An Internet of Things (IoT) electronic hardware system for agricultural environmental monitoring

Abstract

This invention discloses an IoT-based electronic hardware system for agricultural environmental monitoring, aiming to solve problems such as low monitoring accuracy, weak anti-interference capability, and poor adaptability in different agricultural scenarios. The system adopts a core architecture of multispectral coupled sensing, layered anti-interference packaging, dual-mode communication, and neural network self-calibration, making it adaptable to various scenarios such as desert oases, tea gardens, and northern arid lands. By optimizing the sensing unit, strengthening the protective structure, and implementing adaptive power supply and real-time calibration algorithms, it achieves accurate monitoring of multiple dimensions, including temperature and humidity, soil / water quality parameters, and crop status. Its simple structure, convenient installation, strong resistance to interference from wind, sand, freeze-thaw cycles, and high humidity, and stable data transmission provide reliable data support for precise agricultural production management, adapting to various agricultural environmental monitoring needs.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) and agricultural environmental monitoring technology, and in particular to an IoT-based electronic hardware system for agricultural environmental monitoring. Background Technology

[0002] Current IoT-based agricultural environmental monitoring systems largely rely on a combination of traditional physical sensors (temperature, humidity, light intensity, soil conductivity, etc.) and mainstream wireless communication modules (LoRa, 4G / 5G, WiFi). This approach suffers from the following technical limitations, and existing improvement solutions struggle to balance accuracy, interference resistance, low power consumption, and scenario adaptability: The monitoring dimensions are limited and the correlation is weak: existing systems mostly focus on collecting environmental physical parameters and do not couple crop physiological responses with environmental parameters, resulting in monitoring data that cannot directly reflect the actual impact of the environment on crop growth. At the same time, there are few simultaneous monitoring schemes for trace harmful gases (such as ammonia and hydrogen sulfide) in the field and soil organic matter, and many of them have cross-sensitivity issues. The detection accuracy is seriously affected by environmental interference. According to statistics, the data error rate of traditional gas sensors can reach 25%-30% in complex field environments.

[0003] Insufficient anti-interference capability and high maintenance cost: There are multiple interference factors in the field, such as electromagnetic interference (agricultural machinery, power lines), dust adhesion, and temperature drift. Traditional sensors mostly use single shielding or filtering technology, which has limited anti-interference effect. Moreover, the sensor probe is easily blocked by dust and corroded by water vapor, which leads to a decrease in detection sensitivity. For example, the detection accuracy of photoelectric sensing devices without dust protection measures can drop to below 80% after two dust removal cycles in the field, requiring frequent manual maintenance and calibration.

[0004] The contradiction between communication coverage and power consumption is prominent: remote farmland often lacks base station signals. Although LoRa communication has low power consumption, its transmission distance is limited and its resistance to obstruction is weak. Satellite communication has excessive power consumption, making it difficult to achieve long-term battery life. Existing systems have not optimized communication links for complex terrain in fields (ditches, tree obstruction), resulting in a high data transmission interruption rate. At the same time, the continuous operation of multiple modules further shortens the equipment's battery life cycle.

[0005] The self-calibration mechanism is imperfect: Traditional sensors mostly rely on periodic manual calibration, which has a long calibration cycle, high cost, and cannot correct errors caused by temperature drift, device aging, etc. in real time. After long-term operation, the data accuracy will decrease significantly. For example, the error of traditional temperature sensors without drift compensation can reach ±2℃, which is far beyond the needs of precision agriculture monitoring.

[0006] To address the aforementioned technical pain points, existing technologies mostly adopt single-dimensional improvements (such as only optimizing anti-interference packaging or only increasing communication power), failing to form a collaborative optimization scheme across the entire "sensing-anti-interference-communication-calibration" chain. Furthermore, they do not incorporate the indirect monitoring needs of crop physiological responses, resulting in insufficient innovation and limited scenario adaptability, making it difficult to meet the long-term, low-cost, and high-precision requirements of modern agricultural precision monitoring. Summary of the Invention

[0007] This invention aims to address the technical problems of existing IoT-based agricultural environmental monitoring electronic hardware systems, such as limited monitoring dimensions, weak anti-interference capabilities, contradictions between communication coverage and power consumption, and imperfect self-calibration. It provides an IoT-based agricultural environmental monitoring electronic hardware system based on multispectral coupled sensing and scattering communication, which enables accurate, long-term, and low-power monitoring of multi-dimensional parameters in complex field environments, improves the system's adaptability in remote areas, reduces maintenance costs, and provides reliable data support for precision agriculture decision-making.

[0008] The technical solution of this invention is implemented as follows: An IoT-based electronic hardware system for agricultural environmental monitoring includes a core control module, a multispectral coupling sensing module, a layered anti-interference packaging module, a dual-mode communication module, an adaptive power supply module, and a neural network self-calibration module. The core control module is electrically connected to each of the other modules to realize data interaction and logic control. The layered anti-interference packaging module is wrapped around the multispectral coupling sensing module. The dual-mode communication module, adaptive power supply module, and neural network self-calibration module are all integrated with the core control module through a PCB board. The system collects multi-dimensional environmental and crop-related parameters through a multispectral coupled sensing module. After hierarchical anti-interference processing, preprocessing by the core control module, and real-time correction by the neural network self-calibration module, the data is adaptively uploaded by the dual-mode communication module. The adaptive power supply module provides long-term and stable power supply for the entire system, enabling accurate, low-power, and long-term monitoring in complex field environments.

[0009] Preferably, the multispectral coupled sensing module includes a physical parameter sensing unit, a trace gas sensing unit, and a multispectral crop physiological correlation sensing unit. The physical parameter sensing unit uses temperature, humidity, light, and soil conductivity sensors with gold-plated contacts, and is connected to the core control module via an I2C bus. The trace gas sensing unit adopts a resistive gas sensor array based on metal porphyrin-modified polypyrrole, and outputs signals in conjunction with a signal conditioning circuit and an ADC conversion module. The core control module achieves selective identification of target gases through a built-in SVM model, which is used to solve the cross-sensitivity problem of traditional gas sensors and improve the detection accuracy of trace gases.

[0010] Preferably, the layered anti-interference packaging module includes a dustproof and dust removal submodule, an electromagnetic shielding submodule, and a temperature buffer submodule; The dust prevention and dust removal submodule is equipped with a high-transmittance quartz glass dustproof plate and an ultra-fine fiber dust removal cloth. The dustproof plate is driven to reciprocate through a stepper motor-screw mechanism. The core control module triggers the dust removal action according to the light intensity attenuation threshold of the multispectral sensor to avoid dust from obstructing the sensing accuracy. The electromagnetic shielding submodule adopts a multi-layer structure of aluminum alloy shell + conductive polymer shielding film + wave absorbing material, combined with shielded twisted pair cable and common mode choke, to suppress electromagnetic interference in the field; The temperature buffer submodule constructs a buffer cavity using thermal insulation cotton and incorporates a miniature thermistor to collect temperature data, providing a basis for subsequent temperature drift compensation.

