Intelligent water-soluble fertilizer application method and device

Abstract

The application discloses a kind of water-soluble fertilizer intelligent fertilization method and device, it is related to agricultural fertilization technical field, by dividing plot into multiple parts, and deploying multiple parameter sensor array, the crop environment parameter, soil parameter and crop parameter of each plot are acquired in real time, data is uploaded to cloud intelligent decision module, the module utilizes machine learning algorithm, combines crop growth model and expert knowledge base, and formulates individualized fertilization strategy for each plot, strategy is issued to corresponding edge computing gateway, the gateway will according to the strategy received, real-time control electromagnetic valve module variable control, reach dynamic matching fertilization amount and realize the purpose of multi-region differentiation operation.Electromagnetic valve module executes specific fertilization operation after receiving strategy, while process data is uploaded to edge computing gateway in real time, the intelligent fertilization method can not only improve fertilizer utilization rate, but also reduce resource waste.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of agricultural fertilization, more particularly to a water-soluble fertilizer intelligent fertilization method and device. BACKGROUND

[0002] In the development of modern cities, intelligent technology is increasingly widely used, especially in the field of agriculture. The research and application of intelligent fertilization systems are of great significance for improving crop yield and quality and achieving sustainable development. A water-soluble fertilizer intelligent fertilization method can accurately control the type, amount, and time of fertilization according to the actual needs of crops and the specific conditions of the soil, thereby improving the utilization rate of fertilizers, reducing environmental pollution, and promoting the healthy growth of crops.

[0003] Currently, the traditional fertilization method has the following main problems:

[0004] Lack of targeting: relying on the experience and routine plan of farmers for fertilization, it is impossible to accurately control the soil conditions and crop needs of different plots. This may lead to excessive or insufficient use of fertilizers, causing resource waste and pollution.

[0005] Low efficiency: conventional fertilization plans usually adopt fixed fertilization time and dosage, which cannot monitor crop growth conditions in real time and adjust in time. This reduces the utilization rate of fertilizers and affects crop yield and quality.

[0006] Large environmental burden: excessive fertilization can cause environmental problems such as soil and groundwater, which is not conducive to the sustainable development of agriculture. SUMMARY

[0007] In view of the problems existing in the prior art, the purpose of the present application is to provide a water-soluble fertilizer intelligent fertilization method and device, which can effectively solve the problems mentioned in the background.

[0008] To solve the above problems, the present application adopts the following technical solution:

[0009] A water-soluble fertilizer intelligent fertilization method, comprising the following steps:

[0010] Divide the plots into first plots, second plots, and third plots;

[0011] Deploy a multi-parameter sensor array in each plot to obtain crop environmental parameters, soil parameters, and crop parameters of the plot;

[0012] Upload the collected environmental parameters, soil parameters, and crop parameters data of each plot to the corresponding edge computing gateway for preprocessing, and obtain farmland data through data cleaning, feature extraction, and spatio-temporal alignment;

[0013] Among them, the farmland data includes the same parameters and different parameters;

[0014] The farmland data of each plot is sent to the cloud intelligent decision module. The cloud intelligent decision module formulates a basic fertilization strategy based on the same parameters and a targeted fertilization strategy based on different parameters for different crop varieties and growth stages of different plots based on machine learning algorithms and crop growth models and expert knowledge bases, and implements differentiated fertilization strategies;

[0015] The cloud intelligent decision module issues the formulated fertilization strategy to the edge computing gateway corresponding to the plot, and the edge computing gateway controls the electromagnetic valve module of the first plot, the second plot and the third plot to perform variable real-time regulation and control according to the received strategy, dynamically matches the fertilization amount, and realizes multi-region differentiated operation;

[0016] After the electromagnetic valve module receives the fertilization strategy issued by the edge computing gateway, the fertilization time, frequency and amount of each plot are controlled. In the fertilization process, the multi-parameter sensor array of each plot continuously monitors and uploads the monitoring data to the edge computing gateway in real time. The edge computing gateway uploads the fertilization data and crop growth data of each plot to the cloud intelligent decision module, and the cloud intelligent decision module dynamically adjusts and optimizes the fertilization strategy, and issues the optimized strategy to the edge computing gateway.

[0017] As preferred, the first plot is a clay plot, the second plot is a sandy plot, and the third plot is a loamy plot.

[0018] As preferred, the multi-parameter sensor array deployed in each plot includes:

[0019] An environmental parameter sensor for collecting temperature, humidity, light intensity, and wind speed parameters of the plot;

[0020] A soil parameter sensor for collecting soil moisture, nutrient content, pH value, and conductivity parameters of the plot;

[0021] A crop parameter sensor for collecting plant height, leaf area index, and stem thickness parameters of the crop in the plot.

[0022] As preferred, the edge computing gateway includes a farmland data acquisition method:

[0023] Aligning the data collected by different sensors of the same plot in time sequence to form time series data reflecting the growth state of the crop in the plot;

[0024] Aligning the data collected by the same type of sensors of different plots in spatial position to form spatial data reflecting the growth state difference of crops in different regions;

[0025] Combining the time series data and the spatial data to generate multi-dimensional farmland data containing time and space dimensions.

