An ai-driven agricultural monitoring and prediction system

Abstract

The application relates to the technical field of agricultural monitoring and prediction, in particular to an AI-driven agricultural monitoring and prediction system, which comprises the following: an environment sensor for measuring soil humidity, air temperature and wind speed parameters of a farm; an image acquisition device for acquiring image data of plants; a data processing unit for performing data preprocessing and data filtering functions; an environment-plant interaction model which receives the preprocessed data transmitted by the data processing unit and generates a comprehensive environment-plant interaction index; a central processing unit for generating a prediction model according to the environment-plant interaction index (EPI); and an automatic control unit for receiving the result of the prediction model transmitted by the central processing unit and automatically adjusting irrigation and fertilization operation parameters of the farm according to the result. The application realizes all-round monitoring of crops and farmland environment and ensures rapid data flow and efficient data processing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of agricultural monitoring and prediction, and particularly relates to an AI-driven agricultural monitoring and prediction system. BACKGROUND

[0002] With the rapid rise of global population and food demand, modern agriculture faces great challenges in improving yield and sustainability. Traditional agricultural management methods often rely on farmers' experience and intuitive judgment, which to some extent limits the efficiency and accuracy of agricultural production. In recent years, information technology and artificial intelligence (AI) have been increasingly applied in the field of agriculture to achieve fine management of crop growth environment. However, most existing agricultural monitoring systems still have some limitations.

[0003] Firstly, many existing systems are relatively independent in data collection and processing, lacking a highly integrated and real-time data processing platform. This not only leads to low data processing efficiency, but also may affect the timeliness of decision-making.

[0004] Secondly, although some advanced systems may contain prediction or recommendation functions, these functions are usually based on relatively simple data models, which are difficult to accurately reflect the complex environment-plant interactions, thereby affecting the accuracy of decision-making.

[0005] Finally, existing systems often lack effective emergency response mechanisms when facing extreme environmental conditions or unexpected events. This makes crops more vulnerable to unpredictable factors such as sudden extreme weather.

[0006] Therefore, there is an urgent need for a highly integrated, real-time, and accurate agricultural monitoring and prediction system that can accurately predict crop needs and has a reliable emergency response mechanism to solve the above problems. SUMMARY

[0007] Based on the above purpose, the present application provides an AI-driven agricultural monitoring and prediction system.

[0008] An AI-driven agricultural monitoring and prediction system, the system comprising:

[0009] a set of environmental sensors for measuring soil moisture, air temperature and wind speed parameters of the farm and sending the parameters to the data processing unit;

[0010] one or more image acquisition devices for acquiring image data of plants and sending the image data to the data processing unit;

[0011] a data processing unit for receiving data from the environmental sensors and the image acquisition device, and performing data pre-processing and filtering functions, and then transmitting the pre-processed data to the central processing unit;

[0012] an environment-plant interaction model embedded in the central processing unit for receiving the pre-processed data from the data processing unit, and generating an integrated environment-plant interaction index reflecting the interaction between environmental factors and plant growth conditions through calculation;

[0013] a central processing unit for generating a prediction model based on the "environment-plant interaction index (EPI)", and transmitting the results of the prediction model to the automated control unit;

[0014] an automated control unit for receiving the results of the prediction model from the central processing unit, and automatically adjusting the irrigation and fertilization operation parameters of the farm based on the results, and then sending the adjusted operation parameters back to the central processing unit for optimization.

[0015] Further, the environmental sensor module includes a humidity sensor, a temperature sensor, and a wind speed sensor, which use a 16-bit analog-to-digital converter to convert analog signals to digital signals, and the environmental sensor module also includes a low-power microcontroller unit for timestamping and packaging the digital signals of multiple sensors, and sending these timestamped and packaged digital signals to the data processing unit through Bluetooth 5.0.

[0016] Further, the image acquisition device module includes a high-definition visible light camera and an infrared camera;

[0017] The visible light camera uses a CMOS image sensor with a resolution of 3840x2160 pixels;

[0018] The infrared camera uses a micro-thermoelectric pile as an image sensor with a resolution of 640x480 pixels;

[0019] The image acquisition device module also includes an image processing unit for real-time compression and encryption of image data, and sending these compressed and encrypted image data to the data processing unit through Ethernet.

