Corn and peanut intercropping environment monitoring and feedback regulation system based on internet of things

Abstract

The present application is a corn and peanut intercropping environment monitoring and feedback regulation system in the field of agricultural Internet of Things, aiming to solve the problems of incomplete monitoring, lagging regulation and low resource utilization in traditional intercropping. The process is as follows: first, distribute soil, weather, high-definition camera, vibration sensor, collect initial data such as soil humidity, pH value, air temperature and humidity, insect density, corn stem vibration amplitude; through ZigBee and 4G / 5G network to the cloud platform, do format check and 3σ outlier elimination on numerical data, use YOLOv5 model to identify insect density after image data preprocessing, generate preprocessed data; the cloud platform combines the preprocessed data with the future 24h precipitation, calculates the water, nutrient, pest comprehensive stress index, generates irrigation, fertilization, pesticide control decision; the local control unit drives the equipment to execute, the sensor verifies the effect after regulation, if it still meets the decision trigger judgment, secondary regulation, suitable for fine planting of corn and peanut intercropping.

Description

Technical Field

[0001] This invention relates to the field of agricultural Internet of Things (IoT), specifically to an IoT-based monitoring and feedback regulation system for intercropping corn and peanuts. Background Technology

[0002] With the promotion of agricultural modernization and precision planting concepts, intercropping of corn and peanuts has become an important planting model for improving land utilization and crop yield, leading to a significant increase in demand for field environmental monitoring and real-time control. In the corn-peanut intercropping system, crop growth is affected by multiple factors such as soil, weather, and biology, requiring the real-time collection of a large amount of environmental and crop status data. Moreover, the environmental requirements of crops at different growth stages vary greatly, placing dual demands on monitoring and control technologies for comprehensive perception and precise response.

[0003] Existing corn-peanut intercropping management techniques have significant limitations: traditional manual monitoring is time-consuming, labor-intensive, and results in discontinuous data collection, making it difficult to reflect dynamic changes in the field; single-factor regulation fails to consider the synergistic effects of multiple factors, easily leading to resource waste; and the lack of intelligent decision-making mechanisms, relying on experience-based regulation, makes it difficult to balance crop needs and management efficiency. These problems result in lagging environmental monitoring, low resource utilization efficiency, and untimely pest and disease control in intercropping, restricting the application of corn-peanut intercropping models in scenarios such as yield improvement and quality assurance.

[0004] Based on this, the present invention proposes an Internet of Things-based monitoring and feedback regulation system for corn-peanut intercropping environment, aiming to solve the above pain points and achieve the unity of comprehensive monitoring and precise regulation of the intercropping environment through distributed sensing, cloud-based intelligent analysis and closed-loop regulation strategies. Summary of the Invention

[0005] To address the technical problems mentioned in the background section, the present invention adopts the following technical solution: a corn-peanut intercropping environment monitoring and feedback regulation system based on the Internet of Things, the system comprising: a data acquisition module, a cloud platform, and a local control unit;

[0006] The data acquisition module is used to collect environmental data and crop status data of corn-peanut intercropping as initial data;

[0007] The cloud platform includes a data preprocessing module, a data storage module, a remote access module, and a decision support module.

[0008] The data preprocessing module uses a two-dimensional data preprocessing technique to preprocess the initial data to obtain preprocessed data.

[0009] The decision support module employs intelligent decision-making technology for intercropping dual crops to generate decision instructions for irrigation, fertilization, and pest control in intercropping scenarios.

[0010] The local control unit is used to receive and execute decision instructions issued by the cloud platform;

[0011] The data acquisition module transmits the acquired data to the cloud platform via a ZigBee wireless sensor network.

[0012] Furthermore, the initial data collected by the data acquisition module is represented as follows:

[0013]

[0014] in, Indicates soil moisture; Indicates soil pH value; Indicates soil electrical conductivity; N Indicates the nitrogen content in the soil; This indicates the phosphorus content in the soil; Indicates the potassium content in the soil; Indicates air temperature; Indicates air humidity; Light intensity; Indicates the amplitude of vibration of the corn stalk; Represents crop growth images;

[0015] The initial data is divided into numerical data and image data, the numerical data including... The image data is .

