Self-powered crop growth condition monitoring system based on machine learning

Abstract

The invention provides a self-powered crop growth condition monitoring system based on machine learning. The self-powered crop growth condition monitoring system comprises a self-powered module, a weight sensing module, an intelligent analysis module, a user monitoring terminal, an intelligent irrigation module and an intelligent fertilization module. And the intelligent analysis module comprises processes of data acquisition, feature labeling, feature extraction, model training and state judgment. The weight sensing module converts LED lamp position changes caused by crop growth into weight data through the elastic deformation mechanism and the optical marking assembly, a traditional sensor is not needed, cost is low, and potential safety hazards of a battery are eliminated. According to the system, drip irrigation water drop impact energy is collected through a self-powered module (TENG technology) to achieve autonomous power supply, weight data and a training set are compared and analyzed by combining a machine learning model, and high-precision monitoring of the health state of a single plant is achieved. The user monitoring terminal integrates data through the Internet of Things and visually displays the data, when diseases are detected, the intelligent irrigation and fertilization module is automatically controlled to stop water and fertilizer supply, the agricultural production accuracy and efficiency are improved, and the problems that in the prior art, real-time performance is poor, precision is insufficient, and power supply depends are solved.

Description

Technical Field

[0001] The present invention relates to the field of smart agricultural technology, and specifically to a real-time crop growth monitoring system that integrates machine learning and self-powered technology, aiming to improve the accuracy and efficiency of agricultural production through intelligent means. Background Art

[0002] According to statistics, in 2021, the global agricultural workforce was approximately 873 million, accounting for 27% of the total labor force, yet its contribution to global GDP was only 4%. Crop diseases not only lead to a surge in pesticide use but also significantly reduce agricultural economic returns. While machine learning and IoT technologies are mature in industrial and medical fields, their refined application in agricultural monitoring remains limited.

[0003] Existing crop monitoring technologies primarily rely on remote sensing images or environmental sensors, making it difficult to achieve real-time tracking of individual crops. Data also exhibit significant lag and are susceptible to interference from environmental factors such as climate and soil, resulting in insufficient monitoring accuracy. Notably, triboelectric nanogenerator (TENG) technology has demonstrated unique advantages in environmental energy harvesting in recent years. Through the principles of triboelectric charging and electrostatic induction, it can efficiently harvest ground mechanical energy, such as raindrop impact and leaf vibration, providing a new path for self-powering agricultural IoT devices. However, the current application of TENG technology in agricultural scenarios is still limited to environmental energy harvesting and has yet to be effectively linked to the precise monitoring of crop growth status. Summary of the Invention

[0004] The present invention aims to provide a self-driven crop growth status monitoring system based on machine learning. By introducing TENG technology to construct a self-powered module, the device can autonomously collect ground energy in the environment (such as the impact energy of drip irrigation droplets). At the same time, it uses a weight sensing module to obtain crop growth data, combines machine learning models to complete health status analysis, and realizes data integration and intelligent control through the Internet of Things, solving the problems of poor real-time performance, insufficient accuracy, and dependence on external energy for power supply in existing technologies.

[0005] The system includes a self-powered module (based on TENG technology), a weight sensing module, an intelligent analysis module, a user monitoring terminal, an intelligent irrigation module, and an intelligent fertilization module. The intelligent analysis module includes five core components: data collection, feature labeling, feature extraction, model training, and state determination.

[0006] Furthermore, the self-powered module utilizes a triboelectric nanogenerator (TENG). This captures the mechanical energy generated by the impact of drip irrigation water droplets, converts it into electrical energy through the triboelectric effect, and stores it in an energy storage unit, providing continuous power for the light source component of the weight sensing module. This design eliminates the need for traditional batteries, eliminating the risk of battery leakage, and achieving efficient utilization of environmental energy.

[0007] Furthermore, the weight sensing module integrates an elastic deformation mechanism and an optical marker assembly. When the mechanism deforms due to crop growth, the three micro-LED lights mounted on it shift relative to each other. A camera captures this change in light spacing and converts crop weight information into image data. After algorithmic analysis, a real-time weight curve is generated and transmitted to the intelligent analysis module.

[0008] Furthermore, the intelligent analysis module first performs data collection, obtains weight sequence data of target crops in healthy and diseased states through a sensor network deployed in the field, and constructs a multi-dimensional data set.

[0009] Furthermore, feature annotation is performed, the image data is normalized, and key features such as crop growth stage, disease type, weight fluctuation cycle, etc. are marked to form structured training samples.

[0010] Furthermore, in the feature extraction stage, deep features such as the slope of the weight curve, the mutation point of the growth rate, and the amplitude of periodic fluctuations are extracted as input parameters of the machine learning model.

