Efficient ecological cultivation method for gastrodia elata

Abstract

The invention relates to the field of biological information feature identification, and provides a method for identifying the health state of a plant by deploying a multi-angle multi-spectral image identification sensor to obtain three-dimensional phenotype information of the plant, extracting geometric, texture and spectral feature vectors, and combining a convolutional neural network and a support vector machine fusion model to judge the health state. A growth rate model is constructed based on time sequence analysis so as to identify an abnormal trend; and further matching the phenotypic data with a curative effect database, predicting a curative effect grade through nonlinear regression, generating a personalized optimization scheme of moisture, nutrients and environment regulation and control according to prediction deviation, driving water and fertilizer integration, environment regulation and control and light supplementing equipment to execute closed-loop operation, and realizing system self-learning through effect evaluation feedback. According to the method, accurate cultivation regulation and control from symptom recognition to a curative effect target can be realized, the controllability and consistency of accumulation of medicinal components of gastrodia elata are remarkably improved, and the quality of medicinal materials and the stability of clinical curative effects are guaranteed.

Description

Technical Field

[0001] This application relates to the field of bioinformatics technology, and in particular to a method for efficient and ecological cultivation of Gastrodia elata. Background Technology

[0002] Currently, as an important traditional Chinese medicine, the cultivation of Gastrodia elata mainly relies on traditional experience and manual observation. Cultivators usually assess the growth status by visually observing the plant's appearance. This method is highly subjective and lacks quantitative standards. Because it cannot accurately identify early growth abnormalities, the best time for intervention is often missed, which directly affects the accumulation of medicinal components and the final efficacy of Gastrodia elata.

[0003] In recent years, computer image recognition technology has been initially applied in the agricultural field, but it has mostly focused on crop disease and pest identification or yield prediction. Current technology has not yet established a correlation model between plant image features and medicinal component content, and there is a lack of specialized research on the relationship between different symptoms of Gastrodia elata and its therapeutic effects. A gap still exists between image analysis results and cultivation decisions, making it difficult to achieve efficacy-oriented precision cultivation management.

[0004] Existing methods for optimizing the cultivation of Gastrodia elata mainly focus on the regulation of single environmental factors, failing to integrate multi-dimensional growth information for comprehensive decision-making. Traditional methods cannot achieve a dynamic correlation between growth status and expected therapeutic effects, and cultivation measures lack specificity and foresight. These technical deficiencies lead to unstable cultivation quality of Gastrodia elata and difficulty in effectively guaranteeing its medicinal value. There is an urgent need to develop an intelligent cultivation technology system that can link symptoms and therapeutic effects. Summary of the Invention

[0005] The purpose of this invention is to provide a highly efficient and ecological cultivation method for Gastrodia elata, which can effectively solve the problems mentioned in the background. Traditional Gastrodia elata cultivation relies on manual experience to judge the plant's condition, lacking quantitative monitoring and control methods for the accumulation process of medicinal components, resulting in large fluctuations in the quality of the medicinal material and unstable clinical efficacy. Existing agricultural image recognition technology has failed to establish a quantitative correlation between phenotypic features and efficacy levels, and cultivation management measures are mostly static adjustments, unable to achieve closed-loop optimization based on dynamic growth trends and expected efficacy goals. This invention integrates multi-angle multispectral image acquisition, three-dimensional phenotypic modeling, intelligent health status identification, time-series growth analysis, efficacy prediction, and personalized control decision-making technologies to construct a precise cultivation and control system from symptom identification to efficacy guidance, breaking through the technical bottleneck of traditional cultivation that emphasizes growth while neglecting efficacy.

[0006] To achieve the above objectives, this application provides the following technical solution: A highly efficient and ecological cultivation method for Gastrodia elata, characterized by comprising: Step 1: Deploy image recognition sensors to periodically acquire color and near-infrared images of Gastrodia elata plants. Use multi-angle synchronous acquisition to acquire plant image data, construct complete three-dimensional phenotypic information of the plants, and transmit the acquired image data to the image processing server via wireless network. Step 2: Preprocess the acquired image data, extract multidimensional feature parameters from the segmented image regions, and calculate feature vectors containing three types of features: geometric, texture, and spectral. Step 3: Call the pre-trained plant health status classification model, input the feature vector, and evaluate the health status label of the Gastrodia elata plant through the discriminative architecture that combines convolutional neural network and support vector machine. Step 4: Based on time series analysis, the continuously collected feature vectors are dynamically modeled to construct the plant growth rate curve, identify growth inflection points and abnormal fluctuation ranges, and combined with environmental sensors, analyze the potential causes of growth trend deviation to obtain the current image analysis results of the Gastrodia elata plant. Step 5: Match the current image analysis results with the Gastrodia elata therapeutic effect database, predict the expected therapeutic effect level of the current plant through a nonlinear regression model, divide the optimization requirements into different priorities according to the deviation between the expected therapeutic effect level and the target therapeutic effect level, and retrieve the corresponding priority control strategy combination from the cultivation knowledge base to generate a personalized cultivation optimization plan that includes water management, nutrient allocation and environmental control measures. Step 6: Transform the optimized plan into equipment control commands to drive the integrated water and fertilizer device, environmental control equipment and supplemental lighting system to perform corresponding operations, and collect new image data after a preset time after execution to evaluate the consistency between phenotypic changes and improved efficacy.

[0007] Further, the image recognition sensor mentioned in step 1 includes: The overhead acquisition unit is arranged above the cultivation area, the lateral acquisition unit is installed between the plant rows, and the underground root observation unit is set in the soil surface layer. The top-down acquisition unit uses a multispectral imager and is equipped with simultaneous acquisition functions in the visible light band and near-infrared band. The side-view acquisition unit is equipped with an electric gimbal, which rotates at a preset angle to capture images of the side of the plant canopy. The underground root observation unit adopts a special optical mirror structure, combined with a waterproof and anti-fog device, to achieve non-destructive monitoring of root growth status.

[0008] Furthermore, the preprocessing and feature extraction described in step 2 include: The acquired raw images are converted to HSV color space to increase the distinguishability of color features; A region-growing-based image segmentation algorithm, combined with morphological filtering, is used to separate the plant from the background region. In the segmented image region, geometric feature extraction includes calculating leaf area index, leaf tilt angle distribution, plant height, and canopy projected area; Texture feature extraction employs the gray-level co-occurrence matrix method to calculate contrast, correlation, energy, and homogeneity statistics; Spectral feature extraction includes calculating the normalized vegetation index, photochemical reflectance index, and water stress index.