[0016] Furthermore, the data preprocessing module performs data preprocessing on the initial data as follows:

[0017] The data preprocessing includes preprocessing the numerical data and preprocessing the image data;

[0018] The numerical data undergoes preprocessing, including format validation and outlier removal for each data item. The steps are as follows:

[0019] First, the historical data stored on the cloud platform and the measurement range of the sensors are used to determine each numerical data point. effective range ;

[0020] in, Indicates the first value in the numerical data Item parameters, ; and Let and represent the lower bound and the lower limit of the valid range of the i-th parameter, respectively;

[0021] Then, based on the stated valid range, the data is determined to be valid or invalid, and the determination rule is as follows:

[0022] like It is deemed valid;

[0023] like It was deemed invalid.

[0024] The determination is valid. Remove those deemed invalid. ;

[0025] Secondly, regarding the retention Outliers are removed using the 3σ principle, and the formula is as follows:

[0026]

[0027] Where S represents the normal range, and the range interval is represented as follows:

[0028]

[0029] and This indicates the historical data stored on the cloud platform. The maximum and minimum values; , express The first five valid data values; express The average of the data; express The standard deviation of the data;

[0030] The image data is preprocessed, and the formatted and redundantly cropped image data is input into the YOLOv5 model. The YOLOv5 model then identifies and outputs the insect population density, which includes the corn borer density. And peanut aphid density ;

[0031] The processed numerical data and the insect population density are combined and represented as preprocessed data as follows:

[0032] .

[0033] Furthermore, the data storage module includes:

[0034] Redis real-time database and MySQL historical database;

[0035] The Redis real-time database is used to store the preprocessed data within 10 minutes;

[0036] The MySQL historical database is used to store the initial data, preprocessed data, and decision-making and control records for one year.

[0037] Furthermore, the process by which the decision support module generates decision instructions is as follows:

[0038] Step 1-1: Combine the data from the Redis real-time database with the precipitation data for the next 24 hours obtained from a weather website. , as input data;

[0039] Step 1-2: Using the input data, make decision trigger determinations, including irrigation trigger determination, fertilization trigger determination, and pest control trigger determination;

[0040] Steps 1-3: If the irrigation trigger, fertilization trigger, or pest control trigger is met, the intelligent decision-making technology for intercropping dual crops is used to generate the corresponding decision instructions.

[0041] Steps 1-4: Send the generated decision instructions to the local control unit.

[0042] Furthermore, the decision trigger determination:

[0043] When the condition is met And the rainfall in the next 24 hours Generate irrigation decision instructions;

[0044] in, This indicates the lower limit of suitable soil moisture during the current growth stage of both crops, and is determined by human experience. This indicates the soil moisture content after data preprocessing.

[0045] When the condition is met or or or Generate fertilization decision instructions;

[0046] in, , and These represent the suitable lower limits of soil nitrogen, phosphorus, potassium content, and EC value during the current growth stage of the two crops, respectively, and are determined by human experience; , , and This indicates the nitrogen, phosphorus, and potassium content and soil electrical conductivity in the soil after data preprocessing.

[0047] The pest control decision instruction will be generated if any of the following conditions are met:

[0048] Condition 1: The vibration sensor detects vibration three times consecutively. , It is the acceleration due to gravity;

[0049] Condition 2: The YOLOv5 model identifies the population density of corn borers. More than 3 aphids per plant or peanut aphid density More than 5 heads per plant.

[0050] Furthermore, the generation process of the irrigation decision command, fertilization decision command, and pest control decision command is as follows:

[0051] The irrigation decision instructions include irrigation time and irrigation flow rate;

[0052] The irrigation time The calculation formula is:

[0053]

[0054] in, This indicates the foundation pouring time, set to 15 minutes. Indicates the synergy coefficient of the growth period of two crops; This represents the comprehensive water stress index;

[0055] The irrigation flow rate The calculation formula is:

[0056]

[0057] in, Indicates the flow coefficient; Indicates the magnitude of seasonal fluctuations; Represents the accumulated days of the year, with a value range of [value range missing]. ; Indicates the seasonal sinusoidal fluctuation factor;

[0058] The fertilization decision instructions include the amount of nitrogen fertilizer, the amount of phosphorus fertilizer, and the amount of potassium fertilizer.

