Plant growth environment regulation and control method, device and system and storage medium

Abstract

The invention discloses a plant growth environment regulation and control method, device and system and a storage medium. The method comprises the following steps: acquiring multi-dimensional plant growth data; obtaining a growth prediction model according to the multi-dimensional plant growth data; acquiring environmental parameter data of illumination intensity, temperature and humidity and carbon dioxide concentration in real time; when the environmental parameter data fluctuates severely, an early warning mechanism is triggered, and the working parameters of the artificial light source, the humidifying equipment and the ventilation equipment are automatically adjusted according to the feedback of the growth prediction model. By means of the technical scheme, the growth quality and yield of plants can be effectively improved, manual intervention is reduced, and the planting cost is reduced.

Description

Technical Field

[0001] The present invention belongs to the field of information technology, and in particular relates to a method and device, a system, and a storage medium for regulating a plant growth environment. Background Art

[0002] In the process of developing a universal plant growth environment control system based on AI, many technical difficulties and contradictions are faced. First of all, different plant species have very different requirements for the growth environment. For example, the optimal ranges of parameters such as light intensity, temperature and humidity, carbon dioxide concentration, and nutrient solution ratio are different. Even the same plant's requirements for the environment will change at different growth stages. The system needs to collect and analyze a large amount of plant physiological data, and it must have self-learning capabilities, and continuously optimize the growth model suitable for specific plants as the data accumulates. Secondly, due to the inconsistency of the growth cycles of different plants, when a variety of plants are mixed and planted in a system, it is also necessary to coordinate the growth rhythms of different plants, reasonably allocate resources, and avoid mutual influence. In addition, the dynamic changes in the external environment also bring challenges to system control. How to respond to climate fluctuations, day and night changes and other factors in real time and dynamically adjust parameters is also a major difficulty. In short, the universal plant growth environment control system needs to take into account both applicability and accuracy, and flexibly respond to complex and changeable actual scenarios. It has high requirements for system architecture design, algorithm model selection, and comprehensive integration of software and hardware, and needs to overcome a series of technical difficulties. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method and device, a system and a storage medium for regulating the plant growth environment.

[0004] To achieve the above object, the present invention adopts the following technical solution:

[0005] A method for regulating a plant growth environment, comprising:

[0006] Obtain multi-dimensional plant growth data;

[0007] Obtaining a growth prediction model according to the multi-dimensional plant growth data;

[0008] Real-time collection of environmental parameter data such as light intensity, temperature and humidity, and carbon dioxide concentration;

[0009] When the environmental parameter data fluctuates drastically, the early warning mechanism is triggered, and the working parameters of the artificial light source, humidification equipment, and ventilation equipment are automatically adjusted according to the feedback of the growth prediction model.

[0010] Preferably, the multi-dimensional plant growth data includes: light intensity, temperature and humidity, carbon dioxide concentration, and nutrient solution ratio.

[0011] Preferably, based on the multi-dimensional plant growth data type, a machine learning algorithm is used to train a growth prediction model for a specific plant.

[0012] The present invention also provides a plant growth environment control device, comprising:

[0013] An acquisition module, used to acquire multi-dimensional plant growth data;

[0014] A training module, used for obtaining a growth prediction model according to the multi-dimensional plant growth data;

[0015] The acquisition module is used to collect real-time environmental parameter data such as light intensity, temperature and humidity, and carbon dioxide concentration;

[0016] The control module is used to trigger the early warning mechanism when the environmental parameter data fluctuates violently, and automatically adjust the working parameters of the artificial light source, humidification equipment, and ventilation equipment according to the feedback of the growth prediction model.

[0017] Preferably, the multi-dimensional plant growth data includes: light intensity, temperature and humidity, carbon dioxide concentration, and nutrient solution ratio.

[0018] Preferably, the training module uses a machine learning algorithm to train a growth prediction model for a specific plant based on the multi-dimensional plant growth data type.

[0019] The present invention also provides a plant growth environment regulation system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a plant growth environment regulation method when executed by the processor.

[0020] The present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the plant growth environment regulation method when running.

[0021] The present invention establishes a plant growth data model by collecting multidimensional parameters such as light intensity, temperature and humidity, and carbon dioxide concentration, and uses a machine learning algorithm to train a growth prediction model for specific plants. A variety of sensors are deployed in the plant growth environment to monitor environmental parameters in real time. When environmental parameters fluctuate violently, the system triggers an early warning mechanism and automatically adjusts the working parameters of artificial light sources, humidification equipment, ventilation equipment, etc. according to the feedback of the growth prediction model to maintain the best growth environment. The present invention can effectively improve the growth quality and yield of plants, reduce manual intervention, and reduce planting costs through intelligent environmental monitoring and control. At the same time, the prediction and automatic adjustment functions of the system can respond to environmental changes in a timely manner and improve the stability and reliability of plant planting. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0023] Figure 1 The present invention is a flow chart of a method for regulating plant growth environment based on AI. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below in combination with the drawings and specific embodiments.

[0025] like Figure 1 As shown, a method for regulating plant growth environment according to an embodiment of the present invention includes:

[0026] Step S101, obtaining parameters such as light intensity, temperature and humidity, carbon dioxide concentration, and nutrient solution ratio, and establishing a multidimensional plant growth data model.

[0027] By deploying light intensity sensors, temperature and humidity sensors, carbon dioxide concentration sensors and other equipment in the plant growth environment, multi-dimensional growth environment parameters such as light intensity parameters, temperature parameters, humidity parameters, and carbon dioxide concentration parameters in the plant growth process can be obtained in real time. According to the obtained multi-dimensional growth environment parameters such as light intensity parameters, temperature parameters, humidity parameters, and carbon dioxide concentration parameters, combined with the results of plant growth status evaluation, a mathematical model between plant growth environment parameters and plant growth status is established through a multivariate linear regression algorithm to obtain a multi-dimensional plant growth data model. The support vector machine algorithm is used to train a plant growth prediction model with multi-dimensional growth environment parameters such as light intensity parameters, temperature parameters, humidity parameters, and carbon dioxide concentration parameters as input and plant growth status evaluation results as output. The nutrient solution ratio parameters applied during plant growth are obtained and used as an input dimension of the multi-dimensional plant growth data model. The data dimension is reduced through the principal component analysis algorithm to optimize the training efficiency and prediction accuracy of the model. At different stages of plant growth, according to the prediction results of the multi-dimensional plant growth data model, the parameters such as light intensity, temperature and humidity, and carbon dioxide concentration of the plant growth environment are dynamically adjusted to provide the optimal environment for plant growth. The environmental parameter data collected from each stage of plant growth, the applied nutrient solution ratio data, the plant growth status assessment data, etc. are uniformly stored in the database to form a comprehensive plant growth data set. Data mining and analysis are performed on the plant growth data set to extract the regular knowledge hidden in the data, optimize the multidimensional plant growth data model, improve the model's interpretability and generalization capabilities, and provide data support for the intelligent management of plant factories.