[0020] Further, the data processing unit includes an FPGA hardware platform that performs Fourier transform to extract frequency domain features from environmental data and image data, and uses principal component analysis running on a multi-core CPU for dimensionality reduction and noise filtering, and uses Apache Kafka message queue for efficient data exchange with other modules.

[0021] Further, the environment-plant interaction model calculates an environment-plant interaction index (EPI) formula as follows:

[0022] EPI = a S env + b S plant

[0023] wherein S env is a score from the environment sensor, calculated as S env = w1 H + w2 T + w3 W, wherein H, T, W are the measured values of humidity, temperature, and wind speed, respectively, and w1, w2, w3 are the respective weights; S plant is a plant score from the image acquisition device, calculated as S plant = k1 I visible + k2 I infrared , wherein I visible , I infrared are the feature values of the visible light and infrared images, respectively, and k1, k2 are the respective weights, and the weight parameters a and b are optimized through historical data and gradient descent algorithm.

[0024] Further, the central processing unit includes a random forest prediction model, which includes 500 decision trees, each with a depth of 20, and the random forest model performs prediction according to the environment-plant interaction index (EPI) and sends the prediction results to the automated control unit through a RESTful API.

[0025] Further, the automated control unit includes a set of electromagnetic valves and nozzles, which are controlled by a PLC that receives the prediction results from the central processing unit and adjusts the on-off state of the electromagnetic valves and the spray angle of the nozzles according to the prediction results to automatically perform irrigation and fertilization operations.

[0026] Further, the data processing unit further includes a data caching submodule that temporarily stores the raw data from the environment sensor module and the image acquisition device module using a distributed key-value storage system.

[0027] Further, the environment-plant interaction model further includes an adaptive learning mechanism that dynamically adjusts the weight parameters a and b, as well as the sub-weights w1, w2, w3, k1, k2, according to historical data and the current environment-plant interaction index.

[0028] Advantages of the present application:

[0029] This invention achieves comprehensive monitoring of crops and farmland environment through multiple modules such as environmental sensor module, image acquisition device module, and data processing unit. The data processing unit has high real-time performance and can receive and quickly process large amounts of data from various sensors and image acquisition devices. By using an FPGA hardware platform and principal component analysis (PCA) module, as well as a distributed key-value storage system (such as Redis), the system ensures rapid data flow and efficient data processing.

[0030] The design of the Environment-Plant Interaction Model (EPI-Model) and the central processing unit in this invention further enhances the prediction accuracy of the system. In particular, the adaptive learning mechanism uses reinforcement learning algorithms to dynamically adjust the model parameters, improving the accuracy and sensitivity of the prediction. As the system continuously acquires data from practical applications, the prediction model will be continuously optimized. The automated control unit receives the prediction results and intelligently adjusts the irrigation and fertilization parameters, thereby achieving the goal of precision agriculture. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a schematic diagram of the system modules in an embodiment of the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0034] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0035] like Figure 1 As shown, an AI-driven agricultural monitoring and forecasting system includes:

[0036] A set of environmental sensors is used to measure soil moisture, air temperature and wind speed parameters on the farm and send these parameters to the data processing unit;

[0037] One or more image acquisition devices are used to acquire image data of plants and send the image data to a data processing unit;

[0038] The data processing unit is used to receive data from environmental sensors and image acquisition devices, perform data preprocessing and data filtering functions, and then transmit the preprocessed data to the central processing unit.

[0039] The environment-plant interaction model is embedded in the central processing unit. It receives preprocessed data from the data processing unit and generates a comprehensive environment-plant interaction index through calculation. This index reflects the interaction between environmental factors and plant growth status.

[0040] The central processing unit generates a prediction model based on the "Environment-Plant Interaction Index (EPI)" and transmits the results of the prediction model to the automated control unit.

[0041] The automation control unit receives the results of the prediction model from the central processing unit, automatically adjusts the farm's irrigation and fertilization operation parameters based on the results, and then sends the adjusted operation parameters back to the central processing unit for optimization.

