System and method for preventing and controlling citrus diseases and insect pests

Through the citrus pest management system combined with the Internet of Things and machine learning, the use of drones and automatic spraying robots to achieve precise application of medicines, solving the problems of low efficiency and limited automation applications of citrus pest management in the existing technology, and improving the pest management Accuracy and efficiency reduce pesticide use and environmental pollution.

CN118968351AInactive Publication Date: 2024-11-15GANZHOU CITRUS SCI RES INST
View PDF 0 Cites 10 Cited by

Patent Information

Application Number
CN202410973998.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-19
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing citrus pest management methods are labor-intensive and inefficient, making them difficult to achieve precise prevention and control, and are limited in automation and intelligent applications, and face technical challenges such as real-time, security, processing accuracy and execution accuracy of data transmission.

Method used

Using IoT technology to ensure real-time transmission and security of data, combined with machine learning to generate real-time prevention and control strategies, use drones and automatic spraying robots to achieve automated execution, and collect data through environmental sensors, plant physiological sensors and image collection equipment, and use time series analysis The model and convolutional neural network perform data processing and analysis to generate accurate prevention and control strategies.

Benefits of technology

It improves the accuracy and efficiency of pest management, reduces the excessive use of pesticides and environmental pollution, achieves accurate pest identification and prediction, improves the adaptability and effectiveness of prevention and control, and supports sustainable agricultural practices.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118968351A_ABST
    Figure CN118968351A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of pest control, and discloses a citrus pest control system and method, and the system comprises a data collection module which is used for collecting environment data, plant health data and image data, and an environment sensor, a plant physiological sensor and an image collection device which are disposed in a citrus garden; the data transmission module is used for transmitting the data collected by the data collection module to a server through an Internet of Things gateway; and the data processing and analysis module is used for processing a time sequence analysis model of environment data and a convolutional neural network model of a disease and insect pest image. Through cooperation of the data collection, transmission and processing module, environment and plant physiological data can be analyzed in real time, the development trend of plant diseases and insect pests can be recognized and predicted in an early stage by using a convolutional neural network model, and the intelligent decision support module generates an accurate prevention and control strategy according to the data, improves the response precision, enables prevention and control measures to adapt to environment changes, and improves the prevention and control efficiency. And timely and effective pest and disease management is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of pest control, and in particular to a system and method for preventing and controlling citrus pests and diseases. Background Art

[0002] In modern agricultural production, pest and disease management is one of the key factors to ensure crop health and yield. Especially for fruits such as citrus, due to their high value and sensitivity to pests and diseases, effective pest and disease control strategies are particularly important. Traditional pest and disease management methods mainly rely on manual inspection and regular spraying of chemical pesticides. This method is not only labor-intensive and inefficient, but also difficult to achieve precise prevention and control, which can easily lead to excessive use of pesticides and pose a potential threat to the environment and human health.

[0003] With the development of information technology and automation technology, intelligent agriculture has become a trend. Especially in the field of pest management, the use of advanced sensor technology and data analysis methods can achieve early identification and positioning of pests and diseases.

[0004] However, although some existing solutions can automatically collect data, they often lack effective data integration and analysis capabilities and cannot provide real-time decision support. The application of automatic execution systems is not widespread and still requires manual intervention and adjustment. In addition, existing technologies still face some technical challenges in realizing the automation and intelligence of pest control, such as the real-time and security of data transmission, the accuracy and efficiency of data processing, and the precision and reliability of the execution mechanism. Summary of the invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a system and method for preventing and controlling citrus pests and diseases, which uses the Internet of Things technology to ensure the real-time transmission and security of data, and uses machine learning to optimize decision support and generate real-time prevention and control strategies. The automated execution module (drones and robots) accurately applies pesticides, and the implementation and feedback module monitors the execution effect and automatically adjusts it, improving the accuracy and efficiency of pest and disease management.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a system for preventing and controlling citrus pests and diseases, comprising:

[0007] Data collection module, which includes environmental sensors, plant physiological sensors and image collection equipment deployed in the citrus orchard to collect environmental data, plant health data and image data;

[0008] A data transmission module is used to transmit the data collected by the data collection module to a server through an Internet of Things gateway;

[0009] Data processing and analysis module, which is used to process time series analysis models of environmental data and convolutional neural network models for identifying pest and disease images;

[0010] An intelligent decision support module, used to generate a prevention and control strategy based on the analysis results of the data processing and analysis module;

[0011] Automated execution modules, including drones and autonomous spraying robots for implementing control strategies.

