Seedling raising facility environment regulation and control method based on AI algorithm

Abstract

The invention discloses a seedling raising facility environment regulation and control method based on an AI algorithm, and relates to the technical field of environment regulation and control, and the method comprises the steps: S1, data collection and preprocessing; s2, establishing an AI prediction model; s3, seedling simulation and scheme making; s4, real-time environment regulation and control; and S5, data feedback and model optimization. According to the seedling culture facility environment regulation and control method based on the AI algorithm, seedling culture simulation can be carried out before seedling culture by establishing the AI prediction model, and the optimal seedling culture scheme can be made by simulating growth conditions under different seedling culture conditions in a virtual environment. Meanwhile, actual seedling culture data are monitored and collected in real time, so that seedling culture demand changes are quickly responded, the environment conditions in the seedling culture facility are automatically adjusted to the environment conditions most suitable for seedling culture, accurate regulation and control are achieved, it is guaranteed that crops are in the optimal growth state all the time, and the success rate and efficiency of seedling culture are improved.

Description

Technical Field

[0001] The present invention relates to the field of environmental regulation technology, and specifically to a seedling facility environmental regulation method based on an AI algorithm. Background Art

[0002] Environmental regulation in seedling nurseries is a key technology in modern agricultural seedling production. It involves precisely controlling environmental factors within the facility to meet the needs of seedling growth. These factors include, but are not limited to, temperature, humidity, light, gas concentrations (such as CO2), and soil conditions. By regulating the environment in seedling nurseries, a suitable growth environment is created for the seedlings, ensuring their healthy and rapid growth. In practice, appropriate regulatory measures are selected based on the seedling type, growth stage, and specific conditions of the facility.

[0003] However, in actual operation, the environmental control of existing seedling facilities is generally of low intelligence level. Usually, technical personnel make judgments on the growth of seedlings based on previous seedling cultivation experience, and then regulate the environment of the seedling facilities. This makes seedling cultivation susceptible to the influence of human subjective factors, making it difficult to achieve precise control, which is not conducive to optimizing the seedling cultivation effect. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for controlling the environment of seedling facilities based on AI algorithm to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for controlling the environment of a seedling facility based on an AI algorithm, comprising the following steps:

[0006] S1. Data collection and preprocessing: Collect environmental parameter data within the nursery facility and external meteorological data, and clean the collected data (remove outliers, fill in missing values, etc.), standardize or normalize them for subsequent algorithm processing;

[0007] S2. Establish an AI prediction model: Based on historical data, use machine learning algorithms (such as regression analysis, decision trees, neural networks, etc.) to establish a prediction model between environmental parameters and seedling cultivation effects;

[0008] S3. Seedling Simulation and Program Development: Different environmental parameter combinations are input into the prediction model to simulate the seedling effect. Based on the simulation results, the optimal seedling environment control program is developed;

[0009] S4. Real-time environmental control: According to the established control plan, the environmental parameters in the seedling raising facility are adjusted in real time, and the actual seedling raising effect is monitored and compared with the predicted results to adjust the control plan;

[0010] S5. Data feedback and model optimization: Collect data from the actual seedling cultivation process as new training samples. When the model performance declines, use the new data to update and optimize the prediction model.

[0011] Furthermore, in step S1, environmental parameter data are collected in real time through a sensor network deployed in the seedling facility, including temperature and humidity (key factors affecting crop growth and development), light intensity (directly affecting crop photosynthesis, and thus affecting its growth rate and quality), CO2 concentration (CO2 is an important raw material for crop photosynthesis, and its concentration directly affects the efficiency of photosynthesis), soil moisture, soil pH, and culture medium nutrient content; the external meteorological data directly calls the monitoring data of the meteorological observation station.

[0012] Furthermore, in step S1, Python's Pandas and NumPy libraries are used for cleaning, and MinMaxScaler or StandardScaler is used for standardization or normalization. The specific operations are as follows:

[0013] 1) Data cleaning

[0014] Remove outliers: Draw a time series graph or scatter plot of the data to visually inspect and identify outliers that significantly deviate from the normal data range. Alternatively, use statistical methods, such as calculating the mean, standard deviation, median, and quartiles, to define the range of normal data. Any data point outside this range is considered an outlier. For identified outliers, delete them or replace them with other reasonable values ​​(such as the mean, median, the previous valid value, or a value predicted based on data trends).

[0015] Filling missing values: For a small number of or occasional missing values, we can use the previous valid value, the next valid value, the mean, or the median for simple filling. For time series data with a large number of missing values, we can use linear interpolation, polynomial interpolation, or spline interpolation to fill in the missing values.