[0026] As preferred, the same parameter-based basic fertilization strategy and the different parameter-based targeted fertilization strategy are formulated, and the differentiated fertilization strategy is implemented, specifically including:

[0027] The same parameters and different parameters of different plots are acquired by a multi-parameter sensor array;

[0028] The same parameters are compared with preset parameter thresholds;

[0029] If the same parameters are greater than or less than the preset parameter threshold interval, the fertilization strategy one is adjusted;

[0030] If the same parameters are within the preset parameter threshold interval, the fertilization strategy two is used for normal fertilization;

[0031] The fertilization strategy three is adjusted for different parameters;

[0032] The fertilization strategy two has a higher priority than the fertilization strategy one when the same parameters are within the preset parameter threshold interval;

[0033] The fertilization strategy one has a higher priority than the fertilization strategy three when the same parameters are greater than or less than the preset parameter threshold interval.

[0034] As preferred, the machine learning algorithm uses the following formula:

[0035] F(i, j) = α × M(i, j) + β × E(i, j) + γ × S(i, j) + δ × W(i, j)

[0036] Wherein, F(i, j) represents the fertilization strategy of the i-th plot at the j-th growth stage;

[0037] M(i, j) represents the fertilization recommendation of the crop growth model for the i-th plot at the j-th growth stage;

[0038] E(i, j) represents the fertilization recommendation of the expert knowledge base for the i-th plot at the j-th growth stage;

[0039] S(i, j) represents the influence factor of the soil properties of the i-th plot on the fertilization strategy;

[0040] W(i, j) represents the influence factor of the meteorological factors of the i-th plot at the j-th growth stage on the fertilization strategy;

[0041] α, β, γ, δ are weight coefficients of the crop growth model, the expert knowledge base, the soil properties, and the meteorological factors on the fertilization strategy, and α + β + γ + δ = 1.

[0042] As preferred, the specific method for the edge computing gateway to control the electromagnetic valve module for variable real-time regulation includes:

[0043] According to the fertilization strategy issued by the cloud intelligent decision module, the target fertilization amount of each plot is determined, which is converted into control instructions for controlling the opening degree and opening duration of the electromagnetic valve;

[0044] Using the PID control algorithm, the corresponding relationship between the valve opening degree and the fertilization amount is calculated in real time according to the flow coefficient of the electromagnetic valve and the pipeline pressure, forming a valve opening degree-fertilization amount control curve;

[0045] Through the feedback mechanism of the electromagnetic valve, the actual opening degree of the valve is monitored in real time, and the actual opening degree is compared with the target opening degree. According to the deviation value, the valve opening duration is dynamically adjusted to ensure the consistency of the actual fertilization amount and the target fertilization amount;

[0046] During the fertilization process, if the crop growth or soil condition of a certain plot is found to have changed significantly, the edge computing gateway can autonomously adjust the fertilization amount of the plot within a certain range according to the preset regulation rules, and report the adjustment result to the cloud intelligent decision module for recording and analysis.

[0047] As preferred, the specific steps of the PID control algorithm include:

[0048] The deviation value between the target fertilization amount and the actual fertilization amount is taken as the input of the controller, and the output of the control amount is calculated through proportional, integral, and differential operations;

[0049] The results of the three operations are weighted and summed to obtain the final output value of the PID controller, which is converted into the corresponding valve opening adjustment instruction, thereby realizing accurate control of the fertilization amount.

[0050] The application also provides a water-soluble fertilizer intelligent fertilization device, which comprises:

[0051] Multi-parameter sensor array: deployed in each plot for collecting environmental parameters, soil parameters, and crop parameters;

[0052] Edge computing gateway: receiving data from the sensor array and performing preprocessing, including data cleaning, feature extraction, and spatio-temporal alignment;

[0053] Electromagnetic valve module: controlling the fertilization time, frequency, and amount according to the fertilization strategy received by the edge computing gateway;

[0054] Cloud intelligent decision module: based on machine learning algorithms, combining crop growth models and expert knowledge bases to develop differentiated fertilization strategies;

[0055] PID control algorithm: used to adjust the opening degree and opening duration of the electromagnetic valve in real time to match the target fertilization amount;

[0056] Communication interface: for data transmission between sensor array, edge computing gateway and cloud intelligent decision module;

[0057] Feedback mechanism: monitors the actual opening of the electromagnetic valve and compares it with the target opening to ensure the accuracy of fertilization;

[0058] Regulation rule storage unit: stores the preset fertilization regulation rules for the edge computing gateway to adjust the fertilization amount autonomously under certain conditions.

[0059] Compared with the prior art, the advantages of the present application are:

[0060] 1. The intelligent fertilization system of water-soluble fertilizer in the present application can collect multi-dimensional data such as farmland environment, soil and crop growth conditions in real time by deploying a multi-parameter sensor array, and can greatly improve the data quality and timeliness by means of data preprocessing and feature extraction with the edge computing gateway. At the same time, the cloud intelligent decision module utilizes machine learning algorithms, crop growth models and expert knowledge bases to develop differentiated and refined fertilization strategies for different soil types, crop varieties and growth stages. The electromagnetic valve module further combines the PID control algorithm to realize variable real-time regulation and dynamically match the fertilization amount. This intelligent fertilization system has significant benefits in improving quality and yield, saving fertilizer and water, and provides an innovative solution for the green and efficient development of agriculture.