[0059] The formula for calculating the amount of nitrogen fertilizer applied is as follows:

[0060]

[0061] in, Indicates soil bulk density; This indicates the nitrogen content among the available nutrients in the soil; Indicates the crop nutrient utilization rate coefficient; This indicates the nitrogen content of urea; This indicates the suitable nitrogen content for the current growth stage of both crops, determined by local experience data or experimental results. For example, the suitable nitrogen content is 1.5% during the jointing stage of maize and the flowering stage of peanut. Indicates the comprehensive nutritional stress index;

[0062] The calculation methods for phosphate fertilizer application rate and potash fertilizer application rate are the same as those for nitrogen fertilizer application rate;

[0063] The pest control decision instructions include the dosage of pesticide. The calculation formula is:

[0064]

[0065] in, This indicates the basic dosage of pesticides per acre; Indicates the growth period coefficient of two crops; Indicates the drug efficacy coefficient; Indicates the integrated pest stress index;

[0066] Furthermore, the local control unit includes:

[0067] Irrigation control module, fertilization control module, and sprayer control module;

[0068] The irrigation control module drives the drip irrigation system and adjusts the irrigation flow and irrigation time as needed; the fertilization control module drives the variable fertilizer pump and dynamically regulates the fertilizer solution supply and concentration; the sprayer control module drives the insecticidal sprayer.

[0069] Compared with the prior art, the advantages of the present invention are as follows:

[0070] 1. This invention employs a dual-dimensional data preprocessing technique. For numerical data, outliers are removed through format verification and the 3σ principle. For image data, after format unification and redundancy cropping, the data is input into the YOLOv5 model to identify insect population density. Compared with traditional single data processing methods, this significantly improves data accuracy and effectiveness, and solves the problems of outlier interference and low image data utilization in the original monitoring data.

[0071] 2. This invention employs a distributed deployment of soil sensors, meteorological sensors, high-definition cameras, and vibration sensors to cover multi-dimensional monitoring of soil, weather, and crop status. Combined with the design of adding nodes at the edge of the plot, it solves the problems of limited monitoring range and incomplete data collection in traditional manual monitoring or single sensor monitoring, providing complete data support for precise regulation.

[0072] 3. This invention adopts intelligent decision-making technology for intercropping and dual-crop synergy, designs a weighted calculation model of multi-dimensional stress indices of water, nutrition and pests, and generates irrigation, fertilization and pesticide application decisions based on the rainfall in the next 24 hours. The control effect is executed and verified by a local control unit, which solves the problems of traditional control relying on experience, not considering the synergistic effect of multiple factors and lacking effect feedback. Attached Figure Description

[0073] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0074] Figure 1 This is a system framework diagram of the present invention;

[0075] Figure 2 This is a flowchart of the system of the present invention;

[0076] Figure 3 This is a graph of the air temperature and soil moisture data collected in this invention. Detailed Implementation

[0077] To achieve the above objectives, the present invention provides an Internet of Things-based monitoring and feedback regulation system for the intercropping environment of corn and peanuts. (See the system framework diagram below.) Figure 1 See system flowchart Figure 2 The specific implementation process includes:

[0078] Step 1: Sensor Deployment

[0079] A distributed deployment approach is adopted, with a monitoring node set up every 5-10 meters in the field according to the row spacing of corn-peanut intercropping. Each monitoring node covers a crop area of ​​10-20 square meters to ensure no blind spots in data collection. Simultaneously, 1-2 additional monitoring nodes are set up at the edges and corners of the plots. These monitoring nodes include:

[0080] Soil sensor: Installed 20-30cm below the soil surface, used to monitor various indicators in the crop root zone, including soil moisture. Soil pH value Soil electrical conductivity value Nitrogen content in soil N Phosphorus content and potassium content ;

[0081] Weather sensor: Mounted on a 1.5-meter-high bracket, it collects various environmental parameters, including air temperature. air humidity and light intensity ;

[0082] High-definition cameras: one per 50 square meters to capture images of crop growth. ;

[0083] Vibration sensors: Installed at the base of the corn stalks, one per 20 square meters, to monitor the vibration amplitude of the corn stalks. ;

[0084] The environmental and crop status data collected for corn-peanut intercropping are used as initial data and are represented as follows:

[0085]

[0086] The initial data is divided into numerical data and image data, the numerical data including... The image data is ;

[0087] The curves of the collected air temperature data and soil moisture data are shown below. Figure 3 As shown.

[0088] Step 2: Data Transmission

[0089] Using a ZigBee wireless sensor network to aggregate node data, the numerical data in the initial data is uploaded to the cloud platform in JSON format every minute via a 4G / 5G network; the image data in the initial data is uploaded to Alibaba Cloud OSS object storage in JPEG compressed format every 10 minutes, and the cloud platform records the OSS storage path of the image.