[0009] Furthermore, the time series analysis method described in step 4 includes: Establish a feature vector time series database to store continuously collected plant phenotypic feature data; The sliding window method is used to smooth the feature vector sequence and eliminate noise interference caused by single-point acquisition error; A growth rate calculation model was constructed, and the daily growth change was obtained by the difference operation of the feature vectors of adjacent time points; Set up an abnormal growth early warning mechanism to trigger an early warning signal when the growth rate is lower than the threshold at multiple consecutive time points. By combining temperature, humidity, and light intensity data collected by environmental sensors, a correlation analysis matrix between environmental factors and growth rate is established.

[0010] Furthermore, the generation of personalized cultivation optimization schemes in step 5 includes: Based on the degree of deviation between the expected efficacy level and the target efficacy, the optimization needs are divided into multiple priority levels; For different priorities, the corresponding combination of control strategies is retrieved from the cultivation knowledge base; The regulatory strategies include water management strategies, nutrient allocation strategies, and environmental control strategies; Based on the aforementioned water management strategy, adjustments to irrigation timing, irrigation volume, and irrigation methods are determined; Based on the nutrient allocation strategy, nitrogen, phosphorus, and potassium ratio schemes and trace element supplementation schemes were formulated for different growth stages. The environmental control strategy includes temperature regulation range, humidity control target, and light management scheme.

[0011] Furthermore, the construction of the cultivation knowledge base includes: Historical data on the cultivation of Gastrodia elata in different production areas were collected, including soil conditions, climate characteristics, management measures, and final efficacy evaluation. Establish a database of the correspondence between cultivation measures and the content of medicinal components; Establish a table of optimal environmental parameter ranges for different growth stages; Summarize the diagnostic features of common growth abnormalities and corresponding control measures; Record the differences in cultivation characteristics and therapeutic effects of different varieties of Gastrodia elata.

[0012] Furthermore, the generation of the device control commands in step 6 includes: Analyze the parameters in the cultivation optimization plan and convert them into control signals that the equipment can execute; The control commands for the integrated water and fertilizer system include the solenoid valve opening duration, fertilizer injection ratio, and mixed liquor concentration setting. The control commands for environmental control equipment include adjustment parameters for fan speed, shading net opening, and humidifier output power; The supplemental lighting system control commands include configuration schemes for supplemental lighting duration, light intensity, and spectrum composition; All control commands are encapsulated through industrial IoT protocols to ensure reliable and real-time transmission.

[0013] Furthermore, the method also includes an effect evaluation and feedback mechanism: At a preset time point after the optimization measures are implemented, plant image data are re-acquired; Calculate the magnitude of change in key phenotypic traits and evaluate the effectiveness of optimization measures; Establish a drug efficacy enhancement evaluation model and analyze the correlation between changes in phenotypic characteristics and the accumulation of medicinal components; The results of the effectiveness evaluation are fed back into the cultivation knowledge base to optimize the accuracy of subsequent decisions; Based on statistical analysis of the results of multiple executions, the parameter settings of the optimization scheme are dynamically adjusted.

[0014] Furthermore, the method also includes system calibration and maintenance mechanisms: Regularly perform color calibration and geometric calibration on the image acquisition device to ensure the consistency of the acquired data; Establish an equipment status monitoring system to monitor the working status of each data acquisition unit in real time; Set up a data quality assessment module to automatically identify abnormal data collection and trigger a re-collection mechanism; The plant health status classification model is updated regularly, incorporating new training samples and features, maintaining the Gastrodia elata efficacy database, and supplementing it with the latest research results and clinical validation data.

[0015] Furthermore, the method also includes user interaction and decision support functions: It provides a visual operation interface to display plant growth status, efficacy prediction results and optimization plans, and sets up a manual intervention interface to allow cultivation experts to modify or confirm the optimization plans generated by the system. Generate cultivation management reports, including growth monitoring data, records of optimization measures and effect evaluation results, and provide historical data query function, supporting multi-dimensional retrieval by time, efficacy level and growth status; Establish a decision support knowledge base to provide solutions and expert advice for typical problems.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention achieves comprehensive acquisition of three-dimensional phenotypic information of the above-ground and underground parts of Gastrodia elata plants by constructing a multi-angle, multi-spectral image recognition sensor, overcoming the limitations of traditional observation methods that are limited to appearance. By fusing geometric, textural, and spectral features to construct a high-dimensional feature vector, it significantly improves the accuracy of plant health status identification and early warning capabilities. The use of a convolutional neural network and support vector machine fusion model for health discrimination enhances the robustness of classification for complex physiological abnormalities. A dynamic growth rate model based on time series analysis, combined with environmental factor correlation analysis, enables attribution diagnosis of growth trend deviations. By establishing a nonlinear regression prediction model between phenotypic features and medicinal component content, it achieves, for the first time, a direct mapping from symptom identification to efficacy level. Personalized cultivation optimization plans are generated based on efficacy prediction deviations, driving the coordinated regulation of multi-dimensional equipment such as water, fertilizer, environment, and light, forming a closed-loop optimization system of monitoring-diagnosis-decision-execution-feedback. Through effect evaluation and knowledge base update mechanisms, continuous iterative improvement of system performance is achieved. The overall solution significantly improves the controllability and consistency of medicinal component accumulation during the cultivation of Gastrodia elata, ensuring the stability of the quality of medicinal materials and the reliability of clinical efficacy, and promoting the transformation and upgrading of Chinese medicinal material production towards intelligence, precision, and efficacy-oriented approaches. Attached Figure Description

