Method for precise regulation of growth of polygonatum franchetii by fusing multi-source information

Abstract

The application discloses a Yunnan rhizoma polygonati growth precise regulation and control method fusing multi-source information, and particularly relates to the technical field of intelligent planting of traditional Chinese medicinal materials; multi-source environment information of a target planting area of Yunnan rhizoma polygonati is collected to construct a standardized environment characteristic vector; a weight factor group of environment factors in each stage is calculated based on a growth period physiological response model; a regulation and control influence index is constructed by fusing the characteristic vector and the weight factor, and compared with a growth period sensitive threshold value to identify a high-risk regulation and control area; a regulation and control position and a target factor are determined in combination with historical yield and effective component trends; corresponding regulation and control measure templates in a planting database are matched and precise intervention is implemented to realize closed-loop feedback optimization; the method can realize dynamic perception and intelligent regulation and control of the whole-cycle growth state of Yunnan rhizoma polygonati, improve yield stability and quality consistency, and has good application value.

Description

Technical Field

[0001] This invention relates to the field of intelligent cultivation technology for Chinese medicinal herbs, specifically to a method for precise regulation of the growth of Polygonatum yunnanensis that integrates multi-source information. Background Technology

[0002] Polygonatum kingianum is an important traditional Chinese medicine used for both medicinal and edible purposes. Its underground rhizome has significant tonic effects and is widely used in traditional Chinese medicine preparations, functional foods, and health products. In recent years, driven by market demand, the planting area of ​​Polygonatum kingianum has expanded rapidly. However, due to its highly sensitive physiological characteristics to the growing environment (such as soil moisture, temperature, light, and altitude), the yield and quality of Polygonatum kingianum vary significantly in different regions.

[0003] Currently, the cultivation of *Polygonatum yunnanense* in Yunnan generally relies on manual, experience-based management, failing to achieve precise identification and dynamic intervention of key growth regulators. This is particularly problematic in the following scenarios: in the high-altitude terraced areas of western Yunnan, there is strong sunlight, large diurnal temperature variations, and poor soil water retention; while in the humid and hot hilly areas of southern Yunnan, there is a high incidence of root rot and the hidden risk of waterlogging affecting growth. If the growth environment cannot be precisely controlled based on the microclimate of different regions, multiple environmental factors, and the physiological responses of *Polygonatum yunnanense* at different growth stages, it will lead to large-scale yield reduction, root and stem variations, and even a severe decline in the content of effective components. Summary of the Invention

[0004] The purpose of this invention is to provide a method for precise regulation of the growth of Polygonatum yunnanensis that integrates multi-source information, so as to overcome the shortcomings of the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for precise regulation of the growth of Polygonatum yunnanense that integrates multi-source information, comprising:

[0006] Obtain multi-source environmental information set M for the target planting area of ​​Polygonatum yunnanense, including soil moisture content, diurnal temperature range, effective accumulated temperature, light intensity, atmospheric humidity, soil pH value, historical disease incidence rate and topographic slope of the target area;

[0007] The information set M is normalized to obtain the standardized environmental feature vector V;

[0008] Based on the physiological response model of Polygonatum yunnanense at different growth stages, the weight factor group W of the standardized environmental feature vector V at each growth stage was calculated.

[0009] Based on the calculated weight factor group W and the standardized environmental feature vector V, feature fusion is performed to construct the growth regulation influence index E of Polygonatum yunnanense.

[0010] The calculated growth regulation impact index E of Polygonatum yunnanense is compared with the sensitivity threshold R of the current reproductive stage. If E < R, the current reproductive period is identified as being in a high-risk regulation zone.

[0011] Based on the identification results and spatial distribution data within the planting area, combined with historical yield and effective ingredient concentration trends, the precise control location L and the control target factor F were determined.

[0012] Based on the regulatory location L and the regulatory target factor F, the corresponding regulatory measure template library S is matched from the planting database;

[0013] The control measure template S was applied to location L to achieve precise control over the growth of Polygonatum yunnanense.

[0014] Preferably, the physiological response model based on different growth stages of Polygonatum yunnanense calculates the weight factor group W of the standardized environmental feature vector V at each growth stage, including:

[0015] Based on the growth process of Polygonatum yunnanense, the growth cycle is divided into four stages: germination stage, elongation stage, swelling stage, and maturity stage, and corresponding physiological response model sets are established for each stage. ;

[0016] The obtained standardized environmental feature vector V is input into each physiological response model. In this study, the sensitivity coefficients of various environmental factors to target physiological parameters during different reproductive stages were determined by multifactor regression analysis.