[0012] Preferably, the environmental sensors include a temperature and humidity sensor, a soil moisture sensor, a light sensor, and a wind speed and direction sensor.

[0013] Preferably, the plant physiological sensor includes a chlorophyll fluorescence sensor and a spectral sensor.

[0014] Preferably, the image collection device includes a high-resolution camera and a camera mounted on a drone.

[0015] Preferably, the data processing and analysis module includes an autoregressive difference moving average model applied to environmental data, the formula of which is:

[0016]

[0017] Where p, d, q are the number of autoregressive terms, the number of differences, and the number of moving average terms, respectively. i is the autoregressive coefficient, θ i is the moving average coefficient, L is the lag operator, ∈ t is the error term.

[0018] Preferably, the data processing and analysis module includes a convolutional neural network model for pest and disease image recognition, and its generation formula is:

[0019] Convolutional layer output:

[0020]

[0021] in is the output of the lth convolutional layer, f is the activation function, W (l) is the convolution kernel of the lth layer, * indicates the convolution operation, I (l-1) is the input of the l-1th layer, b (l) is the bias of the lth layer;

[0022] Pooling layer output:

[0023]

[0024] in is the output of the lth pooling layer, where pool is the pooling operation;

[0025] Fully connected layer output:

[0026]

[0027] Among them, fc is the output of the fully connected layer, g is the activation function, and W fc is the weight matrix of the fully connected layer, flatten means flattening the output of the pooling layer into a vector, b fc is the bias of the fully connected layer;

[0028] Output layer:

[0029] O softmax =softmax(W out ·O fc +b out )

[0030] Among them, softmax is the output of the Softmax layer, W out is the weight matrix of the output layer, b out is the bias of the output layer.

[0031] Preferably, the intelligent decision support module includes:

[0032] The data integration module is used to integrate the environmental data analysis results and the pest and disease image recognition results, including a method based on multi-source data fusion. The formula is:

[0033]

[0034] Where F(X t ) is comprehensive data, w i is the weight, X i,t is the value of the i-th data at time t, and n is the number of data sources;

[0035] The risk assessment module is used to assess the risk level of pests and diseases based on comprehensive data, using a weighted risk assessment formula, which is:

[0036]

[0037] Where r is the risk level, w i is the weight of each data source, f i (X t ) is the function output of the ith data source at time t, and n is the number of data sources;

[0038] The strategy generation module is used to generate specific prevention and control strategies based on risk levels, including prevention and control methods, optimal application time, pesticide dosage and spraying frequency. The formula is:

[0039] D=α·R·A

[0040] Where D is the drug dose, α is the dose adjustment coefficient, R is the risk level, and A is the area of ​​the affected region;

[0041] The implementation and feedback module is used to implement prevention and control strategies and collect effect data after implementation to adjust future decisions.

[0042] Preferably, the strategy generation module includes:

[0043] Selection of control methods: Select appropriate chemical, biological or physical control methods based on risk level and environmental conditions;

[0044] Optimal application time: Determine the best time period for spraying the pesticide based on environmental data prediction;

[0045] Calculation of pesticide dosage and spraying frequency: Calculate the appropriate pesticide dosage and spraying frequency based on the crop growth stage, pest and disease type, and risk assessment results.

[0046] Preferably, the implementation and feedback module includes:

[0047] Execute prevention and control strategies: drones and automatic spraying robots receive instructions generated by the decision support module and spray at designated times and areas;

[0048] Real-time monitoring and adjustment: Through the real-time monitoring system, the environmental changes and plant responses during the spraying process are monitored to ensure the spraying effect. If uneven spraying or poor effect is detected, the system can automatically adjust the spraying parameters and instruct repeated spraying;

[0049] Effect data collection: Collect effect data on pest and disease reduction rate and plant health recovery after spraying. Based on the collected data, adjust the model and decision logic to optimize future prevention and control strategies.