[0061] 2. The intelligent fertilization device of water-soluble fertilizer in the present application adopts modular design, with key components such as multi-parameter sensor array, edge computing gateway, electromagnetic valve module, cloud intelligent decision module, PID control algorithm, communication interface, feedback mechanism and regulation rule storage unit. The modules are seamlessly integrated and interact with each other through standard communication interfaces, forming a complete cyber-physical system. The sensor array and edge gateway are responsible for the collection and preprocessing of farmland data, the cloud decision module is responsible for the generation and optimization of fertilization strategies, the electromagnetic valve and PID algorithm are responsible for precise control at the execution level, and the feedback mechanism and regulation rule support local adaptive adjustment. In practical applications, this system shows strong environmental adaptability, crop compatibility and management flexibility. In wheat, rice, corn and other crops, as well as different soil conditions such as clay, sandy soil and loam soil, it has achieved obvious yield increase and efficiency increase. The device is easy to replicate and promote on a large scale, and can be widely used in various agricultural parks, grain bases and facility agriculture, providing power for the digitalization and intelligentization upgrade of modern agriculture. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 Flowchart of the fertilization method in the embodiment of the present application;

[0063] Figure 2 Specific flowchart of the fertilization method in the embodiment of the present application;

[0064] Figure 3 Flowchart of fertilization decision-making in embodiments of the present application;

[0065] Figure 4 Bar chart comparing experimental and control groups in embodiments of the present application;

[0066] Figure 5 Timing chart of water-soluble fertilizer intelligent fertilization system in embodiments of the present application;

[0067] Figure 6 Sensor grid layout in embodiments of the present application;

[0068] Figure 7 Random sensor layout in embodiments of the present application;

[0069] Figure 8 Representative sensor layout in embodiments of the present application;

[0070] Figure 9 Triangular sensor layout in embodiments of the present application. DETAILED DESCRIPTION

[0071] The technical solutions in the embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0072] In the development of modern cities, intelligent technology is increasingly widely used, especially in the field of agriculture. The research and application of intelligent fertilization system are of great significance for improving crop yield and quality and realizing sustainable development. A water-soluble fertilizer intelligent fertilization method can accurately control the type, amount and time of fertilization according to the actual needs of crops and the specific conditions of the soil, thereby improving the utilization rate of fertilizers, reducing environmental pollution, and promoting the healthy growth of crops.

[0073] Currently, traditional fertilization methods often rely on the experience of farmers and conventional fertilization plans, which may result in inaccurate use of fertilizers, wasting resources and increasing environmental burden. Intelligent fertilization systems, through the integration of sensor technology, edge computing and cloud computing technology, can monitor the status of crops and soil in real time, automatically adjust the fertilization strategy according to the data analysis results, and achieve the goal of precision agriculture.

[0074] In order to enable those skilled in the art to better understand the intelligent fertilization method, the intelligent water-soluble fertilizer intelligent fertilization method will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. The working process of the intelligent fertilization method is as follows:

[0075] Figure 1 An intelligent fertilization method for water-soluble fertilizer is provided, comprising the following steps:

[0076] S100: Divide the land into a first land, a second land and a third land, the first land is a clay property land, the second land is a sandy soil property land and the third land is a loam property land;

[0077] S101: Deploy a multi-parameter sensor array in each land to obtain crop environmental parameters, soil parameters and crop parameters of the land: the multi-parameter sensor array deployed in each land includes;

[0078] An environmental parameter sensor is used to collect temperature, humidity, light intensity, wind speed parameters of the land;

[0079] A soil parameter sensor is used to collect soil moisture, nutrient content, pH value, electrical conductivity parameters of the land;

[0080] A crop parameter sensor is used to collect plant height, leaf area index, stem thickness parameters of the crops in the land.

[0081] The multi-parameter sensor array can be arranged as follows:

[0082] Grid arrangement: please participate Figure 6 Divide the land into several grids and arrange a multi-parameter sensor node at the center or intersection of each grid. This arrangement is simple and easy to install and maintain, and is suitable for regular-shaped lands.

[0083] Random arrangement: please participate Figure 7 Randomly select a certain number of positions in the land and arrange multi-parameter sensor nodes. This arrangement is flexible and can be adjusted according to the characteristics and needs of the land, and is suitable for irregularly shaped lands.

[0084] Representative arrangement: please refer to Figure 8 According to the differences in soil, terrain, light and other conditions in the land, select several representative positions to arrange sensor nodes. This arrangement can reduce the number of nodes and reduce costs, but requires pre-analysis of the characteristics of the land.

[0085] Triangular arrangement: please participate Figure 9 Select three vertices in the land to form a triangle and arrange sensor nodes on the edges and inside the triangle. This arrangement can achieve full coverage of the land and facilitate communication and positioning between nodes.

[0086] S102: upload the environment parameters, soil parameters and crop parameters data collected by each plot to the corresponding edge computing gateway for preprocessing, obtain farmland data through data cleaning, feature extraction and space-time alignment;

[0087] The farmland data acquisition method of the edge computing gateway comprises:

[0088] The data collected by different sensors in the same plot are aligned in time sequence to form time series data reflecting the growth state of crops in the plot;

[0089] The data collected by the same type of sensors in different plots are aligned according to the spatial position to form spatial data reflecting the difference in the growth state of crops in different regions;

[0090] The time series data and the spatial data are combined to generate multi-dimensional farmland data containing time dimension and space dimension.

