and pre-sets a plan library covering 10 types of emergency scenarios such as extreme weather and pest outbreaks. When a sudden drop in soil humidity or a sharp increase in temperature is detected, the precise execution module is immediately triggered to suspend the fertilization operation or activate the drainage valve, forming an automated risk control chain of "prediction - response - disposal"; the farmer behavior analysis module analyzes the operation timing characteristics in the farmer's fertilization records through the hidden Markov model, combines the agronomic specifications in the expert system knowledge graph, automatically generates a behavior diagnosis report including dimensions such as fertilization timing deviation and dosage error, and collaborates with the reinforcement learning algorithm to optimize the personalized management plan. The system management unit adopts a hierarchical security architecture to ensure the robustness of the system. The permission management module dynamically assigns role permissions based on the RBAC model, realizes the transmission of sensitive data through the SSL / TLS encrypted channel, and the operation audit log records are refined to the button click level to effectively prevent unauthorized access; the device monitoring module uses the MQTT protocol to collect sensor heartbeat packets in real time, diagnoses anomalies such as device communication interruption and sensor drift based on the expert system rule library, automatically switches to standby devices and generates a disposal plan containing fault codes and repair suggestions; the data security module uses national secret algorithms to encrypt core data such as fertilization prescriptions and plot images end-to-end, and the storage layer uses blockchain evidence storage technology to ensure that the data cannot be tampered with, and reduces the data loss risk to less than 0.01% through a triple redundant backup mechanism. These mechanisms jointly construct an intelligent closed-loop of perception - decision - execution, improving the system response time by 40% and achieving financial-grade data reliability standards, providing a reliable guarantee for sustainable management in complex farmland scenarios.

[0116] The technical scope of the present invention is not limited to the content described above. Those skilled in the art can make various deformations and modifications to the above embodiments without departing from the technical idea of the present invention, and these deformations and modifications should all fall within the protection scope of the present invention.

Claims

1. A soybean planting fertilization management system, characterized in that, The system specifically includes: Multi-source data acquisition unit: Based on sensors and UAV imaging scattered in the field, it can obtain farmland environment, soybean crop growth and agricultural machinery operation data in real time, providing multi-source data input for system decision-making; Intelligent decision-making unit: Based on multi-source data analysis, it generates a precise management plan for soybean crop fertilization, realizing dynamic optimization management of soybean growth fertilization, specifically including: Intelligent diagnosis module: Based on a deep learning model, it detects the nutrient deficiency pathological characteristics of soybean plants in UAV images, combines GLCM texture analysis to analyze the severity of plant nutrient imbalance, and uses element concentration thresholds to determine the nutrient deficiency types of soybean plants, outputs the pathological classification results and nutrient deficiency warnings, triggering the system to generate targeted fertilization prescriptions; Expert system module: Used to call historical agricultural data and knowledge graphs, and provide emergency management suggestions and output soil improvement plans for complex scenarios with concurrent drought / high temperature; Scheme decision-making module: Used to dynamically adjust the nitrogen, phosphorus and potassium element ratios of the fertilization scheme based on the triggered nutrient deficiency warning, and generate a field-level fertilization prescription map in combination with Kriging method, recommending the best fertilization time and dosage; Precision execution module: Used to operate the variable fertilizer applicator to adjust the fertilizer application rate for drip irrigation or sprinkler irrigation fertilization and irrigation according to the fertilization prescription of soybean plants generated by the scheme decision-making module, and convert the decision-making instructions into precise physical operations; User interaction module: Used to provide a human-computer interaction interface for farmers to display targeted fertilization prescriptions, and support farmers to monitor and remotely operate equipment for intervention.

2. The soybean planting fertilization management system according to claim 1, wherein, The multi-source data acquisition unit specifically includes: Soil monitoring module: Used to monitor soil physical and chemical properties including pH value, nitrogen, phosphorus and potassium content, trace elements and soil humidity in real time through multi-parameter sensors. After the collected data is calibrated, it generates a soil fertility distribution map, marks acidified / salinized areas, and provides basic data support for fertilizer ratio; Meteorological monitoring module: Used to deploy a field weather station to collect microclimate data including temperature, humidity, rainfall, wind speed and light intensity, combine satellite remote sensing data to predict drought / flood risk, output drought index, accumulated temperature cumulative value and weather warning for the next 24 hours, triggering an environmental risk warning response; Plant monitoring module: Used to analyze the chlorophyll content of leaves, pathological feature distribution and plant height / stem diameter growth curve by using UAV multispectral imaging, identify the nutrient deficiency pathological types of plants through the ResNet-50 model, and output the pathological heat map and root nodule number statistical results, providing nutrient deficiency pathological feature input for the intelligent diagnosis module; Agricultural machinery data interface: Used to receive real-time operation data of seeders and fertilizer applicators, generate an agricultural machinery operation coverage report in combination with GPS positioning information, and synchronize it to the application management system to record the operation trajectory of agricultural machinery, ensuring the precise execution of variable fertilizer applicators and the water and fertilizer integration system.

3. The soybean planting fertilization management system according to claim 1, characterized in that, The intelligent decision-making unit also includes a simulation engine module and a digital twin module; Simulation engine module: Used to simulate the photosynthesis-transpiration-nutrient absorption process based on the DSSAT model, predict the impact of different fertilization schemes on yield and soil, identify key regulatory factors through parameter sensitivity analysis, and provide optimization parameters for the precision execution module; Digital Twin Module: It is used to construct a 3D model of the field using drone oblique photography, map sensor data in real time, generate a virtual farmland visualization interface and mark abnormal areas for the User Interaction Module to retrieve remote diagnosis.

