A system and method for remote monitoring of plant hydrogen peroxide

Abstract

This invention discloses a remote monitoring system and method for hydrogen peroxide in plants. The system includes: an electrochemical microsensor, the working electrode of which is a micron-sized metal wire modified with gold nanoparticles and graphene oxide; a micro-electrochemical device connected to the sensor; a smartphone communicating with the device for controlling detection and data transmission; and a data management platform communicating with the smartphone. The method involves: inserting the sensor into a leaf, initiating detection via the smartphone to acquire dynamic hydrogen peroxide data, and uploading it to the platform; the platform extracts multidimensional waveform features from the data and inputs these features into a pre-trained machine learning model, outputting a plant stress state discrimination result. This invention achieves remote, continuous, and in vivo monitoring of hydrogen peroxide in plants, and through deep adaptation of electrode materials and machine learning algorithms, enables intelligent and accurate discrimination of plant stress states.

Description

Technical Field

[0001] This invention relates to the field of microelectrode electrochemical detection of plant signal molecules, specifically to a remote monitoring system and method for plant hydrogen peroxide. Background Technology

[0002] Smart agriculture is a new form of agricultural development that utilizes intelligent sensor networks applied to various agricultural activities. This network primarily includes cutting-edge information technologies such as internet big data collection platforms, GIS technology, drone technology, robotics, and intelligent sensors. Intelligent sensors are a key component of these networks, capable of sensing and collecting firsthand information about crop growth. Currently, intelligent sensors are mainly used for detecting various physical quantities in agricultural operations; for example, temperature and humidity sensors can be used to detect temperature and humidity in the air and soil. While there are numerous reports on the application of intelligent sensors in regulating external plant factors, such as various environmental factors and chemical agents, their application in detecting signaling molecules within plants is relatively limited. H2O2, as an important plant signaling molecule, plays a role in plant adaptive responses, making the development of intelligent plant sensors for detecting H2O2 crucial.

[0003] Commonly used methods for detecting H2O2 in plants include colorimetric / spectrophotometric methods, DAB staining methods, fluorescent probe methods, and electrochemical methods. However, these methods have several drawbacks: First, existing H2O2 detection methods (such as colorimetric methods, DAB staining methods, and fluorescent probe methods) require destructive sampling of plants, which is cumbersome, time-consuming, and makes in-situ, continuous monitoring difficult. Second, existing detection systems are mostly limited to data acquisition and lack the ability to deeply interpret complex electrochemical signals. The H2O2 signal generated under plant salt stress is highly dynamic and complex. In traditional methods, researchers often rely on manual interpretation (such as manually measuring peak height and peak width), which is not only time-consuming and labor-intensive but also highly subjective and difficult to guarantee the consistency and accuracy of the interpretation results. Even with the introduction of general machine learning algorithms, it is difficult to deeply adapt to the waveform characteristics generated by specific electrodes, resulting in limited discrimination accuracy. Therefore, a detection method that can accurately and timely reflect the H2O2 content in plants is needed, ideally one that can dynamically track H2O2, a key plant signaling molecule that affects plant growth and development. Summary of the Invention

[0004] Purpose of the invention: This invention provides a remote monitoring system and method for hydrogen peroxide in plants, solving the technical problems faced in the detection of signal molecules and the determination of stress status in plants in the prior art.

[0005] Technical solution: The remote monitoring system for plant hydrogen peroxide of the present invention includes: An electrochemical microsensor, whose working electrode is a micron-sized metal wire modified with gold nanoparticles and graphene oxide, is used to insert into plant leaves and generate an electrochemical signal characterizing the concentration of hydrogen peroxide. A micro-electrochemical device, connected to the electrochemical microsensor, is used to apply an electrical signal and acquire the electrochemical signal; A smartphone is communicatively connected to the microelectrochemical device for controlling the microelectrochemical device, receiving and processing the electrochemical signals, and remotely transmitting the processed data. The data management platform communicates with the smartphone to receive, store, and analyze the data, and identifies the stress state of the plant based on the waveform characteristics of the electrochemical signal using a machine learning model.