[0017] Figure 1 A flowchart illustrating the process of a highly efficient and ecological cultivation method for Gastrodia elata claimed in this embodiment of the invention; Figure 2 A time series analysis flowchart of a highly efficient ecological cultivation method for Gastrodia elata claimed in an embodiment of the present invention; Figure 3 A flowchart illustrating the construction of a cultivation knowledge base for a highly efficient and ecological cultivation method of Gastrodia elata as claimed in an embodiment of the present invention. Figure 4 This is a flowchart illustrating the effect evaluation and feedback process of a highly efficient ecological cultivation method for Gastrodia elata, as claimed in an embodiment of the present invention. Figure 5 This is a system calibration and maintenance flowchart of a highly efficient ecological cultivation method for Gastrodia elata claimed in an embodiment of the present invention; Figure 6 The flowchart illustrates the user interaction and decision support process for a highly efficient and ecological cultivation method of Gastrodia elata, as claimed in this embodiment of the invention. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0019] The terms "first," "second," and "third" in this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of those features. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. All directional indications in the embodiments of this application, such as up, down, left, right, front, back, etc., are only used to explain the relative positional relationships and movements between components in a specific orientation as shown in the accompanying drawings. If the specific orientation changes, the directional indications will change accordingly. Furthermore, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0020] References to embodiments herein mean that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0021] Currently, against the backdrop of high-quality development in the traditional Chinese medicine industry, Gastrodia elata, as an important medicinal plant with significant neuroregulatory and sedative effects, has its cultivation quality directly affecting the stability and controllability of clinical efficacy. Traditional Gastrodia elata cultivation relies primarily on manual experience to judge plant growth status, making management decisions based on visual observation of stem and leaf color, morphology, and other phenotypic characteristics. This method is highly subjective, lacks quantitative evidence, and struggles to identify early physiological abnormalities in a timely manner, leading to delayed intervention and severely impacting the accumulation efficiency and content consistency of key medicinal components such as gastrodin and p-hydroxybenzyl alcohol in Gastrodia elata. However, existing agricultural image recognition technologies mostly focus on pest and disease detection and yield prediction for common grains or cash crops, and have not yet established a quantitative mapping relationship between plant phenotypic characteristics and medicinal component content. Especially in underground tuber-type medicinal materials like Gastrodia elata, its efficacy is highly dependent on the physiological state at specific growth stages. Current technologies lack both the ability to acquire three-dimensional phenotypic data using multimodal images such as color and near-infrared fusion, and the ability to construct a correlation model between symptom presentation and efficacy levels. Furthermore, cultivation regulation is still limited to static adjustment of single environmental factors such as temperature and humidity, failing to achieve multi-dimensional synergistic optimization based on dynamic growth trends and expected therapeutic effects, resulting in a disconnect between management measures and efficacy goals. To address these technical problems, this invention proposes a precise cultivation regulation system that integrates multi-angle multispectral image acquisition, three-dimensional phenotypic modeling, intelligent health status identification, time-series growth analysis, efficacy prediction, and personalized regulatory decision-making. This system, from symptom identification to efficacy guidance, is applied to a highly efficient and ecological cultivation method for Gastrodia elata.

[0022] According to the first embodiment of the present invention, referring to Figure 1 This invention claims protection for a highly efficient and ecological cultivation method for Gastrodia elata, comprising: Step 1: Deploy image recognition sensors to periodically acquire color and near-infrared images of Gastrodia elata plants. Use multi-angle synchronous acquisition to acquire plant image data, construct complete three-dimensional phenotypic information of the plants, and transmit the acquired image data to the image processing server via wireless network. Step 2: Preprocess the acquired image data, extract multidimensional feature parameters from the segmented image regions, and calculate feature vectors containing three types of features: geometric, texture, and spectral. Step 3: Call the pre-trained plant health status classification model, input the feature vector, and evaluate the health status label of the Gastrodia elata plant through the discriminative architecture that combines convolutional neural network and support vector machine. Step 4: Based on time series analysis, the continuously collected feature vectors are dynamically modeled to construct the plant growth rate curve, identify growth inflection points and abnormal fluctuation ranges, and combined with environmental sensors, analyze the potential causes of growth trend deviation to obtain the current image analysis results of the Gastrodia elata plant. Step 5: Match the current image analysis results with the Gastrodia elata therapeutic effect database, predict the expected therapeutic effect level of the current plant through a nonlinear regression model, divide the optimization requirements into different priorities according to the deviation between the expected therapeutic effect level and the target therapeutic effect level, and retrieve the corresponding priority control strategy combination from the cultivation knowledge base to generate a personalized cultivation optimization plan that includes water management, nutrient allocation and environmental control measures. Step 6: Transform the optimized plan into equipment control commands to drive the integrated water and fertilizer device, environmental control equipment and supplemental lighting system to perform corresponding operations, and collect new image data after a preset time after execution to evaluate the consistency between phenotypic changes and improved efficacy.

[0023] Further, the image recognition sensor mentioned in step 1 includes: The overhead acquisition unit is arranged above the cultivation area, the lateral acquisition unit is installed between the plant rows, and the underground root observation unit is set in the soil surface layer. The top-down acquisition unit uses a multispectral imager and is equipped with simultaneous acquisition functions in the visible light band and near-infrared band. The side-view acquisition unit is equipped with an electric gimbal, which rotates at a preset angle to capture images of the side of the plant canopy. The underground root observation unit adopts a special optical mirror structure, combined with a waterproof and anti-fog device, to achieve non-destructive monitoring of root growth status.

[0024] In this embodiment, step 1 involves deploying an image recognition sensor to periodically acquire color and near-infrared images of the Gastrodia elata plant. The sensor uses a multi-angle synchronous acquisition method to acquire plant image data, constructs complete three-dimensional phenotypic information of the plant, and transmits the acquired image data to an image processing server via a wireless network.

[0025] Specifically, the image recognition sensor includes a top-view acquisition unit arranged above the cultivation area, a side-view acquisition unit installed between the plant rows, and an underground root observation unit set in the soil surface layer. The top-view acquisition unit uses a multispectral imager, configured with simultaneous acquisition functions in the visible light and near-infrared bands, to obtain the reflectance spectral characteristics of the plant canopy; The side-view acquisition unit is equipped with a motorized gimbal, which rotates at a preset angle to capture images of the side of the plant canopy, enhancing the integrity of the three-dimensional structure reconstruction. The underground root observation unit adopts a special optical mirror structure, combined with a waterproof and anti-fog device, to achieve non-destructive monitoring of root growth status and avoid plant damage and growth interference caused by excavation.

[0026] The top-view acquisition unit performs a full-area scan every 2 hours, covering all plants in the cultivation area each time, acquiring color images with a resolution of 4096×3072 pixels and near-infrared images of the corresponding resolution. The side-view acquisition unit performs a 360-degree rotation shooting twice a day at 10:00 AM and 3:00 PM, acquiring side images from 8 azimuth angles, each 45 degrees apart. The underground root system observation unit records root expansion images every 12 hours through a transparent observation window buried 5 cm below the soil surface using a high-sensitivity near-infrared camera. All acquisition units are integrated on a unified support structure with adjustable height to accommodate plant heights at different growth stages. The acquired image data is uploaded in real time to the image processing server located in the control center via a Wi-Fi 6 wireless network, using the MQTT over TLS 1.3 transmission protocol to ensure data security and low latency.