[0091] The farmland data includes the same parameters and different parameters;

[0092] The same parameters include: environmental temperature, humidity, light intensity, wind speed, soil moisture, soil pH value, crop height, leaf area index, stem thickness;

[0093] The different parameters include: the first plot is a clay plot, and the soil adhesion and air permeability parameters are specially obtained;

[0094] The second plot is a sandy plot, and the soil particle size and permeability parameters are specially obtained;

[0095] The third plot is a loam plot, and the soil organic matter content and water retention parameters are specially obtained;

[0096] S103: send the farmland data of each plot to the cloud intelligent decision module, and based on the machine learning algorithm, combine the crop growth model and the expert knowledge base, formulate the basic fertilization strategy based on the same parameters, formulate the targeted fertilization strategy based on the different parameters, and implement the differentiated fertilization strategy for different crop varieties and growth stages in different plots;

[0097] The basic fertilization strategy based on the same parameters, the targeted fertilization strategy based on the different parameters, and the implementation of the differentiated fertilization strategy specifically include:

[0098] Obtain the same parameters and different parameters of different plots through a multi-parameter sensor array;

[0099] Compare the same parameters with the preset parameter threshold;

[0100] If the same parameters are greater than or less than the preset parameter threshold interval, adjust the fertilization strategy 1;

[0101] If the same parameters are within the preset parameter threshold interval, the fertilization strategy two is used for normal fertilization.

[0102] Adjust the fertilization strategy three for different parameters.

[0103] Among them, when the same parameters are within the preset parameter threshold interval, the priority of fertilization strategy two is higher than that of fertilization strategy one.

[0104] When the same parameters are greater than or less than the preset parameter threshold interval, the priority of fertilization strategy one is higher than that of fertilization strategy three.

[0105] One of the experimental examples:

[0106] Three different types of soil plots: clay plot (first plot), sandy plot (second plot) and loam plot (third plot), each plot is equipped with a multi-parameter sensor array, which obtains the same parameters (such as temperature, humidity, light, etc.) and different parameters (such as soil stickiness, air permeability of clay plot; soil granularity, permeability of sandy plot; organic matter content, water retention of loam plot).

[0107] These data are sent to the cloud intelligent decision module, which will develop differentiated fertilization strategies based on machine learning algorithms, crop growth models and expert knowledge base.

[0108] Develop basic fertilization strategies based on the same parameters:

[0109] Temperature, humidity and light intensity are the same parameters, and the preset threshold interval is:

[0110] Temperature: 20-30℃

[0111] Humidity: 40%-60%

[0112] Light intensity: 800-1200μmol / (m 2 ·s)

[0113] If any parameter exceeds the preset threshold interval, adjust the fertilization strategy one, for example: reduce the amount of fertilization or change the type of fertilizer.

[0114] If all parameters are within the preset threshold interval, normal fertilization is carried out according to the fertilization strategy two.

[0115] Develop targeted fertilization strategies based on different parameters:

[0116] For clay plot, if the soil stickiness is too high or the air permeability is poor, adjust the fertilization strategy three, for example: increase the amount of organic fertilizer or modifier.

[0117] For sandy plot, if the soil granularity is too coarse or the permeability is too high, adjust the fertilization strategy three, for example: increase the amount of water retention agent.

[0118] For the loam plot, if the organic matter content is too low or the water retention is poor, adjust the fertilization strategy three, for example: increase the amount of organic fertilizer.

[0119] Priority:

[0120] If the same parameters are within the preset threshold interval, the fertilization strategy two has higher priority than the fertilization strategy one.

[0121] If the same parameters exceed the preset threshold interval, the fertilization strategy one has higher priority than the fertilization strategy three.

[0122] For example: in the clay plot, the temperature is 25℃ (within the threshold interval), the humidity is 35% (lower than the threshold interval), the light intensity is 1000 μmol / (m 2 ·s) (within the threshold interval), and the soil viscosity is too high.

[0123] According to the priority, the fertilization strategy one is executed first to adjust the fertilization amount or change the fertilizer type, and at the same time, due to the high soil viscosity, the fertilization strategy three is executed afterwards to increase the amount of organic fertilizer or modifier.

[0124] Wherein, the machine learning algorithm adopts the following formula:

[0125] F(i, j) = a x M(i, j) + β x E(i, j) + γ x S(i, j) + δ x W(i, j) Wherein, F(i, j) represents the fertilization strategy of the i-th plot at the j-th growth stage;

[0126] M(i, j) represents the fertilization recommendation of the crop growth model for the i-th plot at the j-th growth stage;

[0127] E(i, j) represents the fertilization recommendation of the expert knowledge base for the i-th plot at the j-th growth stage;

[0128] S(i, j) represents the influence factor of the soil properties of the i-th plot on the fertilization strategy;

[0129] W(i, j) represents the influence factor of the meteorological factors of the i-th plot at the j-th growth stage on the fertilization strategy;

[0130] α, β, γ, δ are respectively the weight coefficients of the crop growth model, the expert knowledge base, the soil properties and the meteorological factors on the fertilization strategy, and α + β + γ + δ = 1.

[0131] The following takes the clay rice in the tillering stage (j = 1) of the first plot (i = 1) as an example to illustrate the fertilization strategy generation process based on machine learning.

[0132] The known related parameters of the clay rice in the tillering stage of the first plot are as follows:

[0133] The recommended amount of nitrogen, phosphorus and potassium given by the crop growth model M(1,1) = [12, 6, 10] (kg / acre);

[0134] The recommended amount of nitrogen, phosphorus and potassium given by the expert knowledge base E(1,1) = [15, 8, 12] (kg / acre);

[0135] The influence factor of soil properties on fertilization strategy S(1,1) = [0.8, 1.2, 1.0];

[0136] The influence factor of meteorological factors (such as precipitation) on fertilization strategy at the tillering stage W(1,1) = [1.1, 0.9, 1.0].

