Scutellaria baicalensis georgi intelligent topping cultivation method based on environmental factors and estimation algorithm

Abstract

The application discloses a kind of intelligent topping cultivation methods of Scutellaria baicalensis based on environmental factors and estimation algorithm, belong to agricultural technology field.Method includes: S1, the air temperature of planting area, air relative humidity and rainfall, calculate effective accumulated temperature GDD;S2, effective accumulated temperature, air relative humidity and rainfall are input into topping parameter estimation model, model output suggests topping time window and suggested topping reserved height interval;S3, topping operation is executed;S4, axillary bud and topdressing management;S5, record yield quality index after harvesting, and yield quality index is fed back to topping parameter estimation model to correct model parameter.The application relies on multiple environmental factor coupling algorithm and intelligent iteration model, fuses accumulated temperature, humidity, rainfall multiple field environmental parameters, constructs self-adaptive, iterative, expandable intelligent topping decision system, realizes accurate quantization of topping opportunity, effectively improves crop yield and quality, reduces field disease risk.

Description

Technical Field

[0001] This invention belongs to the field of agricultural technology, specifically relating to an intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms. Background Technology

[0002] Scutellaria baicalensis ( Scutellaria baicalensis Scutellaria baicalensis (Georgi) is a perennial herbaceous plant belonging to the Lamiaceae family. Its dried root is used medicinally, and its main active ingredients are flavonoids, represented by baicalin and wogonin. As a major Chinese medicinal herb with abundant resources and wide applications, its standardized cultivation and quality improvement are of great significance for ensuring raw material supply and clinical efficacy. At the cultivation physiology level, there is a dynamic "source-sink" game relationship between the vegetative growth of the above-ground parts of Scutellaria baicalensis and the accumulation of medicinal components in the underground roots, and topping is a key agronomic measure to regulate this relationship. By timely disrupting apical dominance, topping can change the distribution flow of photosynthetic products, promoting their translocation to root sink organs, thereby simultaneously increasing the accumulation of root dry matter and the efficiency of flavonoid secondary metabolite synthesis. In its preliminary research, the applicant systematically analyzed the interaction effect of topping parameters on yield and quality indicators using a two-factor complete experimental design based on time and height (5 topping time levels × 6 retention height levels, for a total of 30 treatment combinations). A relatively optimal parameter combination was initially established, namely, topping is carried out every 15 days starting from early July, with a above-ground retention height of 20 cm.

[0003] However, the technical solutions developed by the aforementioned existing technologies reveal the following deep-seated technical defects when transitioning from controlled experimental environments to the interannually variable field production environments: (1) Failure of temporal regulation based on the fixed phenological hypothesis: Existing methods fix the timing of topping to a specific calendar date, which implicitly assumes that the growth and development rhythm of Scutellaria baicalensis remains constant from year to year. However, the transition from vegetative growth to reproductive growth (budding stage) of Scutellaria baicalensis is nonlinearly regulated by multiple environmental variables such as effective accumulated temperature, photoperiod, and water supply, and the phenological period can drift significantly by several days or even more than ten days from year to year. The fixed calendar topping date often shows a serious mismatch with the actual physiological development stage of the plant, which is essentially a mismatch between mechanical temporal regulation and dynamic biological response process. If the plant is topped before it enters the budding stage, the absolute amount of photosynthetic products supplied by the source organ will be low due to insufficient leaf area; if the topping is delayed after the flowering stage, reproductive growth has consumed a large amount of assimilates, and the window period for the redistribution of "source-sink" by topping has passed, and the effects of increasing yield and quality will be greatly reduced.

