Intelligent water and fertilizer management method and system for crops

Abstract

The present invention relates to an intelligent water and fertilizer management system for crops, comprising: a data acquisition and monitoring module for monitoring soil environmental data and meteorological conditions; a data storage and analysis module for storing all data acquired from the data acquisition and monitoring module; a model customization module for formulating irrigation and fertilization rules according to the different growth stages and characteristics of different crops. Simultaneously, based on the basic rule settings, different random forest models customized for different crops are introduced to optimize irrigation and fertilization strategies. The random forest model discovers the correlation between crop growth and water and fertilizer supply, adjusts the irrigation and fertilization strategies based on the correlation, and optimizes the rules and models based on abnormal information provided by an alarm and fault handling module; and an intelligent control module for automatically adjusting the water-fertilizer ratio and irrigation scheme based on real-time monitoring data and crop growth requirements in combination with a model provided by the model customization module.

Description

Technical Field

[0001] The present invention relates to the technical field of crop management, and in particular to an intelligent water and fertilizer management method and system for crops. Background Art

[0002] Intelligent crop water and fertilizer management uses advanced sensors, controllers, actuators, and other equipment to monitor soil moisture, nutrient status, and crop growth requirements in real time, automatically adjusting irrigation and fertilization rates to achieve precise water and fertilizer supply. This system relies primarily on precise sensing and intelligent control. Various high-precision sensors installed in the field collect real-time field environmental data, transmit this data to an intelligent control system for analysis and decision-making, and ultimately achieve precise water and fertilizer management through actuators.

[0003] Intelligent control algorithms are the core of intelligent water and fertilizer management systems. They must make decisions based on real-time monitoring data and adjust irrigation and fertilization rates. However, water and fertilizer requirements vary significantly across crops, growth stages, and soil conditions. Consider an intelligent water and fertilizer management system where wheat and tomatoes are grown simultaneously. Wheat, as a grain crop, has a relatively long growth cycle and relatively stable water and fertilizer requirements. Tomatoes, on the other hand, as a vegetable crop, have a shorter growth cycle and significantly varying water and fertilizer requirements across different growth stages. Intelligent control algorithms fail to fully account for these differences in water and fertilizer requirements for these two crops and may adopt the same irrigation and fertilization strategies. This can result in wheat yield and quality being affected by excess water and fertilizer in the late growth stages, while tomato growth is hindered by insufficient water and fertilizer during the critical growth period. Therefore, a method and system for intelligent water and fertilizer management of crops is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent water and fertilizer management method and system for crops.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] An intelligent water and fertilizer management system for crops, comprising:

[0007] Data acquisition and monitoring module: uses sensors (such as soil moisture sensors, nutrient sensors, meteorological sensors, etc.) to monitor soil environmental data (soil moisture, soil conductivity, soil pH value, nutrient content) and meteorological conditions (temperature, humidity, light, etc.) in real time, and sends the soil environmental data and meteorological conditions to the data storage and analysis module and the intelligent control module;

[0008] Data Storage and Analysis Module: This module is responsible for storing all data obtained from the Data Acquisition and Monitoring Module, and performs preliminary data cleaning, integration, and preprocessing. It also provides data analysis functions such as data trend analysis and anomaly detection, and provides the analyzed data to the Remote Monitoring and Management Module. At the same time, it provides the stored historical data to the Model Customization Module.

[0009] Model customization module: This module is responsible for formulating irrigation and fertilization rules based on the different growth stages and characteristics of different crops. At the same time, based on the basic rule settings, it introduces different random forest models customized for different crops to optimize irrigation and fertilization strategies. The random forest model discovers the correlation between crop growth and water and fertilizer supply, and adjusts irrigation and fertilization strategies based on the correlation rules. The rules and models are optimized based on the abnormal information provided by the alarm and fault handling module. For crops with long growth cycles and relatively stable water and fertilizer requirements (such as wheat), a relatively stable irrigation and fertilization plan is adopted to ensure that the crops receive a continuous and appropriate supply of water and fertilizer during the growth process. For crops with short growth cycles and large differences in water and fertilizer requirements (such as tomatoes), water and fertilizer supply is increased during their critical growth period (such as flowering and fruiting period) to meet their needs for rapid growth and fruiting.

[0010] Intelligent control module: Based on real-time monitoring data and crop growth needs, combined with the model provided by the model customization module, it automatically adjusts the water-fertilizer ratio and irrigation plan, including the control of intelligent valves, irrigation pipe networks, and water-fertilizer machines. The intelligent control module sends the adjusted water-fertilizer ratio and irrigation plan, as well as the real-time operating status, to the remote monitoring and management module. If the intelligent control module detects an abnormality or equipment failure while adjusting the water-fertilizer ratio and irrigation plan, it will immediately send an alarm to the alarm and fault handling module.