[0006] The method for remote monitoring of hydrogen peroxide and determination of stress status in plants using the system described in this invention includes the following steps: Step S1: Insert the working electrode of the electrochemical microsensor into the plant leaf, and place the reference electrode and counter electrode on the leaf surface; Step S2: Control the micro electrochemical device via smartphone to initiate electrochemical detection and obtain dynamic concentration data of hydrogen peroxide in plant leaves; Step S3: Upload the dynamic concentration data to the data management platform via a smartphone; Step S4: The data management platform extracts features from the received data to obtain multidimensional waveform features; Step S5: Input the extracted multidimensional waveform features into a pre-trained machine learning model, and the model outputs the stress state discrimination result of the plant.

[0007] Preferably, the preparation method of the electrochemical microsensor in step S1 is as follows: a micron-sized metal wire is passed through a hollow sleeve to lead out an interface, the micron-sized metal wire and the hollow sleeve are bonded together with insulating adhesive, a conductive tape is attached to one end of the micron-sized metal wire outside the hollow sleeve as an electrode pin, and the other end is modified with gold nanoparticles and graphene oxide to obtain a microelectrode formed by the modification of the micron-sized metal electrode wire; the microelectrode is used as the working electrode, and the electrochemical sensor is composed of the working electrode, the reference electrode and the counter electrode.

[0008] More preferably, the hollow sleeve is one of a glass tube, a silicone tube, a fiberglass tube, or a ceramic tube, preferably a glass tube. The conductive tape is one of a copper foil tape, a carbon conductive tape, or a copper-nickel fiber tape, preferably a copper foil tape.

[0009] Preferably, the micron-sized metal wire has a diameter of 80~120 μm, and more preferably 100 μm.

[0010] Preferably, the specific steps for H2O2 detection in step S2 include: measuring the current response values ​​of H2O2 solutions of different concentrations using an electrochemical sensor to obtain a series of H2O2 standard solution change curves; placing a microelectrode formed by modifying micron-sized metal wires as the working electrode on the plant tissue to be tested, while the reference electrode and the counter electrode must be placed near the working electrode without touching each other; adding 5 μL of PBS buffer solution to the tested location on the leaf; connecting the electrochemical workstation to a smartphone to detect H2O2 in the plant leaves and obtaining the concentration of H2O2 in the tested plant leaves.

[0011] More preferably, the buffer solution is 0.2 M PBS with pH 7.0. The detection time is 3600 s.

[0012] Preferably, in step S3, the user first logs into the database, then uploads a CSV / Excel file, and selects or creates the target database and table.

[0013] Preferably, in step S4, the upload history is viewed, data analysis and visualization are performed, and finally, charts or data are exported.

[0014] Preferably, in step S5, by integrating a machine learning model, the system can automatically and accurately identify the salt stress state of plants based on the collected electrochemical signals.

[0015] Invention Principle: Electrochemical methods utilize carbon- or metal-particle-modified electrode devices connected to a sample solution to measure potential changes during chemical reactions. Compared to other methods, it offers advantages such as simple operation, high selectivity, and high detection sensitivity. This invention proposes a mobile-enabled intelligent electrochemical microsensor based on a needle-shaped three-electrode system for continuous in vivo monitoring of H2O2 in tomato seedlings under salt stress. A 0.1 mm diameter Au / Go / Pt working electrode, a 0.1 mm diameter Pt counter electrode, and a 0.1 mm diameter Ag / AgCl reference electrode, modified with graphene oxide (GO) and gold nanoparticles (Au), are used. These can be placed within the plant to selectively capture and identify H2O2 within the plant. Compared with traditional electrochemical detection systems, the intelligent electrochemical microsensor developed for mobile phones exhibits excellent response to H2O2 standard solutions. In plant detection, it not only avoids the cumbersome and complicated connection of laboratory equipment and enables mobile phone power supply to solve the problem of inconvenient outdoor power sources, but also features a compact and lightweight design that is easy to carry. It can continuously monitor H2O2 in living tomato leaves in the field and achieve remote data transmission.