[0017] Based on the sensitivity coefficient, each environmental factor is weighted and normalized to obtain the weight factor group W for each growth stage.

[0018] Preferably, the construction of the growth regulation influence index E of Polygonatum yunnanense includes:

[0019] The obtained weight factor group W is linearly weighted and fused with the standardized environmental feature vector V in a one-to-one correspondence manner to obtain the initial value of environmental suitability E′.

[0020] By introducing key physiological parameter thresholds for different growth stages of Polygonatum yunnanense, and determining a standardized regulatory index function f(x) based on the current growth stage, a nonlinear mapping is performed on the initial value of environmental suitability E′ to obtain the growth regulation influence index E of Polygonatum yunnanense.

[0021] Spatial interpolation was performed on the E-values ​​of the growth regulation influence index of Polygonatum yunnanense calculated from multiple regional samples to generate a distribution map of regulation risk in Polygonatum yunnanense planting areas.

[0022] Preferably, the step of comparing the calculated growth regulation influence index E of Polygonatum yunnanense with the sensitivity threshold R of the current reproductive stage includes:

[0023] Construct sensitive thresholds corresponding to the four reproductive stages of Polygonatum yunnanense. Each threshold is determined based on the minimum suitability index of environmental conditions in historical high-quality yield samples;

[0024] Based on the current monitoring time, the reproductive stage of *Polygonatum yunnanense* is matched, and the corresponding sensitive threshold is selected. ;

[0025] The obtained growth regulation influence index E of Polygonatum yunnanense and Compare, when E is less than At that time, it was determined to be in a high-risk control zone.

[0026] Preferably, determining the precise control position L and the control target factor F includes:

[0027] The high-risk area labels are spatially overlaid with the spatial grid data within the planting area to extract areas where abnormal environmental features are clustered as candidate control locations.

[0028] By retrieving the corresponding yield data and effective component concentration data of the candidate locations over the years, a spatial-yield-quality correlation model is constructed to identify sensitive areas of yield decline or component abnormalities.

[0029] Analyze the dominant deviation factors in the current standardized environmental feature vector V of the sensitive area, and combine them with the reproductive period response weight factor group W to determine the regulatory target factor F that has the greatest impact and needs priority intervention.

[0030] Output the precise control position L and the target factor F.

[0031] Preferably, the step of matching the corresponding regulatory measure template library S from the planting database based on the regulatory location L and the regulatory target factor F includes:

[0032] The output target factor F is matched with the intervention factor index in the standardized planting database by keywords to select a set of regulatory measures templates associated with the target factor.

[0033] Based on the soil type, planting density, and historical control feedback results of the area where the control location L is located, the initial screening templates are screened a second time and prioritized.

[0034] Based on the current reproductive period and the urgency of regulation, the optimal regulatory measure template S is selected from the ranked templates using a recommendation algorithm;

[0035] Output a list of control strategies S bound to location L, including soil watering frequency adjustment, light shading ratio setting, and disease prevention spraying plan.

[0036] Preferably, applying the control measure template S to location L includes:

[0037] The control measure template S is converted into control instructions that the equipment can recognize, generating a set of control operation parameters that includes control objectives, operation time, and execution path;

[0038] Based on the geographical coordinates of location L and the operating radius of the operating equipment, the optimal path is planned and the corresponding agricultural execution units are scheduled.

[0039] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0040] 1. This invention integrates multi-source heterogeneous environmental data and combines it with a specific physiological response model of Polygonatum yunnanense during its growth period to construct a standardized feature vector and dynamic weight factor group. It further generates a regulatory influence index and introduces a stage-specific sensitive threshold judgment, thereby realizing a quantitative assessment and dynamic early warning of the environmental suitability status of Polygonatum yunnanense throughout its entire growth cycle. This invention breaks through the problems of relying on experience-based judgment, delayed response, and extensive regulation in traditional planting, and significantly improves the scientificity, accuracy, and timeliness of regulatory decisions.

[0041] 2. This invention constructs a spatial-yield-quality correlation model of planting area, intelligently identifies the control location L and target factor F, and recommends the optimal control measure template S based on the planting database. This enables automatic deployment of control tasks, optimization of operation paths, and feedback closed-loop correction, truly realizing an intelligent closed-loop control system for precise control "according to location, time, and factors". It effectively improves the yield stability and consistency of effective components of Polygonatum yunnanense, and has significant promotion value and industrial application prospects. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0043] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0044] 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.

