Big data-based dynamic monitoring method for health-care plant volatile components

Abstract

This invention belongs to the field of plant physiological dynamic monitoring and decision optimization technology, specifically providing a big data-based method for dynamic monitoring of volatile components in health-promoting plants. The method mainly includes: acquiring time-series data of plant volatile concentrations, environmental parameter sets, and meteorological time-series data; obtaining a spatiotemporally synchronized multidimensional data cube after collaborative preprocessing; and calculating environmental interference components based on the multidimensional data cube to obtain a physiological volatile characteristic set. This application effectively solves the problem in existing technologies where it is difficult to comprehensively capture dynamic changes in volatiles and integrate multidimensional correlation information under large-scale, long-term, and complex environments, leading to data distortion. It achieves accurate understanding of the actual release patterns of volatiles, providing strong support for the evaluation and control decisions of plant health-promoting effects.

Description

Technical Field

[0001] This invention belongs to the field of plant physiological dynamic monitoring and decision optimization technology, and specifically relates to a method for dynamic monitoring of volatile components in health-promoting plants based on big data. Background Technology

[0002] Monitoring of plant volatiles typically employs a combination of point-to-point sampling and laboratory analysis. Instruments such as gas chromatography-mass spectrometry (GC-MS) are used to determine the concentration of volatiles at specific time intervals, while simultaneously recording environmental data such as temperature and humidity. These methods are applied in small-scale, short-term monitoring scenarios, providing fundamental data for understanding the basic composition and release patterns of plant volatiles.

[0003] When the monitoring scope is expanded to a large area and the time span is extended to several weeks or months, and when faced with complex environmental conditions such as sudden changes in light intensity and significant fluctuations in temperature and humidity, existing methods are unable to fully capture the dynamic changes in volatile concentrations, nor can they fully integrate the correlation information between multi-dimensional environmental factors and plant growth status. This results in the volatile data obtained failing to reflect their true release patterns, affecting the accurate assessment of the health benefits of plants, and failing to provide a reliable basis for relevant regulatory decisions. Summary of the Invention

[0004] This application provides a big data-based method for dynamic monitoring of volatile components in health-promoting plants. This method effectively solves the problem in existing technologies where it is difficult to comprehensively capture dynamic changes in volatiles and integrate multi-dimensional related information under large-scale, long-term, and complex environments, leading to data distortion. It enables accurate understanding of the actual release patterns of volatiles and provides strong support for the evaluation and control decisions of plant health-promoting effects.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] Firstly, this application provides a method for dynamic monitoring of volatile components in health-promoting plants based on big data, including:

[0007] We acquire time-series data on plant volatile concentrations, environmental parameter sets, and meteorological time-series data, and obtain a spatiotemporally synchronized multidimensional data cube after collaborative preprocessing.

[0008] Based on the calculation of the environmental disturbance components of the multidimensional data cube, a physiological volatile feature set is obtained.

[0009] Obtain plant physiological parameters; perform correlation analysis between the physiological volatiles feature set and the plant physiological parameters to obtain dynamic decision data containing stress triggering conditions and target volatiles label set.

[0010] The dynamic decision data is fused and multimodal correction is performed to output a spatially distributed volatiles dataset.

[0011] A three-dimensional map of volatiles across the entire domain is generated based on the volatiles dataset, and decision support data is output by combining it with a pharmacological knowledge base.

[0012] Furthermore, the collaborative processing includes:

[0013] Wavelet threshold noise filtering is performed on the time series data of volatile concentrations to obtain the purified volatile concentration sequence.

[0014] Multiple interpolation is performed on the environmental parameter set to obtain the complete environmental parameter matrix.

[0015] The sliding window variance abrupt change feature is extracted from meteorological time series data to obtain a meteorological event label set.

[0016] The purified volatile concentration sequence, parameter matrix, and event tag set are timestamped to output a multidimensional data cube.

[0017] Spatial grid density clustering is performed based on the multidimensional data cube to identify sensor coverage blind spots and generate dynamic coordinate sets.

[0018] Furthermore, the generation of the dynamic coordinate set includes:

[0019] Calculate the sensor distribution density within each spatial grid and mark grids with densities below a threshold as blind zones.

[0020] The dynamic coordinate set is used for the spatial distribution correction.

[0021] Furthermore, based on the environmental disturbance components calculated from the multidimensional data cube, a physiological volatiles feature set is obtained, including:

[0022] A spatiotemporal graph convolutional network model is constructed based on a multidimensional data cube, outputting a set of coupling coefficients between environmental parameters and volatile concentrations.