[0011] Remote monitoring and management module: allows users to view various monitoring data of the irrigation area and the operating status of the equipment in real time through management platforms such as mobile phone apps and computer software, and perform remote operations and management, such as adjusting irrigation and fertilization parameters and viewing historical data. Users adjust irrigation and fertilization parameters through the remote monitoring and management module, and the parameters are sent to the model customization module for update. At the same time, the model customization module sends the customized random forest model and rules to the remote monitoring and management module for display and application;

[0012] Alarm and fault handling module: When an abnormal situation or equipment failure is detected, an alarm message is automatically issued and corresponding fault handling measures are taken. The fault handling measures include the identification of abnormal data, the sending of alarm information (such as SMS, email, APP push, etc.) and the provision of fault handling suggestions. The alarm and fault handling module sends the monitored abnormal information to the model customization module and sends the alarm information to the remote monitoring and management module.

[0013] The above technical solution further includes:

[0014] Furthermore, the data acquisition and monitoring module includes a sensor unit, an Internet of Things transmission device and a data acquisition terminal, which is responsible for directly measuring and collecting soil environmental data and meteorological conditions. The temperature sensors include soil moisture sensors, nutrient sensors, meteorological sensors, etc., which can monitor in real time and convert them into electrical signals or digital signals for subsequent processing and transmission. The Internet of Things transmission device is responsible for transmitting the data collected by the sensor unit to the data acquisition terminal. The data acquisition terminal is responsible for receiving the data sent by the Internet of Things transmission device and performing preliminary processing and storage. The data acquisition terminal sends the processed data to the data storage and analysis module and the intelligent control module.

[0015] Furthermore, the data acquisition and transmission module includes a data acquisition interface unit, a data storage management unit, a data cleaning and integration unit, a data preprocessing unit and a data analysis engine unit. The data acquisition interface unit is responsible for communicating with the data acquisition and monitoring module, receiving the data transmitted by it, and providing data format conversion function to ensure that the received data meets the internal processing requirements of the module. The data storage management unit is responsible for data storage and retrieval, managing the database or data warehouse, ensuring the security and integrity of the data, and providing an efficient data indexing and query mechanism. The data cleaning and integration unit implements a data cleaning algorithm, removes invalid and erroneous data, integrates data from different data sources, and forms a unified data view. The data preprocessing unit preprocesses the data to improve the accuracy of data analysis and provides data smoothing, standardization, dimensionality reduction and other processing functions. The data analysis engine unit implements data analysis, such as trend analysis, anomaly detection, etc.

[0016] Furthermore, the model customization module includes a rule setting unit, a random forest customization unit and an irrigation and fertilization strategy adjustment unit. The rule setting unit receives growth data, environmental data and historical irrigation and fertilization data of different crops, formulates irrigation and fertilization rules according to different growth stages and characteristics of crops, and passes the formulated rules to the irrigation and fertilization strategy adjustment unit. The random forest customization unit receives growth data, environmental data and historical irrigation and fertilization data of different crops, customizes different random forest models for different crops, uses customized models to analyze data, discovers the correlation between crop growth and water and fertilizer supply, and passes the discovered correlation and adjustment suggestions to the irrigation and fertilization strategy adjustment unit. The irrigation and fertilization strategy adjustment unit receives information transmitted by the rule setting unit and the random forest model customization unit, comprehensively considers the results of rule setting and model prediction, and adjusts the irrigation and fertilization strategies.

[0017] Furthermore, the intelligent control module includes a data receiving unit, an intelligent control unit, an equipment control unit and an abnormal data identification unit. The data receiving unit is responsible for receiving data from the sensor network and the equipment control unit, and transmitting the received data to the intelligent control unit and the abnormal data identification unit. The intelligent control unit receives the data transmitted by the data receiving unit, and at the same time, accepts the random forest model provided by the model customization module, formulates the water-fertilizer ratio and irrigation plan, and transmits the control instructions to the equipment control unit. The equipment control unit receives the control instructions transmitted by the intelligent control unit, controls the switching and adjustment of the intelligent valve, irrigation pipeline network and water and fertilizer machine, and provides real-time feedback on the operating status of the equipment and the irrigation and fertilization process to the intelligent control unit and the abnormal data processing and alarm unit. The abnormal data identification unit receives the data transmitted by the data receiving unit, analyzes and judges the data according to preset rules and thresholds, identifies abnormal conditions or equipment failures, and transmits the identification results to the alarm and fault processing module.