[0027] Furthermore, the preprocessing and feature extraction described in step 2 include: The acquired raw images are converted to HSV color space to increase the distinguishability of color features; A region-growing-based image segmentation algorithm, combined with morphological filtering, is used to separate the plant from the background region. In the segmented image region, geometric feature extraction includes calculating leaf area index, leaf tilt angle distribution, plant height, and canopy projected area; Texture feature extraction employs the gray-level co-occurrence matrix method to calculate contrast, correlation, energy, and homogeneity statistics; Spectral feature extraction includes calculating the normalized vegetation index, photochemical reflectance index, and water stress index.

[0028] In this embodiment, specifically for step 2, the preprocessing and feature extraction include color space conversion of the acquired original image, converting the RGB image to the HSV color space, enhancing the distinguishability of color features, and improving the recognition accuracy of physiological abnormalities such as yellowing and wilting. An image segmentation algorithm based on region growing, combined with morphological filtering, is used to effectively separate plants from the background area and reduce missegmentation caused by factors such as soil and shading. In the segmented image region, geometric feature extraction includes calculating leaf area index, leaf tilt angle distribution, plant height, and canopy projection area to reflect the plant's spatial expansion ability and light energy capture efficiency; texture feature extraction uses the gray-level co-occurrence matrix method to calculate contrast, correlation, energy, and homogeneity statistics to characterize changes in the microstructure of the leaf surface. Spectral feature extraction includes calculating the normalized vegetation index, photochemical reflectance index, and water stress index to assess the photosynthetic activity, nitrogen nutrition status, and water stress level of plants.

[0029] During the color space conversion process, the H channel is quantized to 360 levels, and the S and V channels are quantized to 100 levels each to preserve subtle color difference information; the region growing algorithm uses the center point of the plant canopy as the seed point, sets the similarity threshold to 0.15, and iterates until convergence; Morphological filtering uses circular structuring elements with a radius of 3 pixels for opening and closing operations to eliminate holes and burrs. In geometric feature calculations, the leaf area index is obtained by the ratio of the canopy projected area to the actual total leaf area. The leaf tilt angle distribution is obtained by fitting the slope of the leaf edge straight line and converting it to the angle with the horizontal plane. Plant height is determined by joint triangulation of top and side view images, with an accuracy of ±2 mm. In texture feature calculations, the distance parameter of the gray-level co-occurrence matrix is ​​set to 1, and the direction is taken at four angles: 0 degrees, 45 degrees, 90 degrees, and 135 degrees, and the average value is taken. The spectral index calculation formula is as follows: ; Wherein, NIR represents near-infrared reflectance, and Red represents red reflectance; the photochemical reflectance index PRI is defined as (531nm-570nm) / (531nm+570nm)(531nm-570nm) / (531nm+570nm), and the water stress index WSI is defined as NIR / SWIR short-wave infrared bands. The final feature vector has 32 dimensions, of which geometric features account for 8 dimensions, texture features account for 16 dimensions (4 statistics × 4 directions), and spectral features account for 8 dimensions (3 indices + 5 key band reflectances).

[0030] In the above-mentioned optimization method for Gastrodia elata cultivation, step 3 involves calling a pre-trained plant health status classification model, inputting the feature vector, and evaluating the health status label of the Gastrodia elata plant through a discriminative architecture that fuses convolutional neural networks and support vector machines.

[0031] Specifically, the plant health status classification model adopts a discriminative architecture that integrates convolutional neural networks and support vector machines. The convolutional neural network is used to automatically extract deep features from the image, while the support vector machine is used to make high-precision classification decisions in a low-dimensional feature space, thereby improving the accuracy of identifying latent symptoms such as early nutrient deficiency and pathogen infection. During the model training process, a transfer learning strategy is introduced. The model is pre-trained using a large-scale general plant disease dataset and then fine-tuned using Gastrodia elata-specific samples, which significantly reduces the need for labeled data and improves generalization ability.

[0032] The model first inputs the original image into a 5-layer convolutional neural network (CNN), with each layer containing 32 / 64 / 128 / 128 / 256 3×3 convolutional kernels, followed by a global average pooling layer, outputting 512-dimensional deep features. Simultaneously, the 32-dimensional handcrafted feature vector generated in step 2 is concatenated to this deep feature, forming a 544-dimensional fused feature. This fused feature is then input into a linear kernel support vector machine (SVM) for final classification.

[0033] The health status labeling system includes six categories: healthy, mild jaundice, severe jaundice, early wilting, late wilting, and pathogen infection. The model was trained using a dataset containing 12,000 labeled images, with 80% used for training, 10% for validation, and 10% for testing.

[0034] In the transfer learning phase, the CNN part was initialized with the weights of a ResNet-18 backbone network pre-trained on the PlantVillage dataset. In the fine-tuning phase, only the last two convolutional layers and the fully connected layers were updated. The learning rate was set to 0.001, the batch size was 32, and the model converged after 50 training epochs. The model achieved an accuracy of 96.7% on the test set and a recall of over 92% for early symptoms of mild yellowing and early wilting.

[0035] Furthermore, referring to Figure 2 The time series analysis method described in step 4 includes: Establish a feature vector time series database to store continuously collected plant phenotypic feature data; The sliding window method is used to smooth the feature vector sequence and eliminate noise interference caused by single-point acquisition error; A growth rate calculation model was constructed, and the daily growth change was obtained by the difference operation of the feature vectors of adjacent time points; Set up an abnormal growth early warning mechanism to trigger an early warning signal when the growth rate is lower than the threshold at multiple consecutive time points. By combining temperature, humidity, and light intensity data collected by environmental sensors, a correlation analysis matrix between environmental factors and growth rate is established.

[0036] In this embodiment, the time series analysis method includes establishing a feature vector time series database, continuously storing continuously collected plant phenotypic feature data, and forming a complete growth trajectory record. The sliding window method is used to smooth the feature vector sequence, eliminating noise interference caused by single-point acquisition errors and improving data stability; A growth rate calculation model was constructed, and the daily growth variation was obtained by the difference operation of the feature vectors of adjacent time points, thus quantifying the plant growth dynamics. An abnormal growth early warning mechanism was set up. When the growth rate at multiple consecutive time points is lower than a set threshold, an early warning signal is triggered, indicating that there may be environmental stress or physiological disorders. Combined with temperature, humidity and light intensity data collected by environmental sensors, a correlation analysis matrix between environmental factors and growth rate was established to identify key external factors affecting growth.

