Intelligent detection and regulation method based on traditional Chinese medicinal material planting environment
Through the combination of dynamic graph neural network and fuzzy control algorithm, intelligent regulation of the Chinese medicinal materials planting environment is achieved, the problems of multi-factor coupling and physiological signal monitoring are solved, the synthesis efficiency of secondary metabolites and the quality of medicinal materials are improved, and energy consumption is reduced.
Patent Information
- Application Number
- CN202510849334.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-24
- Publication Date
- 2025-10-14
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing technologies lack comprehensive monitoring and integrated consideration of multiple environmental factors in the cultivation of Chinese medicinal materials, and fail to effectively deal with the nonlinear coupling effects between environmental factors and real-time monitoring of plant physiological signals, resulting in inaccurate regulatory strategies and environmental imbalance, affecting the synthesis and quality of medicinal ingredients.
A dynamic graph neural network is used to fuse multimodal perception data to construct a spatiotemporal correlation model between environmental parameters and plant physiological parameters. Multi-objective collaborative control parameters are generated through fuzzy control algorithms, and intelligent control is achieved by combining a closed-loop feedback mechanism to optimize light, temperature, water and fertilizer strategies to adapt to the characteristics of different medicinal materials and environmental changes.
The synthesis efficiency of secondary metabolites of Chinese medicinal materials has been significantly improved, energy consumption has been reduced, and the stability of medicinal material quality and economic benefits have been ensured.
Smart Images

Figure CN120780070A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of Chinese medicinal materials, and in particular to an intelligent detection and control method based on a Chinese medicinal material planting environment. Background Art
[0002] The synthesis of effective ingredients in traditional Chinese medicines is highly dependent on specific environmental conditions, and their quality and yield are directly affected by the dynamic changes in the planting environment. Traditional planting models rely on empirical judgment and are difficult to accurately match the complex environmental parameters required for the growth of traditional Chinese medicines, resulting in insufficient accumulation of effective ingredients or imbalanced biomass. With the large-scale development of the traditional Chinese medicine industry, intelligent detection and regulation of the planting environment have become key technical paths to ensure the quality of medicinal materials. By real-time monitoring of environmental factors and dynamically optimizing regulatory strategies, the synthesis efficiency of secondary metabolites can be significantly improved, while reducing the risk of quality degradation due to environmental fluctuations. However, the existing technical system still has significant limitations when responding to the special growth needs of traditional Chinese medicines.
[0003] Publication No. CN116491381A discloses a method for improving the quality of cultivated Scutellaria baicalensis based on ecological water control measures, addressing the prior art issues of Scutellaria baicalensis germplasm degradation and unstable quality. The method comprises the following steps: after sowing Scutellaria baicalensis, the plant is allowed to grow naturally according to routine daily management; measuring the saturated moisture content of the soil in which the plant is cultivated, and grading the soil moisture environment of the plant based on the soil moisture content into normal moisture, mild drought, moderate drought, and severe drought; monitoring the soil moisture content of the plant in September and October of the following year. When the soil moisture content reaches the moderate drought level, a single application of water at the saturated moisture content is applied. The soil moisture content of the plant is then monitored daily, and the plant is harvested when the soil moisture content reaches the mild or moderate drought level. This invention creatively utilizes artificial water control to significantly increase the content of active ingredients in Scutellaria baicalensis.