[0023] The environmental interference component is calculated based on the set of coupling coefficients to obtain the environmental interference baseline.

[0024] Volatile concentration sequences were extracted from the multidimensional data cube, and the environmental interference baseline was subtracted from the volatile concentration sequences to obtain the physiological volatile characteristic sequences.

[0025] The isolated forest algorithm is used to detect outliers in the feature sequence and output a set of physiological volatile features.

[0026] Furthermore, a correlation analysis is performed between the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data containing stress triggering conditions and target volatiles label sets, including:

[0027] Local weighted smoothing is performed on the time-series data of plant physiological parameters to obtain the steady-state physiological parameter stream.

[0028] Canonical correlation analysis was performed on the physiological volatiles feature set and the steady-state physiological parameter flow to obtain the correlation map between physiology and volatiles.

[0029] Based on the node degree mutation detection stress response threshold of the association graph, a set of stress triggering conditions is obtained.

[0030] When the steady-state physiological parameter flow exceeds the triggering condition, the tracking mode is activated to obtain the target volatile marker set.

[0031] Furthermore, correlation analysis is performed on the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data containing stress triggering conditions and target volatiles label sets, which also includes:

[0032] When the meteorological event label set shows extreme events, the confidence weights of environmental parameters are reduced according to the event level index to obtain environmental-physiological dynamic equilibrium parameters containing a confidence weight matrix.

[0033] Furthermore, the dynamic decision data is fused and multimodal correction is performed to output a spatially distributed corrected volatile data set, including:

[0034] The target volatile label set and dynamic equilibrium parameters are input into the attention mechanism model, and the meteorological disturbance confidence score is output.

[0035] If the confidence score of meteorological disturbance is lower than the preset threshold, the metabolic pathway feature library verification is triggered, and the table of stress volatile components is output to confirm the results.

[0036] The biocompatibility of the composition table is verified based on dynamic equilibrium parameters, and a list of bioavailable volatiles is output.

[0037] A turbulent diffusion model is used to perform spatial compensation on the dynamic equilibrium parameters, and a volatiles dataset is output.

[0038] Furthermore, based on the confidence weight matrix in the dynamic equilibrium parameters, the boundary conditions of the canopy turbulence diffusion model are modified.

[0039] Furthermore, a global volatiles 3D map is generated based on the volatiles dataset, and decision support data is output by combining it with a pharmacological knowledge base, including:

[0040] By integrating the dynamic coordinate set of blind zones and the volatile matter dataset from the integrated monitoring, a blind zone prediction model is constructed using the Kriging interpolation algorithm, and a three-dimensional volatile matter map with no blind zones in the entire region is output.

[0041] Based on the three-dimensional map of volatiles without blind spots, the bioactivity data of compounds in the plant pharmacology database are matched to output a mapping table of health and wellness efficacy components.

[0042] Furthermore, the generation of a global volatiles 3D map based on the volatiles dataset, combined with a pharmacological knowledge base, to output decision support data also includes:

[0043] It receives an expert experience rule base, performs knowledge conflict resolution on the efficacy component mapping table, and outputs a two-way verification dynamic knowledge graph containing credibility scores.

[0044] Based on the dynamic knowledge graph, a spatial planning guidance matrix and a set of plant maintenance optimization schemes are generated.

[0045] Secondly, this application provides a dynamic monitoring system for the volatile components of health-promoting plants based on big data, which includes:

[0046] Data acquisition and preprocessing module: Acquires time-series data of plant volatile concentrations, environmental parameter sets, and meteorological time-series data, and obtains a spatiotemporally synchronized multidimensional data cube after collaborative preprocessing.

[0047] Environmental interference solution module: Based on the multidimensional data cube, the environmental interference components are solved to obtain the physiological volatiles feature set.

[0048] Association analysis and decision generation module: acquire plant physiological parameters; perform association analysis on the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data including stress triggering conditions and target volatiles label set.

[0049] Fusion Correction Module: Fuses the dynamic decision data and performs multimodal correction, outputting a spatially distributed corrected volatile data set.

[0050] The map construction and decision support module generates a three-dimensional map of volatiles across the entire domain based on the volatiles dataset, and outputs decision support data in conjunction with the pharmacological knowledge base.

[0051] Thirdly, this application provides a big data-based dynamic monitoring device for the volatile components of health-promoting plants, which includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program to implement the steps of the big data-based dynamic monitoring method for the volatile components of health-promoting plants as described in the first aspect.

