A wheat canopy nitrogen content in-situ monitoring device and method based on multispectral technology

Abstract

This invention discloses an in-situ monitoring device and method for wheat canopy nitrogen content based on multispectral technology, belonging to the field of agricultural intelligent monitoring technology. It includes a multispectral acquisition module, an intelligent power management module, a data processing and communication module, a cloud service platform, and a mobile terminal APP. This invention collects three-band spectral data of the wheat canopy, calculates vegetation indices after preprocessing, establishes a predictive model for nitrogen content, and combines it with an intelligent device to achieve in-situ real-time monitoring. This enables real-time and accurate monitoring of wheat nitrogen content while reducing power consumption. By combining multispectral technology with intelligent algorithms, it achieves in-situ real-time monitoring of wheat nitrogen content, improving monitoring stability and accuracy. The mobile APP and cloud service platform provide users with convenient data viewing and management tools, providing strong support for refined agricultural management and contributing to the promotion of refined management in agricultural production.

Description

Technical Field

[0001] This invention relates to the field of agricultural intelligent monitoring technology, specifically to an in-situ monitoring device and method for nitrogen content in wheat canopy based on multispectral technology. Background Technology

[0002] Agricultural intelligent monitoring integrates IoT, big data, and AI technologies to achieve real-time, precise monitoring and intelligent decision-making of the farmland environment, significantly improving agricultural production efficiency and sustainability. Nitrogen is a key nutrient element for wheat growth and development, and wheat nitrogen content directly affects yield and quality, involving nitrogen fertilizer application, varietal differences, and environmental interactions. Multispectral technology, with its high spectral resolution, non-destructive nature, and speed, has become an important means of detecting wheat nitrogen content, providing technical support for precision fertilization, yield prediction, and quality control. Traditional methods for detecting nitrogen content in wheat mainly rely on laboratory chemical analysis, which is cumbersome, time-consuming, and costly, and cannot achieve real-time, large-area dynamic monitoring. Existing field monitoring equipment often suffers from high power consumption, poor stability, and weak adaptability, making it difficult to meet the needs of modern agricultural precision management. Therefore, there is an urgent need for an intelligent device and method that can monitor the nitrogen content in the wheat canopy in situ, in real time, and accurately. Summary of the Invention

[0003] This invention provides an in-situ monitoring device and method for wheat canopy nitrogen content based on multispectral technology. It can effectively solve the problems mentioned in the background art, where traditional wheat nitrogen content detection mainly relies on laboratory chemical analysis, which is cumbersome, time-consuming and costly, and cannot achieve real-time, large-area dynamic monitoring. Existing field monitoring equipment often suffers from high power consumption, poor stability and weak adaptability, making it difficult to meet the needs of modern agricultural precision management.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an in-situ monitoring device for nitrogen content in wheat canopy based on multispectral technology, comprising a multispectral acquisition module, an intelligent power management module, a data processing and communication module, a cloud service platform, and a mobile terminal APP; The multispectral acquisition module is used to acquire raw three-band spectral data of the wheat canopy, the intelligent power management module monitors the battery charging and discharging and power in real time, and the data processing and communication module preprocesses and analyzes the acquired raw data and uploads the data to the cloud server through remote wireless communication. The cloud service platform receives and stores monitoring data, runs a nitrogen content prediction model, and the mobile terminal APP displays the wheat canopy vegetation index and leaf nitrogen content prediction values ​​in real time.

[0005] According to the above technical solution, the multispectral acquisition module uses three specific wavelength bands of light source and sensor. When selecting the wavelength of the light source, a customized high-energy light-emitting diode is used as the light source. The detection wavelength was screened twice by integrating sphere. When the light source signal is highly modulated, the wheat canopy monitoring scenario is in the field, which is greatly affected. Pulse width modulation technology is used, and photomultiplier tubes are used as detectors to capture the canopy reflection signal. By designing a high-frequency electronic switch, the circuit can start and stop the LED at high frequency, generating a pulsed light signal, which causes the photovoltaic cell to generate a square wave response signal with the same modulation frequency. When selecting a condenser lens, an integrated recessed three-band six-source sensor optoelectronic system is used, Fresnel lenses are selected for collimation and focusing, and a lower cutoff extinction filter is added to the outer layer for reflected signal acquisition.