[0016] 2) Standardization:

[0017] Convert the data to a distribution with a mean of 0 and a standard deviation of 1 to make data of different dimensions comparable. The formula is: z = (x-μ) / σ, where x is the original data, μ is the mean of the data, and σ is the standard deviation of the data.

[0018] 3) Normalization processing:

[0019] Scale the data to a specific range (usually 0 to 1) to eliminate the dimension effect of the data. The formula is: x' = (x-x_min) / (x_max-x_min), where x is the original data, x_min is the minimum value of the data, and x_max is the maximum value of the data.

[0020] Furthermore, the step S2 specifically includes the following sub-steps:

[0021] S21. Data preparation: Collect environmental parameter data (such as temperature, humidity, light intensity, CO2 concentration, etc.) and corresponding seedling cultivation effect data (such as seedling growth rate, survival rate, health status, etc.) in the seedling cultivation facility from historical records;

[0022] S22. Select a machine learning algorithm: Select an appropriate machine learning algorithm based on the requirements of the nursery facility environmental control and the characteristics of the data;

[0023] S23. Model training: Divide historical data into a training set and a test set. The training set is used to train the model, and the test set is used to evaluate the model's performance. The selected machine learning algorithm is trained using the training set data to obtain a prediction model. During the training process, the model parameters (such as learning rate, number of iterations, neural network structure, etc.) are continuously adjusted to improve the model's prediction accuracy.

[0024] S24. Model Evaluation and Optimization: Use the test set data to evaluate the trained model, calculate indicators such as prediction accuracy, recall rate, and F1 score to assess the model's performance, and optimize the model based on the evaluation results. Optimization methods include adjusting model parameters, selecting a more appropriate machine learning algorithm, and increasing training data. Repeat the model training, evaluation, and optimization process until the model's performance meets the requirements.

[0025] S25. Model deployment: Deploy the trained model to the environmental control system of the seedling facility to achieve real-time prediction and intelligent control.

[0026] Furthermore, in step S22, the machine learning algorithm includes but is not limited to regression analysis, decision tree, and neural network, as follows:

[0027] Regression analysis: Applicable to scenarios where continuous values ​​are predicted, and can establish linear or nonlinear relationships between environmental parameters and seedling cultivation effects;

[0028] Decision tree: Suitable for classification and regression tasks. It uses a tree structure to represent the decision-making process and can intuitively demonstrate the impact of environmental parameters on seedling production.

[0029] Neural network: Suitable for modeling complex nonlinear relationships, and uses multi-layer neural networks to approximate the functional relationship between environmental parameters and seedling cultivation effects.

[0030] Furthermore, in step S3, a simulation interface or software is developed to allow the user to input different parameter combinations for simulation, and an optimization algorithm (such as a genetic algorithm or a particle swarm algorithm) is used to find the optimal solution. The specific operations are as follows:

[0031] Select an optimization algorithm: Based on the complexity and requirements of the problem, choose an appropriate optimization algorithm, such as genetic algorithm or particle swarm optimization. These algorithms can efficiently search the parameter space and find a solution close to the global optimal solution.

[0032] Define an objective function: Based on the simulation results, define an objective function to evaluate the seedling cultivation effect under different environmental parameter combinations. The objective function is the weighted sum of the seedling growth rate, the product of the survival rate, or other comprehensive indicators;

[0033] Set algorithm parameters: According to the requirements of the optimization algorithm, set necessary parameters, such as population size, number of iterations, mutation rate, inertia weight, etc.

[0034] Running the optimization algorithm: Integrating the optimization algorithm into the simulation interface or software allows the user to start the algorithm to search for the optimal solution. The algorithm automatically iterates the search until a preset stopping condition is reached (such as reaching the maximum number of iterations, convergence of the objective function, etc.);

[0035] Output the optimal solution: After the algorithm runs, it outputs the optimal combination of environmental parameters and its corresponding predicted value of seedling cultivation effect. Users can formulate actual seedling cultivation environment control plan based on this information.

[0036] Furthermore, the seedling environment control program includes set values ​​and control strategies for environmental conditions such as temperature, humidity, light intensity, and CO2 concentration.

[0037] Furthermore, in step S4, an automatic control system (such as PLC, intelligent controller) is connected to environmental parameter adjustment equipment (such as heater, humidifier, lighting equipment, etc.) to achieve automatic regulation, and a monitoring interface is developed to display environmental parameters and seedling status in real time to facilitate manual intervention.

[0038] Furthermore, in step S5, a data feedback mechanism is established to ensure that new data can be collected in a timely manner and used for model updating, specifically by setting regular data upload and update cycles, or uploading new data in real time according to actual needs.

[0039] Furthermore, in step S5, online learning or incremental learning technology is used to achieve dynamic updating of the model. These technologies allow the model to perform parameter adjustment and model optimization in real time when receiving new data without retraining the entire model. The online learning technology uses a rolling window method to update the model, using only a subset of data to update the model parameters each time, thereby achieving online learning and incremental learning.