[0050] A method for preventing and controlling citrus pests and diseases comprises the following steps:

[0051] Collect environmental data, plant health data and image data through the data collection module;

[0052] Transmitting data to the server through the data transmission module;

[0053] Apply time series analysis models and convolutional neural network models to process and analyze data in the data processing and analysis module;

[0054] Based on the analysis results, the intelligent decision support module generates a control strategy, including selecting control methods, determining the best application time, calculating the dosage of pesticides and spraying frequency, and implements the control strategy through the automated execution module;

[0055] Collect post-implementation performance data and use that data to adjust future decisions.

[0056] The present invention provides a system and method for preventing and controlling citrus pests and diseases. It has the following beneficial effects:

[0057] 1. The present invention provides a highly accurate monitoring system through the cooperation of a data collection module, a data transmission module, and a data processing and analysis module. It can capture and analyze environmental and plant physiological data in real time, and use advanced convolutional neural network models for image analysis. It can identify pests and diseases at an early stage and accurately predict their development trends. Finally, under the action of an intelligent decision support module, specific prevention and control strategies can be generated based on these data to ensure timely and accurate responses, improve the accuracy of diagnosis and response, and enable prevention and control measures to flexibly adapt to changing environmental conditions.

[0058] 2. The automated execution module of the present invention can utilize drones and automatic spraying robots to accurately apply pesticides according to the prevention and control instructions of the intelligent decision support module, greatly improving the efficiency of operations and reducing labor costs. The implementation and feedback module can monitor the prevention and control process and plant reactions in real time, automatically adjust spraying parameters according to actual results, ensure the best control effect, and significantly improve the adaptability and effectiveness of prevention and control.

[0059] 3. The present invention significantly reduces excessive use of pesticides and environmental pollution by precisely controlling the dosage of pesticides and finely adjusting the spraying strategy. In addition, precise pest and disease management also helps optimize water resource utilization and supports sustainable agricultural practices. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 is a system structure diagram of the present invention;

[0061] Figure 2 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0063] Please see attached Figure 1 The embodiment of the present invention provides a system and method for preventing and controlling citrus pests and diseases, comprising:

[0064] Data collection module, environmental sensors, plant physiological sensors and image collection equipment deployed in the citrus orchard, used to collect environmental data, plant health data and image data. Environmental sensors are used to monitor the microclimate conditions of the farm in real time. Plant physiological sensors are used to monitor key physiological parameters of plants and help assess plant health. Image collection equipment is used to capture images of citrus trees and their fruits to facilitate subsequent identification of pests and diseases through image processing technology;

[0065] The data transmission module is used to transmit the data collected by the data collection module to the server through the IoT gateway. Through the IoT gateway, all collected data are transmitted to the server securely and in real time. The data integration provides a comprehensive data view for pest management, ensuring the timeliness and accuracy of the information;

[0066] Data processing and analysis module, which is used to process the time series analysis model of environmental data and the convolutional neural network model for identifying pest and disease images. The time series analysis model uses the collected environmental data and applies time series analysis to predict high-risk environmental conditions that may appear in the future. The convolutional neural network model is used to analyze plant images obtained from image collection equipment to identify early signs of pests and diseases, enabling the model to learn complex features in images and accurately identify various pests and diseases;

[0067] Intelligent decision support module, used to generate prevention and control strategies based on the analysis results of the data processing and analysis module. The intelligent decision support module automatically generates specific optimal spraying time, pesticide selection, spraying amount and frequency strategies based on the information provided by the data processing and analysis module, which greatly reduces the decision-making burden of farm managers through intelligent operation;

[0068] The automated execution module includes drones and automatic spraying robots for implementing the prevention and control strategies. The automated execution module is used to implement the prevention and control strategies generated by the intelligent decision support module. Drones are used for large-scale rapid spraying, while automatic spraying robots are used for more precise local treatment.

[0069] Please see attached Figure 1In a preferred embodiment of the present invention, the environmental sensor includes a temperature and humidity sensor, a soil moisture sensor, a light sensor, and a wind speed and direction sensor. The temperature and humidity sensor provides real-time data by measuring the temperature and humidity in the air to help agricultural managers monitor environmental conditions. The soil moisture sensor is used to determine the moisture content in the soil, which is a key factor in evaluating irrigation needs and preventing excessive and insufficient irrigation. The soil moisture state directly affects the growth of citrus trees and the potential occurrence of diseases. The light sensor measures the light intensity reaching the surface of citrus plants. Light is an important environmental factor affecting citrus growth, fruit quality and disease resistance. The wind speed and direction sensor provides data on wind speed and direction, which is used to predict and manage the spread of pests and diseases in citrus orchards, especially airborne diseases. By accurately monitoring environmental conditions, the use of water resources and pesticides can be optimized, waste and environmental burden can be reduced, and yield and fruit quality can be improved.