[0006] According to the above technical solution, the intelligent power management module includes a lithium battery, a charging management chip, and a power monitoring circuit; Employing intelligent power gating management technology, it monitors battery charging and discharging and power level in real time and dynamically adjusts the power supply strategy. The casing adopts an IP67 protection rating design and has built-in temperature and humidity sensors to monitor the device's operating environment in real time, triggering a protection mechanism in case of abnormalities.

[0007] According to the above technical solution, the data processing and communication module adopts an ARM processor and a 5G wireless communication module. During the processing, it integrates algorithms such as fractional derivative and continuous wavelet transform to reduce noise and enhance the original spectral data, and calculates key vegetation parameters such as normalized vegetation index and ratio vegetation index. During the communication process, wireless communication technology is used, combined with AES encryption and MD5 hash algorithm to perform dual authentication encryption on the communication data. Advanced device management technology is used to realize automatic device registration, flexible parameter configuration and upgrade functions, and complete the design of the communication module. It adopts the lightweight Modbus protocol and is compatible with IoT communication standards such as MQTT and CoAP.

[0008] According to the above technical solution, the cloud service platform adopts a distributed architecture design, supports simultaneous access by multiple terminals, runs a nitrogen content prediction model, and performs cross-validation by combining multi-view image data from UAVs. The cloud service platform also provides a remote data access interface, supporting concurrent access from multiple terminals; The mobile terminal APP is developed based on Java language and supports Android system. It is a cloud service-based mobile application for in-situ monitoring of wheat canopy. It presents the predicted values ​​of wheat canopy vegetation index and leaf nitrogen content. Users can view it anytime and anywhere through their mobile phones. When presented based on the visualization interface, a 3D GIS map is integrated with a real-time heat map of the monitoring points, and the changing trends of NDVI and RVI indices are displayed through dynamic line graphs. The cloud server adopts a distributed network architecture. In addition, it can provide timely fertilization advice to wheat growers based on monitoring results, recommend fertilization based on nitrogen content prediction, support nitrogen, phosphorus and potassium ratio calculation, and trigger SMS or APP push when nitrogen content deviates from the threshold. It also supports historical data query and trend analysis to assist agricultural production management.

[0009] A method for in-situ monitoring of nitrogen content in wheat canopy based on multispectral technology includes the following steps: Step 1: Data Acquisition and Preprocessing; Step 2, vegetation index calculation; Step 3: Model building and optimization; Step 4: Device parameter adjustment; Step 5: Real-time monitoring and data presentation.

[0010] According to the above technical solution, in step one, three-band raw data of wheat samples are collected, noise reduction preprocessing is performed through an algorithm, and analog-to-digital conversion technology is designed with parallel dual-channel and three-stage pipeline structure to sample the input wheat canopy reflection signal; In the design of the data acquisition circuit, digital signal processing is performed using a high-speed comparator, a delay-locked loop clock source, and an operational amplifier circuit to obtain the wheat canopy reflectance spectrum signal. The schematic diagram of the data acquisition and processing circuit is designed using Protel 99 SE, and the circuit board of the data acquisition module is developed. During the actual data acquisition process, the multispectral acquisition module is activated to periodically collect canopy reflectance spectra. The processing includes fractional-order differentiation of the original spectrum, combined with continuous wavelet transform to eliminate baseline drift and background interference.