[0137] Through training set data and model tuning, the weight coefficients of each influence factor are obtained:

[0138] The weight of the crop growth model α = 0.4;

[0139] The weight of the expert knowledge base β = 0.3;

[0140] The weight of soil properties γ = 0.2;

[0141] The weight of meteorological factors δ = 0.1.

[0142] Substitute the above parameters into the machine learning algorithm, and obtain the first plot, the first growth stage: F(1,1)

[0143] = 0.4 × [12, 6, 10] + 0.3 × [15, 8, 12] + 0.2 × [0.8, 1.2,

[0144] 1.0] + 0.1 × [1.1, 0.9, 1.0]

[0145] = [4.8, 2.4, 4.0] + [4.5, 2.4, 3.6] + [0.16, 0.24, 0.20] + [0.11, 0.09, 0.10]

[0146] = [9.57, 5.13, 7.90]

[0147] After normalization, the fertilization strategy for the first plot of clay paddy rice at the tillering stage is: nitrogen 9.57 kg / acre, phosphorus 5.13 kg / acre, and potassium 7.9 kg / acre. Compared with the use of crop growth model or expert knowledge base alone, this strategy takes into account the influence of soil properties and meteorological factors, and has stronger pertinence and adaptability.

[0148] In addition, the fertilization strategy of the first plot at other growth stages and the second and third plots can also be calculated to form a complete intelligent water-soluble fertilizer fertilization scheme, so as to realize the precision management of crop growth under different soil and weather conditions, maximize the growth potential of crops, and achieve the goal of increasing yield and income while improving fertilizer utilization rate.

[0149] S104: The cloud intelligent decision module issues the formulated fertilization strategy to the edge computing gateway corresponding to the plot, and the edge computing gateway controls the electromagnetic valve module of the first plot, the second plot and the third plot respectively according to the received strategy to realize variable real-time regulation and control, dynamically match the fertilization amount, and realize multi-region differentiated operation;

[0150] S105: After the electromagnetic valve module receives the fertilization strategy issued by the edge computing gateway, the fertilization time, frequency and amount of each plot are controlled. During the fertilization process, the multi-parameter sensor array of each plot continuously monitors and uploads the monitoring data to the edge computing gateway in real time. The edge computing gateway uploads the fertilization data and crop growth data of each plot to the cloud intelligent decision module. The cloud intelligent decision module dynamically adjusts and optimizes the fertilization strategy, and issues the optimized strategy to the edge computing gateway.

[0151] The specific method of the edge computing gateway controlling the electromagnetic valve module to realize variable real-time regulation and control includes:

[0152] According to the fertilization strategy issued by the cloud intelligent decision module, the target fertilization amount of each plot is determined, which is converted into control instructions for controlling the opening degree and opening time length of the electromagnetic valve;

[0153] Using the PID control algorithm, the corresponding relationship between the valve opening degree and the fertilization amount is calculated in real time according to the flow coefficient and pipeline pressure of the electromagnetic valve, and a valve opening degree-fertilization amount control curve is formed:

[0154] The actual opening degree of the electromagnetic valve is monitored in real time through the feedback mechanism of the electromagnetic valve, the actual opening degree is compared with the target opening degree, the valve opening time length is dynamically adjusted according to the deviation value, and the consistency of the actual fertilization amount and the target fertilization amount is ensured;

[0155] During the fertilization process, if it is monitored that the crop growth or soil condition of a plot changes significantly, the edge computing gateway can autonomously adjust the fertilization amount of the plot within a certain range according to the preset regulation and control rule, and report the adjustment result to the cloud intelligent decision module for recording and analysis.

[0156] The specific steps of the PID control algorithm include:

[0157] The deviation value between the target fertilization amount and the actual fertilization amount is taken as the input of the controller, and the output of the control amount is calculated through proportional, integral and differential three kinds of operation modes;

[0158] The proportional operation calculates the control amount according to the size of the deviation value, the larger the deviation value, the larger the control amount; the integral operation calculates the control amount according to the duration of the deviation value, the longer the deviation duration, the larger the control amount; the differential operation calculates the control amount according to the change rate of the deviation value, the faster the change rate of the deviation, the larger the control amount;

[0159] The results of the three operations are weighted and summed to obtain the final output value of the PID controller, which is converted into a corresponding valve opening adjustment instruction, thereby realizing accurate control of the fertilization amount.

[0160] Specifically, the target fertilization amount is set as Q, the actual fertilization amount is set as q, and the deviation value e = Q-q is taken as the input of the PID controller.

[0161] The proportional operation P: Kp x e, where Kp is the proportional coefficient. For example, if Kp is 0.5 and e = 2 kg, the proportional operation output value is 0.5 x 2 = 1 kg.

[0162] The integral operation I: Ki x ∫edt, where Ki is the integral coefficient and ∫edt represents the accumulation of errors. For example, if Ki is 0.2 and e is 1 kg for 1 minute, the integral operation output value is 0.2 x 1 x 1 = 0.2 kg.

[0163] The differential operation D: Kd x de / dt, where Kd is the differential coefficient and de / dt represents the rate of change of the error. For example, if Kd is 0.1 and e increases from 1 kg to 2 kg in 1 minute, the differential operation output value is 0.1 x (2-1) / 1 = 0.1 kg.