[0045] For examples, please refer to Figure 1As shown in this embodiment, the method for precise regulation of the growth of Polygonatum yunnanense that integrates multi-source information includes:

[0046] Obtain multi-source environmental information set M for the target planting area of ​​Polygonatum yunnanense, including soil moisture content, diurnal temperature range, effective accumulated temperature, light intensity, atmospheric humidity, soil pH value, historical disease incidence rate and topographic slope of the target area;

[0047] The information set M is normalized to obtain the standardized environmental feature vector V;

[0048] Based on the physiological response model of Polygonatum yunnanense at different growth stages, the weight factor group W of the standardized environmental feature vector V at each growth stage was calculated.

[0049] Based on the calculated weight factor group W and the standardized environmental feature vector V, feature fusion is performed to construct the growth regulation influence index E of Polygonatum yunnanense.

[0050] The calculated growth regulation impact index E of Polygonatum yunnanense is compared with the sensitivity threshold of the current reproductive stage. If E < R, the current reproductive period is identified as being in a high-risk regulation zone.

[0051] Based on the identification results and spatial distribution data within the planting area, combined with historical yield and effective ingredient concentration trends, the precise control location L and the control target factor F were determined.

[0052] Based on location L and target factor F, the corresponding control measure template library S is matched from the planting database, including soil watering frequency adjustment, light shading ratio setting and disease prevention spraying plan.

[0053] The control measure template S was applied to location L to achieve precise control over the growth of Polygonatum yunnanense.

[0054] In a preferred embodiment of the present invention, a multi-source environmental information set M of the target planting area of ​​*Polygonatum yunnanense* is first obtained to comprehensively characterize the external environmental factors affecting the growth of *Polygonatum yunnanense* in this area. The multi-source environmental information set M includes the following eight categories of key parameters:

[0055] Soil moisture content: By setting soil moisture sensors at different depths, the moisture content data of the top layer (0-20cm) and root layer (20-60cm) of the planting area are collected in real time to reflect the soil's water holding capacity and the risk of drought or waterlogging.

[0056] Diurnal temperature range: By deploying micro-weather station equipment, the maximum difference between daytime and nighttime temperatures is obtained. This parameter directly affects the photosynthetic efficiency and respiratory metabolic rhythm of Polygonatum yunnanensis.

[0057] Effective accumulated temperature: Based on continuous temperature sampling data, the effective accumulated temperature is calculated by accumulating the daily average temperature according to the starting temperature threshold (e.g., 10℃) of the growth period of Polygonatum yunnanense, reflecting the impact of regional heat resources on growth progress;

[0058] Light intensity: The intensity of solar radiation per unit time is obtained through a light sensor to quantify the light energy input intensity of the area, and actual influencing factors such as shading and slope aspect are taken into account.

[0059] Atmospheric humidity: Collect the water vapor content in the air to analyze the strength of regional transpiration and the possibility of high humidity inducing diseases;

[0060] Soil pH: The soil acidity and alkalinity of different plots in the planting area were determined by potentiometric method or chemical analysis to identify the degree of impact on root absorption and microbial activity;

[0061] Historical disease incidence: Based on historical planting data, the frequency and intensity of major diseases of Polygonatum yunnanense (such as root rot and leaf blight) in similar climatic years are statistically analyzed and quantified to provide a basis for disease risk early warning.

[0062] Target area topographic slope: Calculate the slope of the plot using a digital elevation model (DEM) generated by a high-precision UAV or remote sensing image to determine the tendency of water accumulation or loss, thereby assessing the soil and water conservation capacity and suitability level.

[0063] The aforementioned multi-source environmental factors include both real-time dynamic monitoring data and historical static reference information. By associating them with the spatial location of the plot, a high-precision spatiotemporal environmental parameter database is formed, providing basic support for subsequent modeling of the growth regulation of Polygonatum yunnanensis.

[0064] In a preferred embodiment of the present invention, after obtaining the multi-source environmental information set M of the target planting area of ​​Polygonatum yunnanense, in order to eliminate the influence of differences in physical dimensions, value ranges and magnitudes among different environmental factors and improve the effectiveness of model fusion and parameter analysis, the environmental factor data in the information set M are normalized to construct a unified standard data input format, namely, the standardized environmental feature vector V.

[0065] Specifically, for each original environmental factor value in the multi-source environmental information set M (i=1,2,...,8) are normalized using one of the following methods respectively:

[0066] Linear normalization (Min-Max standardization): Applicable to continuous variables within a stable range, such as soil moisture content and light intensity. The normalization formula is: ;in, and These represent the minimum and maximum values ​​of the i-th environmental factor in historical data, respectively. Z-Score standardization (zero-mean normalization): suitable for factors with large distribution fluctuations and requiring preservation of distribution characteristics, such as diurnal temperature range and accumulated temperature. The normalization formula is: ;in, Let be the mean of the i-th environmental factor. Its standard deviation.