[0018] Furthermore, the remote monitoring and management module includes a device management unit, a database unit, and a management platform unit. The device management unit is responsible for the control and management of intelligent valves, irrigation pipe networks, and water and fertilizer machines, receives control instructions sent by the management platform unit, executes corresponding device operations, and transmits the operating status and feedback information of the equipment to the management platform of the remote monitoring and management module. The database unit stores historical data such as monitoring data of the irrigation area, equipment operating status, and user operation records, provides data query and export functions, and supports users to query unit historical data through the management platform. The management platform unit provides a user interaction interface, receives user operation instructions, and transmits them to the device management unit for execution, providing historical data query, alarm and notification functions;

[0019] Furthermore, the alarm and fault handling module includes an alarm information sending unit, a fault handling suggestion providing unit, and a fault recording and tracking unit. The alarm information sending unit receives the identification result transmitted by the abnormal data identification unit, generates abnormal information according to the identification result, and sends the abnormal information to relevant personnel according to the preset alarm method and recipient information, records the sending status of the abnormal information and the recipient feedback, and at the same time passes the abnormal information to the model customization module for model optimization. The fault handling suggestion providing unit receives the equipment fault information transmitted by the abnormal data identification unit, provides corresponding fault handling suggestions according to the fault type and the specific situation of the equipment, and passes the fault handling suggestions to the user or relevant maintenance personnel. The fault recording and tracking unit records each abnormal situation and related information of the fault handling, provides fault tracking and query functions, and facilitates users to view historical fault records.

[0020] Furthermore, the specific steps of the model customization module are:

[0021] Data collection and preprocessing: Collect growth data, environmental data, and historical irrigation and fertilization data of different crops, and preprocess them to ensure data accuracy and consistency;

[0022] Establishment of basic rule base: According to the growth characteristics and needs of different crops, a basic rule base is established, including irrigation rules, fertilization rules, etc.

[0023] Random forest model training: Features related to crop growth and water and fertilizer supply are selected from preprocessed data, such as soil moisture, soil nutrient content, meteorological conditions (such as temperature, humidity, and light intensity), historical irrigation volume, and fertilizer application rates. This is used to determine the target the model needs to predict, such as crop growth rate, yield, and quality, or more directly, to determine whether irrigation or fertilization is needed.

[0024] Constructing decision trees: Random forests consist of multiple decision trees. Each tree is independently constructed based on a subset of the training data (obtained through bootstrap sampling). When constructing each tree, a portion of features are randomly selected from all features as candidate split features.

[0025] Splitting nodes: For each node of the decision tree, the best feature is selected for splitting to maximize the information gain. The information gain is calculated as:

[0026]

[0027] Where D is the dataset, A is the feature, Values(A) is all possible values ​​of feature A, and D v It is a subset of the dataset where the feature A takes the value v;

[0028] Integrated prediction: The final prediction result is determined through a voting mechanism;

[0029] Model evaluation and optimization: Use k-fold cross-validation to evaluate model performance and provide feature importance assessment, which helps understand which features have the greatest impact on model prediction results. Use grid search, random search, and other methods to adjust random forest hyperparameters (such as the number of trees, maximum depth, minimum number of sample splits, etc.) to optimize model performance.

[0030] Furthermore, the model customization module optimizes the rules and model specific steps according to the abnormal information:

[0031] Adjust rules: Analyze and adjust the irrigation and fertilization rules in the basic rule base based on abnormal information. For example, if soil moisture is detected to be persistently low, it may be necessary to increase the irrigation frequency or amount.

[0032] Introducing new rules: In response to newly discovered abnormal situations, new rules are introduced to adapt to crop growth needs;

[0033] Update the dataset: add abnormal data and its processing results (such as manually adjusted irrigation and fertilization strategies) to the training dataset;

[0034] Retrain the model: Use the updated dataset to retrain the random forest model so that it can learn new rules and patterns.

[0035] A control method for an intelligent water and fertilizer management system for crops, comprising the following steps:

[0036] Data collection: Soil moisture, soil conductivity, soil pH, nutrient content and meteorological conditions are collected in real time through the data collection and monitoring module. The data is transmitted to the data storage and analysis module and the intelligent control module through the Internet of Things technology;

[0037] Data analysis and decision-making: The intelligent control module uses the model provided by the model customization module to analyze and process the collected data. Based on the analysis results, it automatically selects the optimal irrigation and fertilization strategy and generates corresponding control instructions;

[0038] Instruction issuance: The intelligent control module sends control instructions to the intelligent valve and water and fertilizer machine. The intelligent valve adjusts the flow and ratio of water and fertilizer according to the instructions to ensure accurate irrigation and fertilization;

[0039] Irrigation mode selection: Based on the growth characteristics and needs of crops, the intelligent control module selects the appropriate irrigation mode (such as timed irrigation, remote irrigation, cycle irrigation, etc.). The intelligent control module automatically adjusts the irrigation strategy based on the preset irrigation plan and the actual growth conditions of the crops;

[0040] Remote monitoring: Users can view various monitoring data of the irrigation area and the operating status of the equipment in real time through the remote monitoring and management module. Through the visual interface provided by the platform, users can intuitively understand the irrigation and fertilization status of the farmland;

[0041] Alarm processing: When an abnormal situation or equipment failure is detected, the alarm and fault processing module automatically sends an alarm message. After receiving the alarm message, the user takes measures to ensure the normal growth of crops.