[0070] Please see attached Figure 1 In a preferred embodiment of the present invention, the plant physiological sensor includes a chlorophyll fluorescence sensor and a spectral sensor. By installing the chlorophyll fluorescence sensor near the plant leaves, when the chloroplast is irradiated with light, it absorbs light energy for photosynthesis, and the energy not used for photosynthesis is released in the form of fluorescence. By measuring the intensity and dynamic changes of this fluorescence, the photosynthetic efficiency and health of the plant can be evaluated. The spectral sensor device is installed in a fixed position, and the physiological state and nutritional level of the plant are monitored by analyzing the specific wavelengths of light reflected and absorbed by the plant leaves. By monitoring chlorophyll fluorescence and the spectral characteristics of the leaves, these sensors can provide early warnings before pests and diseases cause significant damage, allowing farm managers to take preventive measures in time to reduce losses. Chlorophyll fluorescence and spectral data can be used to evaluate the photosynthetic activity and nutritional status of plants, helping farm managers optimize irrigation and fertilization plans and improve resource utilization efficiency. Continuous health monitoring and timely pest and disease management help maintain the health of citrus trees, thereby increasing fruit yield and quality.

[0071] Please see attached Figure 1 In a preferred embodiment of the present invention, the image collection device includes a high-resolution camera and a camera mounted on a drone. By mounting the high-resolution camera on a bracket, the growth status of the plants and possible signs of pests and diseases can be monitored in real time by capturing detailed images of the citrus trees and their surroundings. These cameras are usually equipped with high dynamic range sensors that can work under various lighting conditions, ensuring that clear images can be obtained under different weather and light conditions. The drone is equipped with a high-definition camera and flies regularly to cover the entire citrus orchard, providing the ability to monitor large tracts of farmland from the air, making the early detection of pests and diseases more comprehensive and rapid.

[0072] Please see attached Figure 1 In a preferred embodiment of the present invention, the data processing and analysis module includes an autoregressive difference moving average model applied to environmental data, and its formula is:

[0073]

[0074] Where p, d, q are the number of autoregressive terms, the number of differences, and the number of moving average terms, respectively. i is the autoregressive coefficient, θ i is the moving average coefficient, L is the lag operator, ∈ t The error term is used to clean and preprocess the collected data, including filling missing values ​​and removing outliers to ensure data quality. The non-stationary time series data is differentially processed to make it a stationary series to meet the application premise of the autoregressive difference moving average model. By analyzing the autocorrelation function and partial autocorrelation function graph of the data, the model parameters p (number of autoregressive terms), d (number of differences), and q (number of moving average terms) are determined, and the maximum likelihood is used to estimate the autoregressive coefficient φ i and the moving average coefficient θ i , the autoregressive difference moving average model is fitted to the historical environmental data, and the cross-validation test method is used to verify the prediction accuracy and robustness of the model. The fitted autoregressive difference moving average model is used to predict future environmental conditions, providing a scientific basis for the pest and disease prevention and control decision-making of citrus orchards. The prediction results are integrated into the intelligent decision support module to optimize the pest and disease management strategy. The autoregressive difference moving average model is used to accurately predict environmental data, which helps to timely predict environmental conditions that may lead to pests and diseases, so as to take prevention and control measures in advance. Accurate environmental predictions can guide resource allocation in agricultural production, such as reasonable arrangement of irrigation, fertilization and spraying, thereby reducing resource waste and increasing economic benefits. Through efficient data analysis, it supports the comprehensive pest and disease management of citrus orchards and improves the intelligence and precision of agricultural production.