[0067] For discontinuous variables such as historical disease incidence rates, discrete grading mapping can be performed based on expert experience or disease level thresholds, for example, dividing them into three levels: low (0-0.3), medium (0.3-0.7), and high (0.7-1). The normalized environmental factors constitute a dimensionless, standardized vector V of the same scale.

[0068] The standardized environmental feature vector V can be used for subsequent steps such as calculating physiological response weight factors and constructing regulatory influence indices, ensuring the consistency of algorithm inputs and the scientific nature of result interpretation.

[0069] It should be noted that the data boundary values ​​used in the normalization process... , , and Based on local historical monitoring data and long-term climate databases, and dynamically updated via a sliding window during actual deployment, it adapts to sudden weather changes and extreme environmental events.

[0070] In a preferred embodiment of the present invention, considering the significant differences in the response of Polygonatum yunnanense to environmental factors at different growth stages, a corresponding physiological response model is constructed based on a standardized environmental feature vector V, and the weight factor group W of environmental factors at each stage is calculated to achieve targeted regulation. This process specifically includes the following steps:

[0071] First, based on the actual agronomic characteristics of the growth cycle of Polygonatum yunnanense and field observation data over the years, its growth period is divided into four key stages: germination stage ( ): From the time the underground buds begin to sprout until the above-ground stems and leaves unfold; elongation period ( ): Rapid growth of above-ground stems and leaves, increased photosynthetic efficiency; bulking stage ( ): The underground rhizomes expand rapidly, and biomass accumulation accelerates; during the maturity period ( The above-ground parts gradually cease growth, while the underground rhizomes mature. An independent physiological response model is established for each stage. (i=1~4) are used to describe the mathematical relationship between the target growth parameters of *Polygonatum yunnanensis* and environmental factors during this stage. The model uniformly adopts a multiple linear regression model, whose basic structure is as follows: ;in: This represents the target physiological parameters for stage i, such as aboveground stem length, leaf area index, root and stem diameter, and dry matter accumulation. β0~β8 are the factors in the aforementioned standardized environmental feature vector V; β0~β8 are the regression coefficients to be trained; ε is the random perturbation term.

[0072] The model training process was based on the actual field monitoring dataset of Polygonatum yunnanensis, and stepwise regression was used to screen variables to ensure the stability and interpretability of the model.

[0073] The obtained standardized environmental feature vector V is then sequentially input into the model for each reproductive stage. Extract the partial derivative coefficients of each factor with respect to the target physiological parameters. The absolute value of the β coefficients was used as the sensitivity coefficient. The sensitivity coefficient reflects the intensity of the influence of a specific environmental factor on the physiological response of *Polygonatum yunnanensis* at that stage. To ensure that the sensitivity of different factors is compared on a uniform scale, all β coefficients were normalized to their absolute values ​​as follows:

[0074] The normalized weighting factor is obtained by taking the absolute value of the sensitivity coefficient of each environmental factor and dividing it by the sum of the absolute values ​​of the sensitivity coefficients of all factors. That is, the weight of each environmental factor is equal to the absolute value of its sensitivity coefficient ÷ the sum of the absolute values ​​of the sensitivity coefficients of all factors.

[0075] The above processing yields the environmental factor weight vector for the current reproductive stage. ,in This represents the importance of the i-th environmental factor at this stage, with a sum of 1. This weight vector can be used as a weighted fusion input to subsequently construct the growth regulation influence index E of *Polygonatum yunnanense*, in order to model the differential response of different growth stages to environmental changes.

[0076] It is worth noting that the weight factor group W not only has dynamism, but also regional adaptability. During model training, parameters can be recalibrated for different ecological types of regions to enhance the geographical adaptability of practical applications.

[0077] In a preferred embodiment of the present invention, based on the standardized environmental feature vector V obtained in the aforementioned steps and its corresponding weight factor group W, further feature fusion processing is performed to construct the growth regulation influence index E of *Polygonatum yunnanense* during its current growth stage, which is used to dynamically assess the comprehensive influence of the current environment on its growth status. This process specifically includes the following sub-steps:

[0078] The obtained standardized environmental feature vector V={T1′,T2′,...,T8′} is processed element-by-element with the calculated weight factor group W={w1,w2,...,w8}. A linear weighted fusion algorithm is then used to obtain the initial environmental suitability value E′, which is expressed as: the initial environmental suitability value E′ equals the sum of the products of the standardized value of each environmental factor and its corresponding weight value. That is: Among them, the higher the value of E′, the closer the current comprehensive environment is to the optimal suitable state for the target growth stage of Polygonatum yunnanensis.