[0011] Preferably, the dual-mode communication module includes a LoRa communication unit, a scattering communication unit, and a link quality detection unit, all of which are connected to the core control module via an SPI bus; The link quality detection unit monitors the RSSI value and bit error rate of the two communication modes in real time. When the LoRa communication signal strength is ≥-85dBm and the bit error rate is ≤1%, the core control module controls the system to prioritize LoRa mode communication. Otherwise, it automatically switches to scatter communication mode to balance communication coverage in remote areas and system power consumption, and avoid data transmission interruption.

[0012] Preferably, the neural network self-calibration module includes a calibration reference source, an error detection unit, and a neural network operation unit; The calibration reference source has built-in high-precision voltage and temperature references to provide calibration reference signals for each sensing unit. The error detection unit acquires the sensor output signal and the reference signal in real time and calculates the error value. The neural network operation unit has a built-in BP neural network model. It takes the sensor output value and temperature data as input and the calibrated value as output. It adaptively adjusts the calibration frequency according to the environmental stability to correct errors caused by temperature drift and device aging in real time, ensuring long-term monitoring accuracy.

[0013] Preferably, the adaptive power supply module includes a solar panel, a lithium battery pack, a power management chip, and a power monitoring unit; The power management chip enables charge and discharge management and overcharge, over-discharge, and over-temperature protection. The power monitoring unit collects power consumption data of each module in real time and transmits it to the core control module. The core control module adaptively allocates power according to the solar power supply and power consumption data. When the lithium battery voltage is lower than 3.0V, the control system enters a low-power sleep mode to achieve long-term battery life and adapt to remote areas without external power supply scenarios.

[0014] Preferably, the multispectral crop physiological correlation sensing unit uses a near-infrared multispectral sensor with a center wavelength of 730nm, 850nm, and 940nm. It focuses the light reflected from crop leaves through an optical lens and converts it into an electrical signal. It adopts a pulse working mode to indirectly monitor crop water stress and nutrient deficiency status, so that the monitoring data is directly related to crop growth needs, thus making up for the limitations of traditional physical parameter monitoring.

[0015] Preferably, the aluminum alloy shell of the electromagnetic shielding submodule is 1.5mm thick, the conductive polymer shielding film is 0.2mm thick, and ferrite particle absorbing material is filled between the shell and the shielding film. The PCB boards of the core control module and the dual-mode communication module adopt a grounding optimization design to make the system electromagnetic interference suppression ratio ≥80dB, which is used to ensure the stable operation of the system around agricultural machinery and power lines.

[0016] Preferably, the core control module adopts a low-power ARM Cortex-M4 core MCU, integrates an edge computing unit and a real-time clock module, and has a built-in Flash storage unit with a capacity of not less than 1MB; The edge computing unit is used to filter, normalize and fuse multi-sensor data, the real-time clock module is used to trigger timed monitoring and calibration tasks, and the Flash storage unit is used to store calibration parameters, cache monitoring data and configure communication protocols, providing core support for the logical control and data processing of the entire system.

[0017] Preferably, the sensor array of the trace gas sensing unit is adapted to three target gases: ammonia, hydrogen sulfide, and carbon dioxide, with a detection limit ≤0.33 mg / m³. 3 The signal conditioning circuit includes a differential amplifier and filter module, which converts the sensor resistance change signal into a 0-3.3V analog signal. After conversion by the 16-bit ADC module, the signal is transmitted to the core control module. The core control module uses an SVM model to achieve a target gas identification accuracy of ≥97%, which is used to meet the needs of accurate monitoring of trace harmful gases in the field.

[0018] The embodiments of the present invention have the following advantages due to the adoption of the above technical solutions: I. More comprehensive and accurate monitoring dimensions, innovatively solving the problem of the disconnect between traditional monitoring and crop growth needs: Through multispectral coupled sensing modules, the system achieves simultaneous monitoring of physical environment, trace gases, soil physicochemical properties, and crop physiological parameters. Combined with the SVM model, it solves the problem of cross-sensitivity of gas sensors. Multispectral reflectance monitoring indirectly reflects the crop growth status, and the monitoring data is more in line with the actual needs of agricultural production. Experimental verification shows that the detection accuracy of each parameter of this system is ≥93%, and the temperature drift error is ≤±0.5℃, which is significantly better than traditional monitoring systems.

[0019] II. Strong anti-interference capability, adaptable to complex field environments, and reduced maintenance costs: The layered anti-interference packaging module effectively resists multiple interferences such as dust, electromagnetic interference, and temperature fluctuations through triple protection of dust prevention, electromagnetic shielding, and temperature buffering; the dust prevention submodule keeps the sensor probe cleanliness above 90% and the detection accuracy above 93%, eliminating the need for frequent manual cleaning; the electromagnetic shielding design enables the system to work stably near agricultural machinery and power lines, reducing the data error rate to below 5%.

[0020] III. Overcoming the contradiction between communication coverage and power consumption, and adapting to long-term monitoring in remote areas: The dual-mode communication module achieves a balance between low power consumption at close range and non-line-of-sight transmission in remote areas through intelligent link switching. The scattering communication mode solves the communication problem in remote areas without base stations, and the power consumption is only 1 / 5 of that of traditional satellite communication. The data packet compression and power adaptive allocation design enable the system to achieve continuous maintenance-free operation for 3 months under solar power conditions, far exceeding the existing system.

[0021] IV. Improved self-calibration mechanism to ensure long-term monitoring accuracy: The adaptive self-calibration module based on BP neural network realizes real-time automatic calibration, eliminating the need for manual periodic calibration and reducing calibration costs by more than 80%; the calibration frequency is adaptively adjusted according to environmental stability, ensuring calibration accuracy while reducing calibration power consumption. After long-term operation, the data accuracy can still be maintained within ±1%, meeting the long-term monitoring needs of precision agriculture.

[0022] V. Novelty: Existing technologies have not combined scattering communication with LoRa dual-mode for agricultural environmental monitoring, have not achieved the coupled monitoring of multispectral crop physiological correlation sensing with trace gases and physical parameters, and have not formed a full-link collaborative optimization scheme of "sensing-anti-interference-communication-calibration-power supply"; the core technology combination of this solution has few applications in the field of agricultural monitoring and is highly feasible.

[0023] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application 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 some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1This is a schematic diagram of the core system architecture of the present invention; Figure 2 This is a schematic diagram of the workflow of the core module of the present invention. Detailed Implementation

[0026] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0027] It is important to note that terms such as "first," "second," "symmetric," and "array" are used only to distinguish between descriptive and positional descriptions and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, features specified with terms such as "first" or "symmetric" may explicitly or implicitly include one or more of that feature; similarly, when the quantity of certain features is not limited by words such as "two" or "three," it should be noted that such features also explicitly or implicitly include one or more features. In this invention, unless otherwise explicitly specified and limited, terms such as "installation," "connection," and "fixation" should be interpreted broadly; for example, they can refer to a fixed connection, a detachable connection, or an integral molding; they can refer to a mechanical connection, a direct connection, a welding connection, or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the accompanying drawings and specific circumstances.