[0004] (2) The risk coupling between environmental effects and agronomic operations is not incorporated into the decision-making model: Topping creates wounds on the main stem of the plant, and the healing efficiency and pathogen infection risk constitute a dynamic balance driven by the microenvironment. After topping, the tissues around the wound are exposed, and the field microenvironment with high humidity, high temperature and poor ventilation will significantly prolong the formation period of wound callus tissue, providing an extended invasion window for soil-borne and wind-borne pathogens such as Fusarium wilt and stem rot, thus increasing the risk of disease systems. On the other hand, interannual accumulated temperature changes regulate carbon assimilation capacity and water use efficiency by affecting photosynthetic rate and transpiration rate. In years with low accumulated temperature, the retained height is too low, which will excessively weaken the photosynthetic source area, resulting in a negative balance of carbon supply in the roots. Existing technologies lack an integrated framework for quantitative perception and risk threshold determination of key environmental factors such as temperature, humidity and water, and cannot achieve intelligent decision-making for environmental risk avoidance during the topping operation window, resulting in the inability to achieve a dynamic optimization balance between the goal of increasing yield and the goal of stable yield control.

[0005] (3) The open-loop static parameter output lacks continuous evolution capability: From the perspective of information theory, the existing technical solution is a one-way, open-loop knowledge transfer system. The optimal parameter combination obtained based on the experimental field in a specific year and location, once solidified and released, is completely disconnected from the massive feedback data generated in subsequent production practice. The four-dimensional high-dimensional dataset of measured yield, quality indicators, environmental data and topping operation records over the years contains gradient information on the direction of parameter optimization, but the existing technology cannot use this valuable data for machine learning or empirical regression-based parameter self-correction and model iteration, which results in the recommended parameters not being able to continuously approach the global or local optimization under complex and ever-changing production conditions. The technology itself lacks the evolutionary function driven by human or data, forming a vicious cycle of "experience solidification - diminishing returns - difficulty in technology accumulation".

[0006] Therefore, providing a spatiotemporal parameter adaptive decision-making method for dynamic topping of Scutellaria baicalensis based on real-time environmental factors, which has a coupled decision kernel that integrates a real-time phenological prediction model and an environmental risk threshold model, and can continuously iterate and optimize the decision kernel through a feedback learning mechanism using historical production measurement data, has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0007] The purpose of this invention is to provide an intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms. This method aims to solve the systemic technical problems in existing technologies, such as the mismatch between fixed-date topping and dynamic biological processes, the failure to incorporate environmental risk factors into topping decisions leading to high yield reduction and high disease risk, and the stagnation of parameter optimization due to the lack of data-driven closed-loop iteration. To achieve the above objectives, the technical solution adopted by the present invention is as follows: This invention discloses an intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms, comprising: S1. Environmental data collection and processing: During the growth period of Scutellaria baicalensis, the air temperature, relative humidity and rainfall in the planting area were collected, and the effective accumulated temperature (GDD) from the date of planting of Scutellaria baicalensis to the current date was calculated. S2. Estimation of topping parameters: Input the effective accumulated temperature, relative humidity and rainfall into the pre-built topping parameter estimation model. The model outputs a suggested topping time window and a suggested topping retention height range. S3. Topping operation: Within the recommended topping time window, control the plant height to the recommended topping height range, and remove the top inflorescences, flower buds and tender shoots. S4. Axillary bud and topdressing management: After topping, remove excess axillary buds within the preset time period and apply topdressing based on the growth stage. S5. Harvesting and Feedback Update: After harvesting, record the yield and quality indicators and feed them back to the topping parameter estimation model to correct the model parameters.

[0008] In some embodiments of the present invention, in step S1, the air temperature includes the daily maximum temperature T. max Daily minimum temperature T min Daily average temperature T 均 ; The formula for calculating effective accumulated temperature (GDD) is as follows: ; Among them, T 基础 The basic temperature for the growth of Scutellaria baicalensis is defined as 5℃~10℃. T 均 The calculation method is as follows: ; When T 均 <T 基础 At that time, the effective accumulated temperature reading for that day was 0; The topping parameter estimation model uses a preset threshold range of 700-1150℃ for the effective accumulated temperature (GDD) as one of the criteria for triggering the topping time window. Preferably, the rainfall is cumulative, calculated as the total rainfall from the last topping operation or from the date of planting to the current date.