[0042] The environmental sensor module includes a humidity sensor, a temperature sensor, and a wind speed sensor. The humidity sensor, temperature sensor, and wind speed sensor use a 16-bit analog-to-digital converter to convert analog signals into digital signals. The environmental sensor module also includes a low-power microcontroller unit, which is used to timestamp and package the digital signals from multiple sensors and send these tagged and packaged digital signals to the data processing unit via Bluetooth 5.0.

[0043] The image acquisition module includes a high-definition visible light camera and an infrared camera;

[0044] The visible light camera uses a CMOS image sensor with a resolution of 3840x2160 pixels;

[0045] The infrared camera uses a micro thermopile as an image sensor with a resolution of 640x480 pixels.

[0046] The image acquisition device module also includes an image processing unit, which is used to compress and encrypt image data in real time and send the compressed and encrypted image data to the data processing unit via Ethernet.

[0047] The data processing unit includes an FPGA (Field Programmable Gate Array) hardware platform that performs Fourier transforms to extract frequency domain features from environmental and image data. The data processing unit uses principal component analysis, runs on a multi-core CPU, and is used for data dimensionality reduction and noise filtering. The data processing unit uses the Apache Kafka message queue to achieve efficient data exchange with other modules.

[0048] The formula for calculating the Environment-Plant Interaction Index (EPI) using the Environment-Plant Interaction Model is as follows:

[0049] EPI = α·S env +β·S plant

[0050] Among them, S env It is a score from environmental sensors, calculated as S env = w1·H + w2·T + w3·W, where H, T, and W are the measured values ​​of humidity, temperature, and wind speed, respectively, and w1, w2, and w3 are their respective weights; S plant This is a plant score from the image acquisition device, calculated as S. plant =k1·I visible +k2·I infrared , where I visible ,I infrared These are the feature values ​​of the visible light and infrared images, respectively. k1 and k2 are their respective weights. The weight parameters α and β are obtained by optimizing historical data and gradient descent algorithm.

[0051] The central processing unit includes a random forest prediction model, which consists of 500 decision trees, each with a depth of 20. The random forest model makes predictions based on the environment-plant interaction index and sends the prediction results to the automated control unit via a RESTful API.

[0052] The automated control unit includes a set of solenoid valves and nozzles, which are controlled by a PLC. The PLC receives prediction results from the central processing unit and adjusts the on / off state of the solenoid valves and the spray angle of the nozzles according to the prediction results to automatically perform irrigation and fertilization operations.

[0053] The data processing unit also includes a data caching submodule, which uses a distributed key-value storage system (e.g., Redis) to temporarily store raw data from the environmental sensor module and the image acquisition device module. This data caching submodule works in conjunction with the FPGA hardware platform and the principal component analysis (PCA) module to achieve efficient data reading and processing, further improving the system's response speed.

[0054] The environment-plant interaction model also includes an adaptive learning mechanism. This mechanism dynamically adjusts the weight parameters α and β, as well as the sub-weights w1, w2, w3, k1, k2, based on historical data and the current environment-plant interaction index. This adaptive learning mechanism uses reinforcement learning algorithms to optimize the model parameters according to the actual crop growth status, thereby improving the accuracy of predictions.

[0055] The automated control unit also features an emergency response subsystem. This subsystem consists of a set of independent humidity and temperature sensors, and an independent power source (e.g., a solar panel). In the event of a failure in the central processing unit or other modules, the emergency response subsystem can operate independently, automatically controlling solenoid valves and sprinklers to provide basic irrigation and temperature control, ensuring that crops are not damaged by system failure.

[0056] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention is limited to these examples; within the framework of the invention, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of the different aspects of the invention as described above, which are not provided in detail for the sake of brevity.