[0052] Fourthly, this application provides a storage medium storing computer program instructions, which are read and executed by a processor to perform the steps of the dynamic monitoring method for volatile components of health-promoting plants based on big data as described in the first aspect.

[0053] The beneficial effects of this invention are:

[0054] This application utilizes time-series data on plant volatile concentrations, followed by collaborative preprocessing to obtain a spatiotemporally synchronized multidimensional data cube. It then resolves environmental interference, correlates plant physiological parameters, and generates a three-dimensional map after fusion and correction. Combined with a pharmacological knowledge base, it outputs decision data, effectively solving the problem of data distortion in existing technologies due to the difficulty in comprehensively capturing dynamic changes in volatiles and integrating multidimensional correlation information under large-scale, long-term, and complex environments. This approach achieves accurate understanding of the actual release patterns of volatiles, providing strong support for evaluating and regulating the health benefits of plants.

[0055] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description and the drawings. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 A flowchart illustrating the present invention provides a method for dynamic monitoring of volatile components in health-promoting plants based on big data. Detailed Implementation

[0058] To address the problems raised in the background technology, this application obtains a multi-dimensional data cube through preprocessing of multiple types of data, solves for interference to obtain a feature set, combines physiological parameter analysis to obtain decision data, corrects the output dataset, generates a spectrum, and combines a knowledge base to output supporting data, which can accurately monitor volatiles.

[0059] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] In some embodiments, such as Figure 1 As shown, this application provides a method for dynamic monitoring of volatile components in health-promoting plants based on big data, including:

[0061] S1. Acquire time-series data of plant volatile concentrations, environmental parameter sets, and meteorological time-series data, and obtain a spatiotemporally synchronized multidimensional data cube after collaborative preprocessing.

[0062] S2. Based on the multidimensional data cube, the environmental disturbance components are solved to obtain the physiological volatiles feature set.

[0063] S3. Obtain plant physiological parameters; perform correlation analysis between the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data containing stress triggering conditions and target volatiles label set.

[0064] S4. Integrate dynamic decision data and perform multimodal correction to output a spatially corrected volatile dataset.

[0065] S5. Generate a global 3D volatile map based on the volatile dataset, and output decision support data by combining it with the pharmacological knowledge base.

[0066] In some embodiments, the time-series data of plant volatile concentrations in S1 are acquired through a sensor network deployed in a health-promoting plant community. This network consists of an array of electrochemical gas sensors that collect the concentration values ​​of specified volatile organic compounds (VOCs) at fixed time intervals to form a raw time-series data stream. Each data point includes a timestamp, spatial coordinates, and the measured concentration value of the corresponding compound (such as α-pinene or limonene).

[0067] The environmental parameter set is collected synchronously by IoT devices installed in the same space as the sensor network. The parameter categories include three basic environmental indicators: temperature, humidity, and light intensity, forming a multi-dimensional parameter set. Each parameter includes the collection time, spatial location, and measured value.

[0068] Meteorological time-series data are continuously acquired through anemometers connected to meteorological stations, collecting wind speed and wind direction angle data to form a continuous time series. Each data point includes a timestamp, wind speed value, and wind direction angle, which is used to characterize the impact of atmospheric motion on the diffusion of volatiles.

[0069] Collaborative processing includes:

[0070] S11. Perform wavelet threshold noise filtering on the time series data of volatile concentrations to obtain the purified volatile concentration sequence.

[0071] In the specific implementation process, the sym4 wavelet basis function is selected for three-level decomposition, and the threshold of the detail coefficients in each level is set according to... ;in, The threshold representing the detail coefficients of the j-th level wavelet decomposition. Represents the standard deviation of noise. This represents the number of detail coefficients in the j-th level wavelet decomposition.

[0072] For example, when processing menthol limonene data, a timestamp-aligned purification concentration sequence is output at 1-minute intervals.

[0073] S12. Perform multiple interpolation on the environmental parameter set to obtain the complete environmental parameter matrix.

[0074] During implementation, several sets (e.g., 5 sets) of independent imputation datasets are generated, and the mean is used to fill in the imputation to form a complete environmental parameter matrix.

[0075] For example, when humidity data for a certain area is missing at 10:00 AM, five estimated values ​​are generated using the temperature-humidity covariance and then averaged. The output includes minute-level complete temperature, humidity, and illumination data for each monitoring point.