[0042] The present invention has the following beneficial effects:

[0043] 1. This invention develops more refined irrigation and fertilization rules tailored to the different growth stages and characteristics of crops like wheat and tomatoes. Furthermore, customized machine learning models are used for different crops to automatically discover the correlation between crop growth and water and fertilizer supply. Irrigation and fertilization strategies are dynamically adjusted based on crop type and growth stage.

[0044] 2. In the present invention, rules and models are optimized based on abnormal information. Through the optimized rules and models, the system can more accurately control the amount and time of irrigation and fertilization, thereby meeting the growth needs of crops and improving crop yield and quality. In addition, after introducing new rules and retraining the model, it can better adapt to different environments and crop types. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a system block diagram of an intelligent water and fertilizer management method and system for crops proposed by the present invention;

[0046] Figure 2 This is a method step diagram of an intelligent water and fertilizer management method and system for crops proposed by the present invention. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0048] See also Figure 1-Figure 2 As shown, the present invention is an intelligent water and fertilizer management system for crops, comprising:

[0049] Data acquisition and monitoring module: uses sensors (such as soil moisture sensors, nutrient sensors, meteorological sensors, etc.) to monitor soil environmental data (soil moisture, soil conductivity, soil pH value, nutrient content) and meteorological conditions (temperature, humidity, light, etc.) in real time, and sends the soil environmental data and meteorological conditions to the data storage and analysis module and the intelligent control module;

[0050] Data storage and analysis module: responsible for storing all data obtained from the data acquisition and monitoring module, and performing preliminary data cleaning, integration and preprocessing, providing data analysis functions such as data trend analysis and anomaly detection, and providing the analyzed data to the remote monitoring and management module;

[0051] Model Customization Module: This module is responsible for developing irrigation and fertilization rules based on the different growth stages and characteristics of different crops. Furthermore, based on the basic rule settings, it introduces customized random forest models for different crops to optimize irrigation and fertilization strategies. The random forest model discovers the correlation between crop growth and water and fertilizer supply, and adjusts irrigation and fertilization strategies accordingly. The module also optimizes rules and models based on abnormal information provided by the alarm and fault handling module.

[0052] Intelligent control module: Based on real-time monitoring data and crop growth needs, combined with the model provided by the model customization module, it automatically adjusts the water-fertilizer ratio and irrigation plan, including control of intelligent valves, irrigation pipe networks, and water and fertilizer machines. The intelligent control module sends the adjusted water-fertilizer ratio and irrigation plan, as well as the real-time operating status, to the remote monitoring and management module. If the intelligent control module detects an abnormality or equipment failure while adjusting the water-fertilizer ratio and irrigation plan, it will immediately send an alarm to the alarm and fault handling module.

[0053] Remote Monitoring and Management Module: This module allows users to view various monitoring data and equipment operating status of the irrigation area in real time through mobile phone apps, computer software, and other management platforms, and to perform remote operations and management, such as adjusting irrigation and fertilization parameters and viewing historical data. Users adjust irrigation and fertilization parameters through the remote monitoring and management module, and the parameters are sent to the model customization module for update. At the same time, the model customization module sends the customized random forest model and rules to the remote monitoring and management module for display and application.

[0054] Alarm and fault handling module: When an abnormal situation or equipment failure is detected, an alarm message is automatically issued and corresponding fault handling measures are taken. The fault handling measures include the identification of abnormal data, the sending of alarm information (such as SMS, email, APP push, etc.) and the provision of fault handling suggestions. The alarm and fault handling module sends the monitored abnormal information to the model customization module and sends the alarm information to the remote monitoring and management module.

[0055] In one embodiment, for the above-mentioned data acquisition and monitoring module, the data acquisition and monitoring module includes a sensor unit, an Internet of Things transmission device and a data acquisition terminal, which is responsible for directly measuring and collecting soil environmental data and meteorological conditions. The sensors include soil moisture sensors, nutrient sensors, meteorological sensors, etc., which can monitor in real time and convert them into electrical signals or digital signals for subsequent processing and transmission. The Internet of Things transmission device is responsible for transmitting the data collected by the sensor unit to the data acquisition terminal. The data acquisition terminal is responsible for receiving the data sent by the Internet of Things transmission device and performing preliminary processing and storage. The data acquisition terminal sends the processed data to the data storage and analysis module and the intelligent control module.