[0028] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] like Figure 1-2 As shown, the present invention provides an electronic hardware system for monitoring the agricultural environment via the Internet of Things, including a core control module, a multispectral coupling sensing module, a layered anti-interference packaging module, a dual-mode communication module, an adaptive power supply module, and a neural network self-calibration module. Each module is integrated and connected via a PCB board. The core control module is electrically connected to the other modules to realize data interaction and logic control. I. Core Control Module The core control module adopts a low-power ARM Cortex-M4 core microcontroller (MCU), integrating an edge computing unit and a real-time clock module; the MCU has a built-in Flash storage unit (capacity not less than 1MB) for storing sensor calibration parameters, monitoring data cache, and communication protocol configuration; the edge computing unit is used to realize the preprocessing (filtering, normalization, fusion) of multi-sensor data and communication link switching decisions, and the real-time clock module is used to trigger timed monitoring and calibration tasks to reduce continuous operation power consumption.

[0030] The core function of the core control module is to receive the raw data collected by the multispectral coupled sensing module, perform data preprocessing and feature extraction through the edge computing unit, determine the working mode of the dual-mode communication module according to the communication link quality, trigger the real-time calibration task of the neural network self-calibration module, control the power distribution of the adaptive power supply module, upload the processed data to the terminal platform through the communication module, and receive control commands issued by the platform (such as calibration parameter updates and monitoring frequency adjustments).

[0031] II. Multispectral Coupled Sensing Module To address the issues of single monitoring dimensions and cross-sensitivity in traditional methods, this module employs a coupled design of "physical parameter sensing + trace gas sensing + multispectral crop physiological correlation sensing," specifically including: Physical parameter sensing unit: It adopts a high-precision temperature and humidity sensor (DS18B20 improved version, temperature range -40℃-85℃, error ±0.1℃), light sensor (TSL2591, range 0-68800lux), and soil conductivity sensor (EC-5 improved version), which are connected to the core control module through I2C bus; the sensor probe adopts gold-plated contacts to reduce the impact of soil corrosion.

[0032] Trace gas sensing unit: Employs a resistive gas sensor array based on metalloporphyrin-modified polypyrrole (compatible with three target gases: ammonia, hydrogen sulfide, and carbon dioxide), combined with a MEMS microcantilever beam structure to improve detection sensitivity (detection limit ≤ 0.33 mg / m³). 3 The signal conditioning circuit (including differential amplification and filtering modules) converts the resistance change signal into a 0-3.3V analog signal, which is then converted by the ADC module (16-bit) and transmitted to the core control module. To solve the cross-sensitivity problem, the sensor array output signal is feature extracted and then selectively identified by the support vector machine (SVM) model built into the core control module, with an identification accuracy of ≥97%.

[0033] Multispectral crop physiology correlation sensing unit: Employs a near-infrared multispectral sensor (center wavelength 730nm, 850nm, 940nm), which focuses the reflected light from crop leaves through an optical lens and converts the reflected light signal into an electrical signal. This unit indirectly reflects the physiological state of crops, such as water stress and nutrient deficiency, by monitoring changes in the near-infrared reflectivity of crop leaves, thus overcoming the deficiency that traditional physical parameters cannot be directly correlated with crop growth needs. The sensor adopts a pulsed working mode to reduce power consumption.

[0034] The working logic of this module is as follows: The core control module triggers each sensing unit to work at regular intervals through the real-time clock module (the monitoring frequency can be configured through the terminal platform, with a default of once every 30 minutes). The physical parameter sensing unit and the trace gas sensing unit collect data synchronously, while the multispectral sensing unit collects data once every 2 hours (matching the time scale of crop physiological response). The collected raw data is transmitted to the core control module after signal conditioning and ADC conversion. The edge computing unit performs data fusion and removes redundant information and outliers.

[0035] III. Layered Anti-interference Packaging Module To address the multiple disturbances encountered in the field, this module employs a layered design of "dust prevention and removal + electromagnetic shielding + temperature buffering" to adapt to complex field environments. Specifically, it includes: Dustproof and dust removal submodule: Addressing the issue of dust accumulation on sensor probes, and referencing the principles of dust isolation and friction dust removal, a transparent dustproof plate (made of high-transmittance quartz glass, 2mm thick) is installed on the outside of the probes of the multispectral sensing unit and the physical parameter sensing unit to isolate the probes from the external environment. A dust removal cloth (made of microfiber) is attached to the outside of the dustproof plate, and a stepper motor-screw mechanism drives the dustproof plate to reciprocate up and down (stroke 8mm) to achieve frictional removal of dust from the surface of the dustproof plate. The core control module monitors the light intensity attenuation of the multispectral sensor (when the light intensity is lower than 70% of the initial value), triggering the dust removal mechanism to work, thus preventing dust from obstructing and affecting detection accuracy.

[0036] Electromagnetic shielding submodule: It adopts a multi-layer shielding structure, with an outer aluminum alloy shell (1.5mm thick) and an inner conductive polymer shielding film (0.2mm thick). The space between the shell and the shielding film is filled with absorbing material (ferrite particles). The sensor signal line uses shielded twisted pair cable, and a common mode choke is set at the interface to reduce electromagnetic interference (suppression ratio ≥80dB). The PCB board of the core control module and the communication module adopts a grounding optimization design to reduce ground loop current and common mode noise, reducing the system noise level by more than 50%.

[0037] Temperature buffer submodule: Thermal insulation cotton (glass wool, 5mm thick) is placed between the sensor probe and the housing to construct a temperature buffer cavity; a miniature thermistor is embedded in the buffer cavity to monitor the temperature around the probe in real time, and the temperature data is transmitted to the core control module to provide a basis for temperature drift compensation; through this design, the sensor's operating temperature range is stabilized between -30℃ and 75℃, avoiding the impact of drastic fluctuations in ambient temperature on detection accuracy.

[0038] The working logic of this module is as follows: the dust removal submodule automatically triggers the dust removal action based on the light intensity attenuation threshold; the electromagnetic shielding submodule and the temperature buffer submodule work continuously to provide a stable working environment for each sensing unit; the core control module receives the temperature data from the temperature buffer submodule in real time to provide input parameters for the temperature drift compensation of the subsequent self-calibration module.

[0039] IV. Dual-mode communication module To address the conflict between communication coverage and power consumption in remote areas, this module adopts a "scatter communication + LoRa dual-mode" design, specifically including a scatter communication unit, a LoRa communication unit, and a link quality detection unit, all of which are connected to the core control module via an SPI bus. LoRa communication unit: uses SX1278 chip, working frequency 433MHz, adjustable transmission power (1-20dBm), transmission distance ≥3km in unobstructed scenarios, suitable for low-power communication in short distances and areas with base station coverage. Scattering communication unit: It adopts a low-power communication chip based on the principle of atmospheric scattering (independently optimized design, operating frequency 915MHz), and uses atmospheric scattering to achieve non-line-of-sight transmission. In obstructed and remote scenarios, the transmission distance is ≥5km and the transmission power is ≤15dBm, which solves the problem of high power consumption in traditional satellite communication. Link quality detection unit: Real-time monitoring of signal strength (RSSI) and bit error rate (BER) of the two communication modes. When the LoRa communication signal strength is ≥-85dBm and the bit error rate is ≤1%, the LoRa communication mode is used first. When the LoRa communication signal strength is <-85dBm or the bit error rate is >1%, the system automatically switches to the scatter communication mode. At the same time, the link quality data is transmitted to the core control module in real time for communication mode switching decisions and power adjustment.