[0028] Furthermore, light intensity sensors, temperature and humidity sensors, carbon dioxide concentration sensors and other equipment are deployed in the plant growth environment to collect data every 10 minutes to obtain multi-dimensional growth environment parameters such as light intensity parameters, temperature parameters, humidity parameters, and carbon dioxide concentration parameters during plant growth. A mathematical model between plant growth environment parameters and plant growth status is established through a multivariate linear regression algorithm to obtain a multi-dimensional plant growth data model, and the determination coefficient R of the model is 2Reached 85 or more. Using the support vector machine algorithm, with multi-dimensional growth environment parameters such as light intensity parameters, temperature parameters, humidity parameters, and carbon dioxide concentration parameters as input, and the results of plant growth status evaluation as output, a plant growth prediction model was trained, and the prediction accuracy of the model reached more than 90%. The ratio parameters of N, P, K and other nutrient elements applied during plant growth were obtained, and the dimensions were reduced by the principal component analysis algorithm. The variance contribution rate of the first three principal components exceeded 80%, optimizing the training efficiency and prediction accuracy of the model. At different stages of plant growth, according to the prediction results of the multi-dimensional plant growth data model, the parameters such as light intensity, temperature and humidity, and carbon dioxide concentration of the plant growth environment are dynamically adjusted. The light intensity is maintained at 10000-50000 lux, the temperature is maintained at 20-30℃, the humidity is maintained at 60%-80%, and the carbon dioxide concentration is maintained at 800-1200ppm. The environmental parameter data collected at each stage of plant growth, the applied nutrient solution ratio data, the plant growth status evaluation data, etc. are uniformly stored in the time series database InfluxDB, and the data is automatically backed up every day. The association rule algorithm is used to perform data mining analysis on the plant growth data set, extract the regular knowledge hidden in the data, optimize the multidimensional plant growth data model, reduce the model's mean absolute percentage error (MAPE) by more than 10%, improve the model's interpretability and generalization ability, and provide data support for the intelligent management of plant factories.

[0029] The mathematical model between plant growth environment parameters and plant growth status was established by multivariate linear regression algorithm to ensure the determination coefficient R 2 Reach above 0.85.

[0030] Obtain the environmental parameter data and corresponding plant growth status data during plant growth, and preprocess the data, including removing outliers, missing values, etc. Based on the preprocessed data, use the multiple linear regression algorithm to construct a mathematical model between plant growth environmental parameters and plant growth status. When constructing the model, select appropriate independent variables and dependent variables, the independent variables are plant growth environmental parameters, and the dependent variables are plant growth status indicators. Through the multiple linear regression algorithm, obtain the regression equation coefficients and intercepts, and establish a mathematical model between plant growth environmental parameters and plant growth status. Use the determination coefficient R 2 Evaluate the goodness of fit of the constructed mathematical model to determine whether the model meets the preset requirement of 85 or above. If the determination coefficient R2 does not reach 85 or above, adjust the independent variable selection and data processing methods in the multivariate linear regression model and rebuild the model until the requirements are met. Apply the constructed mathematical model of plant growth environment parameters and plant growth status to the actual plant growth prediction and regulation, predict the plant growth status according to the changes in environmental parameters, and guide the optimization of plant growth environment conditions.

[0031] Furthermore, during the growth of plants, environmental parameter data such as temperature, humidity, and light intensity are collected through sensors, and growth status data such as plant height, leaf area, and biomass are recorded. The collected data are cleaned, outliers are removed using the 3σ principle, and missing values ​​are filled with mean values. The processed data are divided into training sets and test sets in a ratio of 8:2. Temperature, humidity, and light intensity are selected as independent variables, and plant height is selected as the dependent variable. The mathematical model is constructed using the multivariate linear regression algorithm. The coefficients and intercepts of the regression equation are solved by the least squares method, and the model is obtained: plant height = 5×temperature + 3×humidity + 2×light intensity-10. The coefficient of determination R is calculated using the test set data. 2 The result is 87, which meets the model fitting goodness requirement. The model is applied to actual production. When the temperature is 25℃, the humidity is 60%, and the light intensity is 10000lux, the predicted plant height is 17cm. Based on this, the environmental parameters are adjusted to optimize plant growth conditions and improve yield and quality.

[0032] By using the support vector machine algorithm and taking multi-dimensional growth environment parameters as input, the accuracy of the plant growth status prediction model reached over 90%.

[0033] The multidimensional environmental parameter data during the plant growth process is obtained as the input of the support vector machine algorithm. The obtained multidimensional parameter data is preprocessed, including data cleaning, normalization and other operations to improve the data quality. According to the preprocessed parameter data, the support vector machine algorithm is used to construct a plant growth state prediction model. The constructed prediction model is evaluated by cross-validation and other methods to obtain the prediction accuracy of the model. If the model prediction accuracy does not reach more than 90%, the parameters of the support vector machine algorithm are adjusted and the prediction model is rebuilt. The prediction model is continuously iterated and optimized until the model prediction accuracy reaches more than 90%. The constructed plant growth state prediction model is applied to actual production, and the growth state of the plant is predicted based on the real-time collected environmental parameter data to provide decision support for optimizing the growth environment.

[0034] Furthermore, during the plant growth process, sensors are used to collect multidimensional environmental parameter data such as temperature, humidity, light intensity, and carbon dioxide concentration in real time, every 5 minutes, as input to the support vector machine algorithm. The collected data is cleaned to remove outliers, such as data with a temperature exceeding 50°C or below -10°C. The data is then normalized to its maximum and minimum values, and the data is mapped to the [0,1] interval to improve data quality. Using the preprocessed data, the Gaussian kernel function is selected to construct a support vector machine prediction model, and the prediction accuracy of the model on the test set is evaluated through 5-fold cross validation. If the accuracy is less than 90%, the penalty coefficient C and kernel function parameter γ of the support vector machine are adjusted, and the model is retrained. Iterate continuously until the optimal parameter combination is found, so that the prediction accuracy reaches more than 95%. The trained model is deployed to the plant growth monitoring system, and environmental parameter data is automatically collected every 5 minutes, and substituted into the model for prediction. If the prediction result is "abnormal", an alarm is immediately triggered to notify the management personnel to adjust the environmental parameters, optimize the plant growth environment, and improve yield and quality. At the same time, the prediction results will be compared and analyzed with the actual growth status, and the prediction model will be continuously improved and updated to make it adapt to more plant varieties and growth environments.