[0009] In some embodiments of the present invention, the topping parameter estimation model is constructed using any one of multiple linear regression, weighted scoring, decision tree, or random forest algorithms. The topping parameter estimation model uses the historical dataset of the Scutellaria baicalensis two-factor topping experiment as the training set, and the root dry weight and baicalin content as the target variables for training. During the model training process, the optimization objective is to maximize the root dry weight and baicalin content, and the corresponding suggested topping time window and suggested topping retention height range are matched and output to determine the core parameters of the model.

[0010] In some embodiments of the present invention, in the topping parameter estimation model, the weights of each input factor are: effective accumulated temperature 45% to 55%, relative humidity 20% to 30%, and rainfall 15% to 25%; when a weighted scoring algorithm is used to construct the model, the sum of the weights of all factors is 100%.

[0011] In some embodiments of the present invention, when the relative humidity of the air in the planting area is higher than 85% for 5 consecutive days, the suggested topping time window output by the topping parameter estimation model is automatically extended by 3 to 7 days to reduce the risk of infection of the topping wounds on the plants.

[0012] In some embodiments of the present invention, the output parameters of the topping parameter estimation model satisfy the following: the recommended topping time window is any continuous period from the budding stage of Scutellaria baicalensis to two weeks before the full bloom stage; the recommended topping height range is 15 cm to 50 cm. Preferably, under the optimal accumulated temperature conditions, it is recommended that the height range for topping be limited to 15 cm to 25 cm; the optimal accumulated temperature conditions are: effective accumulated temperature GDD of 750 to 950℃, average relative humidity of the air in the past 5 days ≤ 80%, and cumulative rainfall in the past 15 days ≤ 25 mm; under the optimal accumulated temperature conditions, it is recommended that the height range for topping be limited to 15 cm to 25 cm.

[0013] In some embodiments of the present invention, in step S3, the plant must retain at least two pairs of functional leaves during the topping operation; Preferably, the functional leaf is defined as: a mature leaf with normal leaf color, no disease or pest infestation, and a leaf area of ​​not less than 50% of the standard leaf area; Preferably, in step S4, the preset time period for removing excess axillary buds is 7 to 14 days after topping; More preferably, the thinning standard is: only 2 to 3 strong axillary buds at the top of a single Scutellaria baicalensis plant are retained, and all other axillary buds are removed; In some embodiments of the present invention, in step S5, the product quality indicators include root dry weight and baicalin content.

[0014] Compared with the prior art, the present invention has the following beneficial effects: This invention is scientifically designed and ingeniously conceived. Compared with traditional fixed-pattern crop topping techniques, this invention relies on a multi-environmental factor coupling algorithm and an intelligent iterative model, overcoming the shortcomings of traditional manual topping on fixed dates, single parameters, poor adaptability, and inability to adapt to climate fluctuations. This invention integrates multiple field environmental parameters such as accumulated temperature, humidity, and rainfall to construct an adaptive, iterative, and scalable intelligent topping decision-making system. It inherits and upgrades previous experimental results, achieving precise quantification of topping timing, effectively improving crop yield and quality, and reducing field disease risks. Simultaneously, it possesses algorithmic universality, scenario adaptability, and large-scale implementation capabilities, demonstrating significant overall technological advancement, practicality, and industrialization value. Details are as follows: (1) It can dynamically adapt to interannual climate fluctuations and avoid the risk of yield reduction. This invention collects real-time data on effective accumulated temperature, relative humidity, and rainfall in the planting environment, and uses an estimation model to achieve adaptive dynamic adjustment of the topping time window. It breaks through the limitations of the traditional fixed calendar date topping method, effectively solves the problem of mismatch between calendar time and actual phenological period of plants caused by climate differences in different years, greatly reduces the risk of crop yield reduction caused by climate fluctuations, and ensures stable planting yield.