[0164] The results of P, I, and D are added to obtain the output u of the PID controller = Kp x e + Ki x ∫edt + Kd x de / dt. Assuming Kp = 0.5, Ki = 0.2, Kd = 0.1, e = 2 kg and remains for 1 minute, then u = 0.5 x 2 + 0.2 x 2 x 1 + 0.1 x 0 = 1.4 kg.

[0165] The output value u = 1.4 kg is converted into an adjustment amount of the valve opening, which is added to the previous valve opening to obtain a new valve opening. The electromagnetic valve operates according to the new opening for 1 minute to realize control of the fertilization amount in this period.

[0166] After 1 minute, the next control period is entered, and steps 1-6 are repeated until the entire fertilization process is completed. Through continuous dynamic adjustment by the PID algorithm, the actual fertilization amount q continuously approaches the target value Q, ensuring the accuracy of fertilization.

[0167] Specific experimental example:

[0168] There are three different soil plots in a certain smart agricultural park, each of which is planted with tomatoes. The first plot is clay, with an area of 5 mu; the second plot is sandy soil, with an area of 3 mu; and the third plot is loam soil, with an area of 4 mu.

[0169] According to the soil properties, the farmland is divided into three plots.

[0170] Five multi-parameter sensor arrays are installed in each plot to collect environmental parameters such as air temperature, air humidity, and light intensity, soil parameters such as soil moisture, nutrient content, and pH value, and crop parameters such as tomato height and leaf area index. The sensors collect data every 30 minutes.

[0171] Each sensor array uploads the collected data to the corresponding edge computing gateway in real time. The gateway cleans the data, removes outliers, extracts the average value of each hour as the characteristic value of that period, and aligns the data of different sensors in the same plot by time and the data of the same type of sensors in different plots by space to generate structured farmland data.

[0172] The edge gateway sends the processed farmland data to the cloud intelligent decision module. Based on machine learning algorithms, the decision module combines tomato growth models and agricultural expert knowledge bases to develop differentiated fertilization plans for the soil characteristics and tomato growth stages of the three plots. For example, for the clay plot, fertilize once every 7 days, with 4 cubic meters of fertilizer water each time; for the sandy plot, fertilize once every 3 days, with 1 cubic meter of fertilizer water each time; for the loam plot, fertilize once every 5 days, with 2 cubic meters of fertilizer water each time.

[0173] The intelligent decision module issues the fertilization strategies for the three plots to the corresponding edge gateway, which controls the electromagnetic valves of each plot according to the strategy. For example, at the preset fertilization time, the edge gateway issues an instruction to open the electromagnetic valve of the clay plot to 0.8 opening for 30 minutes, the electromagnetic valve of the sandy plot to 0.4 opening for 15 minutes, and the electromagnetic valve of the loam plot to 0.6 opening for 20 minutes to complete the fertilization task.

[0174] During the fertilization process, the sensor arrays in each plot continue to collect data and upload them, and the gateway sends the fertilization data and crop growth data to the cloud decision module. If it is found that the crop growth in the sandy plot is declining, the decision module adjusts the fertilization frequency to once every 2 days and issues it to the gateway for execution.

[0175] Through the above steps, the precise dynamic control of fertilization time, frequency, and water quantity for different soil plots is realized, and finally the yield and quality of tomatoes in the three plots are improved, with an increase of 12% in the clay plot, 8% in the sandy plot, and 10% in the loam plot.

[0176] Experiment 1: Comparison of different soils for the same crop

[0177] Select wheat as the test crop, and plant it in clay, sandy soil, and loam soil. Set up experimental groups (using intelligent fertilization) and control groups (conventional fertilization) for each soil type, and compare the yield and quality differences of wheat.

[0178] Experiment two: comparison of different crops in the same soil

[0179] Select loam soil as the test soil and plant rice, corn, and soybeans. Set up experimental groups (using intelligent fertilization) and control groups (conventional fertilization) for each crop, and compare the yield and quality differences of the three crops.

[0180] Experiment three: comparison of different management of the same crop in the same soil

[0181] Select corn as the test crop and sandy soil as the test soil. Set up the following three groups of tests:

[0182] Group A: conventional fertilization + conventional irrigation

[0183] Group B: conventional fertilization + intelligent irrigation

[0184] Group C: intelligent fertilization + intelligent irrigation

[0185] Compare the yield, quality, fertilizer utilization rate, and water use efficiency of crops under A, B, and C modes.

[0186] Experiment one: comparison of the same crop (wheat) in different soils

[0187] Clay experimental group: average yield per mu 850 kg, 18% higher than the control group, protein content increased by 1.5%;

[0188] Sandy soil experimental group: average yield per mu 720 kg, 12% higher than the control group, protein content increased by 1.2%;

[0189] Loam soil experimental group: average yield per mu 800 kg, 15% higher than the control group, protein content increased by 1.3%;

[0190] Experiment two: comparison of different crops in the same soil (loam soil)

[0191] Rice experimental group: average yield per mu 650 kg, 12% higher than the control group, amylose content increased by 2%;

[0192] Corn experimental group: average yield per mu 720 kg, 15% higher than the control group, oil content increased by 1.8%;

[0193] Soybean experimental group: average yield per mu 280 kg, 10% higher than the control group, protein content increased by 1.5%;