[0079] Since different growth stages exhibit nonlinear response characteristics to environmental changes, a standardized regulatory index function f(x) is further introduced on the basis of E′, and a nonlinear mapping is performed on it to construct a growth regulation influence index E of Polygonatum yunnanense that has more physiological response characteristics.

[0080] The function f(x) can be established by fitting historical physiological data of Polygonatum yunnanense, preferably using the Sigmoid function or an exponential activation function, for example: Where: e is the natural constant; K is the curve slope adjustment coefficient, representing the sensitivity of Polygonatum yunnanense to environmental deviations at this stage; The environmental suitability threshold for this growth stage represents the minimum environmental suitability required for *Polygonatum yunnanense* to achieve normal growth at this stage. It can be obtained by historical regression of field measured yield and environmental data, and has both stage and regional characteristics.

[0081] To enhance the application value of this regulatory index in large-scale planting areas, spatial interpolation algorithms (such as the Inverse Distance Weighted Method (IDW) or Kriging interpolation) were used to spatially expand the growth regulation influence index E calculated from multiple sampling points, generating a complete regulatory risk distribution map of the Yunnan Polygonatum planting area. Specific implementation methods include:

[0082] Environmental factor data from multiple monitoring locations within the region are acquired using a fixed grid or GPS-based point deployment. The E-value is calculated for each monitoring point. Geospatial interpolation is then performed on all E-values ​​to generate a two-dimensional control index layer with color gradients, used to visualize high-risk and low-risk areas. This layer can be overlaid with an agricultural GIS system to help growers quickly identify potential stress hotspots and achieve precise control location positioning.

[0083] In a preferred embodiment of the present invention, to achieve dynamic identification of growth risks at different growth stages of Polygonatum yunnanense, a risk status determination is made based on the obtained growth regulation influence index E of Polygonatum yunnanense, combined with the current growth stage. This step specifically includes:

[0084] First, a set of sensitive thresholds for the four key reproductive stages of Polygonatum yunnanense was constructed. ,in: Corresponding to the budding stage; Corresponding elongation period; Corresponding to the expansion period; Corresponding to the maturity stage.

[0085] Each threshold This represents the minimum regulatory influence index value required for *Polygonatum yunnanense* to maintain normal growth during the corresponding stage. It is set based on measured data from high-quality production areas over the years. The specific implementation method is as follows: Historical planting samples with achieved yield targets and normal physiological conditions are selected; for each growth stage, the minimum stable value of the *Polygonatum yunnanense* growth regulation influence index E is calculated in these samples; this value is then taken as the specific sensitivity threshold for that stage. This threshold is incorporated into the threshold library R. This dynamic threshold library is regionally adaptable and stage-specific, and can be updated regularly according to different ecological planting areas.

[0086] Based on planting date, accumulated temperature, or field observation indicators (such as stem length and leaf area index), determine the current growth stage of *Polygonatum yunnanense*, and automatically extract the corresponding stage threshold by indexing and matching it with the threshold library R. .

[0087] For example: if the accumulated temperature exceeds 500 degrees Celsius per day and the aboveground parts enter a rapid growth phase, then the matching stage is the elongation phase, and the threshold is extracted. ;

[0088] If the underground rhizome begins to swell, it is matched to the swelling stage, and the corresponding threshold is [value missing]. .

[0089] The fertility period determination algorithm can be implemented based on logical rules or machine learning models to ensure the accuracy and real-time nature of the stage division.

[0090] The calculated growth regulation impact index E of *Polygonatum yunnanense* is compared with the sensitivity threshold of the current growth stage to determine whether a high-risk regulation state has been entered. The comparison method is as follows: if the growth regulation impact index E of *Polygonatum yunnanense* is less than the threshold... This indicates that the overall environmental suitability of the current region is insufficient, and the physiological state of *Polygonatum yunnanense* may be restricted, thus classifying the current region as a high-risk control area. For example: if the current stage is the expansion phase (matching...) =0.68), while the growth regulation influence index of Polygonatum yunnanense E=0.54, therefore E< The system outputs a risk status of "high".

[0091] like If the condition is met, the system is judged to be in a "normal" state. This logic can be implemented by setting a conditional expression, and it can run on edge devices or cloud systems.

[0092] The above comparison results are converted into standardized regulatory risk level labels, for example, setting three states: high risk ( Medium risk Low risk or normal (); This label serves as the output of the judgment result and is simultaneously transmitted to the subsequent regulation area identification module to determine the intervention location L and the regulation factor F, thus achieving a closed-loop information chain.