[0090] Step 3: Cloud Platform Deployment

[0091] The cloud platform adopts a microservice architecture design, which includes a data preprocessing module, a data storage module, a remote access module, and a decision support module;

[0092] (1) Data preprocessing module

[0093] The initial data is used as input, and a two-dimensional data preprocessing technique is used to preprocess the data to obtain preprocessed data that can be used for decision generation.

[0094] The data preprocessing includes preprocessing the numerical data and preprocessing the image data;

[0095] The numerical data undergoes preprocessing, including format validation and outlier removal for each data item. The steps are as follows:

[0096] First, the historical data stored on the cloud platform and the measurement range of the sensors are used to determine each numerical data point. effective range ;

[0097] in, Indicates the first value in the numerical data Item parameters, ; and Let and represent the lower bound and the lower limit of the valid range of the i-th parameter, respectively;

[0098] Then, based on the stated valid range, the data is determined to be valid or invalid, and the determination rule is as follows:

[0099] like It is deemed valid;

[0100] like It was deemed invalid.

[0101] The determination is valid. Remove those deemed invalid. ;

[0102] Secondly, regarding the retention Outliers are removed using the 3σ principle, and the formula is as follows:

[0103]

[0104] Where S represents the normal range, and the range interval is represented as follows:

[0105]

[0106] and This indicates the historical data stored on the cloud platform. The maximum and minimum values; , express The first five valid data values; express The average of the data; express The standard deviation of the data;

[0107] The image data is preprocessed, and the image data, after being standardized in format and redundantly cropped, is input into the YOLOv5 model. The YOLOv5 model then identifies and outputs the insect population density. The specific process is as follows:

[0108] The first step is to standardize the format and remove redundancy from the image data;

[0109] First, the image data Img Convert to a 640×640 pixel format compatible with YOLOv5 models;

[0110] Then, areas without crops in the image are removed, and valid areas containing only corn or peanuts are retained;

[0111] The second step is to deploy the YOLOv5 model, which involves the following steps:

[0112] First, the model is trained: images containing corn borers and peanut aphids are selected from historically collected images. The selected images are labeled to generate bounding box coordinates and category labels, and a pest image dataset of 5,000 images is constructed. The dataset covers different growth stages, including the seedling and jointing stages of corn, and the sowing and flowering stages of peanuts, as well as different shading conditions and different light conditions.

[0113] Next, the dataset images are processed, including horizontal flipping, random scaling, and random brightness adjustment;

[0114] The horizontal flipping involves randomly selecting 50% of the data and mirroring it along the vertical axis to simulate the distribution of pests in different locations.

[0115] The random scaling refers to the random scaling of the image. R times, ;

[0116] The random brightness adjustment involves randomly increasing or decreasing the image brightness, expressed by the following formula:

[0117]

[0118] in, The range of values ​​is ; This represents the original brightness value of the image; The truncation function restricts the calculated brightness value to the range [0, 255]. This indicates the new brightness value of the image;

[0119] Then, the processed pest image dataset is used to train the YOLOv5 model;

[0120] Finally, output the insect population density; the insect population density includes the corn borer density. And peanut aphid density ;

[0121] The processed numerical data and the insect population density are combined and represented as preprocessed data as follows:

[0122] .

[0123] (2) Data storage module

[0124] Data storage is performed using Redis and MySQL databases;

[0125] The Redis database, as a real-time database, stores the preprocessed data within 10 minutes and deletes the preprocessed data older than 10 minutes.

[0126] The MySQL database, as a historical database, is used to store the initial data, preprocessed data, and decision-making and control records for one year.

[0127] (3) Remote access module

[0128] Real-time data push is implemented based on the WebSocket protocol, supporting multiple terminals online simultaneously. The monitoring interface is developed using the Vue.js framework, implementing a visual dashboard design for real-time display.