[0037] The feature vector time series database establishes independent records for each Gastrodia elata plant, with timestamps accurate to the minute. The sliding window length is set to 5 time points, or 10 hours, and a Savitzky-Golay filter is used for smoothing, with a polynomial order of 2. Growth rate calculation uses first-order differencing for key geometric features such as plant height and leaf area index, with units of mm / day or m² / day. Anomaly warning thresholds are dynamically set based on historical data from the same period, typically the mean minus 1.5 times the standard deviation. The environmental sensor network includes air temperature and humidity sensors, light intensity sensors, and soil temperature and humidity sensors, with data acquisition frequency every 10 minutes. Correlation analysis uses the Pearson correlation coefficient matrix to calculate the correlation between each environmental factor and the growth rate; an absolute value greater than 0.6 is considered a strong correlation. The current image analysis results include health status labels, growth rate values, anomaly warning indicators, and a list of major environmental influencing factors.

[0038] Furthermore, the generation of personalized cultivation optimization schemes in step 5 includes: Based on the degree of deviation between the expected efficacy level and the target efficacy, the optimization needs are divided into multiple priority levels; For different priorities, the corresponding combination of control strategies is retrieved from the cultivation knowledge base; The regulatory strategies include water management strategies, nutrient allocation strategies, and environmental control strategies; Based on the aforementioned water management strategy, adjustments to irrigation timing, irrigation volume, and irrigation methods are determined; Based on the nutrient allocation strategy, nitrogen, phosphorus, and potassium ratio schemes and trace element supplementation schemes were formulated for different growth stages. The environmental control strategy includes temperature regulation range, humidity control target, and light management scheme.

[0039] In this embodiment, a treatment efficacy prediction model is constructed and executed: The data foundation is the Gastrodia elata therapeutic efficacy database, which was constructed through long-term, multi-location cultivation experiments. Each data record is associated with the plant's temporal phenotypic feature vector throughout its entire growth cycle, namely the geometric, textural, and spectral features extracted in step 2, the corresponding environmental data logs, the implemented cultivation management measures, and the content of core medicinal components such as gastrodin, which was accurately determined by high-performance liquid chromatography after harvesting, serving as a quantitative indicator of therapeutic efficacy.

[0040] Model Training: A random forest regression algorithm is used as the core nonlinear regression model. A subset of time-series phenotypic features most highly correlated with gastrodin content, selected in the early stages (e.g., the average normalized vegetation index (NDVI) during tuber enlargement, the average daily increase in plant height within a specific time period, and the mean of leaf texture homogeneity indices), are used as input features. Supervised learning is performed with the measured percentage of gastrodin content as the regression objective. Cross-validation and feature importance ranking are conducted before model training to ensure the model's prediction accuracy and generalization ability.

[0041] Therapeutic efficacy rating mapping: This maps the continuous gastrodin content values ​​predicted by the model to discrete therapeutic efficacy ratings. For example, the following mapping rules can be configured: gastrodin content <0.2% is mapped to Level 1 (poor efficacy); 0.2% - 0.4% to Level 2 (moderate efficacy); 0.4% - 0.6% to Level 3 (good efficacy); 0.6% - 0.8% to Level 4 (excellent efficacy); and >0.8% to Level 5 (superior efficacy). These rating thresholds can be dynamically configured based on the Chinese Pharmacopoeia standards, the characteristics of different Gastrodia elata varieties, or specific commercial needs.

[0042] Optimize requirements analysis and prioritization: The system calculates the absolute difference between the expected therapeutic effect level of the current plant and the preset target therapeutic effect level, for example, level 4.

[0043] Based on the magnitude of the deviation, optimization requirements are automatically categorized into three priority levels: High priority: Deviation ≥ 2 levels. For example, a predicted level of 2 but a target level of 4 indicates a serious obstacle affecting the efficacy of treatment, requiring immediate comprehensive intensive intervention.

[0044] Medium priority: Deviation = Level 1. This indicates that the growth status is basically normal, but there is clear room for optimization, and a targeted adjustment plan needs to be developed.

[0045] Low priority: Deviation = Level 0. This indicates that current management measures are effective, and the status quo can be maintained or preventative adjustments can be made.

[0046] Intelligent generation of personalized cultivation optimization plans: Strategy Retrieval: The system retrieves information from a structured cultivation knowledge base based on the determined priority. The knowledge base exists in the form of IF (condition) THEN (strategy) rules and a case library. Conditions include: priority, growth stage, major abnormal symptoms such as excessive water stress index, and deviation of key environmental factors. For example, a rule might be: IF Priority = 'High' AND Growth Stage = 'Tuber Enlargement Stage' AND Major Limiting Factor = 'Potassium Deficiency' THEN Recommended Strategy Combination = {Increase K2O concentration in nutrient solution to X%, lower nighttime temperature by Y℃, and moderately control watering}.

[0047] The retrieved strategy combinations are synthesized into specific, quantifiable operational plans: Water management strategy: If the water stress index, such as the NDVI derivative index, is high and the predicted therapeutic effect is low, a plan to increase the irrigation frequency by 20% while keeping the amount of water per irrigation unchanged will be generated; if the plant growth is too vigorous and the therapeutic effect is insufficient, a plan to reduce the irrigation frequency and implement intermittent mild water stress may be generated.

[0048] Nutrient formulation strategy: To address deviations in predicted treatment efficacy, the fertilizer formula will be adjusted primarily in the mid-to-late stages. For example, for high-priority plants in the growth spurt stage, the program will significantly increase the potassium fertilizer ratio, such as adjusting N:P2O5:K2O to 1:1.2:2.5, and adding appropriate amounts of trace elements such as boron and zinc.

[0049] Environmental control strategy: Based on real-time environmental sensor data, control instructions are generated. For example, to promote secondary metabolism, the plan may include controlling the daytime temperature at 22-25℃ and reducing the nighttime temperature to 12-15℃, maintaining a diurnal temperature range of more than 10℃; maintaining the relative humidity of the air at 75%-85%; and turning on the supplemental lighting system on cloudy or rainy days, increasing the proportion of blue light in the spectrum to 35%, and providing supplemental lighting for 4 hours per day.