[0076] S13. Extract the sliding window variance mutation features from the meteorological time series data to obtain the meteorological event label set.

[0077] Set a time window of a specified duration (e.g., 30 minutes), and when the window variance... ;in, This represents the variance of wind speed within the current sliding window. The average variance of wind speed for the same historical period. The standard deviation represents the variance of wind speed during the same historical period.

[0078] S14. Align the purified volatile concentration sequence, parameter matrix, and event tag set with timestamps to output a multidimensional data cube.

[0079] The alignment process is based on UTC time: all data are resampled to minute-by-minute time points through cubic spline interpolation, and the spatial coordinates are uniformly converted to WGS84 latitude and longitude.

[0080] For example, wind speed data sampled at the second level and environmental parameters sampled at the minute level are unified into a dataset of average values ​​per minute. The output multidimensional data cube data structure includes a three-dimensional spatial index and multiple attribute fields: [time stamp, longitude, latitude, volatile concentration value, temperature value, humidity value, light value, meteorological event status code].

[0081] S15. Perform spatial grid density clustering based on multidimensional data cubes to identify sensor coverage blind spots and generate dynamic coordinate sets.

[0082] In some embodiments, the generation of the dynamic coordinate set in S15 includes:

[0083] Calculate the sensor distribution density within each spatial grid and mark grids with densities below a threshold as blind zones.

[0084] Divide the monitoring area into grids of a specified area (e.g., 1 meter × 1 meter): Calculate the number of sensors N within a specified radius (e.g., 50 meters) of the center point of each grid. When N < 1, mark the grid as a blind zone.

[0085] For example, if there are no sensor devices within 50 meters around grid number G-1024, its center coordinates (X=102.35, Y=35.67) are recorded, and the output dynamic coordinate set contains the Cartesian coordinates (X, Y) of all blind grids and grid area data.

[0086] In some embodiments, S2 calculates the environmental disturbance components based on a multidimensional data cube to obtain a physiological volatiles feature set, including:

[0087] S21. Construct a spatiotemporal graph convolutional network model based on a multidimensional data cube, and output a set of coupling coefficients between environmental parameters and volatile concentrations.

[0088] Multidimensional data cubes are processed by constructing a spatiotemporal graph convolutional network model. This model contains three graph convolutional layers and temporal recurrent units. The inputs are timestamp sequences, spatial coordinates and corresponding volatile concentration values, temperature values, humidity values ​​and light values. The output is a set of coupling coefficients between environmental parameters and volatile concentration.

[0089] For example, when processing limonene concentration data, the model outputs a coupling coefficient of 0.78 for temperature and -0.32 for humidity and 0.15 for illumination, forming a coefficient matrix that includes the weights of each parameter.

[0090] S22. Solve the environmental interference components based on the coupling coefficient set to obtain the environmental interference baseline.

[0091] The environmental parameter values ​​(temperature, humidity, and light intensity) in the multidimensional data cube are weighted and summed with their corresponding coupling coefficients to generate an environmental disturbance baseline.

[0092] For example, at a certain moment, when the temperature is 28℃, the humidity is 65%, and the light intensity is 1200 lux, the calculated environmental interference baseline is: 0.78×28+(-0.32)×65+0.15×1200=181.04ppb.

[0093] S23. Extract the volatile concentration sequence from the multidimensional data cube, subtract the environmental interference baseline from the volatile concentration sequence to obtain the physiological volatile characteristic sequence.

[0094] For example, subtracting the baseline of 181.04 ppb from the original limonene concentration of 192 ppb yields a physiological characteristic value of 10.96 ppb.

[0095] Physiological characteristic values ​​across multiple time dimensions constitute a physiological volatiles characteristic sequence, which can reflect the level of volatiles released by the plant itself.

[0096] S24. Use the isolated forest algorithm to detect outliers in the feature sequence and output the physiological volatiles feature set.

[0097] Construct a specified number (e.g., 100) of isolation trees, detect data points that deviate from the mainstream distribution through random splitting paths, and output a set of physiological volatile features; for example, physiological feature values ​​that are more than three standard deviations can be marked as data points that deviate from the mainstream distribution.

[0098] For example, if a feature value suddenly increases to 58.3 ppb at a certain moment, and this exceeds three times the standard deviation, it is replaced with the mean of the preceding and following moments, 34.2 ppb, to form a smoothed correction sequence.