[0129] Environmental parameters: soil moisture, soil pH, soil electrical conductivity, nitrogen, phosphorus and potassium content in the soil, air temperature and humidity and light intensity at each monitoring node;

[0130] Crop status: Real-time preview of Img data captured by a high-definition camera and insect population density identified by the YOLOv5 model;

[0131] (4) Decision support module

[0132] The process by which the decision support module generates decision instructions is as follows:

[0133] Step 1-1: Combine the data from the Redis real-time database with the precipitation data for the next 24 hours obtained from a weather website. , as input data;

[0134] Step 1-2: Using the input data, make decision trigger determinations, including irrigation trigger determination, fertilization trigger determination, and pest control trigger determination;

[0135] The conditions for triggering the irrigation decision are as follows:

[0136] And the rainfall in the next 24 hours ;

[0137] Among them, among them, This indicates the lower limit of suitable soil moisture during the current growth stage of both crops, and is determined by human experience. This indicates the soil moisture content after data preprocessing.

[0138] The conditions for triggering the fertilization determination are as follows:

[0139] or or or ;

[0140] in, , and These represent the suitable lower limits of soil nitrogen, phosphorus, potassium content, and EC value during the current growth stage of the two crops, respectively, and are determined by human experience; , , and This indicates the nitrogen, phosphorus, and potassium content and soil electrical conductivity in the soil after data preprocessing.

[0141] The conditions for triggering the pest control determination are as follows:

[0142] Condition 1: The vibration sensor detects vibration three times consecutively. , It is the acceleration due to gravity;

[0143] Condition 2: The YOLOv5 model identifies the population density of the corn borer. More than 3 aphids per plant or peanut aphid density More than 5 heads per plant;

[0144] For those that meet the trigger criteria, intelligent decision-making technology for intercropping and dual-crop collaboration is used to generate corresponding decision instructions, including the generation of irrigation decision instructions, fertilization decision instructions, and pest control decision instructions.

[0145] Steps 1-3: The steps for generating the decision instructions are as follows:

[0146] The first step is to generate irrigation decision instructions, which include irrigation time and irrigation flow rate;

[0147] The irrigation time The calculation formula is:

[0148]

[0149] in, This indicates the foundation pouring time, set to 15 minutes. This represents the synergy coefficient of the two crops' growth periods, determined from historical data. This represents the comprehensive water stress index;

[0150] The water stress index The calculation formula is as follows:

[0151]

[0152] in, , and These are the weighting coefficients; Indicates the soil moisture stress index; Indicates the meteorological moisture stress index; Indicates the plant's water stress index;

[0153] The calculation formula is:

[0154]

[0155] in: This indicates the suitable soil moisture for the current growth stage of both crops, determined by local experience data or experimental results, and is expressed as volumetric water content. During the seedling stage, the suitable soil moisture content is 15%-20%. This indicates the soil moisture content after pretreatment;

[0156] The calculation formula is:

[0157]

[0158] in, This indicates the amount of water lost through evaporation. This is the precipitation impact coefficient, when there is no precipitation. =1; When the precipitation is 5-10mm; =0.5; When precipitation is greater than 10mm =0.2;

[0159] The calculation formula is:

[0160]

[0161] in, This indicates the vibration amplitude of the corn stalk after pretreatment; , indicating the vibration threshold. It is the acceleration due to gravity; This represents the comprehensive crop water stress index;

[0162] The irrigation flow rate The calculation formula is:

[0163]

[0164] in, Indicates the flow coefficient; Indicates the magnitude of seasonal fluctuations; Represents the accumulated days of the year, with a value range of [value range missing]. ; Indicates the seasonal sinusoidal fluctuation factor;

[0165] The second step is to generate fertilization decision instructions, which include nitrogen fertilizer application rate, phosphorus fertilizer application rate and potassium fertilizer application rate.

[0166] The formula for calculating the amount of nitrogen fertilizer applied is as follows:

[0167]

[0168] in, Indicates soil bulk density; This indicates the nitrogen content among the available nutrients in the soil; Table of crop nutrient utilization coefficients; This indicates the nitrogen content of urea; This indicates the suitable nitrogen content for the current growth stage of both crops, determined by local experience data or experimental results. For example, the suitable nitrogen content is 1.5% during the jointing stage of maize and the flowering stage of peanut. Indicates the comprehensive nutritional stress index;

[0169] The calculation methods for phosphate fertilizer application rate and potash fertilizer application rate are the same as those for nitrogen fertilizer application rate;

[0170] The comprehensive stress index The calculation formula is as follows:

[0171]

[0172] in, The nitrogen stress index in soil; The stress index for phosphorus in soil; This indicates the potassium stress index in the soil. The stress index represents the pH level in the soil; Indicating soil The stress index;

[0173] The The calculation formula is:

[0174]

[0175] in, This indicates the nitrogen content in the pretreated soil;

[0176] Phosphorus, potassium and The stress index of the value is calculated using the same principle as that of nitrogen stress index;

[0177] The The calculation formula is:

[0178]

[0179] in, This indicates the electrical conductivity of the soil after pretreatment; This represents the minimum threshold for soil electrical conductivity. Based on historical data analysis, the minimum threshold for soil electrical conductivity is set at 0.5 mS / cm.