[0057] This invention is intended to cover all such substitutions, modifications, and variations falling within the broad scope of the claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. An AI-driven agricultural monitoring and forecasting system, characterized in that, The system includes: A set of environmental sensors is used to measure soil moisture, air temperature and wind speed parameters on the farm and send these parameters to the data processing unit; One or more image acquisition devices are used to acquire image data of plants and send the image data to a data processing unit; The data processing unit is used to receive data from environmental sensors and image acquisition devices, perform data preprocessing and data filtering functions, and then transmit the preprocessed data to the central processing unit. The environment-plant interaction model is embedded in the central processing unit. It receives preprocessed data from the data processing unit and generates a comprehensive environment-plant interaction index through calculation. This index reflects the interaction between environmental factors and plant growth status. The central processing unit generates a prediction model based on the "environment-plant interaction index" and transmits the results of the prediction model to the automated control unit. The automation control unit receives the results of the prediction model from the central processing unit, automatically adjusts the irrigation and fertilization operation parameters of the farm based on the results, and then sends the adjusted operation parameters back to the central processing unit for optimization. The formula for calculating the computational environment-plant interaction index is as follows: EPI=α·S env +β·S plant Among them, S env It is a score from environmental sensors, calculated as S env = w1·H + w2·T + w3·W, where H, T, and W are the measured values ​​of humidity, temperature, and wind speed, respectively, and w1, w2, and w3 are their respective weights; S plant This is a plant score from the image acquisition device, calculated as S. plant =k1·I visible +k2·I infrared , where I visible ,I infrared These are the feature values ​​of the visible light and infrared images, respectively. k1 and k2 are their respective weights. The weight parameters α and β are obtained by optimizing historical data and gradient descent algorithm.

2. The AI-driven agricultural monitoring and forecasting system according to claim 1, characterized in that, The environmental sensor module includes a humidity sensor, a temperature sensor, and a wind speed sensor. The humidity sensor, temperature sensor, and wind speed sensor use a 16-bit analog-to-digital converter to convert analog signals into digital signals. The environmental sensor module also includes a low-power microcontroller unit, which is used to timestamp and package the digital signals from multiple sensors, and send these tagged and packaged digital signals to the data processing unit via Bluetooth 5.

0.

3. The AI-driven agricultural monitoring and forecasting system according to claim 2, characterized in that, The image acquisition device module includes a high-definition visible light camera and an infrared camera; The visible light camera uses a CMOS image sensor with a resolution of 3840x2160 pixels; The infrared camera uses a micro thermopile as an image sensor with a resolution of 640x480 pixels. The image acquisition device module also includes an image processing unit, which is used to compress and encrypt image data in real time and send the compressed and encrypted image data to the data processing unit via Ethernet.

4. The AI-driven agricultural monitoring and forecasting system according to claim 3, characterized in that, The data processing unit includes an FPGA hardware platform that performs Fourier transforms to extract frequency domain features from environmental and image data. The data processing unit uses principal component analysis, runs on a multi-core CPU, and is used for data dimensionality reduction and noise filtering. The data processing unit uses the Apache Kafka message queue to achieve efficient data exchange with other modules.

5. The AI-driven agricultural monitoring and forecasting system according to claim 1, characterized in that, The central processing unit includes a random forest prediction model, which consists of 500 decision trees, each with a depth of 20. The random forest model makes predictions based on the environment-plant interaction index and sends the prediction results to the automated control unit via a RESTful API.

6. The AI-driven agricultural monitoring and forecasting system according to claim 5, characterized in that, The automated control unit includes a set of solenoid valves and nozzles, which are controlled by a PLC. The PLC receives prediction results from a central processing unit and adjusts the on / off state of the solenoid valves and the spray angle of the nozzles according to the prediction results to automatically perform irrigation and fertilization operations.

7. The AI-driven agricultural monitoring and forecasting system according to claim 6, characterized in that, The data processing unit also includes a data caching submodule, which uses a distributed key-value storage system to temporarily store raw data from the environmental sensor module and the image acquisition device module.

8. The AI-driven agricultural monitoring and forecasting system according to claim 7, characterized in that, The environment-plant interaction model also includes an adaptive learning mechanism, which dynamically adjusts the weight parameters α and β, as well as the sub-weights w1, w2, w3, k1, k2, based on historical data and the current environment-plant interaction index.

Images