[0011] Furthermore, a deep learning architecture is adopted in the model training stage to construct a crop health status classification model through a combination of multi-layer convolutional layers, pooling layers and fully connected layers, and the backpropagation algorithm is used to optimize the weight parameters of each layer.

[0012] Furthermore, the health level of individual crops is finally output through the status determination module, and the identification result is wirelessly transmitted to the user monitoring terminal.

[0013] Furthermore, the user monitoring terminal supports access from smart devices such as mobile phones and tablets, providing a real-time visualization of crop status across the entire area. It integrates data query, model parameter adjustment, and remote control capabilities. When diseased crops are detected, the terminal automatically sends an early warning signal to the intelligent irrigation and fertilization modules.

[0014] Furthermore, upon receiving the early warning signal, the intelligent irrigation and fertilization module immediately suspends the supply of water and fertilizer to the diseased plants, reducing resource waste while suppressing the spread of the disease through a physical isolation mechanism.

[0015] The beneficial effects of the present invention are: 1. The present invention proposes a self-powered crop growth monitoring system based on machine learning. It uses a self-driven power supply method to convert the mechanical energy generated by dripping water droplets from the drip irrigation device into electrical energy and store it to power the equipment without relying on batteries. While saving energy and costs, it eliminates the potential risks of battery leakage and spontaneous combustion, thereby improving the safety of the system in the agricultural field. 2. The present invention adopts a self-powered crop growth status monitoring system based on machine learning, which uses machine learning as an aid. By collecting data on local crops in different growth states, a data set is established in a big data manner, and the growth data of the same crops in the same growth environment are used for comparative analysis, which greatly improves the accuracy of the entire system. 3. The present invention provides a self-powered crop growth monitoring system based on machine learning, which uses Internet of Things technology to integrate and manage data. The user monitoring terminal summarizes the crop health status judgment results obtained by the intelligent judgment module and displays them to the user in the form of the Internet of Things for viewing at any time, making it convenient for the user to monitor all crops. At the same time, if a large error is found in the judgment result, the data can be fed back to the terminal to modify the data model to further improve accuracy. The system is easy to maintain and can adapt to different environmental requirements. 4. The present invention employs a self-powered crop growth monitoring system based on machine learning, which uses an intelligent irrigation module and an intelligent fertilization module. Upon receiving warning signals from a user monitoring terminal, both modules stop supplying water and fertilizer to diseased crops, thereby saving water and fertilizer and, to a certain extent, curbing the spread of diseases. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Architecture diagram of the self-powered crop growth status monitoring system based on machine learning.

[0017] Figure 2 Data processing flow chart of the intelligent analysis module of the self-powered crop growth status monitoring system based on machine learning.

[0018] Figure 3 Structural block diagram of the energy supply and terminal modules of the self-powered crop growth monitoring system based on machine learning. Specific implementation cases

[0019] To help those skilled in the art better understand the technical solutions of the present invention, the actual application scenarios of the system are now explained in detail with reference to the accompanying drawings. Against the backdrop of a high proportion of the global agricultural population (reaching 27% in 2021) and rising costs for crop disease control, the present invention has constructed an intelligent crop growth monitoring system that integrates triboelectric nano-power generation technology with machine learning algorithms. The system is dedicated to breaking through the bottleneck of traditional monitoring methods in real-time tracking and precise diagnosis of individual crops, and promoting the upgrading of agricultural production towards low manpower dependence and high response efficiency.

[0020] like Figure 1 As shown in the figure, the whole system is divided into self-power module, weight weighing module, intelligent judgment module, user monitoring terminal, intelligent irrigation module and intelligent fertilization module; Figure 2As shown, the intelligent judgment module includes several processes: data collection, data labeling, feature extraction, model construction and result judgment.

[0021] like Figure 1 and Figure 3 As shown in the figure, the self-powered module includes an energy capture unit (TENG), an energy collection unit, and an energy storage unit. When water droplets from the smart drip irrigation device impact the TENG surface, they generate electric charge through triboelectric charging. After rectification and filtering by the energy collection unit, the charge is stored in the energy storage unit, which then powers the LED lights and camera of the weight-weighing module. This completes a self-closed-loop energy supply from "water droplet mechanical energy to electrical energy" without the need for traditional batteries.

[0022] like Figure 1 and Figure 2 As shown, the weight-weighing module consists of a mechanical device with an internal spring. Three LED bulbs are mounted on top and bottom of the device, with the bottom fixed to the crop stem. As the crop grows and gains weight, the spring stretches, increasing the vertical spacing of the LED bulbs. Cameras deployed between the rows capture the changes in the position of the LED bulbs. Using an image algorithm, they calculate the geometric center spacing of the bulb groups and infer the crop weight, enabling dynamic monitoring of the weight of individual crops.