[0040] The working logic of this module is as follows: The core control module dynamically switches the communication mode and adjusts the transmission power based on the monitoring data of the link quality detection unit; the data transmission adopts a packet transmission mechanism, and only the pre-processed valid data (compression ratio 3:1) is uploaded to further reduce communication power consumption; when neither of the two communication modes can be transmitted normally, the data is temporarily stored in the Flash storage unit of the core control module and resent after communication is restored.

[0041] V. Adaptive Power Supply Module To achieve long-lasting battery life, this module adopts a "solar energy + lithium battery + adaptive power distribution" design, specifically including a solar panel (10W, 12V), a lithium battery pack (3.7V, 10000mAh), a power management chip (improved TP4056), and a power monitoring unit. The power management chip is used to manage the charging and discharging of solar panels and lithium battery packs, and has overcharge, over-discharge and over-temperature protection functions. The power monitoring unit monitors the power consumption status of each module in real time and transmits the monitoring data to the core control module. The core control module adaptively allocates power based on the power consumption data and solar power supply status: when solar power supply is sufficient during the day, it prioritizes solar power supply and charges the lithium battery pack at the same time; when solar power supply is insufficient at night or on cloudy or rainy days, it switches to lithium battery power supply and reduces the power consumption of non-core modules (such as reducing the monitoring frequency of the multispectral sensing unit and reducing the transmission power of the communication module); when the lithium battery voltage is lower than 3.0V, the system enters a low-power sleep mode, keeping only the core control module and the link quality detection unit working, waiting for solar charging to resume.

[0042] VI. Neural Network Self-calibration Module To address the shortcomings of traditional manual calibration, this module employs an adaptive self-calibration mechanism based on a BP neural network, integrating a calibration reference source, an error detection unit, and a neural network computation unit. Specifically, it includes: Calibration reference source: Built-in high-precision voltage reference (REF5040, accuracy ±0.02%) and temperature reference (LM335, accuracy ±0.5℃) provide calibration reference signals for each sensing unit; Error detection unit: Real-time acquisition of output signals from each sensor and reference signals from the calibration reference source, and calculation of error values ​​(including systematic error, temperature drift error, and random error). Neural network operation unit: Built-in trained BP neural network model (input layer is sensor output value and temperature value, hidden layer is 12 neurons, output layer is calibrated value), the weight parameters of the neural network are updated in real time through error data from the error detection unit to realize real-time calibration of sensor output signal; the calibration cycle can be adaptively adjusted. When the temperature fluctuation is large (≥5℃ / h) or the error value is ≥1%, the calibration frequency is increased to once every 1 hour; when the environment is stable, the calibration frequency is reduced to once every 6 hours.

[0043] The working logic of this module is as follows: The core control module periodically triggers the error detection unit to collect the sensor output signal and the reference signal, and calculates the error value; the neural network operation unit uses the error data and temperature data to correct the sensor output signal in real time through a BP neural network model. The corrected calibration parameters are stored in the Flash storage unit for subsequent monitoring data calibration reference; at the same time, the calibration results are fed back to the core control module in real time. If the error is still ≥2% after 3 consecutive calibrations, a fault alarm signal is sent to the terminal platform through the communication module to prompt manual maintenance.

[0044] VII. Overall System Workflow The overall workflow of the system of this invention follows a causal and coherent logic of "triggering - acquisition - anti-interference processing - data preprocessing - self-calibration - communication - power supply adaptation", and the specific steps are as follows: Initialization phase: After the system is powered on, the core control module starts the initialization program, configures the working parameters of each module (sensing frequency, communication protocol, calibration benchmark, etc.), the link quality detection unit starts the communication link scan, and the adaptive power supply module switches to the appropriate power supply mode (solar or lithium battery). Monitoring trigger phase: The core control module triggers the multispectral coupling sensor module to start data acquisition according to the timed instructions from the real-time clock module (once every 30 minutes by default); if the terminal platform issues an instant monitoring instruction, the core control module responds to the instant instruction first and starts the acquisition task. Data acquisition and anti-interference processing stage: The physical parameters, trace gases, and raw data of the multispectral coupled sensing module are collected synchronously by the multispectral sensing unit. The layered anti-interference packaging module performs dust removal, electromagnetic shielding, and temperature buffering in real time to avoid external interference affecting the accuracy of the raw data. The collected raw data is transmitted to the core control module after signal conditioning circuit (amplification and filtering) and ADC conversion. Data preprocessing stage: The edge computing unit of the core control module preprocesses the raw data, including digital filtering (Kalman filtering algorithm, filtering accuracy ≥95%), data normalization (conversion error ≤±0.5%), data fusion (weighted average method, fusion efficiency ≥90%), removing outliers and redundant information, and extracting effective data features; Self-calibration phase: The neural network self-calibration module starts synchronously, the error detection unit collects the sensor output signal and the reference signal of the calibration reference source, and calculates the error value; the neural network operation unit uses the error data and the temperature data of the temperature buffer submodule to perform real-time calibration on the preprocessed data through the BP neural network model to correct errors caused by temperature drift, device aging, etc. During the communication transmission phase: The core control module dynamically switches between LoRa / scatter communication modes based on the monitoring data from the link quality detection unit, and uploads the calibrated valid data packets after compression to the terminal platform; if communication is interrupted, the data is temporarily stored in the Flash storage unit and resent after communication is restored; at the same time, it receives control commands (such as monitoring frequency adjustment and calibration parameter update) issued by the terminal platform. Power supply adaptation phase: The power monitoring unit of the adaptive power supply module monitors the power consumption of each module and the solar power supply status in real time. The core control module adaptively allocates power according to the monitoring data and adjusts the working status of each module (normal working mode or low power mode) to ensure the long-term battery life of the system. Cyclic Phase: The system operates in a cyclical manner according to the above steps to achieve continuous monitoring of the agricultural environment; the core control module performs fault self-checks on each module periodically (once a day), and if a fault is detected (such as sensor damage or communication module failure), an alarm signal is immediately sent through the communication module.

[0045] In this embodiment, the present invention operates as follows: First, the system is fixed in a designated location in the field (1.5-2m high) using a pole, with the sensor probe facing the crop growth area. The solar panel is fixed at a 45° angle to the top of the pole to ensure sufficient sunlight. The adaptive power supply module completes power-on initialization first, and the power management chip automatically detects the power supply status, prioritizing solar power and charging the lithium battery pack. If solar power is insufficient at night or on cloudy or rainy days, it switches to lithium battery power mode. At the same time, the core control module starts the overall initialization program, configuring the operating parameters of each module (default physical parameters and trace gas monitoring frequency of 30 minutes / time, multispectral sensor monitoring frequency of 2 hours / time, communication protocol adaptation dual-mode switching logic, etc.). The link quality detection unit simultaneously starts scanning the LoRa and scattering communication links, and the neural network self-calibration module loads preset calibration benchmark parameters and a BP neural network model, completing the preparation for the entire system to start.