[0011] According to the above technical solution, in step two, based on the preprocessed spectral data, the normalized vegetation index and the ratio vegetation index are calculated to reflect the canopy chlorophyll content and coverage, and a comprehensive feature vector is constructed by combining other auxiliary parameters. In step three, a mathematical model of vegetation index and laboratory nitrogen content is established, the model is optimized by combining deep learning algorithm, and cross-validation is performed using multi-view image data. When establishing the initial model, a large number of raw three-band data of wheat samples were collected. The raw spectra were first processed by fractional derivative, followed by noise reduction and preprocessing algorithms using continuous wavelet transform. The normalized vegetation index and ratio vegetation index of the three-band reflectance combination were calculated. The nitrogen content of wheat leaves was determined by laboratory chemical analysis. A mathematical model of vegetation index and nitrogen content was established. The quantitative relationship between wheat canopy reflectance spectrum and leaf nitrogen content was fitted to the experimental data. Finally, a suitable vegetation index was selected to construct the prediction model. Furthermore, using drones equipped with hyperspectral cameras, images of the canopy of different wheat varieties were captured to obtain multi-view image data. Deep learning algorithms were then used to automatically extract and classify the images to predict the nitrogen content of wheat. The multi-view data from the UAV is compared with the ground-based measured data to cross-validate the model's prediction results.

[0012] According to the above technical solution, in step four, the monitoring device is tested and its technical parameters are adjusted under different environmental conditions; Field verification was conducted in wheat-growing areas. The monitoring device was tested under different environmental conditions for a long period of time to adjust technical parameters, analyze the fluctuation of the output data of the monitoring station, identify the factors affecting stability, and optimize it. Field trials were conducted on different wheat varieties. Field trials were conducted under different weather conditions to record the effects of light source intensity, luminous efficiency, beam detection angle, integration time, and temperature and humidity changes on the data. At the same time, technical parameters were adjusted to optimize the signal-to-noise ratio and dynamic range. Furthermore, during the optimization of device stability, the stability of the device was evaluated through repeated measurements. The Kalman filter algorithm was used to suppress random errors, and calibration was performed for different wheat varieties to establish a variety-specific compensation model.

[0013] According to the above technical solution, in step five, data is collected in situ by the device, analyzed by the cloud platform, and the monitoring results are displayed in real time on the mobile terminal APP. The collected data is uploaded to the cloud platform via 5G network, where the optimized prediction model is run in real time to calculate the predicted value of nitrogen content and accelerate big data processing using a distributed computing framework. The mobile app displays vegetation index, nitrogen content prediction, and fertilization suggestions in real time in the form of charts, and also has a threshold warning function.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention collects three-band spectral data of wheat canopy, calculates vegetation indices after preprocessing, establishes a predictive model for nitrogen content, and combines it with intelligent devices to achieve in-situ real-time monitoring. This enables real-time and accurate monitoring of wheat nitrogen content. By combining multispectral technology with intelligent algorithms, in-situ real-time monitoring of wheat nitrogen content is achieved, reducing power consumption and improving monitoring stability and accuracy. At the same time, a mobile APP and cloud service platform provide users with convenient data viewing and management tools, providing strong support for precision agricultural management and helping to promote the precision management of agricultural production. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0016] In the attached diagram: Figure 1 This is a schematic diagram of the monitoring device of the present invention; Figure 2 This is a flowchart of the monitoring method of the present invention; Figure 3 This is a schematic diagram of the internally recessed three-band six-source structure of the present invention; Figure 4 This is a schematic diagram of the remote wireless communication of the present invention; Figure 5 This is a monitoring result diagram of the mobile terminal APP of this invention. Detailed Implementation

[0017] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0018] Example: Figure 1 As shown, the present invention provides a technical solution, an in-situ monitoring device for nitrogen content in wheat canopy based on multispectral technology, which realizes real-time, accurate and low-power monitoring of nitrogen content in wheat, including a multispectral acquisition module, an intelligent power management module, a data processing and communication module, a cloud service platform and a mobile terminal APP; The multispectral acquisition module is used to collect raw three-band spectral data of the wheat canopy. The intelligent power management module adopts intelligent power gating management technology to monitor the battery charging and discharging and power in real time. The data processing and communication module preprocesses and analyzes the collected raw data and uploads the data to the cloud server through remote wireless communication. The cloud service platform receives and stores monitoring data, runs nitrogen content prediction models, provides data support for mobile terminals, and the mobile terminal APP displays wheat canopy vegetation index and leaf nitrogen content prediction values ​​in real time, providing monitoring suggestions to users. Monitoring devices are installed in the field at different growth stages of wheat. The collection angle and height are adjusted. After the device is started, the multispectral acquisition module collects canopy spectral data at regular intervals, the intelligent power management module dynamically manages the power supply, and the data processing module performs noise reduction preprocessing on the raw data, calculates the vegetation indices NDVI and RVI, and uploads the data to the cloud service platform. The cloud service platform calls the optimized prediction model to calculate the predicted value of nitrogen content and feeds it back to the mobile terminal APP. Users can view real-time data and management suggestions through the APP.