[0004] However, this invention mainly focuses on the impact of soil moisture on the quality of Scutellaria baicalensis medicinal materials, and carries out graded regulation of the moisture environment. However, the growth and synthesis of medicinal ingredients of Chinese medicinal materials are not only affected by soil moisture, but also by the combined effects of multiple environmental factors such as temperature, light, CO2 concentration, soil conductivity, ultraviolet rays, etc. The existing technical systems including this patent often lack comprehensive monitoring and comprehensive consideration of these environmental factors; secondly, there is a dynamic coupling effect between environmental parameters, such as high temperature may accelerate the evaporation of soil moisture, and reduced humidity may inhibit the activity of certain microorganisms, indirectly affecting the pH value of the soil, etc. This patent only regulates soil moisture and does not consider the coupling effects of other environmental parameters, which may lead to inaccurate control strategies and trigger chain environmental imbalances. In addition, existing technical systems often lack in-situ perception and real-time monitoring of plant physiological signals, and cannot adjust the control strategy in time according to the physiological state of the plant, resulting in delayed or ineffective control. In summary, in response to the special growth needs of Chinese medicinal materials, the existing technology system, in addition to soil moisture regulation, also needs to be improved and enhanced in terms of comprehensive monitoring of environmental factors, consideration of nonlinear coupling mechanisms, regulation of dynamic coupling effects, physiological signal monitoring and feedback, and the ability to respond to sudden climate events. Summary of the Invention
[0005] To solve the above problems, the present invention provides an intelligent detection and control method based on the Chinese medicinal material planting environment. By integrating multimodal perception data through a dynamic graph neural network, the intelligent control of the Chinese medicinal material planting environment is realized, the light, temperature, water and fertilizer synergy strategy is accurately optimized, the synthesis efficiency of secondary metabolites is significantly improved and energy consumption is reduced, and it adapts to the characteristics of different medicinal materials and environmental changes.
[0006] In order to achieve the above-mentioned purpose, the technical solution of the present invention is as follows: Based on the intelligent detection and control method of the Chinese medicinal material planting environment, the method comprises the following steps:
[0007] S1, multimodal data acquisition: deploying several sensors of different modalities in the planting area to collect environmental parameters and plant physiological parameters in real time;
[0008] S2, dynamic neural network modeling: construct a multi-parameter coupling relationship model based on dynamic graph neural network, perform spatiotemporal correlation analysis on environmental parameters and plant physiological parameters, and generate a dynamic weight matrix of nonlinear coupling strength between parameters;
[0009] S3, fuzzy control and rolling optimization: Based on the dynamic weight matrix, the fuzzy control algorithm is combined to generate multi-objective collaborative control parameters;
[0010] S4, closed-loop feedback and model iteration: drives the actuator to control the environment, and corrects the dynamic weight matrix and control parameters in real time through the online feedback mechanism to form a closed-loop control.
[0011] Further, in S1, the environmental parameters include ultraviolet band intensity collected by a multi-band ultraviolet sensor, soil conductivity spatial distribution obtained by a distributed soil probe, and CO2 concentration.
[0012] The plant physiological parameters include leaf stomatal conductance and leaf temperature gradient.
[0013] Further, in S1, the collection of plant physiological parameters further includes: extracting leaf surface reflectance spectrum through a hyperspectral imaging sensor, and identifying the opening and closing state of stomata by combining a convolutional neural network; and capturing the distribution of leaf temperature gradient by an infrared thermal imager, and reconstructing the overall temperature field change of the plant by a spatial interpolation algorithm.
[0014] Further, in S2, the construction of the dynamic graph neural network model includes: training a graph structure generator according to historical data to dynamically generate connection edge weights between parameter nodes, and analyzing the lag correlation between parameters through time series sliding window analysis; introducing an attention mechanism to feature enhance ultraviolet band intensity, temperature sudden change events, and soil conductivity mutations, and generate adaptive weight distribution of parameter coupling relationship.
[0015] Further, in S3, the implementation of the fuzzy control algorithm includes: establishing a fuzzy rule base based on a multi-objective optimization function, the function taking the synthesis efficiency of secondary metabolites as the optimization objective, and the constraint conditions including environmental parameter coupling stability and actuator energy consumption threshold; quantifying the contribution of each parameter in the dynamic weight matrix by a membership function, and generating executable control instruction sequences by a defuzzification algorithm;
[0016] The multi-objective collaborative control parameter generation adopts a rolling horizon optimization strategy, which adjusts the priority of the control instruction according to the weather forecast data within the next 6 hours; when a sudden mutation in ultraviolet band intensity is detected, the band matching control of the shading net and the light supplement lamp is activated preferentially, and the irrigation operation is delayed to avoid humidity oscillation.