[0016] However, high-performance sensor hardware alone is insufficient to accurately determine the salt stress state of plants from H2O2 concentration data. The H2O2 signals produced by plants under salt stress are highly dynamic and complex, with waveform characteristics containing crucial information about the stress level, but the raw data itself is difficult to interpret directly. Traditionally, researchers rely on manual feature extraction and interpretation of collected current or voltage signals, such as manually measuring peak height, peak width, and response time, and then combining this with empirical thresholds to determine the stress state. This manual approach is not only time-consuming, labor-intensive, and inefficient, but also highly dependent on the operator's subjective experience, making it difficult to guarantee the consistency and accuracy of the interpretation results, and failing to meet the real-time analysis needs of large-scale data generated by continuous monitoring in field environments. With the development of machine learning technology, some studies have begun to attempt to input sensor-collected signals into general machine learning models for classification after simple processing. However, this loosely coupled sensor + algorithm model still struggles to fully extract deep features from the signals, resulting in limited discrimination accuracy.

[0017] This invention achieves intelligent identification of salt stress in tomatoes through deep coupling of electrode modification, signal acquisition, feature extraction, and machine learning models. Experimental verification shows that this method produces a synergistically enhanced technical effect compared to the traditional combination of general electrochemical sensors and general machine learning algorithms. The core of this invention lies in the deep matching of electrode materials, signal characteristics, and machine learning algorithms—the H2O2 signal generated after specific electrode modification has unique waveform characteristics, such as its peak shape, attenuation curve, and response time, which are significantly different from other electrodes in the time and frequency domain. The back-end algorithm is specifically designed with feature extraction and classification models for this waveform; both must be used in conjunction to accurately identify salt stress.

[0018] Experimental results show that the technical effect of this invention does not stem from a simple combination of modules, but rather from the deep adaptation and synergistic effect between electrode materials, signal feature extraction, and machine learning algorithms. These three must work together as an organic whole to achieve high-precision discrimination of tomato salt stress; replacing any one component with a generic solution will disrupt the inherent matching relationship within the system, leading to a significant deterioration in discrimination performance. The gain brought about by this deep coupling surpasses the simple summation of the independent contributions of each module, demonstrating a synergistic enhancement effect where the whole is greater than the sum of its parts. Specifically, this includes the following three aspects: First, the selective modification of the electrode material and the specific signal output. This invention employs a needle-shaped three-electrode system modified with gold nanoparticles and graphene oxide (working electrode diameter 0.1 mm Au / Go / Pt electrode, counter electrode diameter 0.1 mm Pt electrode, reference electrode diameter 0.1 mm Ag / AgCl electrode), which can be placed in plants to achieve selective capture and recognition of H2O2. Experiments show that this electrode has a good response to H2O2 standard solutions, a wide detection range, and basically no response or a response signal that is significantly different from H2O2 to common interfering substances in plants (such as salicylic acid, auxins, etc.), exhibiting excellent specificity. More importantly, the H2O2 signal generated by this electrode has unique waveform characteristics (such as specific peak shape, response time, decay curve, etc.), providing high-quality data raw materials for subsequent in-depth algorithmic analysis.

[0019] Secondly, a portable intelligent hardware platform was constructed. Smartphones handle functions such as connecting to the micro-electrochemical device, power supply, command control, data analysis, result display, and data transmission, solving the problems of inconvenient outdoor power sources and bulky equipment in field environments. A web-based interactive data management platform enables end-to-end management of data uploading, storage, and analysis, providing an efficient and convenient solution for remote monitoring and data sharing.

[0020] Third, deep integration of machine learning algorithms for specific waveforms. This invention does not simply graft general machine learning algorithms onto sensor data, but rather designs a feature extraction and classification model specifically for the waveform characteristics of H2O2 signals generated by Au / Go / Pt electrodes. This algorithm can automatically capture key information about the stress level contained in the waveform, transforming the raw electrochemical signal into a high-precision discrimination of the tomato salt stress state.

[0021] The technical effect of this invention does not stem from a simple superposition of the electrode, hardware, and algorithm modules, but rather from the synergistic enhancement effect resulting from the deep coupling of the three. The essence of this synergistic effect lies in the fact that the selective modification of the electrode material determines that the generated H2O2 signal has specific waveform characteristics, and the machine learning algorithm is specifically designed for this characteristic waveform—there is an inherent and inseparable matching relationship between the two. If other electrode materials are used (such as unmodified platinum electrodes or electrodes modified only with GO), even with the feature extraction algorithm specific to this invention, the algorithm's analytical capabilities will be rendered ineffective because the electrode cannot generate a specific waveform that matches the algorithm's design goals, inevitably affecting the accuracy of the discrimination results. Conversely, if the Au / GO / Pt electrodes remain unchanged, but a general machine learning algorithm (such as a standard classifier not optimized for waveform characteristics) is used instead, the algorithm will be unable to effectively capture the deep features contained in the waveform, leading to the loss of a large amount of key information, and the discrimination accuracy will also be difficult to guarantee.