[0019] like Figure 3 As shown, based on the above technical solution, the multispectral acquisition module uses three specific wavelength bands of light source and sensor. The light source intensity is adjustable. When selecting the light source wavelength, customized high-energy light-emitting diodes (LEDs) with center wavelengths of 680nm, 720nm, and 760nm are used as the light source to achieve all-weather detection. The wavelength is further screened and detected by integrating sphere to ensure that the center wavelength of the LEDs of the same wavelength arranged in a ring is consistent, thereby improving the accuracy of the characteristic signal of the emitted light source from the source. When the light source signal is modulated at high frequency, the wheat canopy monitoring scenario is in the field, which is greatly affected by factors such as sunlight and wind speed. Pulse width modulation (PWM) technology is used, along with a high-sensitivity, high-resolution photomultiplier tube as a detector to capture the canopy reflection signal. By designing a high-frequency electronic switch, the circuit can start and stop the LED at high frequency to generate a pulsed light signal, which causes the photovoltaic cell to generate a square wave response signal with the same modulation frequency, thereby increasing the data acquisition frequency. By optimizing the integration time, the amount of data acquired within a certain period of time can be adjusted, reducing the influence of external interference factors. When selecting a condenser lens, the current crop canopy diagnostic photosensitive lens uses a fisheye lens. A fisheye lens is an extreme lens with a short focal length, a wide angle of view, and thick lenses. The light passing through it is easily distorted and has a large light loss. The integrated recessed three-band six-source sensor photoelectric system uses Fresnel lenses to collimate and focus the light path, which improves the light intensity and uniformity, while reducing the loss caused by light scattering. The outer layer of the reflected signal acquisition layer is equipped with a lower cutoff extinction filter to eliminate interference from sunlight with wavelengths of 670nm and below. Design an efficient and stable light source driving circuit to ensure the stability and consistency of the light source output intensity.

[0020] Based on the above technical solution, the intelligent power management module includes a lithium battery, a charging management chip, and a power monitoring circuit. The lithium battery provides power, and the charging management chip and power monitoring circuit are combined to support dynamic charging and discharging management. At the same time, it uses intelligent power gating management technology to monitor the charging and discharging of the battery and the power level in real time, and dynamically adjusts the power supply strategy to optimize energy consumption, extend the battery life, dynamically grasp the power status, ensure that the device can work continuously, and extend the battery life. The casing is designed with an IP67 protection rating, providing waterproof, dustproof, and electromagnetic interference-resistant protection. It is suitable for harsh field environments and easy to install and replace in farmland. The built-in temperature and humidity sensor monitors the equipment's operating environment in real time and triggers a protection mechanism in case of abnormalities.

[0021] like Figure 4 As shown, based on the above technical solution, the data processing and communication module adopts an ARM processor and a 5G wireless communication module. During the processing, algorithms such as fractional derivative and continuous wavelet transform are integrated to reduce noise and enhance the original spectral data, and to calculate the key vegetation parameters of the normalized vegetation index NDVI and the ratio vegetation index RVI. During communication, Bluetooth Low Energy short-range wireless communication technology is used, combined with AES encryption and MD5 hash algorithm to perform dual authentication encryption on communication data to prevent tampering and ensure data security. Advanced device management technology is used to efficiently realize automatic device registration, flexible parameter configuration and convenient upgrade functions to complete the design of the communication module. It adopts the lightweight Modbus protocol and is compatible with IoT communication standards such as MQTT and CoAP, which facilitates flexible integration and communication between different devices, enables interconnection between devices, and facilitates remote real-time transmission of monitoring data. Meanwhile, the protocol has good cross-platform compatibility, can be implemented on various operating systems and devices, which is conducive to the networking and interoperability of wheat nitrogen monitoring devices, facilitates seamless connection between monitoring devices and various cloud platforms, and is compatible with multi-platform cloud service access.