[0046] Subsequently, the core control module triggers the monitoring task according to the timed instructions from the real-time clock module. If the terminal platform issues an immediate monitoring instruction, it prioritizes responding to the immediate instruction and starting the acquisition. After receiving the trigger instruction, the multispectral coupled sensing module simultaneously starts data acquisition with the physical parameter sensing unit (temperature, humidity, light, and soil conductivity sensors) and the trace gas sensing unit (metalporphyrin-modified polypyrrole sensor array). The physical parameters are transmitted to the I2C bus via gold-plated contacts. The trace gas sensor array converts the gas concentration into a resistance change signal, which is then converted into a 0-3.3V analog signal by the signal conditioning circuit (differential amplification + filtering) and then converted into a digital signal by the 16-bit ADC module. At the same time, the multispectral crop physiological correlation sensing unit starts in pulse working mode, focusing the reflected light from crop leaves through an optical lens, collecting reflected light signals at three center wavelengths of 730nm, 850nm, and 940nm and converting them into electrical signals. All raw acquisition data are synchronously transmitted to the core control module.

[0047] During data acquisition, the layered anti-interference packaging module continuously plays a protective role: the dustproof and dust removal submodule monitors the light intensity signal of the multispectral sensor in real time. If the light intensity attenuates to 70% of the initial value, the core control module triggers the stepper motor-screw mechanism to drive the quartz glass dustproof plate to reciprocate up and down (stroke 8mm). The dust on the surface of the dustproof plate is removed by friction through the outer ultra-fine fiber dust removal cloth, avoiding dust from obstructing the detection accuracy; the electromagnetic shielding submodule uses multiple layers of protection, including an aluminum alloy shell, a conductive polymer shielding film, and ferrite absorbing material, along with shielded twisted-pair cables and a common-mode choke, to suppress electromagnetic interference generated by agricultural machinery and power lines in the field, ensuring stable signal transmission (electromagnetic interference suppression ratio ≥80dB); the temperature buffer submodule maintains the temperature stability around the probe through a glass wool insulation layer, and a miniature thermistor collects the temperature data of the buffer cavity in real time and transmits it to the core control module, providing data support for subsequent temperature drift compensation.

[0048] After receiving the raw data, the core control module's built-in edge computing unit preprocesses the data: it uses a Kalman filter algorithm (filtering accuracy ≥95%) to remove random interference noise, normalizes the data format through data normalization (conversion error ≤±0.5%), and then uses a weighted average method to complete multi-dimensional data fusion (fusion efficiency ≥90%) to extract effective data features. At the same time, it calls the SVM model to selectively identify target gases for trace gas data (identification accuracy ≥97%), distinguishes the concentration data of three target gases: ammonia, hydrogen sulfide, and carbon dioxide, and removes redundant information and outliers.

[0049] After preprocessing, the core control module triggers the neural network self-calibration module to start real-time calibration: the error detection unit collects the preprocessed output signals of each sensor and compares them with the reference signals provided by the calibration reference source (REF5040 voltage reference + LM335 temperature reference) to calculate the system error, temperature drift error and random error; the neural network operation unit takes the sensor output values ​​and the temperature data collected by the temperature buffer submodule as inputs, and updates the weight parameters in real time through the BP neural network model (input layer - 12 hidden neurons - output layer) to calibrate and correct the data. If the current temperature fluctuation is ≥5℃ / h or the error value is ≥1%, the calibration frequency is automatically increased to 1 hour / time, and when the environment is stable, the calibration frequency is maintained at 6 hours / time. The corrected calibration parameters are stored in the Flash storage unit, and the calibrated data is output as valid monitoring data.

[0050] After valid data is generated, the core control module determines the operating mode of the dual-mode communication module based on the monitoring data (RSSI value and bit error rate) fed back in real time by the link quality detection unit: if the LoRa communication signal strength is ≥-85dBm and the bit error rate is ≤1%, the LoRa communication mode is used first (transmission power is adjusted to 5dBm); if the LoRa communication link is unstable (signal strength <-85dBm or bit error rate >1%), it automatically switches to the scattering communication mode (transmission power is adjusted to 12dBm) to achieve non-line-of-sight transmission using atmospheric scattering; data transmission adopts a packet compression mechanism (compression ratio 3:1) to upload valid monitoring data (physical environmental parameters, trace gas concentrations, crop physiological correlation parameters) to the terminal platform, while receiving control commands issued by the platform (such as monitoring frequency adjustment, calibration benchmark update, etc.); if neither communication mode can transmit normally, the data is temporarily stored in the Flash storage unit of the core control module and automatically resent after the communication link is restored.

[0051] During this process, the adaptive power supply module continuously and dynamically adapts to power consumption requirements: the power monitoring unit collects power consumption data of each module in real time and transmits it to the core control module. The core control module allocates power based on the power supply status of the solar panels—when solar power is sufficient during the day, priority is given to ensuring the normal operating power of each module, while charging the lithium battery pack; when solar power is insufficient at night or on cloudy or rainy days, the power consumption of non-core modules is appropriately reduced (such as lowering the monitoring frequency of the multispectral sensing unit and optimizing the transmission power of the communication module as needed); if the lithium battery voltage drops below 3.0V, the system automatically enters a low-power sleep mode, keeping only the core control module and the link quality detection unit working, and resumes normal operation when the solar charging raises the lithium battery voltage back to above 3.3V.

[0052] The system operates in a cycle according to the above-mentioned process of "trigger-acquisition-anti-interference-preprocessing-self-calibration-communication-power supply adaptation" to achieve continuous monitoring of the agricultural environment. The core control module regularly starts a full-module fault self-check once a day. If abnormal conditions such as sensor damage, communication module failure, or calibration error ≥2% for 3 consecutive times are detected, a fault alarm signal is immediately sent to the terminal platform through the currently available communication link to prompt manual maintenance and ensure long-term stable operation of the system.