[0075] Please see attached Figure 1 In a preferred embodiment of the present invention, the data processing and analysis module includes a convolutional neural network model for pest and disease image recognition, and its generation formula is:

[0076] Convolutional layer output:

[0077]

[0078] in is the output of the lth convolutional layer, f is the activation function, W (l) is the convolution kernel of the lth layer, * indicates the convolution operation, I (l-1) is the input of the l-1th layer, b (l)It is the bias of the lth layer. It collects image data from citrus orchards, including images of healthy plants and plants affected by pests and diseases. It also preprocesses the images, such as scaling and normalization, to adapt to the network input requirements. It sets multiple convolutional layers, each layer uses a different number and size of convolution kernels to capture image features from simple to complex. It extracts image features by sliding multiple convolution kernels on the input image.

[0079] Pooling layer output:

[0080]

[0081] in It is the output of the lth pooling layer. Pool is a pooling operation. The pooling layer is introduced after the convolution layer to reduce the spatial size of the convolution layer output, enhance the feature abstraction ability of the network, and reduce the number of parameters and computational complexity.

[0082] Fully connected layer output:

[0083]

[0084] Among them, fc is the output of the fully connected layer, g is the activation function, and W fc is the weight matrix of the fully connected layer, flatten means flattening the output of the pooling layer into a vector, b fc It is the bias of the fully connected layer. After flattening the output of the previous layer, multiple fully connected layers are placed at the end of the network to integrate and learn high-level features.

[0085] Output layer:

[0086] O softmax =softmax(W out ·O fc +b out )

[0087] Among them, softmax is the output of the Softmax layer, W out is the weight matrix of the output layer, b out It is the bias of the output layer. The output layer is responsible for converting the output of the fully connected layer into a probability distribution, normalizing the probability of each category using the Softmax function, and outputting the final classification result.

[0088] Please see attached Figure 1 In a preferred embodiment of the present invention, the intelligent decision support module includes:

[0089] The data integration module is used to integrate the environmental data analysis results and the pest and disease image recognition results, including a method based on multi-source data fusion. The formula is:

[0090]

[0091] Where F(X t ) is comprehensive data, w i is the weight, X i,t is the value of the i-th data at time t, n is the number of data sources, and the data integration module integrates information from multiple data sources and uses a weighted method to improve the comprehensiveness and accuracy of data analysis. A software platform is designed to automatically collect and process information from various data sources. The platform standardizes the data sources, fuses the data according to the preset weights, and generates a comprehensive data output for subsequent risk assessment and strategy formulation;

[0092] The risk assessment module is used to assess the risk level of pests and diseases based on comprehensive data, using a weighted risk assessment formula, which is:

[0093]

[0094] Where r is the risk level, w i is the weight of each data source, f i (X t ) is the function output of the i-th data source at time t, n is the number of data sources, and an algorithm is developed to calculate the output function of the rate of change of plant health indicators and the frequency of occurrence of abnormal meteorological data. Combining these function outputs and corresponding weights, the system can evaluate and update the risk level of pests and diseases in real time;

[0095] The strategy generation module is used to generate specific prevention and control strategies based on risk levels, including prevention and control methods, optimal application time, pesticide dosage and spraying frequency. The formula is:

[0096] D=α·R·A

[0097] Where D is the dosage of the agent, α is the dosage adjustment coefficient, R is the risk level, and A is the area of ​​the affected area. According to different risk levels, the system will recommend different optimal application times and dosages of the agent. These recommendations are sent directly to ground operators, drones, and spraying robots through the user interface;

[0098] The implementation and feedback module is used to implement prevention and control strategies and collect post-implementation effect data, adjust future decisions, monitor the implementation of prevention and control measures through sensors and image data, and collect data on pest and disease response and plant recovery. This information is fed back to the decision support system for optimizing future prevention and control strategies. The implementation and feedback module is responsible for implementing strategies and collecting post-implementation effect data.

[0099] Please see attached Figure 1In a preferred embodiment of the present invention, the strategy generation module includes:

[0100] Selection of control methods: Select appropriate chemical, biological or physical control methods based on risk level and environmental conditions. The strategy generation module selects the most appropriate control method based on the risk level of pests and diseases and the current environmental conditions obtained by the risk assessment module.

[0101] Optimal application time: Based on environmental data prediction, determine the best time period for spraying pesticides. Using information collected from environmental sensors and meteorological data, the model predicts weather conditions in the next few days. Combined with the life cycle and activity pattern of pests and diseases, the best time period for spraying pesticides is selected. Spraying in the morning when the temperature is lower or the wind speed is lower can improve the pesticide's deposition efficiency and control effect.