[0023] like Figure 2 As shown in the figure, the intelligent judgment module first collects data and obtains the weight sequence data of crops in health, different diseases (such as root rot and leaf spot) and growth stages (germination and heading) through the field sensor network, and constructs a multi-dimensional data set containing environmental parameters.

[0024] like Figure 2 As shown, the intelligent judgment module continues to label the data, normalizes the image data, and marks the crop health level, onset time, pathology type (such as bacterial disease) and weight abnormal fluctuation nodes to form structured training samples.

[0025] like Figure 2 As shown in the figure, the intelligent judgment module continues to extract features, selecting artificially designed features such as the weight value at a specific growth moment, the weight change rate per unit time (curve slope), and the number of days from germination to seedling stage, and combining them with the spatial features of the LED lamp spacing (such as the geometric center offset) automatically extracted by machine learning as model input parameters.

[0026] like Figure 2 As shown in the figure, the intelligent judgment module continues to build the model, designs a deep learning model consisting of convolutional layers, pooling layers, and fully connected layers, initializes the weight parameters of each layer, and optimizes the validation set classification accuracy through the back propagation algorithm to ≥95%.

[0027] like Figure 2As shown in the figure, the intelligent judgment module finally makes a judgment on the result, calculates the health probability value of a single crop plant through the trained model, and sends it to the user monitoring terminal in the form of a digital signal.

[0028] like Figure 1 As shown, the user monitoring terminal receives data via a wireless transmission module and displays global crop status using visual charts (health heat maps and weight trend curves). It supports mobile phone and host access. Users can dynamically adjust model parameters, query historical data, and send zone warning signals to the intelligent irrigation and fertilization module.

[0029] like Figure 1 As shown, after receiving the warning signal, the intelligent irrigation module and the intelligent fertilization module suspend the water and fertilizer supply to the partition where the diseased plants are located through the solenoid valve and the fertilizer pump, reducing resource waste. At the same time, the spread of the disease is suppressed by physically cutting off the supply, achieving precise prevention and control.

[0030] All features disclosed in this specification, or steps in all methods or processes disclosed, except for mutually exclusive features and / or steps, may be combined in any manner. Any feature disclosed in this specification (including any appended claims and abstract), unless otherwise stated, may be replaced by an equivalent or similar alternative feature. That is, unless otherwise stated, each feature is merely an example of a set of equivalent or similar features.

[0031] The above technical solution only reflects one technical solution of this technical solution. Some changes that may be made to certain parts thereof by technical personnel in this technical field all reflect the principles of this design system and should also fall within the scope of protection of this patent.

Claims

1. A self-powered crop growth monitoring system based on machine learning, characterized in that: The system includes a self-power supply module, a weight measurement module, and a terminal processing module. The system collects environmental energy for power supply through the self-power supply module, and combines weight measurement data with the terminal processing module to achieve real-time monitoring and intelligent regulation of crop health status. The self-powered module includes an energy capture unit and an energy storage unit. The energy capture unit uses a triboelectric nanogenerator (TENG) to capture the mechanical energy generated by the impact of drip irrigation water droplets and convert it into electric charge. The energy storage unit is used to store electrical energy and power other modules. The weight measurement module includes a weighing module, a light source module, a visual acquisition module, and a data preprocessing unit. The weighing module senses the weight changes caused by crop growth through an elastic deformation mechanism, the light source module reflects structural deformation through changes in the position of micro LED lights, the visual acquisition module captures images through a camera and calculates weight data, and the data preprocessing unit performs noise reduction and normalization on the image data.

2. The monitoring system according to claim 1, characterized in that The terminal processing module includes an intelligent analysis unit, a wireless transmission module, a user interaction terminal, an abnormality warning unit, and an execution control unit; The intelligent analysis unit is used for data collection, feature extraction, model training and status determination, constructing a multi-dimensional data set and outputting the crop health grade based on the deep learning model; The wireless transmission module is used to realize wireless transmission of data signals; The user interaction terminal is used to visually display the crop status and support user operations; The abnormal warning unit is used to send an early warning signal when a disease is detected; The execution control unit is used to control the intelligent irrigation module and the intelligent fertilization module to suspend the supply of water and fertilizer to the diseased plants.

3. The monitoring system according to claim 1, wherein: The feature extraction of the intelligent analysis unit includes the slope of the weight curve, the growth rate mutation point, and the spatial characteristics of the LED lamp spacing; the deep learning model includes a convolution layer, a pooling layer and a fully connected layer, and the parameters are optimized through the back propagation algorithm.

Images