[0099] In some embodiments, S3 involves correlation analysis between the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data containing stress triggering conditions and target volatiles label sets, including:

[0100] S31. Perform local weighted smoothing on the time series data of plant physiological parameters to obtain the steady-state physiological parameter stream.

[0101] A Gaussian kernel function is used to perform a neighborhood-weighted average for each data point to eliminate transient fluctuations and output a steady-state physiological parameter stream.

[0102] The steady-state physiological parameter stream is a smooth physiological parameter sequence aligned with timestamps. For example, the original sequence of chlorophyll fluorescence parameters [45,47,120,46,48] is processed so that the outlier 120 is corrected to 47, and the output steady-state sequence [45,47,47,46,48].

[0103] S32. Perform canonical correlation analysis between the physiological volatiles feature set and the steady-state physiological parameter flow to obtain the correlation map between physiology and volatiles.

[0104] Canonical correlation analysis was performed between the physiological volatiles feature set and the steady-state physiological parameter flow: the covariance matrix of the two sets of variables was calculated, the eigenvector corresponding to the maximum correlation coefficient was solved, and the correlation graph between physiological parameters and volatiles was output. The correlation graph includes nodes (physiological parameters / volatiles) and connecting edges (correlation coefficients).

[0105] For example, the correlation coefficient between limonene concentration and chlorophyll fluorescence was found to be 0.91, and the output spectrum contained the edge ["chlorophyll fluorescence-limonene", 0.91].

[0106] S33. Based on the node degree mutation detection of the association graph, the stress response threshold is obtained to obtain the set of stress triggering conditions.

[0107] The system monitors the number of connections (degrees) of each node in real time. When the degree exceeds the trigger condition (such as three times the standard deviation of the historical mean), a set of stress trigger conditions is generated. The trigger conditions include a combination of threshold values ​​of related parameters.

[0108] For example, the "limonene-chlorophyll fluorescence" node has a normal value of 3 (standard deviation 2.5). When the value suddenly increases to 12, the output trigger conditions are ["chlorophyll fluorescence ≤ 45RFU", "limonene concentration ≥ 80ppb"].

[0109] S34. When the steady-state physiological parameter flow exceeds the triggering condition, the tracking mode is activated to obtain the target volatile marker set.

[0110] When the steady-state physiological parameter flow exceeds any condition in the stress triggering condition set, the tracking mode is activated: the trigger node is located in the association map, the volatiles directly associated with it are marked, and a target volatile label set is generated.

[0111] For example, if the chlorophyll fluorescence value drops to 40 RFU (breaking the 45 RFU threshold), the associated volatile compound "limonene" is labeled, and the label set record is output [time="2023-06-01 14:00", target compound="limonene"].

[0112] In some embodiments, step S3 involves correlation analysis between the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data containing stress triggering conditions and target volatiles label sets, and further includes:

[0113] When the meteorological event label set shows extreme events, the confidence weights of environmental parameters are reduced according to the event level index to obtain environmental-physiological dynamic equilibrium parameters containing a confidence weight matrix.

[0114] In some embodiments, S4 fuses dynamic decision data and performs multimodal correction to output a spatially distributed corrected volatile dataset, including:

[0115] S41. Input the target volatile label set and dynamic equilibrium parameters into the attention mechanism model and output the meteorological disturbance confidence score.

[0116] The target volatile label set and environmental-physiological dynamic equilibrium parameters are input into an attention mechanism model (such as a Transformer encoder). The interaction between the weight assignment of the target volatile labels and the confidence of environmental parameters is calculated through a self-attention mechanism, and the meteorological disturbance confidence score is output.

[0117] For example, given an input tag set ["limonene", "α-pinene"] and a weight matrix [temperature: 0.7, humidity: 0.3], the output score is 0.65 (range [0,1]).

[0118] S42. If the confidence score of meteorological disturbance is lower than the preset threshold, the metabolic pathway feature library verification is triggered, and the table of stress volatile components is output to confirm the output.

[0119] The preset threshold can be determined through historical data statistics. The threshold is 0.7. If the confidence score of meteorological interference is lower than 0.7, the metabolic pathway feature library verification is triggered: matching the known reaction chain of the target volatile in the plant metabolic pathway, such as the terpene synthase pathway, and outputting a table of confirmed stress volatile components.

[0120] For example, after verifying the legitimacy of limonene in the stress metabolic pathway, the component table ["limonene", "stress response pathway: terpene synthase"] is output.

[0121] S43. Perform biocompatibility verification on the component table based on dynamic equilibrium parameters and output a list of bioavailable volatiles.