[0035] By using association rule algorithm to conduct data mining analysis on plant growth data sets, we obtained an optimization effect in which the mean absolute percentage error (MAPE) of the model was reduced by more than 10%.

[0036] Obtain a plant growth data set, preprocess the data, remove missing values ​​and outliers, and perform data normalization. Use the Apriori association rule algorithm, set appropriate support and confidence thresholds, and mine frequent item sets and association rules from the plant growth data set. Based on the mined association rules, build a plant growth prediction model, and select machine learning algorithms such as decision trees and random forests. Divide the data set into a training set and a test set, use the training set data to train the prediction model, and evaluate it on the test set. Calculate the mean absolute percentage error (MAPE) of the model on the test set as an evaluation indicator of model performance. If the MAPE is higher than the preset threshold, adjust the support and confidence thresholds of the association rules, re-mine the association rules, and update the prediction model. Continuously iterate and optimize the model until the MAPE is reduced by more than 10%, and obtain a plant growth prediction model that meets business needs.

[0037] Furthermore, first, the growth data within one year was obtained from the plant growth monitoring system, totaling 10,000 records. The data was preprocessed using Python's Pandas library, features with missing values ​​exceeding 20% ​​were deleted, and outliers were replaced by medians. Then the data was normalized using the minimum-maximum normalization method to map the values ​​to the [0,1] interval. Next, the Apriori association rule algorithm was used to mine frequent item sets and association rules on the data. By setting the minimum support to 05 and the minimum confidence to 8, 50 association rules were mined. According to the association rules, the decision tree algorithm was selected to construct a plant growth prediction model. The data set was divided into a training set and a test set using the ten-fold cross validation method, and the decision tree model was trained using the training set data, and the performance was evaluated on the test set. After calculation, the MAPE of the model on the test set was 15%, which was higher than the preset 10% threshold. Therefore, the minimum support of the Apriori algorithm was adjusted to 03, the minimum confidence was adjusted to 75, the association rules were mined again, and the decision tree model was updated based on the new association rules. Through continuous iterative optimization, we finally obtained a plant growth prediction model with a MAPE of 8%, which meets business needs. This model can predict the growth of plants in the next week based on plant growth environment factors such as temperature, humidity, light, etc., and provide decision support for agricultural production.

[0038] Step S102: Based on the multi-dimensional plant growth data model, a machine learning algorithm is used to train a growth prediction model for a specific plant.

[0039] According to the multidimensional data related to plant growth, a data model describing the plant growth process is established. The data model contains various factors that affect plant growth and the relationship between them. According to the established plant growth data model, the data preprocessing technology is used to clean and transform the original data to obtain a normalized data set suitable for machine learning algorithm training. According to the characteristics of the data set, a suitable machine learning algorithm, such as support vector machine, random forest, etc., is selected. By setting the relevant parameters of the algorithm, a growth prediction model for a specific plant is constructed. The normalized data set is divided into a training set and a test set. The prediction model is trained using the training set data. The model parameters are optimized through multiple iterations so that the model can fit the training data well. The performance of the trained prediction model is evaluated using the test set data. The prediction accuracy and generalization ability of the model are comprehensively judged using methods such as cross-validation to further optimize the model. According to the model evaluation results, the data preprocessing method and the selection of the machine learning algorithm are adjusted, and the model hyperparameters are optimized. The prediction model is retrained to obtain a model with better performance. The trained plant growth prediction model is applied to actual agricultural production. According to the prediction results of the model, agricultural operations are optimized, scientific planting is guided, and yield and quality are improved.

[0040] Furthermore, when establishing a plant growth data model, data from multiple dimensions such as temperature, humidity, light, and soil nutrients can be considered, and correlation analysis can be used to determine the degree of association between the various factors. For example, the correlation coefficient between temperature and light reaches 8, indicating that the two have a relatively consistent effect on plant growth. When cleaning the original data, outliers can be filtered by setting a threshold, such as removing data points that exceed 3 times the standard deviation, and then using methods such as maximum and minimum value normalization to convert the data into values ​​between 0 and 1, so that it is suitable for machine learning algorithm processing. When selecting an algorithm, parameters can be optimized by grid search and other methods, such as setting the penalty coefficient C of the support vector machine to 0, selecting the Gaussian kernel for the kernel function, and then dividing the data set into a training set and a test set at a ratio of 8:2. The model was cross-validated with a 10-fold cross validation using the training set, and the average accuracy was taken as the model performance indicator, and it was continuously iterated and optimized. The evaluation results of the test set showed that the average absolute error of the model was 0.5 and the determination coefficient reached 95, indicating that the model had good prediction performance. In practical applications, the model can be used to predict plant growth trends in the next seven days, and dynamically adjust agricultural operations such as watering and fertilizing based on the prediction results to achieve precise management, which is expected to increase yields by more than 5%.

[0041] According to the complexity of multidimensional factors of plant growth, correlation analysis is used to determine the degree of association between various factors and obtain the key factors affecting plant growth.

[0042] According to the multi-dimensional data collected during the plant growth process, data preprocessing technology is used to clean and standardize the data to eliminate the influence of data noise and outliers. The correlation coefficient matrix between each growth factor is calculated by the correlation analysis algorithm to quantitatively evaluate the linear correlation between each factor. The association network model of plant growth factors is constructed according to the correlation coefficient matrix, and the community discovery algorithm is used to identify the closely related core factor community in the network. The core factor community is analyzed for significance and importance sorting to automatically determine the key factor combination affecting plant growth. The decision tree algorithm is used to establish a mapping model between the key factor combination of plant growth and the growth state to achieve plant growth prediction based on key factors. The feature selection algorithm is used to select the key factor subset with the best prediction effect from the original high-dimensional growth factors to obtain a simplified plant growth model. The cross-validation method is used to evaluate the prediction performance of the simplified model, and the model parameters are continuously iterated to optimize the accuracy and generalization ability of plant growth prediction.

[0043] Furthermore, during the plant growth process, sensors collected data in multiple dimensions such as temperature, humidity, and light intensity. First, the median filter algorithm was used to clean the data and remove outliers; then the minimum-maximum normalization method was used to map the data to the [0,1] interval to eliminate the dimension effect. Then, the Pearson correlation coefficient was used to calculate the correlation between the growth factors to obtain a 10×10 symmetric matrix. Based on the matrix, the association network model was constructed, and the Louvain algorithm was used to identify three closely related core factor communities. The key factor combination composed of temperature, humidity, and light intensity was screened out by chi-square test and information gain rate calculation. The C5 decision tree algorithm was used to establish a classification model of key factors and plant growth status, and the node splitting was performed by the information entropy criterion, and the pruning strategy was used to avoid overfitting. The recursive feature elimination method was used to select 5 optimal feature subsets from the 15 original growth factors to obtain a simplified growth prediction model. The model was cross-validated with 10 folds, and the average prediction accuracy reached more than 85%. Finally, grid search was used to optimize the decision tree parameters such as the number of split nodes and the number of leaf node samples. After five rounds of iterations, the prediction accuracy of the model was further improved to 90%, achieving efficient and accurate plant growth prediction.