[0022] The electrodes determine the uniqueness of the signal—the Au / Go / Pt modification layer endows the H2O2 response signal with unique time-frequency domain characteristics (such as specific peak shapes, response curves, and attenuation patterns); the algorithm is the dedicated decoder for interpreting these signals—a specially designed feature extraction and classification model can accurately capture these features and map them to the salt stress state. Both must work together as an organic whole to achieve accurate discrimination of the salt stress state of tomatoes. Severing the intrinsic connection between them, whether replacing the electrodes or changing the algorithm, will disrupt the inherent matching relationship within the system, leading to a significant deterioration in discrimination performance.

[0023] The gains brought about by this integrated hardware (electrode) and software (algorithm) design go beyond the simple summation of the independent contributions of each module—the electrodes provide high-quality data raw materials, and the algorithm enables deep information mining. The combination of the two is not a simple addition of functions, but a qualitative change: from data acquisition to intelligent state judgment. This synergistic enhancement effect, where the whole is greater than the sum of its parts, constitutes the core innovation of this invention that distinguishes it from existing technologies.

[0024] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: (1) The present invention uses a three-electrode system to detect changes in H2O2 content in plant leaves, which can realize fixed-point in vivo monitoring and collect dynamic change information of H2O2 in plant tissues, providing a new method for studying hormone changes in plants.

[0025] (2) The present invention uses a micron-level modified microelectrode as the working electrode and an electrochemical sensor based on the working electrode to detect H2O2 in plant leaves in situ in real time. Compared with other detection methods, it eliminates the complex and time-consuming processing process, reduces the possible impact on the measurement results, and reflects the changes in H2O2 content in plant leaves more timely. It is of great significance for further in-depth research on H2O2. Attached Figure Description

[0026] Figure 1 A represents the current response of the Au / Go / Pt electrode to the continuous addition of 10 µL of 50 mM~1000 mM H2O2 in 10 mL PBS solution using the computer-CHI1240C system via the It method of this invention, at an application potential of -0.5 V. Figure 1 B shows the linear relationship between the ampere response and H2O2 concentration in the range of 50 µM to 1000 µM, as shown in Figure A of this invention. Figure 1C represents the current response of continuously adding 10 µL of 50 mM~1000 mM H2O2 to 10 mL PBS solution at an application potential of -0.6 V using a mobile phone intelligent detection system. Figure 1 D represents the linear relationship between the ampere response and H2O2 concentration in the range of 50 µM to 1000 µM as shown in Figure C of this invention.

[0027] Figure 2 A represents the anti-interference capability of the computer-CHI1240C system used in this invention to detect H2O2. In 10 mL of PBS solution, under an applied potential of -0.5 V, 10 µL of 500 mM H2O2, SA, IAA, MeJA, ABA, AA, H2MA, Succinic Aicd, CA, and H2O2 were continuously added. Figure 2 B represents the anti-interference capability of the intelligent detection of H2O2 using a mobile phone in this invention. In 10 mL of PBS solution, under an applied potential of -0.5 V, 10 µL of 500 mM H2O2, SA, IAA, MeJA, ABA, AA, H2MA, Succinic Aicd, CA, and H2O2 were continuously added.

[0028] Figure 3 A is a schematic diagram of the in vivo continuous electrochemical monitoring of H2O2 in tomato leaves based on a mobile phone intelligent electrochemical microsensor according to the present invention. Figure 3 B represents the It detection curve of H2O2 in tomato leaves by a mobile phone intelligent micro-sensor under 0.4 M NaCl stress and control treatments according to the present invention.

[0029] Figure 4 This is a diagram illustrating the overall architecture of the machine learning model integrated in this invention. Detailed Implementation

[0030] The following describes a method for rapid detection of hydrogen peroxide in plant leaves based on microelectrodes, using specific embodiments.