[0017] Further, in S3, when the ultraviolet band intensity increases by more than 50% within 10 minutes, the band matching control of the shading net and the light supplement lamp is activated preferentially.
[0018] The rolling horizon optimization strategy delays the irrigation operation by at least 40 minutes to avoid humidity oscillation according to the weather forecast data within the next 6 hours.
[0019] Further, in S4, the implementation of the online feedback mechanism includes: deploying a metabolite prediction model to predict the accumulation rate of secondary metabolites according to real-time environmental parameters and plant physiological parameters after control; if the predicted rate is lower than the preset threshold, the retraining process of the dynamic graph neural network model is triggered to update the parameter coupling relationship and the control parameters.
[0020] Furthermore, the retraining process further includes: collecting the control feedback data of the last 72 hours, and using the time backpropagation algorithm to update the weight parameters of the graph structure generator; the updated control parameters are eliminated through the lag compensation algorithm to eliminate the actuator response delay, and the compensation coefficient α is 0.6.
[0021] Furthermore, in S4, the closed-loop control includes verification of the parameter coupling relationship between feedback and model iteration, wherein the verification method steps include:
[0022] S41, model validation: set up a control planting area and compare the coupling coefficient difference between UV intensity and stomatal conductance in the dynamic weight matrix between the experimental group and the control group in real time;
[0023] S42, calibration trigger: When the coupling coefficient deviation exceeds 15%, the sensor calibration procedure is automatically executed to recalibrate the conductivity measurement baseline value of the soil probe.
[0024] Furthermore, in S42, when the coupling coefficient deviation exceeds 15%, the following calibration operations are performed synchronously:
[0025] The wavelength of the spectral acquisition accuracy of the hyperspectral imaging sensor was calibrated to control its spectral resolution error within the range of ±2nm; the distributed soil probe was calibrated at three points using standard conductivity fluid to update the baseline parameters of the conductivity-voltage conversion curve; the historical coupling relationship data of the last 24 hours stored in the dynamic graph neural network was cleared, and the reinforcement weight coefficient of the attention mechanism was reinitialized.
[0026] The above scheme has the following beneficial effects:
[0027] 1. This scheme uses hyperspectral imaging, infrared thermal imaging and multi-band sensing technology to synchronously collect environmental parameters (UV band, soil conductivity) and plant physiological signals (stomatal conductance, leaf temperature gradient), breaking through the limitations of traditional methods that only rely on general parameters such as temperature and humidity, and constructing a full-dimensional monitoring system of "environmental stimulation-physiological response-metabolic feedback". Based on the spatiotemporal correlation analysis of dynamic graph neural networks, it analyzes the nonlinear relationship between multiple parameters (such as the hysteresis effect between UV pulses and stomatal opening and closing, and the dynamic impact of temperature and humidity gradients on soil microbial activity), revealing the implicit correlation between environmental factors and the synthesis of secondary metabolites, and providing a scientific basis for precise regulation.
[0028] 2. This solution uses a dynamic weight matrix to quantify parameter coupling strength and incorporates a fuzzy control algorithm to generate multi-objective coordinated commands. For example, this approach links light band switching with irrigation timing. This avoids the cascading imbalances caused by traditional independent threshold controls, such as ventilation-induced CO2 loss and humidity fluctuations inhibiting transpiration. This solution integrates future meteorological data with real-time sensor signals, dynamically adjusting command priorities through a rolling optimization strategy. For example, when a sudden change in UV intensity is detected, the shade net and fill light are prioritized for wavelength matching, delaying irrigation to avoid humidity fluctuations and ensuring precise intervention during the metabolic window.
[0029] 3. This solution deploys a metabolite prediction model to evaluate in real time the impact of regulatory strategies on the synthesis rate of secondary metabolites. When the predicted value deviates from the threshold, it triggers retraining of the dynamic graph neural network and updates the parameter coupling relationship model, achieving a closed-loop control system of "monitoring-regulation-verification-optimization." Through historical data training and online learning mechanisms, it adapts to the metabolic characteristics of different medicinal herb varieties and changes in the growing environment, such as soil aging and climate change, breaking through the generalization bottleneck of traditional fixed strategies.