[0044] Through anomaly detection and normalization of multidimensional data, data preprocessing techniques are used to obtain a standardized data set suitable for machine learning algorithms.

[0045] According to the dimensions and attributes of the data, statistical methods are used to detect outliers in multidimensional data and remove abnormal data. According to the numerical range and distribution of different data attributes, appropriate normalization methods are selected for processing to map the data to a unified scale space. Feature subsets that are effective for machine learning algorithms are screened out from the normalized data set through feature selection and feature extraction techniques. According to the data distribution characteristics of the feature subset, data transformation methods are used to smooth the feature data to improve the standardization and uniformity of the data. The preprocessed normalized data set is divided into training sets and test sets, and a suitable machine learning algorithm is selected for model training. The performance of the trained model is evaluated by cross-validation and other methods, and the model is optimized and improved based on the evaluation results. The trained and optimized machine learning model is applied to actual multidimensional data processing scenarios to predict and analyze the data and provide data support for business decisions.

[0046] Furthermore, when performing outlier detection on multidimensional data, distance-based methods such as Mahalanobis distance can be used. By calculating the Mahalanobis distance between each data point and the center of the data set, data points whose distance exceeds a threshold (such as 3 times the standard deviation) are considered outliers and removed. For different data attributes, appropriate normalization methods can be selected. For example, for attributes with a value range between 0 and 1, min-max normalization can be performed directly; for attributes with a normal distribution, z-score standardization can be performed first, and then min-max normalization can be performed to map the data to the [0,1] interval. When selecting features, the Pearson correlation coefficient between the feature and the target variable can be calculated, and features with a correlation coefficient greater than 6 can be selected as valid feature subsets. For the data distribution of feature subsets, smoothing processing can be performed through methods such as Box-Cox transformation to make the data more standardized and unified. The preprocessed data is divided into training set and test set in a ratio of 8:2. The support vector machine (SVM) algorithm is selected for model training, and the model performance is evaluated by 5-fold cross validation. The hyperparameters such as the kernel function and penalty coefficient of SVM are continuously adjusted until the accuracy of the model on the test set reaches more than 95%. Finally, the optimized SVM model is applied to the actual multidimensional data processing scenario to predict and analyze the data and provide data support for business decisions, such as predicting the customer's future purchase intention based on the customer's historical purchase behavior and personal attributes, providing a basis for precision marketing.

[0047] Through model training and hyperparameter optimization, cross-validation and other methods are used to determine the accuracy and generalization ability of the specific plant growth prediction model.

[0048] Obtain growth data of specific plants, including environmental factors, growth status, etc., as a data set for model training and validation. Preprocess the data, including missing value processing, outlier processing, data normalization, etc., to improve data quality. According to the needs of plant growth prediction, select appropriate machine learning algorithms, such as decision trees, support vector machines, neural networks, etc. Divide the data set into training sets and validation sets. The training set is used for model training, and the validation set is used for model performance evaluation. Train the model and optimize the model performance by adjusting hyperparameters, such as learning rate, regularization parameters, etc. Use cross-validation methods, such as k-fold cross-validation, to evaluate the performance of the model on different data subsets and select the model with the best performance. Evaluate the accuracy and generalization ability of the model on an independent test set to ensure that the model can achieve good results in practical applications.

[0049] Furthermore, when obtaining plant growth data, IoT sensors can be used to collect data on environmental factors such as temperature, humidity, and light in real time, and image recognition technology can be used to analyze growth status parameters such as plant leaf color and size. The collected data is cleaned, the maximum likelihood estimation method is used to fill missing values, and outliers are identified and removed using methods such as box plots, and the data is scaled to between 0 and 1 through minimum-maximum normalization. According to the needs of plant growth prediction, the long short-term memory neural network (LSTM) algorithm is selected, which can effectively process time series data and capture long-term dependencies in the process of plant growth. The data set is divided into a training set and a validation set in a ratio of 8:2. Hyperparameters such as learning rate, number of hidden layers, and regularization coefficient are tuned through grid search and Bayesian optimization methods, and the model performance is continuously iterated and optimized. Using 5-fold cross validation, the training set is divided into 5 subsets, and 4 of the subsets are used as training data and 1 subset is used as validation data in turn. The model is trained and evaluated, and the model with the best average performance is finally selected. On an independent test set, evaluation indicators such as root mean square error (RMSE) and mean absolute error (MAE) were used to evaluate the prediction accuracy of the model to ensure that the model can accurately predict the growth trend of plants in the next 7 days, with the average error controlled within 5%, providing reliable decision support for smart agriculture.

[0050] Step S103, deploying a variety of sensors in the plant growth environment to collect environmental parameter data such as light intensity, temperature and humidity, and carbon dioxide concentration in real time.

[0051] According to the pre-established plant growth environment model, the target environmental parameter threshold in the model is obtained; through the deployment of a variety of sensors in the plant growth environment, the environmental parameter data including light intensity, ambient temperature, ambient humidity and carbon dioxide concentration are collected in real time; the collected environmental parameter data is transmitted to the data processing module, and the data is preprocessed, including data cleaning, data normalization, etc.; the preprocessed environmental parameter data is input into the pre-trained machine learning model, and the growth state of the plant under the current environmental parameters is obtained through the model prediction; the predicted plant growth state is compared with the target environmental parameter threshold to determine whether the current environment is suitable for plant growth; if the judgment result is no, according to the predicted plant growth state, a reinforcement learning algorithm is used to learn and optimize the current environmental parameters to make them as close as possible to the target threshold in the model; according to the optimized environmental parameters obtained by the reinforcement learning algorithm, environmental conditioning equipment such as ventilation equipment, humidification equipment, supplementary lighting equipment, etc. are controlled to adjust the plant growth environment and improve the growth state of the plant.