[0022] like Figure 5 As shown, based on the above technical solution, the cloud service platform adopts a distributed architecture design, supports simultaneous access from multiple terminals, supports the storage and efficient processing of massive monitoring data, runs a nitrogen content prediction model, the prediction model is a linear regression model, and is cross-validated by combining UAV multi-view image data. Meanwhile, the cloud service platform also provides a remote data access interface, supporting concurrent access from multiple terminals, enabling dynamic optimization and updating of model parameters, and improving prediction accuracy; The mobile terminal APP is developed based on the Java language and supports the Android system. It is developed using Java to complete the Android mobile phone program and develop a cloud service-based mobile phone program for in-situ monitoring of wheat canopy. It presents the predicted values ​​of wheat canopy vegetation index and leaf nitrogen content. Users can view it anytime and anywhere through their mobile phones. It provides a visual interface. When presenting the data through the visual interface, a 3D GIS map is integrated with a real-time heat map of the monitoring points, and the changing trends of NDVI and RVI indices are displayed through dynamic line graphs to intuitively present the health status of wheat canopy. The cloud server adopts a distributed network architecture, which improves the scalability and capacity of the network, enabling multiple mobile terminal devices to use the remote server online simultaneously. At the same time, it can also provide timely fertilization suggestions to wheat growers based on monitoring results and push them to farmers' mobile devices. It recommends fertilization based on the predicted nitrogen content, supports the calculation of nitrogen, phosphorus and potassium ratio, and triggers APP push when the nitrogen content deviates from the threshold. In addition, it supports historical data query and trend analysis to assist agricultural production management, and to conduct application and demonstration of monitoring devices in wheat planting bases to verify the performance and practicality of the devices.

[0023] like Figure 2 As shown, a method for in-situ monitoring of nitrogen content in wheat canopy based on multispectral technology includes the following steps: Step 1: Data Acquisition and Preprocessing; Step 2, vegetation index calculation; Step 3: Model building and optimization; Step 4: Device parameter adjustment; Step 5: Real-time monitoring and data presentation.

[0024] Based on the above technical solution, in step one, three-band raw data of wheat samples are collected and noise reduction preprocessing is performed through fractional derivative and continuous wavelet transform algorithms. Data acquisition and processing are the foundation and key to the reliability of raw data. Parallel dual-channel and three-stage pipeline structure design analog-to-digital conversion technology is adopted to perform fast, accurate and multi-round sampling of the input wheat canopy reflection signal. In the design of the data acquisition circuit, high-speed comparators, delay-locked loop clock sources, and high-gain low-noise operational amplifier circuits are used for digital signal processing to obtain high-quality wheat canopy reflectance spectrum signals. The schematic diagram of the data acquisition and processing circuit is designed using Protel 99 SE, the data acquisition module circuit board is developed, and the PCB board is fabricated and functional simulation tests are completed. During the specific collection process, at different growth stages of wheat, including the tillering stage, jointing stage, and grain-filling stage, the device was fixed in the field, and the collection height and angle were adjusted. The collection height was 1.3 meters, and the collection angle was vertically downward. The multispectral acquisition module was activated to collect the canopy reflectance spectrum at regular intervals. The processing includes fractional-order differentiation of the original spectrum to suppress high-frequency noise, and continuous wavelet transform (CWT) to eliminate baseline drift and background interference, thereby improving the signal-to-noise ratio.