[0015] (2) Multi-environmental factor synergistic quantitative decision-making, taking into account both stable yield and disease resistance. This invention constructs a multi-factor coupled topping decision-making framework, accurately quantifying the weight of each environmental factor, of which the weight of effective accumulated temperature is 45%–55%, the weight of relative humidity is 20%–30%, and the weight of rainfall is 15%–25%. The model can intelligently adjust the topping timing according to the environmental conditions. In high humidity environments, it automatically delays the topping time to reduce the probability of crop disease growth; in years with low effective accumulated temperature, it appropriately delays topping to ensure that the plant retains sufficient photosynthetic area, ensuring the efficiency of crop photosynthesis, and achieving dual guarantees of yield and quality.

[0016] (3) Built-in iterative optimization mechanism, with continuously improving accuracy. This invention has a long-term self-optimization capability. After the crop harvest is completed each year, the yield and quality data measured in the field are automatically included in the model training dataset to continuously correct and iterate the core parameters of the model. As the usage cycle increases, the accuracy of the model's estimation of topping parameters continues to improve, and the decision-making effect is optimized year by year.

[0017] (4) Iterative upgrade of previous technical achievements, resulting in a broader scope of protection. This invention is based on the applicant's previous two-factor experimental results and is an upgrade (the optimal experimental combination in the early stage: topping on July 1st, topping height of 20cm, which can achieve a ≥28% increase in crop root dry weight and a ≥32% increase in baicalin content). Through algorithm generalization technology, the traditional discrete experimental optimization parameter points are expanded into a continuous dynamic estimation parameter range, breaking through the limitations of traditional experimental parameters and greatly expanding the scope of technology adaptation and patent protection boundaries.

[0018] (5) High degree of intelligence and systematization, suitable for large-scale planting. The present invention is equipped with an intelligent topping decision system, which integrates multiple functions such as real-time data acquisition from sensors, cloud computing, edge computing and offline operation. It is suitable for deployment needs in different sites, is easy to operate and has strong stability. It can be widely used in large-scale Scutellaria baicalensis crop production areas, and helps to standardize, intelligentize and industrialize crop topping operations. Detailed Implementation

[0019] The technical features of the technical solution provided by the present invention will be further clearly and completely described below with reference to specific embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and 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.

[0020] Example 1 (System operating normally, optimal parameters reproduced, medium accumulated temperature year).

[0021] 1. Overview of the Planting Base This embodiment was implemented at a large-scale Scutellaria baicalensis planting base in Da'an City, Jilin Province. The base is at an altitude of 350 m, the soil type is brown soil, and the pH value is 6.2-6.8. The tested variety is Scutellaria baicalensis (…). Scutellaria baicalensis The Georgi variety "Tongqin No. 1" was planted at a 2-year age with a planting density of 30 cm between plants and 20 cm between rows. The planting date was April 20, 2025.

[0022] 2. Step S1: Environmental Data Acquisition Four meteorological sensor nodes were evenly deployed within the test site to collect the following environmental parameters in real time: daily maximum temperature (°C), daily minimum temperature (°C), daily average temperature (°C), relative humidity (%), and daily rainfall (mm). Each sensor uploaded data to the cloud server every 30 minutes.

[0023] Based on base temperature T base The effective accumulated temperature GDD is calculated at 8℃ using the following formula: ; in T max The highest temperature of the day, T min This is the lowest temperature of the day. Among them, T... 基础 The baseline temperature for Scutellaria baicalensis growth is defined as 5℃~10℃. In the Northeast production area (average annual temperature ≤5℃), T is taken as... 基础 =8℃; In the North China and Huanghuai production areas (average annual temperature 6℃~10℃), take T 基础=5℃~7℃; in the southwestern plateau production area (altitude >1500 m), take T 基础 =10℃; the same temperature is used consistently throughout the entire growth period at the same production base. 基础 Values ​​are set to ensure the comparability of interannual accumulated temperature data; T 均 = (T) max +T min ) ÷ 2, when T 均 <T 基础 The effective accumulated temperature on that day was 0. The GDD threshold range triggered by topping was 700–1150℃, corresponding to the period from the budding stage to two weeks before the full bloom of Scutellaria baicalensis.