[0052] Furthermore, according to the pre-established plant growth environment model, the target environmental parameter thresholds in the model are obtained. For example, the model determines that the optimal plant growth temperature range is 22-28°C, the humidity range is 60%-80%, the light intensity range is 10000-50000 lux, and the carbon dioxide concentration range is 600-1000ppm. By deploying high-precision digital temperature and humidity sensors, light intensity sensors, and carbon dioxide concentration sensors in the plant growth environment, the environmental parameter data is collected in real time, and the sampling frequency is once per minute. The collected environmental parameter data is transmitted to the data processing server through a wireless transmission module, the data is cleaned, the abnormal values ​​beyond the normal range are removed, and the minimum-maximum normalization method is used to normalize the data and map the data to the [0,1] interval. The pre-processed environmental parameter data is input into the machine learning model pre-trained with the BP neural network algorithm. The model input layer includes 4 neurons of light intensity, temperature, humidity, and carbon dioxide concentration, the hidden layer contains 8 neurons, and the output layer is the plant growth state. After processing by the sigmoid activation function, the predicted value of the plant growth state is obtained. The predicted value is compared with the target environmental parameter threshold. When the predicted plant growth state is lower than 8, the current environment is judged to be unsuitable for plant growth. The Q-learning reinforcement learning algorithm is used to calculate the reward value of the environmental parameter according to the predicted plant growth state. When the plant growth state is lower than 8, the reward value of the environmental parameter is negative, and when the plant growth state is higher than 8, the reward value of the environmental parameter is positive. The optimization strategy of the environmental parameters is learned through the Q-learning algorithm, and the optimal combination of environmental parameters is found through continuous attempts and learning to achieve the best plant growth state. According to the optimized environmental parameters obtained by the reinforcement learning algorithm, the environmental conditioning equipment is controlled to adjust the environment. For example, when the ambient temperature is lower than 22°C, the heating equipment is controlled to work so that the temperature rises to the optimal range; when the carbon dioxide concentration is lower than 600ppm, the carbon dioxide generator is controlled to work to increase the carbon dioxide concentration in the environment; when the light intensity is lower than 10000lux, the fill light is controlled to work to increase the light required for plant growth, etc. By accurately regulating various environmental parameters, the plant growth environment is always in the optimal state, thereby improving the growth of plants and increasing yield and quality.

[0053] According to the predicted plant growth status, the Qlearning algorithm is used to optimize the environmental parameters and obtain the optimal combination of control environment regulation equipment.

[0054] The environmental parameter data and plant growth status data during the plant growth process are obtained, and a mapping relationship model between environmental parameters and plant growth status is established. According to the established mapping relationship model, the long short-term memory neural network algorithm is used to predict the growth status of plants under different environmental parameter combinations. For the predicted plant growth status, the Q-learning reinforcement learning algorithm is used to guide the agent learning through the reward function to explore the optimal environmental parameter combination. The optimal environmental parameter combination learned by the Q-learning algorithm is used as the control instruction of the environmental adjustment device to adjust the environmental parameters in real time. In the process of adjusting the environmental parameters, the plant growth status data is continuously collected to evaluate the adjustment effect, forming a closed-loop feedback optimization of environmental parameters-plant growth status. If the evaluation result shows that the adjustment effect does not meet expectations, the newly collected data is input into the prediction model and the reinforcement learning model, and the training and optimization are performed again until the plant growth requirements are met. Through the precise regulation of environmental parameters and the continuous learning and optimization of the model, the dynamic adaptation of environmental conditions during plant growth is achieved, the healthy growth of plants is promoted, and the yield quality is improved.

[0055] Furthermore, during the plant growth process, environmental parameter data (such as temperature, humidity, light intensity, etc.) and plant growth status data (such as plant height, number of leaves, chlorophyll content, etc.) were collected every hour through sensors, and 30 days of data were collected cumulatively. The Pearson correlation coefficient was used to analyze the correlation between environmental parameters and plant growth status, and environmental parameters with a correlation coefficient greater than 6 were selected as the input features of the model. The long short-term memory neural network algorithm was used to establish a nonlinear mapping relationship model between environmental parameters and plant growth status. The input of the model was the combination of environmental parameters, and the output was the corresponding predicted value of plant growth status. The collected data was divided into training set and test set in a ratio of 8:2. The average absolute error of the model on the test set after 500 epochs was 0.5. Based on the prediction model, the Q-learning reinforcement learning algorithm was used to explore the optimal combination of environmental parameters. The environmental parameters were discretized, and each parameter was divided into 5 levels to form a state space of environmental parameter combinations. The action space of the agent was defined as increasing or decreasing one level for each environmental parameter. The reward function was set to the higher the predicted plant growth state, the higher the reward value. The intelligent agent continuously interacts with the environment, uses the greedy strategy to select actions, updates the Q value table, and converges to the optimal combination of environmental parameters after 10,000 iterative learning. The optimal combination of environmental parameters is converted into control instructions and passed to the environmental adjustment device to dynamically adjust the plant growth environment in real time. During the adjustment process, plant growth status data is collected every 2 hours, and the improvement rate of plant growth status before and after adjustment is calculated. When the improvement rate is lower than 05, the closed-loop feedback mechanism is triggered, the newly collected data is merged into the original data set, and the prediction model and reinforcement learning model are trained again to adapt to the dynamic changes of plant growth and continuously optimize the control strategy. Through the precise control of environmental parameters and continuous learning of the model, the plants are always in the best growth environment, the average yield is increased by 15%, and the quality pass rate reaches more than 98%.

[0056] The light, temperature, humidity and CO2 data are processed by the BP neural network model to obtain the predicted value and judgment result of the plant growth status.

[0057] According to the environmental factors related to plant growth, data such as light intensity, temperature, humidity and carbon dioxide concentration are obtained, and necessary data preprocessing is performed, including data cleaning, normalization and other operations, to obtain a standardized data set. The BP neural network model is used to construct a neural network structure for predicting plant growth conditions by setting the number of neurons in the input layer, hidden layer and output layer, as well as parameters such as activation function and learning rate. The preprocessed environmental factor data set is divided into a training set and a test set, where the training set is used to train the BP neural network model. The model parameters are optimized through multiple iterations so that it can fit the complex nonlinear relationship between plant growth conditions and environmental factors. The trained BP neural network model is tested and evaluated. By calculating the prediction accuracy, mean square error and other indicators of the model on the test set, the generalization ability and prediction performance of the model are judged, and parameter tuning and model optimization are performed when necessary. The optimized BP neural network model is applied to the actual prediction of plant growth conditions. By collecting data such as light, temperature, humidity and carbon dioxide concentration in real time, the data are input into the model for prediction, and the predicted value of plant growth conditions is obtained. By comparing the predicted value with the preset plant growth threshold, it is determined whether the current growth status of the plant is normal. If the predicted value exceeds the normal range, the early warning mechanism is triggered to prompt the user to take appropriate measures to adjust. The plant growth environment data and prediction results are continuously monitored, and the changing trend of plant growth status is displayed through data visualization and other methods. The model is retrained and optimized regularly to adapt to the dynamic changes in the plant growth environment and improve the accuracy and reliability of the prediction.