[0030] 4. This solution is based on the instruction optimization of the multi-objective genetic algorithm. While ensuring the efficiency of secondary metabolite synthesis, it constrains the energy consumption threshold of the actuator, achieving a dual improvement in green planting and economic benefits.
[0031] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] Figure 1 This is a schematic diagram of the method steps of an embodiment of the method for intelligent detection and control of Chinese medicinal materials planting environment according to the present invention. DETAILED DESCRIPTION
[0033] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.
[0034] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.
[0035] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on the specific circumstances.
[0036] The following is further described in detail through specific implementation methods:
[0037] Example:
[0038] As attached Figure 1 As shown: Based on the intelligent detection and control method of the Chinese medicinal material planting environment, the method includes the following steps:
[0039] S1, multimodal data acquisition: Several sensors of different modalities are deployed in the planting area to collect environmental parameters and plant physiological parameters of the planting area in real time. In S1, environmental parameters include UV band intensity collected by multi-band UV sensors, spatial distribution of soil conductivity and CO2 concentration obtained by distributed soil probes, and plant physiological parameters include leaf stomatal conductance and leaf surface temperature gradient.
[0040] S2, dynamic neural network modeling: construct a multi-parameter coupling relationship model based on a dynamic graph neural network. The construction of the dynamic graph neural network model includes: training the graph structure generator according to historical data, dynamically generating the connection edge weights between parameter nodes, and analyzing the lagged correlation between parameters through the time series sliding window of the connection edge weights; introducing the attention mechanism to enhance the features of ultraviolet band intensity, temperature sudden change events and soil conductivity mutations, performing spatiotemporal correlation analysis on environmental parameters and plant physiological parameters, and generating a dynamic weight matrix of the nonlinear coupling strength between each parameter.
[0041] S3, Fuzzy Control and Rolling Optimization: Based on the dynamic weight matrix, a fuzzy control algorithm is combined to generate multi-objective coordinated control parameters. The implementation of the fuzzy control algorithm includes: establishing a fuzzy rule base based on a multi-objective optimization function. The function uses the efficiency of secondary metabolite synthesis as the optimization objective, and the constraints include the coupling stability of environmental parameters and the energy consumption threshold of the actuator; quantifying the contribution of each parameter in the dynamic weight matrix through membership functions, and using a defuzzification algorithm to generate an executable control instruction sequence;
[0042] The generation of multi-objective collaborative control parameters adopts the introduction of a rolling time domain optimization strategy, which adjusts the priority of control instructions according to the weather forecast data in the next 6 hours. When the ultraviolet band intensity is detected to increase by more than 50% within 10 minutes, the band matching control of the shading net and the fill light is activated first, and the irrigation operation is delayed to avoid humidity oscillation. The rolling time domain optimization strategy delays the irrigation operation for at least 40 minutes based on the weather forecast data in the next 6 hours to avoid humidity oscillation.
[0043] S4, closed-loop feedback and model iteration: drives the actuator to control the environment, and uses the online feedback mechanism to modify the dynamic weight matrix and control parameters in real time to form a closed-loop control;
[0044] The implementation of the online feedback mechanism includes: deploying a metabolite prediction model to predict the accumulation rate of secondary metabolites based on real-time environmental parameters and regulated plant physiological parameters; if the predicted rate falls below a preset threshold, the dynamic graph neural network model is retrained to update the parameter coupling relationship and regulation parameters;
[0045] The retraining process further includes: collecting the control feedback data of the last 72 hours and using the time backpropagation algorithm to update the weight parameters of the graph structure generator; the updated control parameters are then used to eliminate the actuator response delay through the hysteresis compensation algorithm, with the compensation coefficient α set to 0.6;
[0046] Closed-loop control includes verification of the parameter coupling relationship between feedback and model iteration, where the verification method steps include:
[0047] S41, model validation: set up a control planting area and compare the coupling coefficient difference between UV intensity and stomatal conductance in the dynamic weight matrix between the experimental group and the control group in real time;
[0048] S42, calibration trigger: When the coupling coefficient deviation exceeds 15%, the sensor calibration procedure is automatically executed to recalibrate the conductivity measurement baseline value of the soil probe and simultaneously perform the following calibration operations:
[0049] The wavelength of the spectral acquisition accuracy of the hyperspectral imaging sensor was calibrated to control its spectral resolution error within the range of ±2nm; the distributed soil probe was calibrated at three points using standard conductivity fluid to update the baseline parameters of the conductivity-voltage conversion curve; the historical coupling relationship data of the last 24 hours stored in the dynamic graph neural network was cleared, and the reinforcement weight coefficient of the attention mechanism was reinitialized.