[0031] Example 1

[0032] This embodiment uses the leaf part of tomato seedlings as the experimental subject, and includes the following steps: Step 1: Fabrication of microelectrodes: The microelectrode consists of four parts: a hollow sleeve, a micron-sized metal wire, insulating adhesive, and conductive tape. The micron-sized metal wire has a diameter of 100 μm. The micron-sized metal wire is passed through the hollow sleeve and led out to the interface. Both ends of the micron-sized metal wire are bonded to the hollow sleeve with insulating adhesive (it should be noted that the ends of the micron-sized metal wire must be left with appropriate length for subsequent operations). Conductive tape is attached to one end of the micron-sized metal wire outside the hollow sleeve, and the other end is trimmed to an appropriate length. It is then modified with gold nanoparticles and graphene oxide to obtain a microelectrode formed by modifying the micron-sized metal electrode wire.

[0033] Step 2: Construction of the electrochemical sensor: The microelectrode is used as the working electrode, and the electrochemical sensor is composed of the working electrode, the reference electrode, and the counter electrode. The reference electrode is Ag / AgCl, and the counter electrode is an unmodified micron-sized metal electrode wire.

[0034] Step 3: Detection of hydrogen peroxide: Preparation of the H2O2 standard curve: First, H2O2 was dissolved in 0.2 M PBS buffer (pH=7.0) to prepare a stock solution. Then, H2O2 solutions of different concentrations (50 mM~1000 mM) were prepared using buffer solution. A three-electrode system consisting of a working electrode (Pt, d=100 μm), a counter electrode (Pt, d=100 μm), and a reference electrode (Ag / AgCl, d=0.1 mm) was placed in 10 mL of the prepared PBS solution. The amperometric It curve of the microelectrode was recorded using a traditional computer-CHI1240C system. Qualitative and quantitative determinations of different concentrations of H2O2 were performed. Figure 1 As shown in Figure A, a linear relationship graph between current and H2O2 concentration was plotted using the obtained data, as follows: Figure 1 As shown in Figure B. Using the mobile phone-HY1550 system, with an applied potential of -0.6 V, the amperometric It curves were measured by continuously adding 50–1000 mM H₂O₂ solution to PBS solution. Figure 1 -C). Compared to the previous PC-CHI1240C system, the current response value is 1000 times larger, which may be due to differences in internal mechanisms caused by different workstations. With increasing H2O2 concentration, the reduction current of the microelectrode increases significantly, and its current response change has a good linear relationship with the H2O2 concentration. Figure 1 -D).

[0035] Other signaling molecules in plants did not significantly interfere with the electrode: In 10 mL of PBS buffer, 10 μL of 500 mM H₂O₂, 500 mM SA, 500 mM IAA, 500 mM MeJA, 500 mM ABA, 500 mM AA, 500 mM H₂MA, 500 mM succinic acid, 500 mM CA, and 500 mM H₂O₂ solutions were added sequentially. The results showed that only the addition of H₂O₂ caused a significant change in the current response, indicating that the microelectrode did not significantly interfere with the determination of H₂O₂. Figure 2 As shown in Figure A. Figure 2 -B uses a mobile phone-based intelligent detection system to detect the interference of microelectrodes on H2O2. (And...) Figure 2 Compared to the previous system, the intelligent mobile phone detection system did not significantly interfere with the determination of H2O2, but the current response value increased upon the addition of H2O2. These results indicate that the standard curve (It curve) and interference curve (It curve) of the intelligent mobile phone electrochemical microsensor for H2O2 detection are not significantly different from those of the traditional computer-CHI1240C system. Therefore, the intelligent mobile phone detection system can be used for the determination of H2O2 in actual samples.

[0036] Intelligent live continuous monitoring of H2O2 in tomato seedlings using a mobile phone: A mobile phone-HY1550 system was used to continuously monitor H2O2 in live plants using a mobile phone. H2O2 was detected using the time-current method (It method) on an electrochemical microsensor. The initial potential was -0.6 V, the sampling interval was 0.1 s, the settling time was 0 s, and the running time was 3600 s, yielding the concentration of H2O2 in the tomato leaf tissue.

[0037] Figure 3 Figure A shows a schematic diagram of placing the working electrode, counter electrode, and diameter reference electrode in a tomato leaf, and then dropping 5 μL of 0.2M PBS (pH 7.0) onto the electrode surface for electrochemical detection.