[0050] Solution output: The final optimized solution is a structured set of instructions that clearly lists the specific parameters, execution sequence and expected goals of each control measure, providing direct input for equipment control in step 6.

[0051] Closed-loop verification and knowledge base self-evolution: The generated optimization solution is not a static endpoint. After step 6 is executed, the system will restart steps 1 to 5 after a preset evaluation period, such as 7 days.

[0052] The effectiveness of the optimization program is evaluated by comparing changes in phenotypic characteristics before and after optimization, such as whether growth rate rebounds, whether spectral indices improve, and whether the new round of efficacy prediction level is improved.

[0053] The evaluation results will serve as feedback data to optimize the knowledge base: successful solutions and their associated conditions will be reinforced or stored as new cases, while ineffective solutions will be marked or adjusted, thereby enabling continuous self-iteration and optimization of cultivation decision-making rules, making the system more accurate as it is used over time.

[0054] Furthermore, referring to Figure 3 The construction of the cultivation knowledge base includes: Historical data on the cultivation of Gastrodia elata in different production areas were collected, including soil conditions, climate characteristics, management measures, and final efficacy evaluation. Establish a database of the correspondence between cultivation measures and the content of medicinal components, and formulate a table of the optimal environmental parameter ranges for different growth stages; We compiled the diagnostic features and corresponding control measures for common abnormal growth symptoms, and recorded the differences in cultivation characteristics and therapeutic effects of different varieties of Gastrodia elata.

[0055] In this embodiment, generating a personalized cultivation optimization plan includes dividing the optimization needs into multiple priority levels based on the degree of deviation between the expected efficacy level and the target efficacy level, with higher deviations corresponding to higher priorities, triggering more active regulatory interventions. For different priorities, corresponding combinations of control strategies are retrieved from the cultivation knowledge base to ensure the scientific nature and operability of the measures; the control strategies include water management strategies, nutrient allocation strategies and environmental control strategies. Water management strategies determine the adjustment plan for irrigation timing, irrigation amount and irrigation method to avoid over-irrigation leading to root hypoxia or drought stress inhibiting the synthesis of medicinal components; Nutrient allocation strategies include nitrogen, phosphorus, and potassium ratio schemes and micronutrient supplementation schemes for different growth stages, with a focus on increasing potassium supply during the tuber enlargement period to promote gastrodin accumulation. Environmental control strategies include temperature regulation range, humidity control targets, and light management schemes to maintain suitable diurnal temperature range and relative humidity, and to optimize the biosynthetic pathways of secondary metabolites.

[0056] The Gastrodia elata therapeutic efficacy database contains 5000 historical samples. Each sample records phenotypic characteristics, environmental data, cultivation measures, and the final gastrodin content determined by HPLC. The efficacy is graded into 5 levels: Level 1 <0.2%, Level 2 0.2%-0.4%, Level 3 0.4%-0.6%, Level 4 0.6%-0.8%, and Level 5 >0.8%. The nonlinear regression model uses a random forest regressor. The input is the 32-dimensional feature vector from the current image analysis results output in step 4, and the output is the expected percentage of gastrodin content, which is then mapped to the efficacy level.

[0057] The deviation is calculated as |Target Level - Prediction Level|, with a deviation ≥ 2 indicating high priority, a deviation = 1 indicating medium priority, and a deviation = 0 indicating low priority. The construction of the cultivation knowledge base includes collecting historical data on Gastrodia elata cultivation from different production areas, covering soil conditions, climate characteristics, management measures, and final efficacy evaluations, forming a multi-dimensional dataset; establishing a database of the correspondence between cultivation measures and the content of medicinal components, clarifying the impact of operations such as fertilization type and irrigation frequency on key indicators such as gastrodin and p-hydroxybenzyl alcohol; developing a table of optimal environmental parameter ranges for different growth stages to guide the dynamic control of factors such as temperature, humidity, and light; compiling diagnostic characteristics and corresponding control measures for common abnormal growth symptoms to form a standardized response process; and recording the cultivation characteristics and efficacy differences of different varieties of Gastrodia elata to support variety-specific optimization. For example, for plants in the tuber enlargement stage with a predicted therapeutic effect of level 2 and a target of level 4, the system generates a high-priority optimization plan: increase the amount of potassium fertilizer applied to 1.5 times the normal amount, reduce the amount of irrigation by 20% to induce mild water stress, reduce the nighttime temperature by 3°C to increase the diurnal temperature range, and increase the proportion of blue light in the supplemental lighting system to 40%.

[0058] Furthermore, the generation of the device control commands in step 6 includes: Analyze the parameters in the cultivation optimization plan and convert them into control signals that the equipment can execute; The control commands for the integrated water and fertilizer system include the solenoid valve opening duration, fertilizer injection ratio, and mixed liquor concentration setting. The control commands for environmental control equipment include adjustment parameters for fan speed, shading net opening, and humidifier output power; The supplemental lighting system control commands include configuration schemes for supplemental lighting duration, light intensity, and spectrum composition; All control commands are encapsulated through industrial IoT protocols to ensure reliable and real-time transmission.

[0059] In this embodiment, the optimization scheme is converted into equipment control commands to drive the integrated water and fertilizer device, environmental control equipment and supplemental lighting system to perform corresponding operations. After execution, new image data is collected after a preset time to evaluate the consistency between phenotypic changes and efficacy enhancement.

[0060] Specifically, the generation of the equipment control commands includes parsing various parameters in the cultivation optimization scheme and converting them into control signals that the equipment can execute; The control commands for the integrated water and fertilizer system include the setting of solenoid valve opening time, fertilizer pumping ratio, and mixed solution concentration to achieve precise variable fertilization; the control commands for the environmental control equipment include adjustment parameters for fan speed, shading net opening, and humidifier output power to dynamically regulate the microclimate inside the greenhouse. The supplemental lighting system control commands include configuration schemes for supplemental lighting duration, light intensity, and spectrum composition, to supplement insufficient light on cloudy or rainy days, and to adjust the ratio of red to blue light to influence secondary metabolism; All control commands are encapsulated through industrial IoT protocols to ensure reliable and real-time transmission, and to guarantee the timeliness and stability of system response.