[0194] Among them, experiment three: the same crop (corn) and the same soil (sand) are compared under different management

[0195] Group A (conventional fertilization + irrigation): average yield per mu 650 kg, fertilizer utilization rate 35%, water use efficiency 60%;

[0196] Group B (conventional fertilization + intelligent irrigation): average yield per mu 700 kg, fertilizer utilization rate 38%, water use efficiency 75%;

[0197] Group C (intelligent fertilization + intelligent irrigation): average yield per mu 780 kg, fertilizer utilization rate 50%, water use efficiency 85%;

[0198] From the above data comparison, it can be seen that:

[0199] The intelligent fertilization system has achieved more than 10% yield increase in clay, sand and loam soils, and the product quality (such as protein content) has also been improved to varying degrees, indicating that the technology has strong soil adaptability. Among them, the effect of clay and loam is the most significant.

[0200] The intelligent fertilization system also has a significant promoting effect on the yield and quality of rice, corn and soybeans, with a yield increase of 10%-15% and a quality index improvement of 1-2 percentage points, indicating that the technology has good yield-increasing effect on different crops.

[0201] The integration of intelligent fertilization and intelligent irrigation can significantly improve fertilizer utilization rate and water use efficiency while increasing yield. Group C has a 20% increase in yield per mu, a 15% increase in fertilizer utilization rate, and a 25% increase in water use efficiency compared with Group A, which reflects the superiority of integrated water and fertilizer intelligent management.

[0202] In addition, another embodiment of the application is disclosed: a water-soluble fertilizer intelligent fertilization device, comprising:

[0203] Multi-parameter sensor array: deployed in each plot for collecting environmental parameters, soil parameters and crop parameters;

[0204] Edge computing gateway: receiving data from the sensor array and preprocessing, including data cleaning, feature extraction and spatio-temporal alignment;

[0205] Electromagnetic valve module: controlling fertilization time, frequency and amount according to the fertilization strategy received by the edge computing gateway;

[0206] Cloud intelligent decision module: based on machine learning algorithm, combining crop growth model and expert knowledge base to develop differentiated fertilization strategy;

[0207] PID control algorithm: used to adjust the opening degree and opening time of the electromagnetic valve in real time to match the target fertilization amount;

[0208] Communication interface: used for data transmission between the sensor array, edge computing gateway and cloud intelligent decision module;

[0209] Feedback mechanism: monitors the actual opening degree of the electromagnetic valve and compares it with the target opening degree to ensure the fertilization accuracy;

[0210] Regulation rule storage unit: stores the preset fertilization regulation rules for the edge computing gateway to autonomously adjust the fertilization amount under specific conditions.

[0211] It should be noted that the present application is not only applicable to fertilization and pesticide application, but also applicable to irrigation, integrated pesticide and fertilizer application, fertilizer spreading and pesticide application control methods.

[0212] The above is only the preferred specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can make equivalent replacements or changes according to the technical solution and the improvement concept of the present application within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A smart fertilization method for water-soluble fertilizers, characterized in that, Includes the following steps: The land parcel is divided into three parcels: Parcel 1, Parcel 2, and Parcel 3; Parcel 1 is a clay soil parcel, Parcel 2 is a sandy soil parcel, and Parcel 3 is a loam soil parcel. A multi-parameter sensor array is deployed in each plot to acquire crop environmental parameters, soil parameters, and crop parameters for that plot. Environmental, soil and crop parameter data collected from various plots are uploaded to the corresponding edge computing gateway for preprocessing. Farmland data is obtained through data cleaning, feature extraction and spatiotemporal alignment. The method for acquiring farmland data in the edge computing gateway includes: spatiotemporal alignment, which includes aligning data collected by different sensors on the same plot of land according to a time series to form time-series data reflecting the crop growth status of the plot; Data collected by similar sensors from different plots are aligned according to spatial location to form spatial data reflecting the differences in crop growth status in different areas; The farmland data includes both data with the same parameters and data with different parameters. The same parameters include: ambient temperature, humidity, light intensity, wind speed, soil moisture, soil pH, crop height, leaf area index, and stem thickness; Different parameters include: the first plot is a clay-based plot, so soil viscosity and permeability parameters were specifically obtained; The second plot is a sandy soil plot, and soil particle size and permeability parameters were specifically obtained; The third plot is a loamy soil plot, and parameters such as soil organic matter content and water retention were specifically obtained; Data from farmland plots is sent to the cloud-based intelligent decision-making module. Based on machine learning algorithms, combined with crop growth models and expert knowledge bases, the cloud-based intelligent decision-making module formulates basic fertilization strategies based on the same parameters and targeted fertilization strategies based on different parameters for different crop varieties and growth stages, thus implementing differentiated fertilization strategies. The cloud-based intelligent decision-making module distributes the formulated fertilization strategy to the edge computing gateway of the corresponding plot. Based on the received strategy, the edge computing gateway controls the electromagnetic valve modules of the first, second, and third plots to adjust variables in real time, dynamically match the amount of fertilizer applied, and realize differentiated operations in multiple areas. The specific method for the edge computing gateway to control the electromagnetic valve module for real-time variable regulation includes: Based on the fertilization strategy issued by the cloud-based intelligent decision-making module, the target fertilization amount for each plot is determined and converted into control commands for controlling the opening degree and duration of the electromagnetic valves. Using the PID control algorithm, the relationship between valve opening and fertilizer application is calculated in real time based on the flow coefficient of the solenoid valve and the pipeline pressure, forming a valve opening-fertilizer application control curve. The feedback mechanism of the electromagnetic valve monitors the actual opening degree of the valve in real time, compares the actual opening degree with the target opening degree, and dynamically adjusts the valve opening time according to the deviation value to ensure the consistency between the actual fertilizer application amount and the target fertilizer application amount. After receiving the fertilization strategy from the edge computing gateway, the electromagnetic valve module controls the fertilization time, frequency, and amount of fertilizer for each plot. During the fertilization process, the multi-parameter sensor array of each plot continuously monitors the process and uploads the monitoring data to the edge computing gateway in real time. The edge computing gateway then uploads the fertilization data and crop growth data of each plot to the cloud-based intelligent decision-making module. The cloud-based intelligent decision-making module dynamically adjusts and optimizes the fertilization strategy and sends the optimized strategy back to the edge computing gateway.