[0122] Specifically, the rationality verification includes the following steps:

[0123] 1. Extract the temperature confidence level from the confidence weight matrix. .

[0124] 2. Calculate the effective temperature range ;in, This represents the optimal temperature range for a compound, such as the 30-40℃ baseline range for limonene.

[0125] 3. Verify if the current temperature is within the range. Inside.

[0126] For example, if the current temperature is 25°C and within the range of 21-28°C, output the list of bioavailable volatiles: [compound="limonene", availability="high", effective temperature range="21-28°C"].

[0127] S44. Use a turbulent diffusion model to perform spatial compensation on the dynamic equilibrium parameters and output a volatiles dataset.

[0128] Specifically, the Pasquill-Gifford turbulent diffusion model can be used to perform spatial compensation for the dynamic equilibrium parameters. The compensation method is as follows:

[0129] 1. Extract humidity confidence level from confidence weight matrix .

[0130] 2. Corrected diffusion coefficient ;in, Represents the baseline diffusion coefficient. The humidity-sensitive factor can be determined by fitting historical diffusion data of volatiles from the target plant, specifically the slope of a linear regression between the rate of change in humidity and the coefficient of variation of concentration distribution.

[0131] 3. Substitute into the diffusion equation to solve for the concentration distribution: Solve for the distribution of volatile concentration C at spatial location x and time t.

[0132] In some embodiments, the boundary conditions of the canopy turbulence diffusion model are modified based on the confidence weight matrix in the dynamic equilibrium parameters.

[0133] The specific implementation process includes: firstly, parsing the data structure of the confidence weight matrix, which contains temperature confidence weights. Humidity reliability weight Lighting credibility weight There are three weight values, each ranging from [0,1]. The lower the value, the worse the reliability of the corresponding environmental parameter.

[0134] For example, when the meteorological event tag set detects a rainstorm event, the humidity confidence weight... The value was reduced from the baseline of 1.0 to 0.3.

[0135] Subsequently, the standard boundary condition parameters of the canopy turbulence diffusion model were extracted: boundary layer height. and reference diffusion coefficient .

[0136] The lower the temperature confidence level, the more significant the atmospheric stability and boundary layer compression caused by high-temperature events. Dynamic adjustments are made based on the confidence weight matrix. .

[0137] The lower the humidity confidence level, the greater the increase in the diffusion coefficient due to the intensified turbulent mixing caused by humidity interference. Revise the specific reference formula: ;in, This represents a humidity-sensitive factor, which can be determined by fitting a linear regression slope between the humidity change rate and the coefficient of variation of the concentration distribution using historical data.

[0138] In some embodiments, S5 generates a global volatile three-dimensional map based on the volatile dataset and outputs decision support data in conjunction with a pharmacological knowledge base, including:

[0139] S51. By fusing the dynamic coordinate set of blind zones and the volatile matter dataset from the monitoring, a blind zone prediction model is constructed using the Kriging interpolation algorithm, and a three-dimensional map of volatile matter without blind zones is output.

[0140] A blind zone inference model is constructed using the Kriging interpolation algorithm. The blind zone inference model takes the volatile concentration values ​​of known monitoring points in the volatile dataset (such as limonene concentration of 1.8 ppm in grid G1 and 2.1 ppm in grid G2) as input data and the center coordinates of the blind zone grid in the dynamic coordinate set of the monitoring blind zone (such as coordinates X=102.35, Y=35.67 in grid G-1024) as the interpolation target point. The spatial autocorrelation weight is calculated by the semivariance function, and the three-dimensional volatile spectrum of the whole domain without blind zones is output.

[0141] For example, the estimated limonene concentration for blind grid G-1024 is 1.95 ppm. The output three-dimensional map data structure contains the volatile concentration values ​​for all spatial locations (including the original blind area), in the format of [longitude, latitude, concentration value], such as [102.35, 35.67, 1.95 ppm].

[0142] S52. Based on the three-dimensional map of volatiles without blind spots, match the bioactivity data of compounds in the plant pharmacology database, and output a mapping table of health and wellness efficacy components.

[0143] Retrieve the health and wellness efficacy attributes corresponding to the concentration values ​​of each volatile compound in the spectrum, and output a health and wellness efficacy component mapping table.

[0144] The pharmacological properties database includes compound names, bioactivity intensity, efficacy type (e.g., sedation / anti-inflammatory), and effective concentration thresholds.