[0058] Furthermore, in the plant growth status prediction system, the light intensity, temperature, humidity and carbon dioxide concentration of the plant growth environment are first collected through sensors. The collected data may contain noise and outliers, and data cleaning is required to remove abnormal data points. Then the data is normalized, and the data of different dimensions are mapped to the [0,1] interval to eliminate the influence of dimensional differences. Normalization can use the maximum and minimum value normalization method, that is, x'=(x-xmin) / (xmax-xmin), where x is the original data, xmin and xmax are the minimum and maximum values ​​of the feature, respectively. Next, a BP neural network model is constructed, the input layer contains 4 neurons of light intensity, temperature, humidity and carbon dioxide concentration, the hidden layer can be set to 2 layers, each layer contains 8 and 6 neurons respectively, the output layer is the predicted value of plant growth status, the sigmoid activation function is used, and the learning rate is set to 01. The preprocessed data set is randomly divided into a training set and a test set in a ratio of 8:2. The training set is used to train the model. The number of iterations is set to 1000 times. The back propagation algorithm is used to optimize the model parameters and minimize the mean square error loss function. After the training is completed, the prediction performance of the model is evaluated on the test set, and indicators such as the mean square error and the coefficient of determination are calculated. If the model performance is not ideal, it can be optimized by adjusting parameters such as the number of hidden layer neurons and the learning rate to improve the model's fitting ability and generalization ability. The optimized model can be applied to the actual prediction of plant growth conditions. The system collects sensor data in real time, cleans and normalizes the data, and then inputs it into the trained BP neural network model to obtain the predicted value of plant growth conditions. The threshold range of normal growth conditions is preset, such as temperature between 18℃ and 30℃, humidity between 60% and 80%, and carbon dioxide concentration between 800ppm and 1200ppm. If the predicted value exceeds the threshold range, the early warning mechanism is triggered, and the system automatically sends an alarm message to the user, prompting him to take adjustment measures, such as ventilation, supplementary lighting, humidification, etc. The system can also use data visualization technology to display the changing trends of plant growth environment parameters and prediction results in real time, so that users can understand the dynamic changes of plant growth. At the same time, new environmental data and growth status feedback data are collected regularly to retrain and optimize the model, continuously improve the accuracy and reliability of predictions, and adapt to the dynamic changes of plant growth environment. Through continuous data accumulation and model iteration, the plant growth status prediction system can provide strong decision-making support for modern agricultural production, optimize resource allocation, and improve yield and quality.

[0059] Data cleaning and normalization techniques are used to process environmental data collected by sensors to ensure the accuracy and consistency of prediction input data.

[0060] According to the environmental data collected by the sensor, the original data set is obtained, and exploratory data analysis is performed on the original data set. The basic characteristics of the statistical data, including data volume, data type, data distribution, etc., are analyzed to determine whether there are data quality problems such as missing values ​​and outliers. If there are data quality problems, the corresponding data cleaning technology is used to process them, including deleting features with too many missing values, interpolating and filling missing values, removing outliers, etc., to obtain a cleaned data set. According to business needs and data characteristics, the normalization method is determined, and the cleaned data is normalized to map features of different dimensions to the same scale. Commonly used normalization methods include minimum-maximum normalization, Z-score normalization, etc., to obtain a normalized data set. Through feature selection and feature extraction technology, feature subsets related to the prediction target are selected or constructed from the normalized data set to reduce data dimensions and improve model training efficiency and generalization ability. Commonly used feature engineering techniques include filtering, wrapping, embedding, etc. Divide the preprocessed data set into a training set and a test set according to a certain ratio, select a machine learning algorithm suitable for the business scenario and data characteristics, such as linear regression, decision tree, support vector machine, etc., train the model on the training set, and use cross-validation and other methods to optimize the model hyperparameters. Evaluate the model performance on the test set, use evaluation indicators such as mean square error and mean absolute error to determine whether the model meets the business requirements. If not, return to step 4 to adjust the features or step 5 to optimize the model. Apply the trained model to the production environment, preprocess and feature engineer the newly collected environmental data, input the model for prediction, obtain the prediction results and make business decisions.

[0061] Furthermore, after obtaining the original data set collected by the environmental sensor, the data was first exploratory analyzed. Through statistics, it was found that the data set contained 10,000 records, each of which contained 8 features such as temperature, humidity, and light intensity, and the data types were all continuous values. Among them, there were 5% missing values ​​in the temperature feature, and there were a small number of outliers in the light intensity feature. In view of the data quality problem, the nearest neighbor interpolation method was used to fill the missing values ​​of the temperature feature, and the outliers of the light intensity feature were detected by box plot and replaced with the truncated mean. After the data cleaning was completed, the minimum-maximum normalization method was selected according to business needs to normalize the data, and the value range of each feature was scaled to the [0,1] interval. Then, the feature importance evaluation method based on the tree model was used to select the 5 features with the highest correlation with the prediction target to form a feature subset. The processed data was randomly divided into a training set and a test set in a ratio of 8:2. The gradient boosting decision tree algorithm was selected to train the model on the training set, and the model's learning rate, maximum tree depth and other hyperparameters were optimized through grid search. After 10-fold cross validation, the average absolute error of the model on the test set is 0.5, which meets the business accuracy requirements. Finally, the trained model is deployed to the production environment, and the newly collected environmental data is predicted on a regular basis every day, and the prediction report is automatically generated and the business process is triggered.

[0062] Step S104, when the environmental parameter data fluctuates violently, an early warning mechanism is triggered, and the working parameters of the artificial light source, humidification equipment, and ventilation equipment are automatically adjusted according to the feedback of the growth prediction model.

[0063] Obtain environmental parameter data, including temperature, humidity, light intensity, etc., and use them as input to the growth prediction model; determine whether the environmental parameter data fluctuates drastically, and if so, trigger the early warning mechanism; determine the optimized working parameters of the artificial light source, humidification equipment, and ventilation equipment based on the feedback from the growth prediction model; control the switching state and brightness of the artificial light source, and adjust the light intensity to achieve the optimal value of the model feedback; control the switching state and humidification amount of the humidification equipment, and adjust the ambient humidity to achieve the optimal value of the model feedback; control the switching state and ventilation amount of the ventilation equipment, and adjust the air circulation to achieve the optimal value of the model feedback; continuously monitor the environmental parameter data and feed it back to the growth prediction model to form a closed-loop control and maintain a suitable growth environment.