[0053] The following are several other specific embodiments of the application of this invention: Example 1: Monitoring System Adapted for Desert Oasis Farmland This embodiment addresses the challenges of desert oasis farmland characterized by strong winds and sandstorms, large diurnal temperature variations, scarce water resources, the need to monitor soil moisture and crop drought resistance, and inconvenient power supply. Based on the original core architecture, it enhances wind and sand resistance and drought resistance, optimizes water supply and soil monitoring dimensions, and improves the adaptability of solar power supply, as detailed below: I. Hardware Component Selection and Assembly Adjustment The core architecture retains multispectral coupled sensing, layered anti-interference packaging, dual-mode communication, and neural network self-calibration, focusing on optimizing pain points in desert oasis scenarios. The multispectral coupled sensing module adds a soil moisture sensor and a crop drought resistance sensor unit, optimizing multispectral wavelengths for oasis crops such as wheat and corn. This module retains the original temperature, humidity, light, and soil conductivity sensors, adding a soil moisture sensor and a crop drought resistance sensor. The multispectral wavelengths are adjusted to 660nm red light, 730nm near-infrared, and 850nm near-infrared. All sensors feature a wind-resistant and sand-sealed design, and the probes are equipped with dustproof covers. During assembly, the sensor probes are fixed to oasis farmland monitoring piles using wind-resistant reinforced brackets at a height of 1.2 to 1.5 meters. The soil moisture and conductivity sensors are inserted into the soil to a depth of 15 to 20 centimeters, avoiding areas with severe wind and sand erosion. The main unit is fixed to the bottom of the monitoring pile using expansion bolts. The layered anti-interference packaging module enhances wind and sand resistance and impact resistance, optimizing dust and moisture protection for the dry and windy desert environment. The outer shell is made of thick-walled aluminum alloy, with a thickness increased to 3mm. It is encased in a wind-resistant and sand-proof protective sleeve, achieving an IP68 protection rating. The dustproof panel uses 4mm high-transmittance quartz glass, retaining the stepper motor dust removal mechanism with a stroke adjusted to 15mm to address the issue of sand accumulation in strong winds. A windproof base is added to the bottom of the shell, secured with multiple bolts to withstand strong wind impacts. The dual-mode communication module optimizes signal penetration, adapting to open desert environments with sandstorm obstruction, and enhances the module's wind and sand resistance. The LoRa chip uses an improved wind-resistant and sand-proof version of the SX1278, with a transmission power of 12dBm. The scattering communication chip's operating frequency is adjusted to 915MHz. The antenna is a wind-resistant and waterproof external antenna with a gain increased to 5dBi, fixed to the top of the monitoring pile above vegetation. The adaptive power supply module optimizes the solar panel to adapt to the strong sunlight and low-angle characteristics of the desert, increasing the lithium battery capacity and enhancing its high-temperature resistance. The solar panel is a 25W monocrystalline silicon high-temperature resistant type, with an adjustable tilt angle of 20 to 70 degrees. The lithium battery pack uses high-temperature resistant lithium iron phosphate batteries with a capacity increased to 18000mAh. The power management chip uses an improved high-temperature resistant version to ensure normal operation in environments up to 60 degrees Celsius. The solar panel is fixed to the top of the monitoring pile facing south, and the lithium battery pack is installed in the heat-insulated cavity inside the main unit. The core control and self-calibration module adds calibration logic linking crop drought resistance and soil moisture content, and optimizes the data fusion algorithm under high-temperature and sandy conditions. The original MCU is retained, and soil moisture content and drought resistance parameters are added to the input layer of the BP neural network model. The PCB board is coated with a high-temperature resistant and wind-resistant sand-resistant coating, installed in the heat-insulated area inside the main unit, and sealed and protected with the sensor interface.

[0054] II. Software Configuration and Parameter Setting Adjustment The software configuration and parameter settings are optimized for desert oasis farmland scenarios. Regarding monitoring frequency, physical parameters such as temperature, humidity, and light intensity are monitored every 25 minutes; soil moisture content and conductivity every 30 minutes; crop drought resistance and multispectral sensing every hour; and trace gases, primarily water vapor and ammonia volatilized from the soil, are monitored every 40 minutes. For anti-interference parameters, the dust removal trigger threshold is adjusted to 50% of the initial light intensity attenuation. Given the strong winds and sand accumulation in deserts, the electromagnetic interference suppression ratio is set at no less than 90dB, adapting to interference from oasis irrigation equipment and small agricultural machinery. Regarding communication parameters, scattering communication mode is prioritized. In the open desert environment with no tall obstructions, LoRa is used as a backup, with a switching threshold of scattering communication RSSI less than -100dBm. For self-calibration parameters, the calibration frequency is increased when temperature fluctuations are no less than 6℃ per hour or the error is no less than 1.5%. Drift compensation logic for high-temperature and windy sand environments has been added, adapting to the characteristics of large day-night temperature differences and wind and sand interference. In terms of power supply parameters, the low power consumption threshold is adjusted to 2.7V to extend the battery life in high-temperature environments. Solar charging has the highest priority, and the lithium battery deep discharge protection threshold is 2.4V. During high-temperature periods, the power consumption of non-core modules is automatically reduced.

[0055] III. In this embodiment, the specific operation of the present invention... First, the system is fixed to the oasis farmland monitoring pile at a height of 1.4 meters using a windproof base and multiple bolts. Soil moisture and conductivity sensors are inserted 18 centimeters deep into the soil. Dustproof covers are added to the sensor probes to avoid areas with severe wind and sand erosion. A solar panel is fixed to the top of the monitoring pile at a 65-degree angle facing south to ensure sufficient sunlight. After the adaptive power supply module is powered on, the power management chip detects the ambient temperature and power supply status, prioritizing solar power and charging the 18000mAh high-temperature resistant lithium battery pack. If solar power is insufficient during high-temperature periods or windy / sandy weather, it switches to lithium battery power mode and automatically activates the heat insulation and cooling mechanism to ensure normal operation of all modules. The core control module starts the initialization program, loading the parameter configuration adapted for desert oasis crop monitoring and the optimized BP neural network model, including high-temperature drift compensation and drought resistance calibration logic. The link quality detection unit simultaneously scans the scattering and LoRa communication links, prioritizing the scattering communication mode by default, completing the startup preparation.

[0056] Subsequently, the core control module triggers monitoring tasks according to the timed instructions from the real-time clock module, monitoring soil parameters every 30 minutes and temperature, humidity, and light intensity every 25 minutes. If the oasis management platform issues an immediate monitoring instruction, such as a spot check on crop drought resistance, it will respond first. After receiving the instruction, the multispectral coupled sensing module simultaneously collects oasis environmental data using temperature, humidity, and light sensors, collects soil moisture and salinity data using soil moisture and conductivity sensors, identifies drought resistance status by detecting changes in the reflectance spectrum of crop leaves, collects reflected light signals at wavelengths of 660nm, 730nm, and 850nm to correlate with crop growth, and collects water vapor and ammonia concentration data to reflect soil evaporation and fertility status. All raw data are transmitted to the core control module after signal conditioning and ADC conversion.

[0057] During data acquisition, the layered anti-interference packaging module works continuously: the dust removal submodule monitors the light intensity of the multispectral sensor in real time. If the light intensity decreases to 50% of the initial value, the core control module triggers the stepper motor-screw mechanism to drive the 4mm quartz glass dustproof plate to reciprocate 15mm, clearing the surface sand accumulation; the electromagnetic shielding submodule uses a thick-walled aluminum alloy shell and absorbing materials to suppress electromagnetic interference generated by irrigation equipment and agricultural machinery, ensuring signal stability; the temperature buffer submodule, in conjunction with the heat-insulating protective sleeve of the shell, maintains the probe temperature stability, providing accurate temperature data for self-calibration in high-temperature environments.

[0058] The core control module preprocesses data through an edge computing unit, fusing environmental parameters, soil data, crop drought resistance data, and multispectral data to remove outliers. It then triggers a neural network self-calibration module, combining temperature data, soil moisture content, and drought resistance parameters to perform real-time calibration using a BP neural network, correcting errors caused by high-temperature drift and wind and sand interference. The calibrated data is then compressed and uploaded to the oasis management platform via a scattering communication module. If the scattering communication signal is weakly affected by strong winds and sand, it automatically switches to LoRa communication backup. Simultaneously, the power monitoring unit provides real-time power consumption data, and the core control module adaptively allocates power, reducing the power consumption of non-core modules during high-temperature periods to ensure long-term lithium battery life. The system cyclically executes the above process. The core control module performs daily fault self-checks; if sensor malfunctions or communication interruptions are detected, it immediately sends alarm signals to ensure accurate monitoring of desert oasis farmland.