[0050] The specific implementation process is as follows: Since the existing detection and control of the Chinese medicinal material planting environment are mostly based on general parameters (such as temperature, humidity, and light), the synthesis of the effective components of Chinese medicinal materials is affected by nonlinear environments such as the diurnal temperature gradient, such as ginseng, salvia miltiorrhiza, and coptis chinensis. This embodiment takes ginseng as an example. Since the accumulation of ginsenosides is greatly affected by environmental factors such as temperature changes and light conditions, in order to verify the effect of the method adopted in this scheme on improving the efficiency of ginsenoside synthesis, the following experiments are designed:
[0051] 1. Control settings:
[0052] Experimental group: This scheme was used for environmental control;
[0053] Control group: Traditional threshold control was used, with only temperature, humidity and light intensity being independently controlled.
[0054] 2. Experimental period:
[0055] During the critical growth period of ginseng from April to August, leaf samples were collected every 15 days to test the ginsenoside content.
[0056] 3. Implementation process:
[0057] Control group: according to traditional monitoring methods.
[0058] Experimental group:
[0059] S1, multimodal data collection: Environmental parameters are collected by deploying a multi-band UV sensor every 10 m2 to detect UV-A / UV-B intensity. Soil conductivity meters are distributed at a soil depth of 20 cm. CO2 concentration sensors are 1.5 cm above the ground. Temperature and humidity sensors are arranged in three gradient layers every 5 m2, at 0.5 m, 1.0 m, and 1.5 m above the ground.
[0060] A hyperspectral imaging sensor was suspended 1m above the plant with a resolution of 0.5nm. The leaf reflectance spectrum was collected every 30 minutes from 10:00 to 14:00 every day. The infrared thermal imager simultaneously captured the leaf surface temperature distribution with a spatial resolution of 0.1℃.
[0061] The hyperspectral data was input into the pre-trained ResNet-18 model, the 750-780nm band features were extracted, and the probability of stomatal opening and closing was output. When the opening degree was greater than 70%, it was judged to be in an open state. The Kriging spatial interpolation algorithm was used on the infrared thermal imaging data to generate a temperature gradient thermodynamic map of the entire plant.
[0062] S2, dynamic graph neural network modeling: Input the environmental parameters of the ginseng planting base in the past three years and the ginsenoside content data of the corresponding harvest period, construct parameter nodes such as temperature, UV-B intensity, and stomatal conductance, and calculate the lagged correlation coefficient through a sliding window. The sliding window length is 24 hours and the step length is 1 hour. For example, it is found that the impact of the sudden increase in UV-B intensity on stomatal conductance has a 4-hour lag effect.
[0063] When the temperature drops by more than 3°C within 5 minutes or the UV-B intensity increases by more than 20 μW / cm within 10 minutes 2 , the weight of the related nodes is automatically increased by 1.5 times.
[0064] S3, fuzzy control and rolling optimization: To optimize the ginsenoside synthesis rate to ≥0.12 mg / g·day, the following constraints were constructed using a fuzzy rule base: temperature-humidity coupling stability |ΔT / ΔRH|<0.3°C / %, and actuator energy consumption ≤2.5 kW·h / day.