[0024] Since the planting date (April 20th), the cumulative effective temperature is as follows: Until July 1st: 780℃; Until July 16: 950℃; Until July 31: 1130℃.

[0025] 3. Step S2: Estimation of topping parameters The collected environmental parameters and field observation data were input into the topping parameter estimation model. This model was constructed based on the random forest algorithm, and its training set was derived from the applicant's two-factor topping experiment dataset conducted from 2022 to 2024, totaling 540 samples (180 treatment plots × 3 years). The data originated from a two-factor complete experiment conducted for three consecutive years (2022-2024) at a Scutellaria baicalensis planting base. The experiment included 5 topping time levels (July 1st, July 16th, July 31st, August 15th, August 30th) × 6 retention height levels (15 cm, 20 cm, 25 cm, 30 cm, 40 cm, 50 cm), resulting in 30 treatment combinations. Each combination had 6 replicate plots, totaling 180 treatment plots. 15 plants were randomly selected from each plot, and the average value was taken, resulting in a total of 540 samples over the 3 years.

[0026] Data preprocessing includes outlier removal, which identifies and removes outliers based on the mean ± 3 standard deviations (3σ principle). If the root dry weight or baicalin content deviates by more than 40% from the mean of this treatment group, it is considered an abnormal experimental operation and the sample is removed. Data preprocessing also includes data normalization: Min-Max normalization is applied to all continuous input features (GDD, humidity, rainfall) to map feature values ​​to the [0, 1] interval; the target variables (root dry weight, baicalin content) retain their original dimensions and are not normalized; the target variables of the model are root dry weight (g / plant) and baicalin content (%); the input features include: current cumulative GDD, average relative humidity of the past 15 days, and cumulative rainfall of the past 15 days.

[0027] In this embodiment, after inputting the above features on July 1st, the model output results are as follows: Recommended window for topping: July 1st to July 10th; Recommended height range: 18 cm to 25 cm; The weights of each input factor are: effective accumulated temperature 45%–55%, relative humidity 20%–30%, and rainfall 15%–25%. When a weighted scoring algorithm is used to construct the model, the sum of all factor weights is 100%. This embodiment uses a preferred weighted scoring algorithm as an example, as explained below: (1) Calculate normalized scores (0-100 points) for each factor daily: ① GDD score: When the cumulative GDD of the day is within the trigger threshold range (700~1150 GDD·℃), the GDD score = (cumulative GDD of the day - 700) / (1150 - 700) × 100; if it exceeds the range, it will be treated as 0 points or 100 points respectively. ② Humidity score: When the relative humidity of the air is ≤75%, the humidity score = 100 points (optimal conditions for sizing); when it is 75% to 85%, the score decreases linearly to 60 points; when it is >85%, the score = 0 points (sizing is not recommended). ③ Rainfall score: 100 points for cumulative rainfall ≤30 mm in the past 15 days; linearly decreases to 60 points for rainfall between 30 and 60 mm; 30 points for rainfall >60 mm. (2) Comprehensive topping suitability score = GDD score × 0.50 + humidity score × 0.25 + rainfall score × 0.20 + other factor scores × 0.05; (3) Output rules: When the number of consecutive days with a comprehensive score of ≥70 points is ≥3 days, this period is the recommended time window for capping; (4) Recommended height output: negatively correlated with GDD level - when GDD is 700-850, the recommended height is 25-40 cm; when GDD is 850-1050, the recommended height is 18-25 cm; when GDD is 1050-1150, the recommended height is 15-20 cm.

[0028] In this embodiment, the model confidence score is defined as follows: using the validation set (20% of all training data, a total of 108 samples) as a baseline, the weighted average coefficient of determination of the two target variables (root dry weight R1², baicalin content R2²) on the validation set is calculated: Confidence score = 0.6 × R1² + 0.4 × R2². Root dry weight is given a higher weight (0.6) because it has a more direct impact on the final economic benefits of planting. In Example 1, the model's R1² (root dry weight) = 0.88, R2² (baicalin content) = 0.85 on the validation set, and the model confidence score = 0.6 × 0.88 + 0.4 × 0.85 = 0.868 ≈ 0.87. The model confidence score meets the set threshold to adjust the model parameters.