[0056] In one embodiment, for the above-mentioned data acquisition and transmission module, the data acquisition and transmission module includes a data acquisition interface unit, a data storage management unit, a data cleaning and integration unit, a data preprocessing unit and a data analysis engine unit. The data acquisition interface unit is responsible for communicating with the data acquisition and monitoring module, receiving the data transmitted by it, and providing data format conversion function to ensure that the received data meets the internal processing requirements of the module. The data storage management unit is responsible for data storage and retrieval, managing the database or data warehouse, ensuring the security and integrity of the data, and providing an efficient data indexing and query mechanism. The data cleaning and integration unit implements a data cleaning algorithm, removes invalid and erroneous data, integrates data from different data sources, and forms a unified data view. The data preprocessing unit preprocesses the data to improve the accuracy of data analysis and provides data smoothing, standardization, dimensionality reduction and other processing functions. The data analysis engine unit implements data analysis, such as trend analysis, anomaly detection, etc.

[0057] In one embodiment, for the above-mentioned model customization module, the model customization module includes a rule setting unit, a random forest customization unit, and an irrigation and fertilization strategy adjustment unit. The rule setting unit receives growth data, environmental data, and historical irrigation and fertilization data of different crops, formulates irrigation and fertilization rules according to different growth stages and characteristics of the crops, and passes the formulated rules to the irrigation and fertilization strategy adjustment unit. The random forest customization unit receives growth data, environmental data, and historical irrigation and fertilization data of different crops, customizes different random forest models for different crops, uses customized models to analyze data, discovers the correlation between crop growth and water and fertilizer supply, and passes the discovered correlation and adjustment suggestions to the irrigation and fertilization strategy adjustment unit. The irrigation and fertilization strategy adjustment unit receives information transmitted by the rule setting unit and the random forest model customization unit, comprehensively considers the results of rule setting and model prediction, and adjusts the irrigation and fertilization strategy.

[0058] In one embodiment, for the above-mentioned intelligent control module, the intelligent control module includes a data receiving unit, an intelligent control unit, a device control unit and an abnormal data identification unit. The data receiving unit is responsible for receiving data from the sensor network and the device control unit, and transmitting the received data to the intelligent control unit and the abnormal data identification unit. The intelligent control unit receives the data transmitted by the data receiving unit, and at the same time, accepts the random forest model provided by the model customization module, formulates the water-fertilizer ratio and irrigation plan, and transmits the control instructions to the device control unit. The device control unit receives the control instructions transmitted by the intelligent control unit, controls the switching and adjustment of the intelligent valve, irrigation pipeline network and water and fertilizer machine, and provides real-time feedback on the operating status of the equipment and the irrigation and fertilization process to the intelligent control unit and the abnormal data processing and alarm unit. The abnormal data identification unit receives the data transmitted by the data receiving unit, analyzes and judges the data according to preset rules and thresholds, identifies abnormal conditions or equipment failures, and transmits the identification results to the alarm and fault processing module.

[0059] In one embodiment, for the above-mentioned remote monitoring and management module, the remote monitoring and management module includes a device management unit, a database unit and a management platform unit. The device management unit is responsible for the control and management of intelligent valves, irrigation pipe networks and water and fertilizer machines, receives control instructions sent by the management platform unit, and executes corresponding device operations, and transmits the operating status and feedback information of the equipment to the management platform of the remote monitoring and management module. The database unit stores historical data such as monitoring data of the irrigation area, equipment operating status, user operation records, etc., provides data query and export functions, and supports users to query unit historical data through the management platform. The management platform unit provides a user interaction interface, receives user operation instructions, and transmits them to the device management unit for execution, providing historical data query, alarm and notification functions.

[0060] In one embodiment, for the above-mentioned alarm and fault handling module, the alarm and fault handling module includes an alarm information sending unit, a fault handling suggestion providing unit, and a fault recording and tracking unit. The alarm information sending unit receives the recognition result transmitted by the abnormal data identification unit, generates abnormal information based on the recognition result, and sends the abnormal information to relevant personnel according to the preset alarm method and recipient information, records the sending status of the abnormal information and the recipient feedback, and at the same time passes the abnormal information to the model customization module for model optimization. The fault handling suggestion providing unit receives the equipment fault information transmitted by the abnormal data identification unit, provides corresponding fault handling suggestions based on the fault type and the specific situation of the equipment, and passes the fault handling suggestions to the user or relevant maintenance personnel. The fault recording and tracking unit records each abnormal situation and related information of the fault handling, provides fault tracking and query functions, and facilitates users to view historical fault records.

[0061] In one embodiment, for the above-mentioned model customization module, the specific steps of the model customization module are:

[0062] Data collection and preprocessing: Collect growth data, environmental data, and historical irrigation and fertilization data of different crops, and preprocess them to ensure data accuracy and consistency;

[0063] Establishment of basic rule base: According to the growth characteristics and needs of different crops, a basic rule base is established, including irrigation rules, fertilization rules, etc.