[0025] Based on the above technical solution, in step two, the normalized vegetation index (NDVI) and ratio vegetation index (RVI) are calculated based on the preprocessed spectral data to reflect the canopy chlorophyll content and coverage. Normalized Difference Vegetation Index (NDVI1): NDVI1 = (ρ760 - ρ680) / (ρ760 + ρ680) ; NDVI2 = (ρ720 - ρ680) / (ρ720 + ρ680) ; Ratio vegetation indices: RVI1 = ρ760 / ρ680; RVI2 = ρ720 / ρ680; In addition, a comprehensive feature vector is constructed by combining other auxiliary parameters to enhance the amount of information in the model input. Other auxiliary parameters include canopy temperature and soil moisture. Step 3: Establish a mathematical model of vegetation index and laboratory nitrogen content analysis, optimize the model by combining deep learning algorithm, and perform cross-validation through multi-view image data. When establishing the initial model, a large number of raw three-band data of wheat samples were collected. The raw spectra were first processed by fractional derivative, followed by noise reduction and preprocessing algorithms using continuous wavelet transform. The normalized vegetation index (NDVI) and ratio vegetation index (RVI) of the three-band reflectance combination were calculated. The nitrogen content of wheat leaves was determined by laboratory chemical analysis. A mathematical model of vegetation index and nitrogen content was established. The mathematical model was a linear regression model. The quantitative relationship between wheat canopy reflectance spectrum and leaf nitrogen content was fitted to the experimental data. Finally, a suitable vegetation index was selected to construct the prediction model. Furthermore, at different growth stages of wheat, drones equipped with hyperspectral cameras were used to capture images of the canopy of different wheat varieties at different heights and angles, obtaining multi-view image data. Deep learning algorithms were then used to automatically extract and classify the images to predict the nitrogen content of wheat. The model prediction results are cross-validated by comparing multi-view data from UAVs with ground-based measured data, optimizing model parameters, improving prediction accuracy, ensuring model consistency, and evaluating model robustness through K-fold cross-validation. Hyperparameters, including learning rate and number of convolutional kernels, are then adjusted.

[0026] Based on the above technical solution, in step four, the monitoring device is tested under different environmental conditions, the technical parameters are adjusted, and the device stability is optimized. The technical parameters include light source intensity, luminous efficiency, beam detection angle, and integration time. Field verification was conducted in wheat-growing areas. The monitoring device was tested under different environmental conditions for a long period of time to evaluate its stability and reliability. Technical parameters were adjusted, fluctuations in the output data of the monitoring station were analyzed, factors affecting stability were identified and optimized, and field trials were conducted on different wheat varieties to study the adaptability and versatility of the monitoring device. Specifically, field trials were conducted under different weather conditions, including sunny, cloudy, and post-rain, to record the effects of light source intensity, luminous efficiency, beam detection angle, integration time, and temperature and humidity changes on the data. At the same time, technical parameters were adjusted to optimize the signal-to-noise ratio and dynamic range. Furthermore, during the optimization of device stability, the device stability was evaluated through repeated measurements, with the number of repeated measurements being ≥30. The Kalman filter algorithm was used to suppress random errors, and calibration was performed for different wheat varieties to establish a variety-specific compensation model. The wheat varieties included winter wheat and spring wheat.

[0027] Based on the above technical solution, in step five, data is collected in situ by the device, analyzed by the cloud platform, and the monitoring results are displayed in real time on the mobile terminal APP. The collected data is uploaded to the cloud platform via a 5G network, where the optimized prediction model is run in real time to calculate the predicted value of nitrogen content. The distributed computing framework, Spark, is used to accelerate big data processing and ensure millisecond-level response. The mobile app displays vegetation index, nitrogen content prediction, and fertilization suggestions in real time in the form of charts. It also has a threshold warning function that triggers an alarm when the nitrogen content is below the threshold, guiding farmers to take timely and proactive intervention.