[0065] Generate control instructions, based on the detection of a sudden increase in UV-B intensity to 85μW / cm at 11:30 on a certain day 2 For example, the UV-B intensity exceeds the threshold of 50μW / cm 2 , activate rolling time domain optimization, call the weather forecast for the next 6 hours (sunny, ultraviolet rays continue to increase), start the shading net to ensure the shading rate is 60%, and simultaneously open the band of 580-660nm and the intensity of 200μmol / m 2 / s fill light, delay the irrigation plan originally scheduled for 12:00 to 14:30 to avoid transpiration inhibition caused by a sudden increase in humidity, adjust the ventilation unit frequency to 45Hz, and maintain the CO2 concentration at 450-500ppm.
[0066] S4, closed-loop feedback and model iteration: Using the LSTM model, real-time environmental parameters (UV-B intensity, temperature and humidity gradient) and post-regulation stomatal conductance data are input to predict the ginsenoside accumulation rate of 0.15 mg / g·day in the next 24 hours (higher than the threshold of 0.12 mg / g·day). When the predicted value is lower than the threshold of 0.10-0.11 mg / g·day set by the execution feedback module for three consecutive days, the data of the last seven days are collected and the dynamic graph neural network training set is updated. After retraining, the model shows that the coupling weight of soil conductivity and leaf temperature gradient has increased by 22%, generating new control instructions: adjusting the irrigation pH value from 6.8 to 6.5 and increasing the frequency of trace element replenishment (every 48 hours → every 36 hours).
[0067] 4. Experimental results:
[0068] index Experimental group control group Improvement effect Ginsenoside content (mg / g) 4.82±0.31 3.75±0.28 ↑28.5% Environmental control response delay ≤ 5 minutes >30 minutes ↓84% Energy consumption per unit area (kW·h) 2.3 3.1 ↓25.8%
[0069] V. Conclusion
[0070] As can be seen from the experimental results, the technical architecture of dynamic coupling modeling - metabolism-guided regulation - closed-loop feedback optimization has systematically solved the three core problems of unclear parameter coupling mechanism, rigid regulation logic, and insufficient response time in the traditional Chinese medicinal material planting environment control. From "environmental stability maintenance" to "metabolic promotion", the regulation strategy is directly linked to the quality of medicinal materials, and the coordinated control of parameters inhibits environmental imbalance and ensures the steady state of plant physiological state.
[0071] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will readily appreciate that other variations or modifications based on the above descriptions are possible. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.
Claims
1. Based on the intelligent detection and control method of Chinese medicinal materials planting environment, it is characterized by: The following steps are involved: S1, multimodal data acquisition: deploying several sensors of different modalities in the planting area to collect environmental parameters and plant physiological parameters in real time; S2, dynamic neural network modeling: construct a multi-parameter coupling relationship model based on dynamic graph neural network, perform spatiotemporal correlation analysis on environmental parameters and plant physiological parameters, and generate a dynamic weight matrix of nonlinear coupling strength between parameters; S3, fuzzy control and rolling optimization: Based on the dynamic weight matrix, the fuzzy control algorithm is combined to generate multi-objective collaborative control parameters; S4, closed-loop feedback and model iteration: drives the actuator to control the environment, and corrects the dynamic weight matrix and control parameters in real time through the online feedback mechanism to form a closed-loop control.
2. The method for intelligent detection and control of the Chinese medicinal material planting environment according to claim 1, characterized in that: In S1, environmental parameters include the UV band intensity collected by a multi-band UV sensor, the spatial distribution of soil electrical conductivity and CO2 concentration obtained by distributed soil probes; Plant physiological parameters include leaf stomatal conductance and leaf surface temperature gradient.
3. The method for intelligent detection and control of the Chinese medicinal material planting environment according to claim 2, characterized in that: In S1, the collection of plant physiological parameters further includes: extracting the leaf surface reflectance spectrum through a hyperspectral imaging sensor, combining it with a convolutional neural network to identify the opening and closing status of stomata; using an infrared thermal imager to capture the temperature gradient distribution on the leaf surface, and reconstructing the overall temperature field changes of the plant through a spatial interpolation algorithm.