[0093] In this invention, based on the identification results and spatial distribution data within the planting area, combined with historical yield and active ingredient concentration trends, the precise control location L and the control target factor F are determined, specifically including:

[0094] The risk level labels (high, medium, low) are spatially overlaid with the standardized spatial grid data of the planting area. The spatial grid is based on a geographic information system (GIS) and divides the entire planting area into regular grid units of equal area (e.g., one grid every 10 meters × 10 meters). The data is then bound to the environmental factor data of each grid unit.

[0095] By selecting grid cells with a risk level of "high" and combining them with adjacent high-risk grid cells to form cluster blocks, spatial regions with highly concentrated abnormal environmental characteristics are identified as candidate precise control locations L.

[0096] This step uses spatial clustering algorithms (such as DBSCAN density clustering or threshold region growth method) to determine spatial continuity, eliminate isolated outliers, and ensure the feasibility of actual intervention in candidate regions.

[0097] The yield data (yield per unit area) and effective component concentration data (such as polysaccharide content, flavonoids, saponins, etc.) of the plots corresponding to the above candidate control locations L were retrieved in multiple historical years, and a spatial-yield-quality correlation model was constructed to identify sensitive areas with significant yield fluctuations or quality declines.

[0098] The model employs the following steps: establishing time-series yield and quality curves for each candidate location over the years; setting a baseline yield threshold (e.g., below 80% of the mean for two consecutive years) and a component concentration warning value (e.g., an annual average decline in effective components exceeding 15%); and marking grid regions meeting the above conditions as physiologically degraded or quality-abnormal regions, which are then used as the finally confirmed regulatory locations L. This model enables historical source analysis of the regulatory target region, enhancing identification accuracy and avoiding misjudgments of short-term fluctuations.

[0099] For each final determined control position L, extract its current standardized environmental feature vector. And combined with the weighting factor group of this reproductive stage The deviation contribution value Ci for each factor is calculated using the following method:

[0100] For each factor, calculate the absolute value of the difference between its standardized value and the mean of the target region, and then multiply it by the corresponding weight wi to obtain the deviation contribution value. .

[0101] All Ci are sorted, and the top 1-2 factors with the largest contribution values ​​are identified as the dominant deviation factors affecting the current physiological state, i.e., the regulatory target factors F, such as soil moisture content, diurnal temperature range, or disease incidence.

[0102] The identified regulatory position L is combined with its corresponding target factor set F (which can be one or more items) and output as an input parameter to the subsequent regulatory strategy module. This is used to match the appropriate intervention template and achieve intelligent invocation of the regulatory plan. For example: Output If F = {soil moisture content, disease risk}, then the system will match the "drip irrigation humidity control + biological agent spraying" combination strategy in the corresponding template library and execute precise intervention operations.

[0103] In this invention, based on the control location L and the control target factor F, a corresponding control measure template library S is matched from the planting database, including soil watering frequency adjustment, light and shading ratio setting, and disease prevention spraying scheme, specifically including:

[0104] First, the identified target regulatory factors F (such as soil moisture content, disease incidence, light intensity, etc.) are input into the standardized planting database in the system. This database stores intervention template information for various regulatory factors. Each template has a clear index label with the target factor, and the content includes the regulation method, application conditions, applicable stage, regulation intensity level, etc.

[0105] By using keyword matching or tag search, a set of regulatory measure templates S′ associated with the regulatory factor F is selected. For example, if F includes "soil moisture content", then "intelligent drip irrigation humidity regulation scheme" and "mulching evaporation suppression template" are initially screened out; if F includes "disease risk", then "biological agent spraying scheme" and "preventive crop rotation strategy template" are matched out.

[0106] Considering the varying adaptability of the same regulatory factor across different planting areas, a secondary screening and priority ranking of the initially selected template set S′ was conducted, taking into account the environmental and historical characteristics of the plot to which the regulatory location L belongs. This included the following parameters: the degree of matching between soil type (e.g., loam, clay, sandy soil) and the template's recommended soil; the compatibility of planting density with the template's applicable planting method (dense field planting / understory planting); and historical intervention feedback results, i.e., the template's effectiveness score or intervention success rate in the region's historical use. Each factor could be quantified into a scoring dimension, and a comprehensive fit index was calculated using a weighted scoring model, ranking all templates in S′ in descending order. For example: Comprehensive fit index = soil matching score × weight 1 + density adaptation score × weight 2 + historical feedback score × weight 3; where the weights could be set based on experience or an expert system.