[0064] Furthermore, the environmental parameter data in the greenhouse is collected in real time through temperature and humidity sensors, light intensity sensors and other equipment. The data is collected every 5 minutes and uploaded to the cloud server. After receiving the data, the server uses the Kalman filter algorithm to preprocess the data, filter out abnormal values, and compare them with the set threshold. If the change range of parameters such as temperature, humidity, and light intensity exceeds 20%, the early warning mechanism is triggered and a warning SMS is sent to the management personnel. At the same time, the preprocessed data is input into the growth prediction model based on the long short-term memory network (LSTM). The model predicts the growth of crops under the current environmental parameters by analyzing the relationship between historical environmental parameters and crop growth conditions, and gives the optimized environmental parameter setting values. According to the feedback results of the model, the PWM duty cycle of the artificial light source is adjusted through the PID control algorithm to stabilize the light intensity at 500-1000lx; the environmental humidity is maintained at 60%-80% by controlling the working time and humidification amount of the humidifier; the CO2 concentration is maintained at 800-1200ppm by adjusting the wind speed and ventilation time of the ventilation equipment. The adjustment process of environmental parameters is a dynamic closed-loop control process. The system continuously optimizes the adjustment strategy based on the real-time collected environmental parameter data and the results of model feedback, so that crops are always in the best growth environment, thereby achieving the goal of increasing production and income.

[0065] The collected environmental parameter data are preprocessed by the Kalman filtering algorithm to obtain a stable data input model after outlier filtering.

[0066] According to the preset environmental parameter collection frequency, the original environmental parameter data collected by the sensor is continuously obtained and the data is transmitted to the data preprocessing module; in the data preprocessing module, the Kalman filter algorithm is used to filter the acquired original environmental parameter data, and the true value of the environmental parameter is estimated through two steps of state prediction and state update; in the state prediction step, the environmental parameter state at the current moment is predicted according to the optimal estimate value and state transfer matrix at the previous moment, and the predicted estimate value and prediction error covariance matrix are obtained; in the state update step, the Kalman gain is calculated according to the observed value and observation matrix at the current moment, and the predicted estimate value is updated by using the Kalman gain. The optimal estimated value and the optimal error covariance matrix are obtained through correction. The optimal estimated value of the environmental parameters is continuously updated through iterative calculation of the Kalman filter algorithm, which effectively eliminates outliers and noise interference and obtains stable and reliable environmental parameter data. The preprocessed environmental parameter data is input into the machine learning model, and a nonlinear mapping relationship between environmental parameters and target variables is established through model training and optimization to achieve accurate prediction and analysis of environmental status. According to the prediction results of the machine learning model, the working parameters of the environmental control equipment are automatically adjusted to achieve intelligent control of the environment, ensure that the environmental parameters are always within the optimal range, and provide users with a comfortable and healthy environment.

[0067] Furthermore, according to the preset environmental parameter collection frequency, such as collecting temperature, humidity and other parameters every 5 minutes, the original environmental parameter data collected by the DHT11 sensor is continuously obtained, and the data is transmitted to the data preprocessing module through the RS485 bus. In the data preprocessing module, the Kalman filter algorithm is used to filter the obtained original environmental parameter data, and the state transfer matrix A=[[1,0],

[0068] [0,1]], observation matrix H = [[1,0], [0,1]], process noise covariance matrix Q = [[1,0], [0,1]], observation noise covariance matrix R = [[1,0], [0,1]], initial state estimate x0 = [

[25] ,

[60] ], initial error covariance matrix P0 = [[1,0], [0,1]], through the two steps of state prediction and state update, the true value of the environmental parameters is estimated. In the state prediction step, according to the optimal estimate x(k-1) at the previous moment and the state transfer matrix A, the state of the environmental parameters at the current moment is predicted, and the predicted estimate x(k|k-1) = Ax(k-1) and the prediction error covariance matrix P(k|k-1) = AP(k-1)A'+Q are obtained. In the state update step, according to the observation value z(k) and the observation matrix H at the current moment, the Kalman gain K(k) = P(k|k-1)H' / (HP(k|k-1)H'+R) is calculated, and the predicted estimate is corrected by the Kalman gain to obtain the optimal estimate x(k) = x(k|k-1)+K(k)(z(k)-Hx(k|k-1)) and the optimal error covariance matrix P(k) = (IK(k)H)P(k|k-1) at the current moment. Through the iterative calculation of the Kalman filter algorithm, the optimal estimate of the environmental parameters is continuously updated, the outliers and noise interference are effectively eliminated, and stable and reliable environmental parameter data are obtained. The preprocessed environmental parameter data is input into the support vector machine (SVM) model, the Gaussian kernel function is selected, the penalty factor C = 10, and the kernel function parameter g = 1. Through model training and optimization, the nonlinear mapping relationship between environmental parameters such as temperature and humidity and human comfort is established, and the accurate prediction and analysis of environmental comfort is realized. According to the prediction results of the SVM model, the working parameters of environmental control equipment such as air conditioners and humidifiers are automatically adjusted. For example, when the predicted comfort level is lower than 8, the air conditioner temperature is set to 26°C and the humidity is set to 60%. This realizes intelligent control of the environment, ensures that environmental parameters such as temperature and humidity are always within the optimal range, and provides users with a comfortable and healthy environment.

[0069] The long short-term memory network (LSTM) is used to analyze historical data, predict the impact of the current environment on crop growth, and determine the optimization parameters.

[0070] According to the historical environmental data and crop growth data, a long short-term memory network model is constructed, and the historical data is divided into a training set and a test set. The LSTM network is used to train the training set data. By adjusting the network structure and hyperparameters, the model performance is continuously optimized until the expected prediction effect is achieved on the test set. The current environmental data is input into the trained LSTM model to predict the growth of crops under the current environmental conditions. According to the prediction results, it is judged whether the current environment is suitable for crop growth. If the environmental conditions are not ideal, the environmental optimization process is triggered. Through iterative optimization algorithms, such as genetic algorithms or particle swarm optimization algorithms, the optimal parameter combination is searched within the range of environmental parameters to achieve the best predicted crop growth. The optimized environmental parameter settings are applied to the actual agricultural production environment, and the environmental conditions are adjusted through the control system to create good growth conditions for crops. The environmental parameters and crop growth conditions are continuously monitored, and the real-time data is fed back to the LSTM model. The model is updated and optimized online to continuously improve the prediction accuracy and environmental optimization effect.