[0145] For example, when the concentration of limonene is ≥1.5ppm, its "sedative and tranquilizing" effect is matched, and the mapping table entry is output as [compound="limonene", concentration=1.95ppm, effect="sedative and tranquilizing", effect strength="medium"].

[0146] In some embodiments, S5, which generates a global volatile three-dimensional map based on the volatile dataset and outputs decision support data in conjunction with a pharmacological knowledge base, further includes:

[0147] S53. Receive the expert experience rule base, perform knowledge conflict resolution on the efficacy component mapping table, and output a two-way verification dynamic knowledge graph containing credibility scores.

[0148] The expert experience rule base contains logical rules defined by plant pharmacology experts, such as "If the limonene concentration is ≥1.5ppm and the light intensity is >1000 lux, then the confidence level of the sedative effect is +0.3". By fusing the rules with the measured data in the mapping table through the DS evidence theory, a two-way verification dynamic knowledge graph containing confidence scores is output.

[0149] The specific conflict resolution process is as follows:

[0150] 1. Matching rules and mapping table entries, such as rule R1 matching the measured concentration of limonene of 1.8 ppm.

[0151] 2. Calculate the rule support (S) and conflict degree. ; ;in, Represents the number of matching rules. Represents the total number of rules.

[0152] 3. Calculate the credibility score ;in, The rule weights are pre-set by plant pharmacology experts when defining logical rules and are directly stored in the expert experience rule base.

[0153] The final result is a two-way verified dynamic knowledge graph with a data structure of [compound, efficacy, credibility score], such as ["limonene", "sedation", 0.72].

[0154] S54. Generate a spatial planning guidance matrix and a set of plant maintenance optimization schemes based on dynamic knowledge graphs.

[0155] The credibility score in the dynamic knowledge graph is associated with the spatial location to form a gridded decision matrix [coordinates, score, recommended planting density].

[0156] For example, the limonene score for grid G1 (coordinates 120.35, 30.27) is 0.72, and the output matrix entry is ["(120.35, 30.27)", 0.72," 5 plants / ].

[0157] Meanwhile, maintenance strategies are triggered based on the scoring threshold: if the score is ≥0.8, the output is "Increase phosphorus and potassium fertilizer application"; if 0.6≤score<0.8, the output is "Regular irrigation".

[0158] For example, a limonene score of 0.72 trigger scheme ["location=G1","scheme=regular irrigation","cycle=once daily"].

[0159] Secondly, this application provides a dynamic monitoring system for the volatile components of health-promoting plants based on big data, which includes:

[0160] Data acquisition and preprocessing module: Acquires time-series data of plant volatile concentrations, environmental parameter sets, and meteorological time-series data, and obtains a spatiotemporally synchronized multidimensional data cube after collaborative preprocessing.

[0161] Environmental interference solution module: Based on the multidimensional data cube, the environmental interference components are solved to obtain the physiological volatiles feature set.

[0162] Association analysis and decision generation module: acquire plant physiological parameters; perform association analysis on the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data including stress triggering conditions and target volatiles label set.

[0163] Fusion Correction Module: Fuses dynamic decision data and performs multimodal correction, outputting a spatially corrected volatile data set.

[0164] The map construction and decision support module generates a three-dimensional map of volatiles across the entire domain based on the volatiles dataset, and outputs decision support data in conjunction with the pharmacological knowledge base.

[0165] Thirdly, this application provides a big data-based dynamic monitoring device for the volatile components of health-promoting plants, which includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program to implement the steps of the big data-based dynamic monitoring method for the volatile components of health-promoting plants.

[0166] Fourthly, this application provides a storage medium storing computer program instructions, which are read and executed by a processor to perform the steps of a method for dynamic monitoring of volatile components of health-promoting plants based on big data.

[0167] Any references to memory, storage, database, or other media used in the embodiments provided in this invention may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory.