[0107] To further improve the timeliness and effectiveness of regulatory measures, information on the current reproductive stage and the urgency parameter are introduced into the ranking template. The optimal regulatory template S is then determined through a rule engine or recommendation algorithm. The current reproductive stage is provided by the aforementioned reproductive period identification module. The urgency is defined based on the deviation between the previous stage's *Polygonatum yunnanensis* growth regulation impact index E and the sensitivity threshold R; the greater the deviation, the higher the urgency. The following recommendation strategy is preferred: if the urgency is high, templates with rapid action and strong response are selected first. If the urgency is medium to low, long-term regulatory templates with low cost and eco-friendliness are selected first. This rule system can achieve automatic screening based on multi-condition logic, supporting real-time responsive intervention decisions.

[0108] Finally, the output is a control measure template S bound to the control location L, which serves as the input parameter for the subsequent control execution module. Each location point corresponds to one or more control measure items, forming a standardized control strategy list, including soil watering frequency adjustment, light and shading ratio setting, and disease prevention spraying plan; this list can be used to generate operation instructions or directly invoked by intelligent devices such as agricultural robots / irrigation systems.

[0109] In a preferred embodiment of the present invention, to achieve precise execution and dynamic optimization of control measures during the cultivation of *Polygonatum yunnanense*, the aforementioned selected control measure template S is applied to the identified control location L. Through an intelligent operation scheduling system and a real-time feedback mechanism, automated, closed-loop, and precise control of the *Polygonatum yunnanense* growth process is achieved. Specifically, the following steps are included:

[0110] The obtained control measure template S is structured and parsed to transform it into control instructions recognizable by agricultural equipment. This process includes generating the following parameters: control target factors (e.g., soil moisture content ≥25%, disease index ≤0.3); operation time window (e.g., 6:00 AM to 10:00 AM, when the average daily temperature is above 18℃); task execution type (e.g., spraying, irrigation, fertilization); and recommended operation path (based on the coordinate range of L and the boundary of the operation area). The instructions are encapsulated in the form of a task parameter set, including fields such as target value, triggering conditions, operation type, and control duration, and can be directly sent to agricultural execution terminals (e.g., irrigation controllers, spraying drones, inspection robots, etc.) for task docking.

[0111] Based on the spatial distribution information of the control location L and the operational capabilities of the equipment, a path optimization algorithm (such as an improved A* algorithm or genetic algorithm) is used to generate the optimal operational path to reduce operational redundancy and energy consumption. Specifically, the following steps are implemented: extracting the coordinate data of all points L; constructing a regional connectivity graph based on the maximum operating radius, moving speed, and path accessibility constraints of the operating equipment; calculating the optimal solution for the total path length and coverage area, and outputting the control path planning map; and automatically scheduling and matching equipment resources, such as intelligent drip irrigation systems, agricultural drones, and automatic sprayers. This step ensures that the control measures can be implemented efficiently, on demand, and at designated locations, improving the accuracy and controllability of interventions in the growth environment of *Polygonatum yunnanensis*.

[0112] After the control equipment completes its path deployment, specific intervention operations are performed within the control time window. The control effect is monitored in real time via IoT sensing devices. Feedback data includes: changes in soil moisture content (via wireless moisture sensors); temperature, humidity, and disease environmental indicators (via micro-weather stations and disease monitoring equipment); images of *Polygonatum yunnanense* plant status (via agricultural image acquisition terminals); and statistics on equipment operation completion rate and spray dosage. The collected data is labeled by region and timestamp and uploaded to the data center in real time for subsequent analysis and judgment of the control effect. The collected feedback data is compared with the expected indicators set in the control template, and the control effect is evaluated based on an error evaluation model. For example: control deviation = actual measured value – target control value; deviation tolerance range = ±10%; if it exceeds the range, it is considered a control failure. If the control effect does not meet expectations, the system will dynamically adjust relevant template parameters (such as dosage, frequency, and time period) based on historical template effect data and update the template priority ranking to achieve adaptive evolution and optimization of the control strategy.