[0061] The integrated water and fertilizer system uses a Venturi mixer, with a fertilizer injection ratio range of 0.1%-5%. The opening time of the solenoid valve is calculated based on the irrigation volume, with an accuracy of ±5 seconds. In the environmental control equipment, the fan speed adjustment range is 0-3000 rpm, the shading net opening is continuously adjustable from 0% to 100%, and the humidifier output power is 0-500W; the supplementary lighting system adopts an LED array, with the ratio of red light 660nm to blue light 450nm adjustable from 10% to 90%, and the light intensity is 0-1000 μmol / m² / s.

[0062] Control commands are sent to each device controller via the Modbus TCP protocol, with a response delay of less than 1 second. The preset duration after execution is typically 72 hours, at which point steps 1 through 5 are re-executed, forming a closed-loop feedback. The effect evaluation and feedback mechanism includes: calculating the magnitude of changes in key phenotypic characteristics to assess the actual impact of optimization measures on plant growth; establishing a drug efficacy enhancement evaluation model to analyze the correlation between phenotypic characteristic changes and medicinal component accumulation, and verifying the effectiveness of the control measures. The results of the effectiveness evaluation are fed back into the cultivation knowledge base to optimize the accuracy of subsequent decisions; Based on statistical analysis of the results of multiple executions, the parameter settings of the optimization scheme are dynamically adjusted to achieve system self-learning and continuous evolution.

[0063] For example, if after an optimization plan is implemented, the leaf area index decreases but the NDVI increases, and the predicted therapeutic effect improves, then it is considered effective, and the relevant parameter combination is marked as a high-confidence strategy.

[0064] Furthermore, referring to Figure 4 The method also includes an effectiveness evaluation and feedback mechanism: At a predetermined time point after the optimization measures were implemented, plant image data were re-acquired, the changes in key phenotypic characteristics were calculated, and the effectiveness of the optimization measures was evaluated. Establish a drug efficacy enhancement evaluation model and analyze the correlation between changes in phenotypic characteristics and the accumulation of medicinal components; The results of the effectiveness evaluation are fed back into the cultivation knowledge base to optimize the accuracy of subsequent decisions; Based on statistical analysis of the results of multiple executions, the parameter settings of the optimization scheme are dynamically adjusted.

[0065] Furthermore, referring to Figure 5The method also includes system calibration and maintenance mechanisms: Regularly perform color calibration and geometric calibration on the image acquisition device to ensure the consistency of the acquired data; Establish an equipment status monitoring system to monitor the working status of each data acquisition unit in real time; Set up a data quality assessment module to automatically identify abnormal data collection and trigger a re-collection mechanism; The plant health status classification model is updated regularly, incorporating new training samples and features, maintaining the Gastrodia elata efficacy database, and supplementing it with the latest research results and clinical validation data.

[0066] Furthermore, referring to Figure 6 The method also includes user interaction and decision support functions: It provides a visual operation interface to display plant growth status, efficacy prediction results and optimization plans, and sets up a manual intervention interface to allow cultivation experts to modify or confirm the optimization plans generated by the system. Generate cultivation management reports, including growth monitoring data, records of optimization measures and effect evaluation results, and provide historical data query function, supporting multi-dimensional retrieval by time, efficacy level and growth status; Establish a decision support knowledge base to provide solutions and expert advice for typical problems.

[0067] In this embodiment, a system calibration and maintenance mechanism is also included, which regularly performs color calibration and geometric calibration on the image acquisition device, uses standard color cards and calibration boards to eliminate data deviations caused by equipment drift, and ensures the consistency of long-term acquired data; and establishes an equipment status monitoring system to monitor the working status of each acquisition unit in real time and promptly alarm when a fault is detected. A data quality assessment module is set up to automatically identify abnormal data collection and trigger a re-collection mechanism to ensure the reliability of input data; the plant health status classification model is updated regularly, incorporating new training samples and features to improve the model's ability to identify new symptoms. Maintain the Gastrodia elata efficacy database, supplementing it with the latest research findings and clinical validation data to ensure the cutting-edge nature and authority of the knowledge system. Color calibration is performed weekly using the X-Rite ColorChecker standard color chart. Geometric calibration is performed monthly using a checkerboard calibration board; equipment status monitoring is achieved through a heartbeat mechanism, triggering an alarm if the timeout exceeds 30 seconds; data quality assessment is based on indicators such as image clarity, illumination uniformity, and signal-to-noise ratio, and re-sampling is initiated if any indicator falls below the threshold.

[0068] This embodiment also integrates user interaction and decision support functions: It provides a visual operation interface to display plant growth status, efficacy prediction results, and optimization plan implementation progress in real time, and supports graphical data presentation; it also sets up a manual intervention interface to allow cultivation experts to modify or confirm the optimization plan generated by the system, thus retaining human decision-making authority. Generate cultivation management reports, including growth monitoring data, records of optimization measures, and results of effectiveness evaluation, which facilitate archiving and traceability; It provides a historical data query function, supporting multi-dimensional searches by time, efficacy level, growth status, etc., to assist scientific research analysis; Establish a decision support knowledge base to provide solutions to typical problems and expert advice, thereby improving the system's usability and scalability.

[0069] The visual interface uses web front-end technology and supports access from PCs and mobile devices; the manual intervention interface provides controls such as sliders and drop-down menus to adjust key parameters in the optimization plan; the cultivation management report is automatically generated in PDF format and includes a timeline, charts and text descriptions.

[0070] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0071] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units. The above are merely embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0072] The specific embodiments of the invention have been described in detail above, but they are only examples, and this application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of this application. Therefore, all equivalent changes, modifications, and improvements made without departing from the spirit and principles of this application should be covered within the scope of this application.

Claims

1. A highly efficient and ecological cultivation method for Gastrodia elata, characterized in that, include: Step 1: Deploy image recognition sensors to periodically acquire color and near-infrared images of Gastrodia elata plants. Use multi-angle synchronous acquisition to acquire plant image data, construct complete three-dimensional phenotypic information of the plants, and transmit the acquired image data to the image processing server via wireless network. Step 2: Preprocess the acquired image data, extract multidimensional feature parameters from the segmented image regions, and calculate feature vectors containing three types of features: geometric, texture, and spectral. Step 3: Call the pre-trained plant health status classification model, input the feature vector, and evaluate the health status label of the Gastrodia elata plant through the discriminative architecture that combines convolutional neural network and support vector machine. Step 4: Based on time series analysis, the continuously collected feature vectors are dynamically modeled to construct the plant growth rate curve, identify growth inflection points and abnormal fluctuation ranges, and combined with environmental sensors, analyze the potential causes of growth trend deviation to obtain the current image analysis results of the Gastrodia elata plant. Step 5: Match the current image analysis results with the Gastrodia elata therapeutic effect database, predict the expected therapeutic effect level of the current plant through a nonlinear regression model, divide the optimization requirements into different priorities according to the deviation between the expected therapeutic effect level and the target therapeutic effect level, and retrieve the corresponding priority control strategy combination from the cultivation knowledge base to generate a personalized cultivation optimization plan that includes water management, nutrient allocation and environmental control measures. Step 6: Transform the optimized plan into equipment control commands to drive the integrated water and fertilizer device, environmental control equipment and supplemental lighting system to perform corresponding operations, and collect new image data after a preset time after execution to evaluate the consistency between phenotypic changes and improved efficacy.