[0180] The third step is to generate pest control decision instructions, which include the dosage of pesticides, calculated using the following formula:

[0181]

[0182] in, This indicates the basic dosage of pesticides per acre, determined through manual experience. Indicates the growth period coefficient of two crops; Indicates the drug efficacy coefficient; Indicates the integrated pest stress index;

[0183] The integrated pest stress index The calculation formula is as follows:

[0184]

[0185] in, , and Indicates the weighting coefficient; Indicates the pest density stress index; The sensitivity coefficient for the growth period of two crops is determined by human experience. The meteorological condition influence index is determined by human experience.

[0186] The pest density stress index The calculation formula is:

[0187]

[0188] in, and These represent the density of corn borers and peanut aphids, respectively, as identified and output by the YOLOv5 model. and The thresholds for no stress and severe stress in corn borers are determined from historical data; and The thresholds for no stress and severe stress in peanut aphids are determined from historical data;

[0189] Steps 1-4: Send the generated decision instructions to the local control unit.

[0190] Step 4: The local control unit executes the decision.

[0191] The local control unit includes an irrigation control module, a fertilizer application control module, and a sprayer control module;

[0192] The irrigation control module drives the drip irrigation system and adjusts the irrigation flow and irrigation time as needed; the fertilization control module drives the variable fertilizer pump and dynamically regulates the fertilizer solution supply and concentration; the sprayer control module drives the insecticidal sprayer.

[0193] After execution, the monitoring node collects data through sensors to verify the control effect. If the cloud platform detects that the data still meets the decision trigger judgment, it will regenerate the decision instruction and trigger secondary control.

[0194] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A corn and peanut intercropping environment monitoring and feedback regulation system based on Internet of Things, characterized in that, The system comprises a data acquisition module, a cloud platform and a local control unit; The data acquisition module is configured to acquire environment data and crop state data of corn and peanut intercropping as initial data; The cloud platform comprises a data preprocessing module, a data storage module, a remote access module and a decision support module; The data preprocessing module is configured to preprocess the initial data by using a double-dimension data preprocessing technology to obtain preprocessed data; The decision support module is configured to generate irrigation, fertilization and pest control decision instructions for the intercropping scene by using an intelligent decision-making technology for intercropping double-crop cooperation; The local control unit is configured to receive and execute the decision instructions issued by the cloud platform; The data acquisition module transmits the acquired data to the cloud platform through a ZigBee wireless sensor network; In the decision support module, the generation processes of the irrigation decision instruction, the fertilization decision instruction and the pest control decision instruction are as follows: The irrigation decision instruction comprises irrigation time and irrigation flow; The irrigation time The calculation formula is: ； wherein, represents the base irrigation time, set to 15 minutes; represents the double crop growth period coordination coefficient; represents the water comprehensive stress index; represents the data based on the Redis real-time database and the precipitation data in the next 24 hours obtained through the weather network; The irrigation flow The calculation formula is: ； wherein, represents a flow coefficient; represents a seasonal fluctuation amplitude; represents a cumulative day of the year, with a value range of ; represents a seasonal sinusoidal fluctuation factor; The fertilization decision instruction comprises nitrogen fertilizer application amount, phosphorus fertilizer application amount and potassium fertilizer application amount; The nitrogen fertilizer application amount calculation formula is as follows: ； wherein, denotes the soil bulk density; denotes the content of nitrogen in the available soil nutrients; denotes the crop nutrient utilization coefficient; denotes the nitrogen content of urea; denotes the suitable nitrogen content for the current growth period of the double crops, which is determined by local empirical data or test results, and the suitable nitrogen content is 1.5% at the corn jointing stage and the peanut flowering stage; denotes the nutrient comprehensive stress index; denotes the nitrogen content in the soil after data preprocessing; The calculation methods of the phosphorus fertilizer application amount and the potassium fertilizer application amount are the same as that of the nitrogen fertilizer application amount; The de-insectization decision instruction includes a de-insectization drug amount The calculation formula is: ； wherein, represents the base amount of application per mu; represents the double crop growth period coefficient; represents the pesticide efficacy coefficient; represents the comprehensive stress index of insect pests.