[0071] Furthermore, we first selected environmental data and crop growth data from the past five years from the agricultural database, with a total of 10,000 records. The data were arranged in chronological order, with the first 80% of the data used as the training set and the last 20% of the data used as the test set. The LSTM network model was built using the TensorFlow framework, with the number of input layer nodes being the number of environmental parameters and the number of output layer nodes being the number of crop growth indicators. There were three hidden layers, with 256, 128 and 64 nodes in each layer, respectively. The learning rate was set to 0.1, the batch size was set to 32, and 1,000 epochs were trained. The model performance was evaluated on the test set, and when the mean square error was less than 0.5, the model performance was considered to meet expectations. The current environmental data was input into the trained LSTM model to predict crop growth. If the predicted crop yield is lower than the threshold of 8 kg / m 2 , the environment is considered unsuitable for crop growth, triggering the environmental optimization process. Genetic algorithms are used to optimize environmental parameters, with the population size set to 50 and evolution for 100 generations. Individuals are encoded as binary strings of environmental parameters, and new individuals are generated through crossover mutation. Individuals are decoded into environmental parameters and input into the LSTM model for evaluation, with the predicted yield as the fitness value. The individuals with the highest fitness are selected and their environmental parameters are applied to actual agricultural production. The temperature, humidity, light and other parameters are adjusted through the greenhouse control system to achieve the optimized value. Environmental and crop data are collected every hour and input into the LSTM model in real time for prediction. When the error between the predicted value and the actual value exceeds 1, the model is fine-tuned using the data from the last day to update the model parameters. Continuous cycle optimization allows crops to grow under the best environmental conditions and achieve high and stable yields.

[0072] The PID control algorithm is used to adjust the equipment parameters to ensure that the environmental conditions are within the optimal range recommended by the model, thus achieving precise environmental control.

[0073] Obtain the environmental parameter data collected by the environmental sensor in real time and use it as the input of the PID control algorithm. Determine the target value of the PID control algorithm based on the preset optimal range of environmental conditions. Calculate the control amount of the adjustment equipment parameters through the PID control algorithm. Output the control amount to the actuator to adjust the corresponding equipment parameters. Obtain the adjusted environmental parameter data again to determine whether it has reached the optimal range. If it has not reached the optimal range, re-input the new environmental parameter data into the PID control algorithm for the next round of adjustment. Repeat the above process until the environmental parameters are stabilized within the optimal range to achieve precise environmental control.

[0074] Furthermore, the environmental sensor collects environmental parameter data such as temperature, humidity, and carbon dioxide concentration in real time and uses it as the input of the PID control algorithm. According to the optimal environmental conditions for plant growth, such as temperature 25°C, humidity 60%, and carbon dioxide concentration 800ppm, the target value of the PID control algorithm is determined. Through the PID control algorithm, the control quantity for adjusting the parameters of equipment such as air conditioners, humidifiers, and carbon dioxide generators is calculated, such as the cooling capacity of air conditioners, the humidification capacity of humidifiers, and the release amount of carbon dioxide generators. These control quantities are output to the corresponding actuators to adjust the working state of the equipment. The adjusted environmental parameter data is obtained again to determine whether it has reached the optimal range. If the temperature deviates from the target value by ±1°C, the humidity deviates from the target value by ±5%, and the carbon dioxide concentration deviates from the target value by ±50ppm, the new environmental parameter data is re-input into the PID control algorithm for the next round of adjustment. The above process is repeated until the environmental parameters are stabilized within the optimal range to achieve precise environmental control. In the PID control algorithm, the proportional coefficient Kp = 2, the integral coefficient Ki = 5, and the differential coefficient Kd = 1. By adjusting these parameters, the control effect can be optimized, the response speed can be accelerated, the overshoot can be reduced, and the control accuracy can be improved. At the same time, the parameters and target values ​​of the PID control algorithm can be dynamically adjusted in combination with factors such as plant growth conditions and external weather changes to achieve more intelligent and personalized environmental control and provide the best environmental conditions for the healthy growth of plants.

[0075] Embodiment 2:

[0076] The present invention also provides a plant growth environment control device, comprising:

[0077] An acquisition module, used to acquire multi-dimensional plant growth data;

[0078] A training module, used for obtaining a growth prediction model according to the multi-dimensional plant growth data;

[0079] The acquisition module is used to collect real-time environmental parameter data such as light intensity, temperature and humidity, and carbon dioxide concentration;

[0080] The control module is used to trigger the early warning mechanism when the environmental parameter data fluctuates violently, and automatically adjust the working parameters of the artificial light source, humidification equipment, and ventilation equipment according to the feedback of the growth prediction model.

[0081] As an implementation mode of an embodiment of the present invention, the multi-dimensional plant growth data includes: light intensity, temperature and humidity, carbon dioxide concentration, and nutrient solution ratio.

[0082] As an implementation mode of an embodiment of the present invention, the training module uses a machine learning algorithm to train a growth prediction model for a specific plant based on the multi-dimensional plant growth data type.

[0083] Embodiment 3:

[0084] The present invention also provides a plant growth environment regulation system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a plant growth environment regulation method when executed by the processor.

[0085] Embodiment 4:

[0086] The present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the plant growth environment regulation method when running.

[0087] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for regulating plant growth environment, characterized in that: include: Obtain multi-dimensional plant growth data; Obtaining a growth prediction model according to the multi-dimensional plant growth data; Real-time collection of environmental parameter data such as light intensity, temperature and humidity, and carbon dioxide concentration; When the environmental parameter data fluctuates drastically, the early warning mechanism is triggered, and the working parameters of the artificial light source, humidification equipment, and ventilation equipment are automatically adjusted according to the feedback of the growth prediction model.

2. The method for regulating plant growth environment according to claim 1, characterized in that: Multi-dimensional plant growth data includes: light intensity, temperature and humidity, carbon dioxide concentration, and nutrient solution ratio.

3. The method for regulating plant growth environment according to claim 2, characterized in that: According to the multi-dimensional plant growth data type, a growth prediction model for a specific plant is obtained by training with a machine learning algorithm.

4. A plant growth environment control device, characterized in that: include: An acquisition module, used to acquire multi-dimensional plant growth data; A training module, used for obtaining a growth prediction model according to the multi-dimensional plant growth data; The acquisition module is used to collect real-time environmental parameter data such as light intensity, temperature and humidity, and carbon dioxide concentration; The control module is used to trigger the early warning mechanism when the environmental parameter data fluctuates violently, and automatically adjust the working parameters of the artificial light source, humidification equipment, and ventilation equipment according to the feedback of the growth prediction model.

5. The plant growth environment control device according to claim 4, characterized in that: Multi-dimensional plant growth data includes: light intensity, temperature and humidity, carbon dioxide concentration, and nutrient solution ratio.

6. The plant growth environment control device according to claim 5, characterized in that: The training module uses a machine learning algorithm to train a growth prediction model for a specific plant based on the multi-dimensional plant growth data type.

7. A plant growth environment control system, characterized in that: include: A memory and a processor, wherein the memory stores a computer program executed by the processor, and when the computer program is executed by the processor, the plant growth environment regulating method according to any one of claims 1 to 3 is executed.

8. A storage medium, characterized in that: The storage medium stores a computer program, and the computer program, when running, executes the plant growth environment regulating method according to any one of claims 1 to 3.

Images