2. The method according to claim 1, characterized in that, The image recognition sensor described in S2 includes: The overhead acquisition unit is arranged above the cultivation area, the lateral acquisition unit is installed between the plant rows, and the underground root observation unit is set in the soil surface layer. The top-down acquisition unit uses a multispectral imager and is equipped with simultaneous acquisition functions in the visible light band and near-infrared band. The side-view acquisition unit is equipped with an electric gimbal, which rotates at a preset angle to capture images of the side of the plant canopy. The underground root observation unit adopts a special optical mirror structure, combined with a waterproof and anti-fog device, to achieve non-destructive monitoring of root growth status.

3. The method according to claim 1, characterized in that, The preprocessing and feature extraction described in step 2 include: The acquired raw images are converted to HSV color space to increase the distinguishability of color features; A region-growing-based image segmentation algorithm, combined with morphological filtering, is used to separate the plant from the background region. In the segmented image region, geometric feature extraction includes calculating leaf area index, leaf tilt angle distribution, plant height, and canopy projected area; Texture feature extraction employs the gray-level co-occurrence matrix method to calculate contrast, correlation, energy, and homogeneity statistics; Spectral feature extraction includes calculating the normalized vegetation index, photochemical reflectance index, and water stress index.

4. The method according to claim 1, characterized in that, The time series analysis method described in step 4 includes: Establish a feature vector time series database to store continuously collected plant phenotypic feature data; The sliding window method is used to smooth the feature vector sequence and eliminate noise interference caused by single-point acquisition error; A growth rate calculation model was constructed, and the daily growth change was obtained by the difference operation of the feature vectors of adjacent time points; Set up an abnormal growth early warning mechanism to trigger an early warning signal when the growth rate is lower than the threshold at multiple consecutive time points. By combining temperature, humidity, and light intensity data collected by environmental sensors, a correlation analysis matrix between environmental factors and growth rate is established.

5. The method according to claim 1, characterized in that, Step 5, which involves generating a personalized cultivation optimization plan, includes: Based on the degree of deviation between the expected efficacy level and the target efficacy level, the optimization needs are divided into multiple priority levels: high, medium, and low. For different priority levels, corresponding water management strategies, nutrient allocation strategies and environmental control strategies are retrieved from the cultivation knowledge base and combined into a comprehensive control plan; Based on the aforementioned water management strategy, and according to the plant's water stress index and growth stage, adjustments to irrigation timing, irrigation volume, and irrigation method are determined. Based on the nutrient allocation strategy, and according to the predicted therapeutic effect deviation and plant growth rate, nitrogen, phosphorus, and potassium element ratio schemes and micronutrient supplementation schemes are formulated for different growth stages. Based on the aforementioned environmental control strategy, and in conjunction with environmental sensor data and therapeutic goals, the temperature regulation range, humidity control targets, and light management scheme are determined. The light management scheme includes the configuration of supplemental lighting duration, light intensity, and spectral composition.

6. The method according to claim 5, characterized in that, The construction of the cultivation knowledge base includes: Historical data on the cultivation of Gastrodia elata in different production areas were collected, including soil conditions, climate characteristics, management measures, and final efficacy evaluation. Establish a database of the correspondence between cultivation measures and the content of medicinal components; Establish a table of optimal environmental parameter ranges for different growth stages; Summarize the diagnostic features of common growth abnormalities and corresponding control measures; Record the differences in cultivation characteristics and therapeutic effects of different varieties of Gastrodia elata.

7. The method according to claim 1, characterized in that, The generation of device control commands in step 6 includes: Analyze the parameters in the cultivation optimization plan and convert them into control signals that the equipment can execute; The control commands for the integrated water and fertilizer system include the solenoid valve opening duration, fertilizer injection ratio, and mixed liquor concentration setting. The control commands for environmental control equipment include adjustment parameters for fan speed, shading net opening, and humidifier output power; The supplemental lighting system control commands include configuration schemes for supplemental lighting duration, light intensity, and spectrum composition; All control commands are encapsulated through industrial IoT protocols to ensure reliable and real-time transmission.

8. The method according to claim 1, characterized in that, It also includes an effectiveness evaluation and feedback mechanism: At a preset time point after the optimization measures are implemented, plant image data are re-acquired; Calculate the magnitude of change in key phenotypic traits and evaluate the effectiveness of optimization measures; Establish a drug efficacy enhancement evaluation model and analyze the correlation between changes in phenotypic characteristics and the accumulation of medicinal components; The results of the effectiveness evaluation are fed back into the cultivation knowledge base to optimize the accuracy of subsequent decisions; Based on statistical analysis of the results of multiple executions, the parameter settings of the optimization scheme are dynamically adjusted.

9. The method according to claim 1, characterized in that, It also includes system calibration and maintenance mechanisms: Regularly perform color calibration and geometric calibration on the image acquisition device to ensure the consistency of the acquired data; Establish an equipment status monitoring system to monitor the working status of each data acquisition unit in real time; Set up a data quality assessment module to automatically identify abnormal data collection and trigger a re-collection mechanism; The plant health status classification model is updated regularly, incorporating new training samples and features, maintaining the Gastrodia elata efficacy database, and supplementing it with the latest research results and clinical validation data.

10. The method according to claim 1, characterized in that, The method also includes user interaction and decision support functions: It provides a visual operation interface to display plant growth status, efficacy prediction results and optimization plans, and sets up a manual intervention interface to allow cultivation experts to modify or confirm the optimization plans generated by the system. Generate cultivation management reports, including growth monitoring data, records of optimization measures and effect evaluation results, and provide historical data query function, supporting multi-dimensional retrieval by time, efficacy level and growth status; Establish a decision support knowledge base to provide solutions and expert advice for typical problems.

Images