[0038] Figure 3 B shows the ampere-It curves of H2O2 in tomato leaves under high-salt stress and low-salt stress conditions. The results indicate that during a continuous monitoring period of 1 h, the ampere-It curve of the control group showed no significant change in current response, indicating that the H2O2 concentration in tomato leaves without high-salt treatment was stable. However, under high-salt stress, the detected ampere-It curve showed an electrochemical signal of H2O2 starting at 400 s, and continuous current changes occurred throughout the 1 h of monitoring. This suggests that the tomato plants stimulated H2O2 production in response to high-salt stress, and the highest H2O2 concentration produced during this process could reach nearly 60 μM.

[0039] Example 2: Remote monitoring and intelligent diagnostic system based on smartphone and web platform Building upon Example 1, this embodiment further constructs a complete monitoring system capable of remote data transmission and intelligent diagnosis. This system deeply integrates the electrochemical microsensors described in Example 1 with a smartphone and a web-based interactive data management platform, achieving a closed-loop process from signal acquisition to intelligent decision-making. This provides a complete solution for remote, continuous, and in vivo monitoring and data analysis of plant H2O2 in field environments.

[0040] Step 1: System Architecture Design Because the It curve data collected by sensors is large in volume and contains a lot of noise, traditional analysis methods are difficult to process efficiently. Therefore, this invention develops the following web-based data upload and analysis system. The system adopts a B / S architecture, and through the collaborative work of the front-end interactive interface and the back-end database service, it realizes the functions of visual uploading, storage, and analysis of CSV data. The core of the system consists of three parts: a front-end interactive module, a data processing module, and a database service module. The overall architecture is as follows: Figure 4 As shown, the dynamic changes in H2O2 levels in tomato leaves were acquired using a smartphone-based electrochemical microsensor, and the data was saved in CSV format. The CSV format is similar to a table, facilitating subsequent data reading and processing.

[0041] Step 2: Integration of Machine Learning-Based Intelligent Diagnostic Models for Salt Stress To enhance the system's analytical capabilities and upgrade the platform from a simple data management tool to an intelligent diagnostic system with decision support capabilities, this invention further integrates a machine learning-based intelligent salt stress diagnostic function into the platform.

[0042] For each IT data file uploaded via smartphone, the system automatically extracts a set of multidimensional features that characterize the changes in the current signal morphology. These features include, but are not limited to: time-domain statistical features (mean, standard deviation, skewness, kurtosis, extrema, etc.), frequency-domain features (dominant frequency obtained by fast Fourier transform, spectral centroid, high-low frequency energy ratio, etc.), waveform features (peak-to-peak value, crest factor, etc.), and signal stability features (zero-crossing rate, coefficient of variation, local fluctuation intensity, etc.). These features provide a crucial data foundation for subsequently distinguishing between normal plants and salt-stressed plants.

[0043] After the user selects the labeled normal samples and salt-stressed samples from the IT data file through the front-end interface, the system calls the back-end machine learning module and uses a Support Vector Machine (SVM) as the base classifier to train the model. The SVM classifier uses a Radial Basis Function (RBF) kernel, and the performance of the training and test sets is balanced by optimizing the parameters C=10.0 and gamma='scale'.

[0044] Under the same experimental conditions, the SVM algorithm exhibits the best comprehensive diagnostic capability:

[0045] SVM maintains the highest average accuracy while exhibiting the smallest cross-validation standard deviation, indicating that its generalization stability is significantly superior to ensemble learning methods. This is particularly important for scenarios such as plant physiology experiments where it is difficult to obtain a large number of repeated samples—stable predictive performance translates to more reliable diagnostic conclusions.

[0046] After training, the model can be used to automatically determine the state of newly collected IT data files. After the user uploads new data, the system automatically completes feature extraction, model inference, and returns the determination result and confidence level as "normal" or "salt stress".

[0047] This system supports cross-platform access, allowing users to connect via any device with a web browser without requiring local software installation. This zero-client approach reduces new user onboarding time by 60% and facilitates real-time multi-user collaboration. Multiple users can access and analyze data simultaneously, and the intuitive visualization interface offers various chart types to meet diverse analytical needs. The system's modular design allows for flexible expansion and easy feature customization. For data management, users can view and delete historical records, and file preview and table viewing functions further enhance the user experience. Regarding secure connections, the platform integrates advanced password protection and a reliable connection authentication mechanism.