2. The intelligent fertilization method for water-soluble fertilizer according to claim 1, characterized in that: The multi-parameter sensor array deployed in each plot includes: Environmental parameter sensors are used to collect parameters such as temperature, humidity, light intensity, and wind speed of the site. Soil parameter sensors are used to collect soil moisture, nutrient content, pH value, and electrical conductivity parameters of a plot of land. Crop parameter sensors are used to collect parameters such as plant height, leaf area index, and stem thickness of crops in a plot.

3. The intelligent fertilization method for water-soluble fertilizer according to claim 1, characterized in that: By combining time-series data and spatial data, multidimensional farmland data containing both time and spatial dimensions can be generated.

4. The intelligent fertilization method for water-soluble fertilizer according to claim 1, characterized in that: The aforementioned approach of developing a basic fertilization strategy based on the same parameters and a targeted fertilization strategy based on different parameters, implementing differentiated fertilization strategies, specifically includes: The same and different parameters of different plots of land are obtained by using a multi-parameter sensor array; Compare the same parameters with the preset parameter threshold; If the same parameter is greater than or less than the preset parameter threshold range, then adjust fertilization strategy one; If the same parameter is within the preset parameter threshold range, then fertilize normally according to fertilization strategy two; Adjust fertilization strategy three according to different parameters; Among them, when the same parameters are within the preset parameter threshold range, fertilization strategy two has a higher priority than fertilization strategy one; When the same parameter is greater than or less than the preset parameter threshold range, fertilization strategy one has a higher priority than fertilization strategy three.

5. The intelligent fertilization method for water-soluble fertilizer according to claim 4, characterized in that: The machine learning algorithm uses the following formula: F(i,j)=α×M(i,j)+β×E(i,j)+γ×S(i,j)+δ×W(i,j) Where F(i,j) represents the fertilization strategy for the i-th plot at the j-th growth stage; M(i,j) represents the fertilization recommendation of the crop growth model for the i-th plot at the j-th growth stage; E(i,j) represents the fertilization recommendations from the expert knowledge base for the i-th plot at the j-th growth stage; S(i,j) represents the influence factor of soil properties of the i-th plot on fertilization strategy; W(i,j) represents the influence factor of meteorological factors on fertilization strategy for the i-th plot at the j-th growth stage; α, β, γ, and δ are the weighting coefficients of crop growth model, expert knowledge base, soil properties, and meteorological factors on fertilization strategy, respectively, and α+β+γ+δ=1.

6. The intelligent fertilization method for water-soluble fertilizer according to claim 1, characterized in that: During the fertilization process, if significant changes are detected in the crop growth or soil conditions of a certain plot, the edge computing gateway can autonomously adjust the amount of fertilizer applied to that plot within a certain range according to preset control rules, and report the adjustment results to the cloud-based intelligent decision-making module for recording and analysis.

7. The intelligent fertilization method for water-soluble fertilizer according to claim 6, characterized in that: The specific steps of the PID control algorithm include: The deviation between the target fertilization amount and the actual fertilization amount is used as the input of the controller, and the output of the control quantity is calculated through three calculation methods: proportional, integral, and derivative. The results of the three calculations are weighted and summed to obtain the final output value of the PID controller, which is then converted into the corresponding valve opening adjustment command, thereby achieving precise control of the fertilizer application rate.

8. A smart fertilization device for water-soluble fertilizer, used to implement the smart fertilization method for water-soluble fertilizer as described in any one of claims 1-7, characterized in that, include: Multi-parameter sensor array: deployed in various locations to collect environmental, soil, and crop parameters; Edge computing gateway: Receives data from sensor arrays and preprocesses it, including data cleaning, feature extraction, and spatiotemporal alignment; Electromagnetic valve module: Controls the fertilization time, frequency, and amount based on the fertilization strategy received from the edge computing gateway; Cloud-based intelligent decision-making module: Based on machine learning algorithms, combined with crop growth models and expert knowledge base, it formulates differentiated fertilization strategies; PID control algorithm: used to adjust the opening degree and opening duration of the solenoid valve in real time to match the target fertilizer application rate; Communication interface: used for data transmission between sensor arrays, edge computing gateways, and cloud-based intelligent decision-making modules; Feedback mechanism: Monitors the actual opening degree of the solenoid valve and compares it with the target opening degree to ensure fertilization accuracy; Regulation rule storage unit: Stores preset fertilization regulation rules, allowing the edge computing gateway to adjust the fertilization amount autonomously.

Images