[0029] 4. Step S3: Perform the topping operation. Based on the suggested time window output by the model, the first topping will be carried out on July 3, 2025 (sunny day, 8:00-11:00 AM). The procedure is as follows: keep the plant height at 20 cm, remove the top inflorescence, flower buds and tender shoots, and retain the upper 4 pairs of functional leaves (mature leaves with normal leaf color, no pests or diseases, and leaf area ≥ 50% of the standard leaf area).

[0030] Subsequent topping was carried out following the principle of rerunning the model before each topping. The second topping was conducted on July 18th (suggested window: July 16th–25th; suggested height: 20–25 cm; actual height retained: 22 cm), and the third topping was conducted on August 2nd (model suggested window: July 31st–August 9th; suggested height: 18–25 cm; actual height retained: 24 cm), for a total of three toppings. Each topping was spaced 15–17 days apart and performed within the model's suggested time window.

[0031] 5. Step S4: Axillary bud and topdressing management Ten days after each topping operation, axillary buds were observed. A "strong axillary bud" was defined as an axillary bud located in the upper part of the plant, ≥5 cm in length, with at least two unfolded leaves, and free from disease or pest damage. Two strong axillary buds meeting these criteria were retained per plant, and all other axillary buds were removed.

[0032] The timing of topdressing is determined by the model based on the GDD threshold: the first topdressing is applied when the cumulative GDD reaches 1150℃ (approximately August 15th), and the second topdressing is applied when the cumulative GDD reaches 1700℃ (approximately September 15th), applying 12 kg / mu of compound fertilizer with N:P2O5:K2O = 1:1:2 respectively. Other field management procedures are carried out according to DB22 / T 3319-2021.

[0033] 6. Step S5: Harvesting and Feedback Update Harvesting took place on October 15, 2025 (approximately 178 days after transplanting). At harvest, 15 plants were randomly selected from each experimental plot, their rhizomes were separated, washed, and dried at 60℃ to constant weight. The dry weight of the roots (g / plant) was measured. Simultaneously, the content of baicalin (%) was determined by high performance liquid chromatography.

[0034] A control group (CK) was set up separately. No topping treatment was performed during the entire growth period, and other field management was carried out in accordance with DB22 / T3319-2021. Harvesting was carried out on October 15, 2025. The average dry weight of roots and the baicalin content were set as the baseline value (100%).

[0035] The actual test results are as follows: The average dry weight of roots in this embodiment was 42.6 g / plant, which was 31.2% higher than that of the control group (32.5 g / plant). In this embodiment, the average content of baicalin was 9.87 wt%, which was 33.7% higher than that of the control group (7.38%).

[0036] The measured root dry weight and baicalin content data were entered into the system, and the system performed a model feedback update: in addition to the original 540 training samples, the measured data from this embodiment was added as a sample. The change in the target parameter before and after a single correction was 0.04 < 0.10, and the weighted confidence score increased from 0.87 to 0.91 after correction, satisfying the stability constraint. The corrected model was then used in the next growing season.

[0037] Example 2 (High humidity year, automatic delay of topping time).

[0038] 1. Overview of planting bases and years.

[0039] The planting base was the same as in Example 1, implemented in 2025. In that year, the relative humidity was above 88% for eight consecutive days starting from July 1st, and the cumulative rainfall during the same period was 63 mm, making it a typical high-humidity year.

[0040] 2. Step S2 – Model Adaptive Adjustment.

[0041] The model for estimating topping parameters incorporates a high-humidity decision rule: when the relative humidity of the air is ≥85% for 5 consecutive days, the model's suggested topping time window is extended by 3 to 7 days.