[0028] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An in-situ monitoring device for nitrogen content in wheat canopy based on multispectral technology, characterized in that: It includes a multispectral acquisition module, an intelligent power management module, a data processing and communication module, a cloud service platform, and a mobile terminal APP; The multispectral acquisition module is used to acquire raw three-band spectral data of the wheat canopy, the intelligent power management module monitors the battery charging and discharging and power in real time, and the data processing and communication module preprocesses and analyzes the acquired raw data and uploads the data to the cloud server through remote wireless communication. The cloud service platform receives and stores monitoring data, runs a nitrogen content prediction model, and the mobile terminal APP displays the wheat canopy vegetation index and leaf nitrogen content prediction values ​​in real time.

2. The in-situ monitoring device for wheat canopy nitrogen content based on multispectral technology according to claim 1, characterized in that: The multispectral acquisition module uses three specific wavelength bands of light source and sensor. When selecting the light source wavelength, a customized high-energy light-emitting diode is used as the light source. The detection wavelength was screened twice by integrating sphere. When the light source signal is highly modulated, the wheat canopy monitoring scenario is in the field, which is greatly affected. Therefore, pulse width modulation technology is used, and photomultiplier tubes are used as detectors to capture the canopy reflection signal. By designing a high-frequency electronic switch, the circuit can start and stop the LED at high frequency, generating a pulsed light signal, which causes the photovoltaic cell to generate a square wave response signal with the same modulation frequency. When selecting a condenser lens, an integrated recessed three-band six-source sensor optoelectronic system is used, Fresnel lenses are selected for collimation and focusing, and a lower cutoff extinction filter is added to the outer layer for reflected signal acquisition.

3. The in-situ monitoring device for wheat canopy nitrogen content based on multispectral technology according to claim 1, characterized in that: The intelligent power management module includes a lithium battery, a charging management chip, and a power monitoring circuit. Employing intelligent power gating management technology, it monitors battery charging and discharging and power level in real time and dynamically adjusts the power supply strategy. The casing adopts an IP67 protection rating design and has built-in temperature and humidity sensors to monitor the device's operating environment in real time, triggering a protection mechanism in case of abnormalities.

4. The in-situ monitoring device for wheat canopy nitrogen content based on multispectral technology according to claim 1, characterized in that: The data processing and communication module adopts an ARM processor and a 5G wireless communication module. During the processing, it integrates algorithms such as fractional derivative and continuous wavelet transform to reduce noise and enhance the original spectral data, and calculates key vegetation parameters such as normalized vegetation index and ratio vegetation index. During the communication process, wireless communication technology is used, combined with AES encryption and MD5 hash algorithm to perform dual authentication encryption on the communication data. Advanced device management technology is used to realize automatic device registration, flexible parameter configuration and upgrade functions, and complete the design of the communication module. It adopts the lightweight Modbus protocol and is compatible with IoT communication standards such as MQTT and CoAP.

5. The in-situ monitoring device for wheat canopy nitrogen content based on multispectral technology according to claim 1, characterized in that: The cloud service platform adopts a distributed architecture design, supports simultaneous access from multiple terminals, runs a nitrogen content prediction model, and performs cross-validation by combining multi-view image data from UAVs. The cloud service platform also provides a remote data access interface, supporting concurrent access from multiple terminals; The mobile terminal APP is developed based on Java language and supports Android system. It is a cloud service-based mobile application for in-situ monitoring of wheat canopy. It presents the predicted values ​​of wheat canopy vegetation index and leaf nitrogen content. Users can view it anytime and anywhere through their mobile phones. When presented based on the visualization interface, a 3D GIS map is integrated with a real-time heat map of the monitoring points, and the changing trends of NDVI and RVI indices are displayed through dynamic line graphs. The cloud server adopts a distributed network architecture. In addition, it can provide timely fertilization advice to wheat growers based on monitoring results, recommend fertilization based on nitrogen content prediction, support nitrogen, phosphorus and potassium ratio calculation, and trigger SMS or APP push when nitrogen content deviates from the threshold. It also supports historical data query and trend analysis to assist agricultural production management.