[0064] Random forest model training: Features related to crop growth and water and fertilizer supply are selected from preprocessed data, such as soil moisture, soil nutrient content, meteorological conditions (such as temperature, humidity, and light intensity), historical irrigation volume, and fertilizer application rates. This is used to determine the target the model needs to predict, such as crop growth rate, yield, and quality, or more directly, to determine whether irrigation or fertilization is needed.

[0065] Constructing decision trees: Random forests consist of multiple decision trees. Each tree is independently constructed based on a subset of the training data (obtained through bootstrap sampling). When constructing each tree, a portion of features are randomly selected from all features as candidate split features.

[0066] Splitting nodes: For each node of the decision tree, the best feature is selected for splitting to maximize the information gain. The information gain is calculated as:

[0067]

[0068] Where D is the dataset, A is the feature, Values(A) is all possible values ​​of feature A, and D v It is a subset of the dataset where the feature A takes the value v;

[0069] Integrated prediction: The final prediction result is determined through a voting mechanism;

[0070] Model evaluation and optimization: Use k-fold cross-validation to evaluate model performance and provide feature importance assessment, which helps understand which features have the greatest impact on model prediction results. Use grid search, random search, and other methods to adjust random forest hyperparameters (such as the number of trees, maximum depth, minimum number of sample splits, etc.) to optimize model performance.

[0071] In one embodiment, for the above-mentioned model customization module, the model customization module optimizes the rules and model according to the abnormal information in the following specific steps:

[0072] Adjust rules: Analyze and adjust the irrigation and fertilization rules in the basic rule base based on abnormal information. For example, if soil moisture is detected to be persistently low, it may be necessary to increase the irrigation frequency or amount.

[0073] Introducing new rules: In response to newly discovered abnormal situations, new rules are introduced to adapt to crop growth needs;

[0074] Update the dataset: add abnormal data and its processing results (such as manually adjusted irrigation and fertilization strategies) to the training dataset;

[0075] Retrain the model: Use the updated dataset to retrain the random forest model so that it can learn new rules and patterns;

[0076] Assume that during wheat planting, sensors detect that soil moisture remains below a set threshold for an extended period. Based on this abnormality, we can perform the following optimizations:

[0077] Rule optimization: Adjust irrigation rules, increase irrigation frequency or increase irrigation amount to ensure that soil moisture remains within the appropriate range.

[0078] Model retraining: Data from this period (including anomalies and manually adjusted irrigation strategies) is added to the training dataset and the random forest model is retrained. This retraining allows the model to better learn the relationship between soil moisture and irrigation strategies, leading to more accurate future forecasts.

[0079] A management method for an intelligent water and fertilizer management system for crops, comprising the following steps:

[0080] Data collection: Soil moisture, soil conductivity, soil pH, nutrient content and meteorological conditions are collected in real time through the data collection and monitoring module. The data is transmitted to the data storage and analysis module and the intelligent control module through the Internet of Things technology;

[0081] Data analysis and decision-making: The intelligent control module uses the model provided by the model customization module to analyze and process the collected data. Based on the analysis results, it automatically selects the optimal irrigation and fertilization strategy and generates corresponding control instructions;

[0082] Instruction issuance: The intelligent control module sends control instructions to the intelligent valve and water and fertilizer machine. The intelligent valve adjusts the flow and ratio of water and fertilizer according to the instructions to ensure accurate irrigation and fertilization;

[0083] Irrigation mode selection: Based on the growth characteristics and needs of crops, the intelligent control module selects the appropriate irrigation mode (such as timed irrigation, remote irrigation, cycle irrigation, etc.). The intelligent control module automatically adjusts the irrigation strategy based on the preset irrigation plan and the actual growth conditions of the crops;

[0084] Remote monitoring: Users can view various monitoring data of the irrigation area and the operating status of the equipment in real time through the remote monitoring and management module. Through the visual interface provided by the platform, users can intuitively understand the irrigation and fertilization status of the farmland;

[0085] Alarm processing: When an abnormal situation or equipment failure is detected, the alarm and fault processing module automatically sends an alarm message. After receiving the alarm message, the user takes measures to ensure the normal growth of crops.