4. The soybean planting fertilization management system according to claim 3, characterized in that, The Intelligent Diagnosis Module calculates and determines the nutritional status of soybeans based on the Metabolism-Environment Imbalance Index Model, and the formula is as follows: ， Wherein: represents the change amount of soybean plant metabolic indexes including the daily increment of stem diameter, the grain filling rate, and the daily increment of plant height; represents the changes in environmental factors including the daily temperature difference and the fluctuation of soil pH; represents the stress resistance regulation coefficient, with a value of 0.2 for soybeans; represents the morbidity characteristic values of soybean plants including the apical bud necrosis rate and the proportion of leaf color change.

5. The soybean planting fertilization management system according to claim 4, characterized in that, The calculation steps of the Intelligent Diagnosis Module for the nutritional status of soybeans based on the Metabolism-Environment Imbalance Index Model are as follows: (1) Metabolic dynamics: Stem diameter growth rate: , k1 = 0.04, D is the daily increment of stem diameter, and H is the plant height; Grain filling rate: , k2 = 0.03, N is the nitrogen content of grains, and T is the average daily temperature; (2) Environmental impact: Soil pH mutation: , where WHC is the soil water holding capacity; Sudden change in temperature: , where RH is the relative humidity, represents the maximum value of temperature change; (3) Threshold determination: When MEI > 0.7, it is severely imbalanced and requires emergency intervention; When 0.4 ≤ MEI ≤ 0.7, it is moderately imbalanced and requires supplementation of trace elements; When MEI < 0.4, it is mildly imbalanced and requires preventive management.

6. The soybean planting fertilization management system according to claim 3, wherein The Intelligent Diagnosis Module determines whether the soybean plants are lacking fertilizer based on element correlation, and the calculation formula is: ， Wherein: represents the effective concentration of element i; represents the optimal concentration range of element i; represents the functional association strength of element k with respect to i; represents the metabolic distance of element k with respect to i; b represents the environmental interference coefficient, and b = 0.3 during drought.

7. The soybean planting fertilization management system according to claim 6, characterized in that, The calculation steps of the Intelligent Diagnosis Module for determining whether the soybean plants are lacking fertilizer based on element correlation are as follows: (1) Node energy: , c = 0.2; (2)Edge weight calculation: , where T is the soil temperature; (3) Threshold for nutrient deficiency determination: Nitrogen deficiency: END(N) > 0.65; Phosphorus deficiency: END(P) > 0.6; Potassium deficiency: END(K) > 0.55; Calcium deficiency: END(Ca) > 0.7; Magnesium deficiency: END(Mg) > 0.5; Boron deficiency: END(B) > 0.8; Molybdenum deficiency: END(Mo) > 0.75; Iron deficiency: END(Fe) > 0.

85.

8. The soybean planting fertilization management system according to claim 1, characterized in that The User Interaction Module specifically includes: Web Management Module: It is used to overlay the fertilization area and the soil fertility grade layer based on the GIS map engine, dynamically display the yield trend and fertilizer efficiency curve, support farmers to set target yield parameters, trigger the system to recalculate the fertilization prescription, and output monthly fertilization cost statistics and farming operation reports; Mobile APP Module: It is used to bind the farmer's mobile APP, send real-time alerts through the message push service, support remote control of the start and stop of the variable rate fertilizer applicator, and the data manually intervened by the farmer is transmitted back to the intelligent decision-making unit to correct the model parameters and generate an operation confirmation log; Blockchain Traceability Module: It is used to record the fertilization operations of farmers on soybeans in real time and online, and generate NFT traceability tags for downstream enterprises to query.

9. The soybean planting fertilization management system according to claim 1, characterized in that, The system also includes a Feedback Optimization Unit, which is used to provide optimization suggestions for the fertilization plan through closed-loop feedback for the growth situation of soybeans after each fertilization according to the fertilization prescription generated by the system, so as to continuously optimize the system performance. Specifically, it includes: Growth Monitoring Feedback Module: It is used to fuse the actual measured data of the drone and the model prediction value using Kalman filtering, recalibrate the photosynthetic parameters of the DSSAT model every 7 days, and output the model correction coefficient and the adjustment suggestion for the next cycle's fertilization amount; Environmental Risk Warning Module: It is used to predict the rainfall probability in the next 15 days based on the LSTM model, preset at least 10 emergency scenario libraries, and trigger the Precision Execution Module to pause fertilization or start drainage operations when an environmental risk warning is issued; Farmer Behavior Analysis Module: It is used to monitor the fertilization management of farmers in real time, generate farmer behavior reports and optimization suggestions, and cooperate with the Expert System Module to generate an adapted management plan.

10. The soybean planting fertilization management system according to claim 1, wherein The system further includes a system management unit, which is used to perform real-time monitoring and management operations on the operation of the system to ensure the safe and stable operation of the system. Specifically, it includes: Permission management module: It is used to dynamically allocate role permissions using the RBAC model, protect data transmission with SSL / TLS encryption, record operation audit logs, and limit the function access scope of the user interaction module; Device monitoring module: It is used to monitor the online status of sensors or agricultural machinery through heartbeat detection, determine the reasons for device anomalies based on a rule-based expert system, trigger the downgraded mode of the precise execution module, and output device health reports and standby device activation instructions; Data security module: It is used to encrypt the transmission and storage of data generated by soybean planting and fertilization activities, strengthen the confidentiality of data transmission and storage, and reduce the probability of farmland data leakage.

Images