4. The method for intelligent detection and control of the Chinese medicinal material planting environment according to claim 3, characterized in that: In S2, the construction of the dynamic graph neural network model includes: training the graph structure generator based on historical data, dynamically generating the connection edge weights between parameter nodes, and analyzing the lagged correlation between parameters through the time series sliding window; introducing the attention mechanism to enhance the features of ultraviolet band intensity, temperature sudden change events and soil conductivity mutations, and generating adaptive weight distribution of parameter coupling relationships.
5. The method for intelligent detection and control of Chinese medicinal material planting environment according to claim 4, characterized in that: In S3, the implementation of the fuzzy control algorithm includes: establishing a fuzzy rule base based on a multi-objective optimization function, with the secondary metabolite synthesis efficiency as the optimization objective, and constraints including the coupling stability of environmental parameters and the energy consumption threshold of the actuator; quantifying the contribution of each parameter in the dynamic weight matrix through membership functions, and using a defuzzification algorithm to generate an executable control instruction sequence; The generation of multi-objective collaborative control parameters adopts a rolling time domain optimization strategy, which adjusts the priority of control instructions according to the weather forecast data within the next 6 hours. When a sudden change in the ultraviolet band intensity is detected, the band matching control of the shading net and the fill light is activated first, and the irrigation operation is delayed to avoid humidity oscillation.
6. The method for intelligent detection and control of Chinese medicinal material planting environment according to claim 5, characterized in that: In S3, when it is detected that the UV band intensity increases by more than 50% within 10 minutes, the band matching control of the shading net and the fill light is activated first; The rolling time domain optimization strategy delays irrigation operations for at least 40 minutes based on the weather forecast data for the next 6 hours to avoid humidity oscillations.
7. The method for intelligent detection and control of Chinese medicinal material planting environment according to claim 6, characterized in that: In S4, the implementation of the online feedback mechanism includes: deploying a metabolite prediction model to predict the accumulation rate of secondary metabolites based on real-time environmental parameters and regulated plant physiological parameters; if the predicted rate is lower than the preset threshold, the retraining process of the dynamic graph neural network model is triggered to update the parameter coupling relationship and regulation parameters.
8. The method for intelligent detection and control of Chinese medicinal material planting environment according to claim 7, characterized in that: The retraining process further includes: collecting the control feedback data of the last 72 hours, and using the time backpropagation algorithm to update the weight parameters of the graph structure generator; the updated control parameters are eliminated through the lag compensation algorithm to eliminate the actuator response delay, and the compensation coefficient α is 0.
6.
9. The method for intelligent detection and control of Chinese medicinal material planting environment according to claim 8, characterized in that: In S4, the closed-loop control includes verification of the parameter coupling relationship between feedback and model iteration, wherein the verification method steps include: S41, model validation: set up a control planting area and compare the coupling coefficient difference between UV intensity and stomatal conductance in the dynamic weight matrix between the experimental group and the control group in real time; S42, calibration trigger: When the coupling coefficient deviation exceeds 15%, the sensor calibration procedure is automatically executed to recalibrate the conductivity measurement baseline value of the soil probe.
10. The method for intelligent detection and control of Chinese medicinal material planting environment according to claim 9, characterized in that: In S42, when the coupling coefficient deviation exceeds 15%, the following calibration operations are performed synchronously: The wavelength of the spectral acquisition accuracy of the hyperspectral imaging sensor was calibrated to control its spectral resolution error within the range of ±2nm; the distributed soil probe was calibrated at three points using standard conductivity fluid to update the baseline parameters of the conductivity-voltage conversion curve; the historical coupling relationship data of the last 24 hours stored in the dynamic graph neural network was cleared, and the reinforcement weight coefficient of the attention mechanism was reinitialized.
Citation Information
Patent Citations
Method for improving quality of cultivated baical skullcap root medicinal material based on water ecological regulation and control measures
CN116491381A
Cited By
Greenhouse planting regulation and control method for precise application of plant growth regulator
CN121028938A