[0059] Example 2: System Implementation Example Adapted for Tea Garden Ecological Monitoring This embodiment addresses the specific characteristics of tea gardens, including hilly terrain, the need to monitor tea quality-related parameters and pest and disease early warning, high humidity and frequent fog, and minimal interference from agricultural machinery. Based on the original core architecture, it optimizes multispectral adaptation for tea growth, strengthens protection against foggy environments, and adds a new dimension for monitoring the correlation between quality and pests and diseases. Details are as follows: I. Hardware Component Selection and Assembly Adjustment The core architecture retains multispectral coupling sensing, layered anti-interference packaging, dual-mode communication, and neural network self-calibration, focusing on optimizing pain points in tea garden scenarios. The multispectral coupling sensing module adds a tea polyphenol correlation sensing unit and a tea geometrid moth early warning sensing unit, optimizing multispectral wavelengths to suit different tea growth stages. This module retains the original temperature, humidity, light, and soil conductivity sensors, adding a tea polyphenol correlation sensor and a tea geometrid moth early warning sensor, with multispectral wavelengths adjusted to 520nm green light, 680nm red light, and 940nm near-infrared. During assembly, the sensor probes are fixed to monitoring stakes between tea rows using adjustable brackets, at a height of 0.9 to 1.1 meters, facing the middle of the tea tree canopy, avoiding the tea-picking path. The main unit is fixed to the monitoring stakes with clips, protecting the probes from impacts. The layered anti-interference packaging module enhances anti-fog and moisture-proof performance, optimizing the dustproof structure to adapt to tea garden fog and fallen leaf dust. The casing is made of stainless steel with an IP68 protection rating and an external anti-corrosion coating. The dustproof panel uses 3mm high-transmittance quartz glass. The stepper motor dust removal mechanism is retained, and its stroke has been adjusted to 9mm. Anti-fog rubber rings are added to the casing seams to prevent fog from entering the main unit. The dual-mode communication module optimizes signal penetration to adapt to tea garden and tea tree obstructions, and enhances the module's anti-fog and moisture-proof protection. The LoRa chip uses an improved waterproof and anti-fog version of the SX1278, with a transmission power of 7dBm. The scattering communication chip's operating frequency has been adjusted to 868MHz. The antenna is a waterproof and anti-fog external antenna with a gain increased to 4dBi, fixed to the top of the monitoring pile in an unobstructed location. The adaptive power supply module optimizes the solar panel's performance to adapt to the hilly sunlight of the tea garden, and enhances the power supply module's anti-fog and moisture-proof performance. The solar panel is a 12W monocrystalline silicon waterproof type, with an adjustable tilt angle of 30 to 60 degrees. The lithium battery pack is 3.7V 10000mAh and equipped with a waterproof casing. The power management chip is coated with an anti-fog coating. The solar panel is fixed to the top of the monitoring pile facing south, and the lithium battery pack is installed in the waterproof cavity at the bottom of the main unit. The core control and self-calibration module adds calibration logic for tea quality parameters and pest and disease early warning, and optimizes the data fusion algorithm in high humidity and foggy environments. The original MCU is retained, and the input layer of the BP neural network model adds tea polyphenol correlation and tea geometrid moth early warning parameters. The PCB board is coated with a waterproof and anti-fog coating and installed in the waterproof area inside the main unit, with waterproof and insulating protection for the sensor interface.

[0060] II. Software Configuration and Parameter Setting Adjustment The software configuration and parameter settings are optimized for tea garden ecological monitoring. Regarding monitoring frequency, physical parameters such as temperature, humidity, light intensity, and soil conductivity are monitored every 20 minutes; tea polyphenol-related parameters are monitored every 30 minutes; and tea geometrid moth warning and multispectral sensing are monitored every 50 minutes. For anti-interference parameters, the dust removal trigger threshold is adjusted to 62% of the initial light intensity attenuation. Given the ease with which tea garden fog and fallen leaf dust adhere, the electromagnetic interference suppression ratio is set at no less than 80dB to accommodate interference from small tea-picking machines in tea gardens. Regarding communication parameters, LoRa communication is prioritized to accommodate close-range coverage under tea tree shading, with scattering communication as a backup. The switching threshold is set at a LoRa communication RSSI of less than -92dBm. For self-calibration parameters, the calibration frequency is increased when temperature fluctuations are no less than 2.5℃ per hour or when the error is no less than 1%. Drift compensation logic has been added for high-humidity fog environments to adapt to the high humidity characteristics of tea gardens. In terms of power supply parameters, solar power is prioritized, the lithium battery has a low power consumption threshold of 2.9V, and the battery life is no less than 2 months. In rainy or foggy weather, the power consumption of non-core modules is automatically reduced to extend the battery life.

[0061] III. In this embodiment, the specific operation of the present invention... First, the system is secured to monitoring posts between tea rows at a height of 1.0 meter, with the sensor probes facing the middle of the tea tree canopy, avoiding the tea-picking path. Solar panels are fixed to the top of the monitoring posts at a 45-degree angle facing south to ensure sufficient sunlight. After the adaptive power supply module is powered on, the power management chip detects the solar power status, prioritizing solar power and charging the waterproof lithium battery pack. If solar power is insufficient in rainy or foggy weather, it switches to lithium battery power mode. The core control module initiates the initialization program, loading the parameter configuration adapted for tea garden monitoring and the optimized BP neural network model, including tea polyphenol correlation and tea geometrid moth early warning calibration logic. The link quality detection unit simultaneously scans the LoRa and scattering communication links, prioritizing LoRa communication mode by default, completing the startup preparation.

[0062] Subsequently, the core control module triggers monitoring tasks according to the timed instructions from the real-time clock module, monitoring physical parameters every 20 minutes and tea polyphenol-related parameters every 30 minutes. If the tea garden management platform issues an immediate monitoring instruction, such as a tea quality inspection or pest and disease investigation, it will respond first. After receiving the instruction, the multispectral coupled sensing module simultaneously collects tea garden environmental and soil data through temperature, humidity, light, and soil conductivity sensors; the tea polyphenol-related sensor collects tea quality-related parameters; the tea geometrid moth early warning sensor identifies early pest characteristics by detecting signals from tea leaves; and the multispectral sensing unit collects reflected light signals at wavelengths of 520nm, 680nm, and 940nm, linking them to tea growth and quality. All raw data is transmitted to the core control module after signal conditioning and ADC conversion.

[0063] During data acquisition, the layered anti-interference packaging module works continuously: the dust removal submodule monitors the light intensity of the multispectral sensor in real time. If the light intensity decays to 62% of the initial value, the core control module triggers the stepper motor-screw mechanism to drive the 3mm quartz glass dustproof plate to reciprocate 9mm, removing surface fog residue and fallen leaf dust; the electromagnetic shielding submodule uses a stainless steel shell and absorbing materials to suppress electromagnetic interference generated by the tea garden tea picking machine, ensuring signal stability; the temperature buffer submodule maintains the probe temperature stability, providing accurate temperature data for self-calibration, while the waterproof and fog-proof shell structure resists the corrosion of the high humidity and foggy environment of the tea garden.