[0086] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent water and fertilizer management system for crops, characterized in that: include: Data acquisition and monitoring module: monitors soil environmental data and meteorological conditions in real time through sensors, and sends the soil environmental data and meteorological conditions to the data storage and analysis module and the intelligent control module; Data storage and analysis module: responsible for storing all data obtained from the data acquisition and monitoring module, and performing preliminary data cleaning, integration and preprocessing, providing data analysis functions, and providing the analyzed data to the remote monitoring and management module. At the same time, the stored historical data is provided to the model customization module; Model customization module: This module is responsible for developing irrigation and fertilization rules based on the different growth stages and characteristics of different crops. Furthermore, based on the basic rule settings, it introduces customized random forest models for different crops to optimize irrigation and fertilization strategies. The random forest model discovers the correlation between crop growth and water and fertilizer supply, and adjusts irrigation and fertilization strategies accordingly. The module also optimizes the rules and models based on abnormal information provided by the alarm and fault handling module. Intelligent control module: Based on real-time monitoring data and crop growth needs, combined with the model provided by the model customization module, it automatically adjusts the water-fertilizer ratio and irrigation plan, including the control of intelligent valves, irrigation pipe networks, and water-fertilizer machines. The intelligent control module sends the adjusted water-fertilizer ratio and irrigation plan, as well as the real-time operating status, to the remote monitoring and management module. If the intelligent control module detects an abnormality or equipment failure while adjusting the water-fertilizer ratio and irrigation plan, it will immediately send an alarm to the alarm and fault handling module. Remote Monitoring and Management Module: This module allows users to view various monitoring data and equipment operating status of the irrigation area in real time through the management platform, and to perform remote operation and management. Users can adjust irrigation and fertilization parameters through the remote monitoring and management module, and these parameters are sent to the model customization module for update. At the same time, the model customization module sends the customized random forest model and rules to the remote monitoring and management module for display and application. Alarm and fault handling module: When an abnormal situation or equipment failure is monitored, an alarm message is automatically issued and corresponding fault handling measures are taken. The fault handling measures include identifying abnormal data, sending alarm information and providing fault handling suggestions. The alarm and fault handling module sends the monitored abnormal information to the model customization module and sends the alarm information to the remote monitoring and management module.

2. The intelligent water and fertilizer management system for crops according to claim 1, characterized in that: The data acquisition and monitoring module includes a sensor unit, an Internet of Things transmission device and a data acquisition terminal. The sensor unit is responsible for directly measuring and collecting soil environmental data and meteorological conditions. The Internet of Things transmission device is responsible for transmitting the data collected by the sensor unit to the data acquisition terminal. The data acquisition terminal is responsible for receiving the data sent by the Internet of Things transmission device and performing preliminary processing and storage. The data acquisition terminal sends the processed data to the data storage and analysis module and the intelligent control module.

3. The intelligent water and fertilizer management system for crops according to claim 1, characterized in that: The data acquisition and transmission module includes a data acquisition interface unit, a data storage management unit, a data cleaning and integration unit, a data preprocessing unit and a data analysis engine unit. The data acquisition interface unit is responsible for communicating with the data acquisition and monitoring module and receiving the data transmitted by it. The data storage management unit is responsible for storing and retrieving data and managing the database or data warehouse. The data cleaning and integration unit implements a data cleaning algorithm, removes invalid and erroneous data, integrates data from different data sources, and forms a unified data view. The data preprocessing unit preprocesses the data, and the data analysis engine unit implements data analysis.

4. The intelligent water and fertilizer management system for crops according to claim 1, characterized in that: The model customization module includes a rule setting unit, a random forest customization unit and an irrigation and fertilization strategy adjustment unit. The rule setting unit receives growth data, environmental data and historical irrigation and fertilization data of different crops, formulates irrigation and fertilization rules according to different growth stages and characteristics of crops, and passes the formulated rules to the irrigation and fertilization strategy adjustment unit. The random forest customization unit receives growth data, environmental data and historical irrigation and fertilization data of different crops, customizes different random forest models for different crops, uses customized models to analyze data, discovers the correlation between crop growth and water and fertilizer supply, and passes the discovered correlation and adjustment suggestions to the irrigation and fertilization strategy adjustment unit. The irrigation and fertilization strategy adjustment unit receives information transmitted by the rule setting unit and the random forest model customization unit, comprehensively considers the results of rule setting and model prediction, and adjusts the irrigation and fertilization strategy.

5. The intelligent water and fertilizer management system for crops according to claim 4, characterized in that: The intelligent control module includes a data receiving unit, an intelligent control unit, an equipment control unit and an abnormal data identification unit. The data receiving unit is responsible for receiving data from the sensor network and the equipment control unit, and transmitting the received data to the intelligent control unit and the abnormal data identification unit. The intelligent control unit receives the data transmitted by the data receiving unit, and at the same time, accepts the random forest model provided by the model customization module, formulates the water-fertilizer ratio and irrigation plan, and transmits the control instructions to the equipment control unit. The equipment control unit receives the control instructions transmitted by the intelligent control unit, controls the switching and adjustment of the intelligent valve, irrigation pipeline network and water and fertilizer machine, and provides real-time feedback on the operating status of the equipment and the irrigation and fertilization process to the intelligent control unit and the abnormal data processing and alarm unit. The abnormal data identification unit receives the data transmitted by the data receiving unit, analyzes and judges the data according to preset rules and thresholds, identifies abnormal conditions or equipment failures, and transmits the identification results to the alarm and fault processing module.