[0042] The number of days of delay is determined based on the humidity gradient: (1) When the relative humidity of the air is between 85% and 88% (inclusive) for 5 consecutive days, it is recommended to postpone the time window for sprucing the roof by 3 days; (2) If the relative humidity of the air is between 88% and 92% (inclusive) for 5 consecutive days, it is recommended to postpone the time window for sprucing the roof by 5 days; (3) If the relative humidity of the air is higher than 92% for 5 consecutive days, it is recommended to postpone the time window for pruning the roof by 7 days; When the humidity drops below 75% and persists for at least 2 days, the high humidity delay state is lifted, and the topping time window returns to the normal estimation result. The high humidity decision rule is executed independently as the post-processing logic layer of the model output, without changing the main parameters of the estimation model, so as to reduce the infection risk of the topping wound of the plant.

[0043] 3. Step S3 – Topping Operation Execution The first topping was performed on July 11, 2025 (sunny morning), retaining a height of 25 cm. Subsequent toppings will not repeat the model; instead, decisions will be made based on the height range suggested by the first model and the actual height of the new main shoots. Second pruning: July 28th, retaining a height of 22 cm; Third pruning: August 14th, retaining a height of 20 cm.

[0044] The interval between each topping was 17 days, all within the 10-20 day constraint range; the recommended height was gradually reduced (25 cm, 22 cm, 20 cm), reflecting the dynamic decision-making pattern of reducing the recommended height with the negative correlation formula as the accumulated temperature GDD increases.

[0045] 4. Step S5 – Harvesting Results Actual harvesting results: The average dry weight of roots increased by 26.4% compared to the control group (CK, without topping); The baicalin content was 28.9% higher than that of the CK group.

[0046] Disease incidence rate (investigated plant by plant 5-7 days before harvest): The disease incidence rate (mainly root rot and leaf spot) of plants throughout the entire growth period in this example was less than 3%, which was significantly lower than the 15%-20% disease incidence rate of conventional planting in local high-humidity years, indicating that the high-humidity delay strategy effectively reduced the risk of disease.

[0047] Comparative example.

[0048] To directly demonstrate the technological advancement of the "dynamic intelligent topping" method of this invention compared to "traditional fixed parameter topping," the following comparative group was set up in the same planting base, and the experiment was conducted in parallel with Example 1: Comparative Example A (CK, no topping throughout the entire growth period): No topping treatment was performed, and the remaining field management was carried out in accordance with DB22 / T3319-2021. The average dry weight of roots was set as the baseline value (100%), and the baicalin content was set as the baseline value (100%).

[0049] Comparative Example B (Traditional Fixed Parameter Topping): The optimal fixed parameters obtained by the applicant in the previous two-factor experiment were adopted (topping once every 15 days from July 1, retaining a height of 20 cm each time, for a total of 3 toppings), and other field management was carried out in accordance with DB22 / T3319-2021.

[0050]

[0051] Each experiment consisted of 6 replicate plots, with 15 plants randomly selected from each plot for the mean. The results indicate that, compared to traditional fixed-parameter topping (Comparative Example B), the dynamic intelligent topping method of this invention further improves root dry weight by 7.5 percentage points (31.2%–23.7%), baicalin content by 11.1 percentage points (33.7%–22.6%), and reduces plant disease incidence by 2.4 percentage points (4.2%–1.8%), fully demonstrating the technological advancement of this invention.

[0052] Although the present invention has been described in detail with reference to the foregoing embodiments, the scope of protection of the present invention is not limited to the above embodiments. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A method for intelligent topping cultivation of Scutellaria baicalensis based on environmental factors and estimation algorithms, characterized in that, include: S1. Environmental data collection and processing: During the growth period of Scutellaria baicalensis, the air temperature, relative humidity and rainfall in the planting area were collected, and the effective accumulated temperature (GDD) from the date of planting of Scutellaria baicalensis to the current date was calculated. S2. Estimation of topping parameters: Input the effective accumulated temperature, relative humidity and rainfall into the pre-built topping parameter estimation model. The model outputs a suggested topping time window and a suggested topping retention height range. S3. Topping operation: Within the recommended topping time window, control the plant height to the recommended topping height range, and remove the top inflorescences, flower buds and tender shoots. S4. Axillary bud and topdressing management: After topping, remove excess axillary buds within the preset time period and apply topdressing based on the growth stage. S5. Harvesting and Feedback Update: After harvesting, record the yield and quality indicators and feed them back to the topping parameter estimation model to correct the model parameters.