[0064] The core control module preprocesses data through an edge computing unit, integrating environmental parameters, soil data, quality parameters, pest and disease early warning data, and multispectral data to remove outliers. It then triggers a neural network self-calibration module, combining temperature data, tea polyphenol correlations, and tea geometrid moth early warning parameters to perform real-time calibration via a BP neural network, correcting errors caused by high humidity and fog. The calibrated data is then compressed and uploaded to the tea garden management platform via a LoRa communication module. If the LoRa signal is weakly affected by tea tree shading, it automatically switches to scattering communication as backup. Simultaneously, the power monitoring unit provides real-time power consumption data, and the core control module adaptively allocates power, reducing power consumption of non-core modules during rainy and foggy periods to ensure long-lasting lithium battery operation. The system cyclically executes the above process. The core control module performs daily fault self-checks; if it detects sensor anomalies, communication interruptions, or pest and disease early warning signals, it immediately sends an alarm signal, ensuring accurate ecological monitoring and quality and pest and disease control requirements for the tea garden.

[0065] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An Internet of Things (IoT) electronic hardware system for agricultural environmental monitoring, characterized in that, It includes a core control module, a multispectral coupling sensing module, a layered anti-interference packaging module, a dual-mode communication module, an adaptive power supply module, and a neural network self-calibration module; The core control module is electrically connected to each of the other modules to realize data interaction and logic control. The layered anti-interference packaging module is wrapped around the multispectral coupling sensing module. The dual-mode communication module, adaptive power supply module, and neural network self-calibration module are all integrated with the core control module through a PCB board. The system collects multi-dimensional environmental and crop-related parameters through a multispectral coupled sensing module. After hierarchical anti-interference processing, preprocessing by the core control module, and real-time correction by the neural network self-calibration module, the data is adaptively uploaded by the dual-mode communication module. The adaptive power supply module provides long-term and stable power supply for the entire system, enabling accurate, low-power, and long-term monitoring in complex field environments.

2. The system according to claim 1, characterized in that, The multispectral coupled sensing module includes a physical parameter sensing unit, a trace gas sensing unit, and a multispectral crop physiological correlation sensing unit. The physical parameter sensing unit uses temperature, humidity, light, and soil conductivity sensors with gold-plated contacts, and is connected to the core control module via an I2C bus. The trace gas sensing unit adopts a resistive gas sensor array based on metal porphyrin-modified polypyrrole, and outputs signals in conjunction with a signal conditioning circuit and an ADC conversion module. The core control module achieves selective identification of target gases through a built-in SVM model, which is used to solve the cross-sensitivity problem of traditional gas sensors and improve the detection accuracy of trace gases.

3. The system according to claim 1, characterized in that, The layered anti-interference packaging module includes a dustproof and dust removal submodule, an electromagnetic shielding submodule, and a temperature buffer submodule. The dust prevention and dust removal submodule is equipped with a high-transmittance quartz glass dustproof plate and an ultra-fine fiber dust removal cloth. The dustproof plate is driven to reciprocate through a stepper motor-screw mechanism. The core control module triggers the dust removal action according to the light intensity attenuation threshold of the multispectral sensor to avoid dust from obstructing the sensing accuracy. The electromagnetic shielding submodule adopts a multi-layer structure of aluminum alloy shell + conductive polymer shielding film + wave absorbing material, combined with shielded twisted pair cable and common mode choke, to suppress electromagnetic interference in the field; The temperature buffer submodule constructs a buffer cavity using thermal insulation cotton and incorporates a miniature thermistor to collect temperature data, providing a basis for subsequent temperature drift compensation.

4. The system according to claim 1, characterized in that, The dual-mode communication module includes a LoRa communication unit, a scattering communication unit, and a link quality detection unit, all of which are connected to the core control module via an SPI bus. The link quality detection unit monitors the RSSI value and bit error rate of the two communication modes in real time. When the LoRa communication signal strength is ≥-85dBm and the bit error rate is ≤1%, the core control module controls the system to prioritize LoRa mode communication. Otherwise, it automatically switches to scatter communication mode to balance communication coverage in remote areas and system power consumption, and avoid data transmission interruption.

5. The system according to claim 1, characterized in that, The neural network self-calibration module includes a calibration reference source, an error detection unit, and a neural network operation unit. The calibration reference source has built-in high-precision voltage and temperature references to provide calibration reference signals for each sensing unit. The error detection unit acquires the sensor output signal and the reference signal in real time and calculates the error value. The neural network operation unit has a built-in BP neural network model. It takes the sensor output value and temperature data as input and the calibrated value as output. It adaptively adjusts the calibration frequency according to the environmental stability to correct errors caused by temperature drift and device aging in real time, ensuring long-term monitoring accuracy.

6. The system according to claim 1, characterized in that, The adaptive power supply module includes a solar panel, a lithium battery pack, a power management chip, and a power monitoring unit. The power management chip enables charge and discharge management and overcharge, over-discharge, and over-temperature protection. The power monitoring unit collects power consumption data of each module in real time and transmits it to the core control module. The core control module adaptively allocates power according to the solar power supply and power consumption data. When the lithium battery voltage is lower than 3.0V, the control system enters a low-power sleep mode to achieve long-term battery life and adapt to remote areas without external power supply scenarios.

7. The system according to claim 2, characterized in that, The multispectral crop physiology correlation sensing unit uses near-infrared multispectral sensors with center wavelengths of 730nm, 850nm, and 940nm. It focuses the light reflected from crop leaves through an optical lens and converts it into electrical signals. It adopts a pulsed working mode to indirectly monitor crop water stress and nutrient deficiency, so that the monitoring data can be directly correlated with crop growth needs, thus making up for the limitations of traditional physical parameter monitoring.

8. The system according to claim 3, characterized in that, The electromagnetic shielding submodule has an aluminum alloy shell with a thickness of 1.5mm and a conductive polymer shielding film with a thickness of 0.2mm. Ferrite particle absorbing material is filled between the shell and the shielding film. The PCB boards of the core control module and the dual-mode communication module adopt a grounding optimization design, so that the system's electromagnetic interference suppression ratio is ≥80dB, which is used to ensure the stable operation of the system around agricultural machinery and power lines.

9. The system according to claim 1, characterized in that, The core control module adopts a low-power ARM Cortex-M4 core MCU, integrates an edge computing unit and a real-time clock module, and has a built-in Flash storage unit with a capacity of not less than 1MB. The edge computing unit is used to filter, normalize and fuse multi-sensor data, the real-time clock module is used to trigger timed monitoring and calibration tasks, and the Flash storage unit is used to store calibration parameters, cache monitoring data and configure communication protocols, providing core support for the logical control and data processing of the entire system.

10. The system according to claim 2, characterized in that, The trace gas sensing unit's sensor array is compatible with three target gases: ammonia, hydrogen sulfide, and carbon dioxide, with a detection limit ≤0.33 mg / m³. 3 The signal conditioning circuit includes a differential amplifier and filter module, which converts the sensor resistance change signal into a 0-3.3V analog signal. After conversion by the 16-bit ADC module, the signal is transmitted to the core control module. The core control module uses an SVM model to achieve a target gas identification accuracy of ≥97%, which is used to meet the needs of accurate monitoring of trace harmful gases in the field.

Images