6. A method for in-situ monitoring of nitrogen content in wheat canopy according to the device described in claim 1, characterized in that: Includes the following steps: Step 1: Data Acquisition and Preprocessing; Step 2, vegetation index calculation; Step 3: Model building and optimization; Step 4: Device parameter adjustment; Step 5: Real-time monitoring and data presentation.

7. The method for in-situ monitoring of wheat canopy nitrogen content based on multispectral technology according to claim 6, characterized in that: In step one, raw data of wheat samples in three bands are collected, and noise reduction preprocessing is performed through an algorithm. A parallel dual-channel and three-stage pipeline structure is used to design analog-to-digital conversion technology to sample the input wheat canopy reflection signal. In the design of the data acquisition circuit, digital signal processing is performed using a high-speed comparator, a delay-locked loop clock source, and an operational amplifier circuit to obtain the wheat canopy reflectance spectrum signal. The schematic diagram of the data acquisition and processing circuit is designed using Protel 99 SE, and the data acquisition module circuit board is developed. During the actual data acquisition process, the multispectral acquisition module is activated to periodically collect canopy reflectance spectra. The processing includes fractional-order differentiation of the original spectrum, combined with continuous wavelet transform to eliminate baseline drift and background interference.

8. The method for in-situ monitoring of wheat canopy nitrogen content based on multispectral technology according to claim 6, characterized in that: In step two, based on the preprocessed spectral data, the normalized vegetation index and the ratio vegetation index are calculated to reflect the canopy chlorophyll content and coverage, and a comprehensive feature vector is constructed by combining other auxiliary parameters. In step three, a mathematical model of vegetation index and laboratory nitrogen content is established, the model is optimized by combining deep learning algorithm, and cross-validation is performed using multi-view image data. When establishing the initial model, a large number of wheat samples were collected with three-band raw data. The raw spectra were first processed by fractional derivative, followed by noise reduction and preprocessing algorithms using continuous wavelet transform. The normalized vegetation index and ratio vegetation index of the three-band reflectance combination were calculated. The nitrogen content of wheat leaves was determined by laboratory chemical analysis. A mathematical model of vegetation index and nitrogen content was established. The quantitative relationship between wheat canopy reflectance spectrum and leaf nitrogen content was fitted to the experimental data. Finally, a suitable vegetation index was selected to construct the prediction model. Furthermore, using drones equipped with hyperspectral cameras, images of the canopy of different wheat varieties were captured to obtain multi-view image data. Deep learning algorithms were then used to automatically extract and classify the images to predict the nitrogen content of wheat. The multi-view data from the UAV is compared with the ground-based measured data to cross-validate the model's prediction results.

9. The method for in-situ monitoring of wheat canopy nitrogen content based on multispectral technology according to claim 6, characterized in that: In step four, the monitoring device is tested and its technical parameters are adjusted under different environmental conditions. Field verification was conducted in wheat-growing areas. The monitoring device was tested under different environmental conditions for a long period of time to adjust technical parameters, analyze the fluctuation of the output data of the monitoring station, identify the factors affecting stability, and optimize it. Field trials were conducted on different wheat varieties. Field trials were conducted under different weather conditions to record the effects of light source intensity, luminous efficiency, beam detection angle, integration time, and temperature and humidity changes on the data. At the same time, technical parameters were adjusted to optimize the signal-to-noise ratio and dynamic range. Furthermore, during the optimization of device stability, the stability of the device was evaluated through repeated measurements. The Kalman filter algorithm was used to suppress random errors, and calibration was performed for different wheat varieties to establish a variety-specific compensation model.

10. The method for in-situ monitoring of nitrogen content in wheat canopy based on multispectral technology according to claim 6, characterized in that: In step five, data is collected in situ by the device, analyzed by the cloud platform, and the monitoring results are displayed in real time on the mobile terminal APP. The collected data is uploaded to the cloud platform via 5G network, where the optimized prediction model is run in real time to calculate the predicted value of nitrogen content and accelerate big data processing using a distributed computing framework. The mobile app displays vegetation index, nitrogen content prediction, and fertilization suggestions in real time in the form of charts, and also has a threshold warning function.

Images