2. The intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms as described in claim 1, characterized in that, In step S1, the air temperature includes the daily maximum temperature T. max Daily minimum temperature T min Daily average temperature T 均 ; The formula for calculating effective accumulated temperature (GDD) is as follows: ； Among them, T 基础 The basic temperature for the growth of Scutellaria baicalensis is defined as 5℃~10℃. T 均 The calculation method is as follows: ； When T 均 <T 基础 At that time, the effective accumulated temperature reading for that day was 0; The topping parameter estimation model uses a preset threshold range of 700-1150℃ for the effective accumulated temperature (GDD) as one of the criteria for triggering the topping time window. Preferably, the rainfall is cumulative, calculated as the total rainfall from the last topping operation or from the date of planting to the current date.

3. The intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms according to claim 1, characterized in that, The topping parameter estimation model is constructed using any one of the following algorithms: multiple linear regression, weighted scoring, decision tree, or random forest. The topping parameter estimation model was trained using the historical dataset of the Scutellaria baicalensis two-factor topping experiment as the training set, with root dry weight and baicalin content as target variables. During model training, the optimization objectives are to maximize root dry weight and baicalin content. The corresponding suggested topping time window and suggested topping retention height range are matched and output to determine the core parameters of the model.

4. The intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms as described in claim 1, characterized in that, In the topping parameter estimation model, the weights of each input factor are: effective accumulated temperature 45%–55%, relative humidity 20%–30%, and rainfall 15%–25%; when the weighted scoring algorithm is used to construct the model, the sum of the weights of all factors is 100%.

5. A method for intelligent topping cultivation of Scutellaria baicalensis based on environmental factors and estimation algorithms, as described in claim 1, is characterized in that... When the relative humidity of the air in the planting area is higher than 85% for 5 consecutive days, the suggested topping time window output by the topping parameter estimation model will be automatically extended by 3 to 7 days to reduce the risk of infection of the topping wounds on the plants.

6. A method for intelligent topping cultivation of Scutellaria baicalensis based on environmental factors and estimation algorithms, as described in claim 1, is characterized in that... The output parameters of the topping parameter estimation model satisfy the following: the recommended topping time window is any continuous period from the budding stage of Scutellaria baicalensis to two weeks before the full bloom stage; the recommended topping height range is 15 cm to 50 cm.

7. A method for intelligent topping cultivation of Scutellaria baicalensis based on environmental factors and estimation algorithms, as described in claim 6, is characterized in that... The optimal accumulated temperature conditions are: effective accumulated temperature (GDD) of 750–950℃, average relative humidity of the air in the past 5 days ≤80%, and cumulative rainfall in the past 15 days ≤25 mm; under the optimal accumulated temperature conditions, it is recommended that the height of the roof be limited to 15 cm–25 cm.

8. The intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms according to claim 1, characterized in that, In step S3, the plant must retain at least two pairs of functional leaves during the topping operation; Preferably, the functional leaf is defined as: a mature leaf with normal leaf color, free from pests and diseases, and with a leaf area of ​​not less than 50% of the standard leaf area.

9. The intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms according to claim 1, characterized in that, In step S4, the preset time period for removing excess axillary buds is 7 to 14 days after topping; More preferably, the thinning standard is: only 2 to 3 strong axillary buds at the top of a single Scutellaria baicalensis plant are retained, and all other axillary buds are removed.

10. The intelligent topping cultivation method for Scutellaria baicalensis based on environmental factors and estimation algorithms according to claim 5, characterized in that, In step S5, the product quality indicators include root dry weight and baicalin content.

Images