[0102] Calculation of pesticide dosage and spraying frequency: Calculate the appropriate pesticide dosage and spraying frequency based on the crop growth stage, pest and disease type and risk assessment results. The data provided by the risk assessment module is analyzed by the severity of pests and diseases, plant growth stage and risk level, as well as the knowledge base information about the response of specific pests and diseases and crops to pesticides, and the recommended pesticide dosage and spraying frequency are automatically calculated.

[0103] Please see attached Figure 1 In a preferred embodiment of the present invention, the implementation and feedback module includes:

[0104] Execute control strategies: UAVs and automatic spraying robots receive instructions generated by the decision support module and spray at designated times and areas. The system sends spraying instructions to UAVs and spraying robots through wireless communication, including the type of pesticide, dosage, spraying area and time. These devices can accurately control the location and amount of pesticide application to ensure maximum control effect.

[0105] Real-time monitoring and adjustment: Through the real-time monitoring system, the environmental changes and plant responses during the spraying process are monitored to ensure the spraying effect. If uneven spraying or poor effect is detected, the system can automatically adjust the spraying parameters and instruct repeated spraying. Through the sensors and other monitoring equipment equipped on the drones and spraying robots, the environmental changes and plant responses during the spraying process are monitored in real time. The drones and spraying robots are equipped with GPS and sensors to transmit the spraying location and environmental data in real time. The system analyzes this data. If the spraying effect is found to be unsatisfactory, the spraying parameters (such as nozzle angle, flow rate, etc.) are automatically adjusted or the equipment is instructed to repeat spraying a certain area.

[0106] Effect data collection: Collect effect data on pest and disease reduction rate and plant health recovery after spraying. According to the collected data, adjust the model and decision logic to optimize future prevention and control strategies. After spraying, use plant physiological sensors and image collection equipment to regularly monitor plant status and pest and disease activities. The data is transmitted back to the central database and compared with the pre-spraying data to evaluate the treatment effect. Collect and analyze the pest and disease control effect and plant health recovery after spraying, provide feedback to the system for optimizing prevention and control models and strategies.

[0107] The method for controlling citrus pests and diseases described below can be used in conjunction with the system for controlling citrus pests and diseases described above. The principles and technical effects are similar and will not be described in detail here.

[0108] Please see attached Figure 2 , a method for preventing and controlling citrus pests and diseases, comprising the following steps:

[0109] S1. Collect environmental data, plant health data and image data through a data collection module;

[0110] S2, transmitting the data to the server through the data transmission module;

[0111] S3. Apply time series analysis models and convolutional neural network models to process and analyze data in the data processing and analysis module;

[0112] S4. Based on the analysis results, the intelligent decision support module generates a control strategy, including selecting control methods, determining the best application time, calculating the dosage of pesticides and spraying frequency, and implementing the control strategy through the automated execution module;

[0113] S5. Collect post-implementation effect data and adjust future decisions based on this data.

[0114] Although 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 the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A system for preventing and controlling citrus pests and diseases, characterized in that: include: Data collection module, which includes environmental sensors, plant physiological sensors and image collection equipment deployed in the citrus orchard to collect environmental data, plant health data and image data; A data transmission module is used to transmit the data collected by the data collection module to a server through an Internet of Things gateway; Data processing and analysis module, which is used to process time series analysis models of environmental data and convolutional neural network models for identifying pest and disease images; An intelligent decision support module, used to generate a prevention and control strategy based on the analysis results of the data processing and analysis module; Automated execution modules, including drones and autonomous spraying robots for implementing control strategies.

2. A citrus pest control system according to claim 1, characterized in that: The environmental sensors include a temperature and humidity sensor, a soil moisture sensor, a light sensor, and a wind speed and direction sensor.

3. A citrus pest control system according to claim 1, characterized in that: The plant physiological sensor comprises a chlorophyll fluorescence sensor and a spectrum sensor.

4. A citrus pest control system according to claim 1, characterized in that: The image collection equipment includes a high-resolution camera and a camera mounted on a drone.