[0113] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for precise regulation of the growth of Polygonatum yunnanense by integrating multi-source information, characterized by: include: Obtain a multi-source environmental information set M for the target planting area of ​​Polygonatum yunnanense, including soil moisture content, diurnal temperature range, effective accumulated temperature, light intensity, atmospheric humidity, soil pH value, historical disease incidence rate, and topographic slope of the target area; The information set M is normalized to obtain the standardized environmental feature vector V; Based on the physiological response model of Polygonatum yunnanense at different growth stages, the weight factor group W of the standardized environmental feature vector V at each growth stage was calculated. Based on the calculated weight factor group W and the standardized environmental feature vector V, feature fusion is performed to construct the growth regulation influence index E of Polygonatum yunnanense. The construction of the growth regulation influence index E of Polygonatum yunnanense includes: The obtained weight factor group W is linearly weighted and fused with the standardized environmental feature vector V in a one-to-one correspondence manner to obtain the initial value of environmental suitability E′. By introducing key physiological parameter thresholds for different growth stages of Polygonatum yunnanense, and determining a standardized regulatory index function f(x) based on the current growth stage, a nonlinear mapping is performed on the initial value of environmental suitability E′ to obtain the growth regulation influence index E of Polygonatum yunnanense. Spatial interpolation was performed on the E-values ​​of the growth regulation influence index of Polygonatum yunnanense calculated from multiple regional samples to generate a distribution map of regulation risk in Polygonatum yunnanense planting areas. The calculated growth regulation impact index E of Polygonatum yunnanense is compared with the sensitivity threshold R of the current reproductive stage. If E < R, the current reproductive period is identified as being in a high-risk regulation zone. Based on the identification results and spatial distribution data within the planting area, combined with historical yield and effective ingredient concentration trends, the precise control location L and the control target factor F were determined. Based on the regulatory location L and the regulatory target factor F, the corresponding regulatory measure template library S is matched from the planting database; The control measure template S was applied to location L to achieve precise control over the growth of Polygonatum yunnanense.

2. The method for precise regulation of the growth of Polygonatum yunnanense by integrating multi-source information as described in claim 1, characterized in that: The physiological response model based on different growth stages of Polygonatum yunnanense calculates the weight factor group W of the standardized environmental feature vector V at each growth stage, including: Based on the growth process of Polygonatum yunnanense, the growth cycle is divided into four stages: germination stage, elongation stage, swelling stage, and maturity stage, and corresponding physiological response model sets are established for each stage. ; The obtained standardized environmental feature vector V is input into each physiological response model. In this study, the sensitivity coefficients of various environmental factors to target physiological parameters during different reproductive stages were determined by multifactor regression analysis. Based on the sensitivity coefficient, each environmental factor is weighted and normalized to obtain the weight factor group W for each growth stage.

3. The method for precise regulation of the growth of Polygonatum yunnanense by integrating multi-source information according to claim 2, characterized in that: The step of comparing the calculated growth regulation impact index E of Polygonatum yunnanense with the sensitivity threshold R of the current growth stage includes: Construct sensitive thresholds corresponding to the four reproductive stages of Polygonatum yunnanense. Each threshold is determined based on the minimum suitability index of environmental conditions in historical high-quality yield samples; Based on the current monitoring time, the reproductive stage of *Polygonatum yunnanense* is matched, and the corresponding sensitive threshold is selected. ; The obtained growth regulation influence index E of Polygonatum yunnanense and Compare, when E is less than At that time, it was determined to be in a high-risk control zone.

4. The method for precise regulation of the growth of Polygonatum yunnanense by integrating multi-source information according to claim 3, characterized in that: The determination of the precise control location L and the control target factor F includes: The high-risk area labels are spatially overlaid with the spatial grid data within the planting area to extract areas where abnormal environmental features are clustered as candidate control locations. By retrieving the corresponding yield data and effective component concentration data of the candidate locations over the years, a spatial-yield-quality correlation model is constructed to identify sensitive areas of yield decline or component abnormalities. Analyze the dominant deviation factors in the current standardized environmental feature vector V of the sensitive area, and combine them with the reproductive period response weight factor group W to determine the regulatory target factor F that has the greatest impact and needs priority intervention. Output the precise control position L and the target factor F.

5. The method for precise regulation of *Polygonatum yunnanense* growth by integrating multi-source information according to claim 4, characterized in that: The step of matching the corresponding regulatory measure template library S from the planting database based on the regulatory location L and the regulatory target factor F includes: The output target factor F is matched with the intervention factor index in the standardized planting database by keywords to select a set of regulatory measures templates associated with the target factor. Based on the soil type, planting density, and historical control feedback results of the area where the control location L is located, the initial screening templates are screened a second time and prioritized. Based on the current reproductive period and the urgency of regulation, the optimal regulatory measure template S is selected from the ranked templates using a recommendation algorithm; Output a list of control strategies S bound to location L, including soil watering frequency adjustment, light shading ratio setting, and disease prevention spraying plan.

6. The method for precise regulation of the growth of Polygonatum yunnanense by integrating multi-source information as described in claim 5, characterized in that: The step of applying the control measure template S to location L includes: The control measure template S is converted into control instructions that the equipment can recognize, generating a set of control operation parameters that includes control objectives, operation time, and execution path; Based on the geographical coordinates of location L and the operating radius of the equipment, the optimal path is planned and the corresponding agricultural execution units are scheduled.

Images