6. The intelligent water and fertilizer management system for crops according to claim 1, characterized in that: The remote monitoring and management module includes a device management unit, a database unit and a management platform unit. The device management unit is responsible for the control and management of intelligent valves, irrigation pipe networks and water and fertilizer machines, receives control instructions sent by the management platform unit, and performs corresponding device operations, and transmits the operating status and feedback information of the equipment to the management platform of the remote monitoring and management module. The database unit stores historical data of the irrigation area, provides data query and export functions, and supports users to query unit historical data through the management platform. The management platform unit provides a user interaction interface, receives user operation instructions, and transmits them to the device management unit for execution.

7. The intelligent water and fertilizer management system for crops according to claim 5, characterized in that: The alarm and fault handling module includes an alarm information sending unit, a fault handling suggestion providing unit, and a fault recording and tracking unit. The alarm information sending unit receives the identification result transmitted by the abnormal data identification unit, generates abnormal information based on the identification result, sends the abnormal information to relevant personnel, records the sending status of the abnormal information and the feedback of the recipient, and at the same time passes the abnormal information to the model customization module for model optimization. The fault handling suggestion providing unit receives the equipment fault information transmitted by the abnormal data identification unit, provides corresponding fault handling suggestions based on the fault type and the specific situation of the equipment, and passes the fault handling suggestions to the user or relevant maintenance personnel. The fault recording and tracking unit records each abnormal situation and related information of the fault handling, and provides fault tracking and query functions.

8. The intelligent water and fertilizer management system for crops according to claim 4, characterized in that: The specific steps of the model customization module are: Data collection and preprocessing: Collect growth data, environmental data, and historical irrigation and fertilization data of different crops and perform preprocessing; Establishment of basic rule base: Establish a basic rule base based on the growth characteristics and needs of different crops; Random forest model training: Select features related to crop growth and water and fertilizer supply from the preprocessed data to determine the target that the model needs to predict; Constructing decision trees: Random forests consist of multiple decision trees, each of which is independently constructed based on a subset of the training data. When constructing each tree, a subset of features is randomly selected from all features as candidate split features. Splitting nodes: For each node of the decision tree, the best feature is selected for splitting to maximize the information gain. The information gain is calculated as: Where D is the dataset, A is the feature, Values(A) is all possible values ​​of feature A, and D υ It is a subset of the dataset where the feature A takes the value of υ; Integrated prediction: The final prediction result is determined through a voting mechanism; Model evaluation and optimization: Use k-fold cross-validation to evaluate model performance, provide feature importance assessment, and optimize model performance by adjusting the hyperparameters of random forest.

9. The intelligent water and fertilizer management system for crops according to claim 4, characterized in that: The model customization module optimizes rules and models according to abnormal information: Adjust rules: Analyze and adjust the irrigation and fertilization rules in the basic rule base based on abnormal information; Introducing new rules: In response to newly discovered abnormal situations, new rules are introduced to adapt to crop growth needs; Update the data set: add the abnormal data and its processing results to the training data set; Retrain the model: Use the updated dataset to retrain the random forest model so that it can learn new rules and patterns.

10. The management method of an intelligent water and fertilizer management system for crops according to claim 1, characterized in that: The following steps are involved: Data collection: Soil moisture, soil conductivity, soil pH, nutrient content and meteorological conditions are collected in real time through the data collection and monitoring module. The data is transmitted to the data storage and analysis module and the intelligent control module through the Internet of Things technology; Data analysis and decision-making: The intelligent control module uses the model provided by the model customization module to analyze and process the collected data. Based on the analysis results, it automatically selects the optimal irrigation and fertilization strategy and generates corresponding control instructions; Instruction issuance: The intelligent control module sends control instructions to the intelligent valve and water and fertilizer machine. The intelligent valve adjusts the flow and ratio of water and fertilizer according to the instructions. Irrigation mode selection: The intelligent control module selects the appropriate irrigation mode based on the growth characteristics and needs of the crops. The intelligent control module automatically adjusts the irrigation strategy based on the preset irrigation plan and the actual growth conditions of the crops. Remote monitoring: Users can view various monitoring data of the irrigation area and the operating status of the equipment in real time through the remote monitoring and management module; Alarm processing: When an abnormal situation or equipment failure is detected, the alarm and fault processing module automatically issues an alarm message. After receiving the alarm message, the user takes measures to handle it. At the same time, the alarm and fault processing module sends the abnormal information to the model customization module to optimize the rules and models.

Images