[0048] This system utilizes web technology to manage the entire process of data uploading, storage, and analysis, providing an efficient and convenient solution for laboratory data management. Its intelligent salt stress diagnosis function combines machine learning with domain knowledge, making the platform not only a data management tool but also a research assistant with intelligent decision-making capabilities.

[0049] This embodiment presents a remote monitoring and intelligent diagnostic system based on a smartphone and web platform, which successfully achieves remote, continuous, and in vivo monitoring of the dynamic changes in H2O2 concentration in tomato leaves under salt stress. More importantly, by integrating a machine learning model, the system can automatically and accurately identify the salt stress state of plants based on the collected electrochemical signals. The results show that this system not only simplifies the data management process but also provides a new, objective, and efficient method for diagnosing plant stress physiology, offering a novel platform for the development of precision agriculture.

[0050] This invention, based on a highly sensitive portable electrochemical microsensor and a micro-electrochemical workstation, combined with a smartphone and a web-based interactive data management platform, enables remote monitoring of H2O2 content in tomato seedlings. The smartphone connects to the micro-electrochemical device for command control, enabling data analysis, result display, and remote transmission of collected data. Furthermore, by integrating a machine learning model, the system can automatically and accurately identify the salt stress state of plants based on the collected electrochemical signals. This system achieves remote, continuous, and in vivo monitoring and assessment of the dynamic changes in H2O2 concentration in tomato leaves under salt stress conditions, providing a novel platform for the development of precision agriculture.

Claims

1. A remote monitoring system for hydrogen peroxide in plants, characterized in that, include: An electrochemical microsensor, the working electrode of which is a micron-sized metal wire modified with gold nanoparticles and graphene oxide; A miniature electrochemical device connected to the electrochemical microsensor; A smartphone is communicatively connected to the microelectrochemical device. The data management platform communicates with the smartphone.

2. The remote monitoring system according to claim 1, characterized in that, The micron-sized metal wire has a diameter of 80~120 μm.

3. The remote monitoring system according to claim 1, characterized in that, The electrochemical microsensor is a three-electrode system, which also includes a counter electrode and a reference electrode; the counter electrode is an unmodified micron-sized metal wire, and the reference electrode is an Ag / AgCl electrode.

4. The remote monitoring system according to claim 1, characterized in that, The working electrode includes a micron-sized metal wire passing through a hollow sleeve and leading out through an interface. The micron-sized metal wire and the hollow sleeve are bonded together with insulating adhesive. One end of the micron-sized metal wire is attached to the outside of the hollow sleeve with conductive tape as an electrode pin, and the other end is modified with gold nanoparticles and graphene oxide.

5. A method for remote monitoring of hydrogen peroxide and determination of stress status in plants using the system described in any one of claims 1-4, characterized in that, Includes the following steps: Step S1: Insert the working electrode of the electrochemical microsensor into the plant leaf, and place the reference electrode and counter electrode on the leaf surface; Step S2: Control the micro electrochemical device via smartphone to initiate electrochemical detection and obtain dynamic concentration data of hydrogen peroxide in plant leaves; Step S3: Upload the dynamic concentration data to the data management platform via a smartphone; Step S4: The data management platform extracts features from the received data to obtain multidimensional waveform features; Step S5: Input the extracted multidimensional waveform features into a pre-trained machine learning model, and the model outputs the stress state discrimination result of the plant.

6. The method according to claim 5, characterized in that, In step S2, the electrochemical detection is a time-current method with an initial potential of -0.6 V, a sampling interval of 0.1 s, a rest time of 0 s, and a running time of 3600 s.

7. The method according to claim 5, characterized in that, In step S4, the multidimensional waveform features include at least one of time-domain statistical features, frequency-domain features, waveform features, and signal stability features.

8. The method according to claim 5, characterized in that, In step S5, the machine learning model is a support vector machine (SVM).

9. The method according to claim 5, characterized in that, In step S5, the stress state is a salt stress state.

10. The method according to claim 5, characterized in that, The plant in question is a tomato.

Images