[0168] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0169] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic monitoring of volatile components in health-promoting plants based on big data, characterized in that, include: We acquired time-series data on plant volatile concentrations, environmental parameter sets, and meteorological time-series data, and obtained a spatiotemporally synchronized multidimensional data cube after collaborative preprocessing. Based on the calculation of the environmental disturbance components of the multidimensional data cube, a set of physiological volatiles feature sets is obtained; Obtain plant physiological parameters; A correlation analysis was performed between the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data that includes stress triggering conditions and target volatiles label set; By fusing the dynamic decision data and performing multimodal correction, a spatially distributed corrected volatiles dataset is output. By integrating the dynamic coordinate set of blind zones and the volatile matter dataset from the integrated monitoring, a blind zone prediction model is constructed using the Kriging interpolation algorithm, and a three-dimensional map of volatile matter without blind zones is output. Based on the three-dimensional map of volatiles without blind spots, the bioactivity data of compounds in the plant pharmacology database are matched to output a mapping table of health and wellness efficacy components. Receive expert experience rule base, perform knowledge conflict resolution on efficacy component mapping table, and output a two-way verification dynamic knowledge graph containing credibility score; Based on the dynamic knowledge graph, a spatial planning guidance matrix and a set of plant maintenance optimization schemes are generated.

2. The method for dynamic monitoring of volatile components in health-promoting plants according to claim 1, characterized in that, Collaborative processing includes: Wavelet threshold noise filtering was performed on the time series data of volatile concentrations to obtain the purified volatile concentration sequence; Perform multiple interpolation on the environmental parameter set to obtain the complete environmental parameter matrix; The sliding window variance abrupt change feature is extracted from meteorological time series data to obtain a meteorological event label set; The purified volatile concentration sequence, parameter matrix, and event tag set are timestamped and aligned to output a multidimensional data cube; Spatial grid density clustering is performed based on the multidimensional data cube to identify sensor coverage blind spots and generate dynamic coordinate sets.

3. The method for dynamic monitoring of volatile components in health-promoting plants according to claim 2, characterized in that, The generation of dynamic coordinate sets includes: Calculate the sensor distribution density within each spatial grid, and mark grids with densities below a threshold as blind zones; The dynamic coordinate set is used for the spatial distribution correction.

4. The method for dynamic monitoring of volatile components in health-promoting plants according to claim 3, characterized in that, Based on the environmental disturbance components calculated from the multidimensional data cube, a physiological volatiles feature set is obtained, including: A spatiotemporal graph convolutional network model is constructed based on a multidimensional data cube, which outputs a set of coupling coefficients between environmental parameters and volatile concentrations. The environmental interference components are calculated based on the set of coupling coefficients to obtain the environmental interference baseline; Volatile concentration sequences were extracted from the multidimensional data cube, and the environmental interference baseline was subtracted from the volatile concentration sequences to obtain the physiological volatile characteristic sequences. The isolated forest algorithm is used to detect outliers in the feature sequence and output a set of physiological volatile features.

5. The method for dynamic monitoring of volatile components in health-promoting plants according to claim 2, characterized in that, A correlation analysis was performed between the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data containing stress triggering conditions and target volatiles label sets, including: Local weighted smoothing was performed on the time-series data of plant physiological parameters to obtain the steady-state physiological parameter stream; Canonical correlation analysis was performed on the physiological volatiles feature set and the steady-state physiological parameter flow to obtain the correlation map between physiology and volatiles; Based on the node degree mutation detection stress response threshold of the association graph, a set of stress triggering conditions is obtained; When the steady-state physiological parameter flow exceeds the triggering condition, the tracking mode is activated to obtain the target volatile marker set.

6. The method for dynamic monitoring of volatile components in health-promoting plants according to claim 5, characterized in that, A correlation analysis was performed between the physiological volatiles feature set and plant physiological parameters to obtain dynamic decision data containing stress triggering conditions and target volatiles label sets, which also includes: When the meteorological event label set shows extreme events, the confidence weights of environmental parameters are reduced according to the event level index to obtain environmental-physiological dynamic equilibrium parameters containing a confidence weight matrix.

7. The method for dynamic monitoring of volatile components in health-promoting plants according to claim 6, characterized in that, By fusing the dynamic decision data and performing multimodal correction, a spatially distributed corrected volatiles dataset is output, including: The target volatiles label set and dynamic equilibrium parameters are input into the attention mechanism model, and the meteorological disturbance confidence score is output. If the confidence score of meteorological disturbance is lower than the preset threshold, the metabolic pathway feature library verification is triggered, and the table of stress volatile components is output. The biocompatibility of the composition table is verified based on dynamic equilibrium parameters, and a list of bioavailable volatiles is output. A turbulent diffusion model is used to perform spatial compensation on the dynamic equilibrium parameters, and a volatiles dataset is output.

8. The method for dynamic monitoring of volatile components in health-promoting plants according to claim 7, characterized in that, Based on the confidence weight matrix in the dynamic equilibrium parameters, the boundary conditions of the canopy turbulence diffusion model are modified.

Images