2. The system of claim 1, wherein, The initial data acquired by the data acquisition module is represented as follows: ； wherein, represents soil moisture; represents soil pH; represents soil electrical conductivity; N represents nitrogen content in soil; represents phosphorus content in soil; represents potassium content in soil; represents air temperature; represents air humidity; light intensity; represents corn stalk vibration amplitude; represents crop growth image; The initial data is divided into numerical data and image data, the numerical data includes ; the image data is .

3. The system of claim 2, wherein, The data preprocessing process of the data preprocessing module on the initial data is as follows: The data preprocessing comprises preprocessing of the numerical data and preprocessing of the image data; The preprocessing of the numerical data comprises format checking and outlier elimination operations on each item of numerical data, and the steps are as follows: First, the effective range of each numerical data is determined using the historical data stored in the cloud platform and the measurement range of the sensor ;​ wherein, represents the i-th item parameter in the numerical data, ; and respectively represent the lower limit and the lower limit of the effective range of the i-th item parameter.​ Then, the data is determined to be valid or invalid by using the effective range, and the determination rule is as follows: If , the decision is valid; If , the decision is invalid; retained as valid ; rejected as invalid ; Second, the reserved The abnormal value is eliminated by 3σ principle, the formula is: ； Wherein, S represents a normal range, and the range interval is represented as follows: ； and represents the maximum and minimum values of the historical data stored by the cloud platform; , represents the first five valid data values of represents the data mean of represents the data standard deviation of ​​​​ The image data is preprocessed, the image data unified in format and trimmed of redundancy is input into a YOLOv5 model, and a pest density is output by the YOLOv5 model; the pest density includes corn borer pest density and peanut aphid pest density ; The preprocessed numerical data and the insect density are combined as preprocessed data, which is represented as follows: 。 4. The system of claim 3, wherein, The data storage module comprises: A Redis real-time database and a MySQL historical database; The Redis real-time database is configured to store the preprocessed data within 10 minutes; The MySQL historical database is configured to store the initial data, the preprocessed data and decision control records of one year.

5. The system of claim 4, wherein, The decision support module generates decision instructions in the following process: Step 1-1: Data in the Redis real-time database and data on precipitation in the next 24 hours obtained through the weather web as input data ; Step 1-2: decision trigger determination is performed by using the input data, including irrigation trigger determination, fertilization trigger determination and pest control trigger determination; Step 1-3: if the irrigation trigger determination, the fertilization trigger determination or the pest control trigger determination is met, an intelligent decision-making technology for intercropping double-crop cooperation is used to generate corresponding decision instructions; Step 1-4: the generated decision instructions are issued to the local control unit.

6. The system of claim 5, wherein, The decision trigger determination: When the condition is met and the amount of precipitation in the next 24 hours ; irrigation decision instruction generation is performed; wherein, represents the suitable lower limit of the soil moisture at the current growth stage of the double crops, which is determined by artificial experience; represents the soil moisture after data preprocessing; When the condition or or or ; the generation of fertilization decision instructions is performed; wherein, , and represent the suitable lower limits of the current growth stage of the double crop for the nitrogen content, the phosphorus content, the potassium content and the EC value of the soil, respectively, determined by artificial experience; , , and represent the nitrogen content, the phosphorus content, the potassium content and the soil electrical conductivity in the soil after data preprocessing; If any of the following determination conditions is met, the pest control decision instruction is generated, and the determination conditions are as follows: Condition one: the vibration sensor detects three times in succession , is the gravitational acceleration; Condition two: YOLOv5 model identifies corn borer mouthpart density Greater than 3 per plant or peanut aphid density Greater than 5 per plant.

7. The system of claim 1, wherein, The local control unit comprises: An irrigation control module, a fertilization control module and a sprayer control module; The irrigation control module is configured to drive a drip irrigation system to adjust irrigation flow and irrigation time on demand; the fertilization control module is configured to drive a variable rate fertilizer pump to dynamically control fertilizer solution supply amount and concentration; and the sprayer control module is configured to drive a pest control sprayer.

Images