5. The citrus pest control system according to claim 1, characterized in that: The data processing and analysis module includes an autoregressive difference moving average model applied to environmental data, and its formula is: Where p, d, q are the number of autoregressive terms, the number of differences, and the number of moving average terms, respectively. i is the autoregressive coefficient, θ i is the moving average coefficient, L is the lag operator, ∈ t is the error term.

6. A citrus pest control system according to claim 5, characterized in that: The data processing and analysis module includes a convolutional neural network model for pest and disease image recognition, and its generation formula is: Convolutional layer output: in is the output of the lth convolutional layer, f is the activation function, W (l) is the convolution kernel of the lth layer, * indicates the convolution operation, I (l-1) is the input of the l-1th layer, b (l) is the bias of the lth layer; Pooling layer output: in is the output of the lth pooling layer, where pool is the pooling operation; Fully connected layer output: Among them, fc is the output of the fully connected layer, g is the activation function, and W fc is the weight matrix of the fully connected layer, flatten means flattening the output of the pooling layer into a vector, b fc is the bias of the fully connected layer; Output layer: O softmax =softmax(W out ·O fc +b out ) Among them, softmax is the output of the Softmax layer, W out is the weight matrix of the output layer, b out is the bias of the output layer.

7. The citrus pest control system according to claim 1, characterized in that: The intelligent decision support module includes: The data integration module is used to integrate the environmental data analysis results and the pest and disease image recognition results, including a method based on multi-source data fusion. The formula is: Where F(X t ) is comprehensive data, w i is the weight, X i,t is the value of the i-th data at time t, and n is the number of data sources; The risk assessment module is used to assess the risk level of pests and diseases based on comprehensive data, using a weighted risk assessment formula, which is: Where R is the risk level, w i is the weight of each data source, f i (X t ) is the function output of the ith data source at time t, and n is the number of data sources; The strategy generation module is used to generate specific prevention and control strategies based on risk levels, including prevention and control methods, optimal application time, pesticide dosage and spraying frequency. The formula is: D=α·R·A Where D is the drug dose, α is the dose adjustment coefficient, R is the risk level, and A is the area of ​​the affected region; The implementation and feedback module is used to implement prevention and control strategies and collect effect data after implementation to adjust future decisions.

8. A system for controlling citrus pests and diseases according to claim 7, characterized in that: The strategy generation module includes: Selection of control methods: Select appropriate chemical, biological or physical control methods based on risk level and environmental conditions; Optimal application time: Determine the best time period for spraying the pesticide based on environmental data prediction; Calculation of pesticide dosage and spraying frequency: Calculate the appropriate pesticide dosage and spraying frequency based on the crop growth stage, pest and disease type, and risk assessment results.

9. The citrus pest control system according to claim 1, characterized in that: The implementation and feedback module includes: Execute prevention and control strategies: drones and automatic spraying robots receive instructions generated by the decision support module and spray at designated times and areas; Real-time monitoring and adjustment: Through the real-time monitoring system, the environmental changes and plant responses during the spraying process are monitored to ensure the spraying effect. If uneven spraying or poor effect is detected, the system can automatically adjust the spraying parameters and instruct repeated spraying; Effect data collection: Collect effect data on pest and disease reduction rate and plant health recovery after spraying. Based on the collected data, adjust the model and decision logic to optimize future prevention and control strategies.

10. A method for preventing and controlling citrus pests and diseases, characterized in that: Using a citrus pest and disease prevention system according to any one of claims 1 to 9 comprises the following steps: Collect environmental data, plant health data and image data through the data collection module; Transmitting data to the server through the data transmission module; Apply time series analysis models and convolutional neural network models to process and analyze data in the data processing and analysis module; Based on the analysis results, the intelligent decision support module generates a control strategy, including selecting control methods, determining the best application time, calculating the dosage of pesticides and spraying frequency, and implements the control strategy through the automated execution module; Collect post-implementation performance data and use that data to adjust future decisions.

Citation Information

Cited By

  • Multi-sensor fusion unmanned equipment agriculture monitoring method and system

    CN119515029A

  • Traditional Chinese medicine traceability management system

    CN119539836A

  • System for controlling diseases and insect pests through precise spraying of unmanned aerial vehicle

    CN119872889A

  • Agricultural pest risk assessment method and system

    CN120031387A

  • Pest and disease identification method and system based on unmanned aerial vehicle remote sensing

    CN120431496A