[0040] The present invention provides a method for controlling the environment of seedling facilities based on an AI algorithm, which has the following beneficial effects:

[0041] By establishing an AI prediction model, this invention can simulate seedling cultivation before seedlings are grown. By simulating growth under different seedling cultivation conditions in a virtual environment, it helps develop the optimal seedling cultivation plan, effectively reducing the cost and time of actual experiments and improving the scientific nature and feasibility of the plan. At the same time, it monitors and collects actual seedling cultivation data in real time to quickly respond to changes in seedling cultivation needs and automatically adjust the environmental conditions within the seedling cultivation facility to the most suitable environmental conditions for seedling cultivation. This achieves precise control and ensures that crops are always in the optimal growth state, thereby improving the success rate and efficiency of seedling cultivation. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a schematic flow chart of the steps of a method for controlling the environment of a seedling facility based on an AI algorithm of the present invention;

[0043] Figure 2 This is a specific flow chart of step S2 of a method for controlling the environment of seedling facilities based on an AI algorithm of the present invention. DETAILED DESCRIPTION

[0044] The following embodiments of the present invention are described in further detail with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0045] like Figure 1-Figure 2 As shown, a method for controlling the environment of a seedling facility based on an AI algorithm comprises the following steps:

[0046] S1. Data collection and preprocessing:

[0047] Collect environmental parameter data within the seedling facility and external meteorological data. The environmental parameter data are collected in real time through the sensor network deployed in the seedling facility, including temperature and humidity (key factors affecting crop growth and development), light intensity (directly affecting crop photosynthesis, and thus affecting its growth rate and quality), CO2 concentration (CO2 is an important raw material for crop photosynthesis, and its concentration directly affects the efficiency of photosynthesis), soil moisture, soil pH, and culture medium nutrient content; external meteorological data directly calls the monitoring data of the meteorological observation station.

[0048] The collected data is cleaned (removing outliers, filling missing values, etc.), standardized or normalized for subsequent algorithm processing. In actual operation, Python's Pandas and NumPy libraries are used for cleaning, and MinMaxScaler or StandardScaler is used for standardization or normalization. The specific operations are as follows:

[0049] 1) Data cleaning

[0050] Remove outliers: Draw a time series graph or scatter plot of the data to visually inspect and identify outliers that significantly deviate from the normal data range. Alternatively, use statistical methods, such as calculating the mean, standard deviation, median, and quartiles, to define the range of normal data. Any data point outside this range is considered an outlier. For identified outliers, delete them or replace them with other reasonable values ​​(such as the mean, median, the previous valid value, or a value predicted based on data trends).

[0051] Filling missing values: For a small number of or occasional missing values, we can use the previous valid value, the next valid value, the mean, or the median for simple filling. For time series data with a large number of missing values, we can use linear interpolation, polynomial interpolation, or spline interpolation to fill in the missing values.

[0052] 2) Standardization:

[0053] Convert the data to a distribution with a mean of 0 and a standard deviation of 1 to make data of different dimensions comparable. The formula is: z = (x-μ) / σ, where x is the original data, μ is the mean of the data, and σ is the standard deviation of the data.

[0054] 3) Normalization processing:

[0055] Scale the data to a specific range (usually 0 to 1) to eliminate the dimension effect of the data. The formula is: x' = (x-x_min) / (x_max-x_min), where x is the original data, x_min is the minimum value of the data, and x_max is the maximum value of the data.

[0056] S2. Build AI prediction model:

[0057] Based on historical data, machine learning algorithms (such as regression analysis, decision trees, neural networks, etc.) are used to establish a prediction model between environmental parameters and seedling cultivation effects.

[0058] S21. Data preparation: Collect environmental parameter data (such as temperature, humidity, light intensity, CO2 concentration, etc.) and corresponding seedling cultivation effect data (such as seedling growth rate, survival rate, health status, etc.) in the seedling cultivation facility from historical records;

[0059] S22. Select a machine learning algorithm: Select an appropriate machine learning algorithm based on the requirements of the nursery facility environmental control and the characteristics of the data. In this step, the machine learning algorithm includes but is not limited to regression analysis, decision tree, and neural network, as follows:

[0060] Regression analysis: Applicable to scenarios where continuous values ​​are predicted, and can establish linear or nonlinear relationships between environmental parameters and seedling cultivation effects;

[0061] Decision tree: Suitable for classification and regression tasks. It uses a tree structure to represent the decision-making process and can intuitively demonstrate the impact of environmental parameters on seedling production.

[0062] Neural network: Suitable for modeling complex nonlinear relationships, using multi-layer neural networks to approximate the functional relationship between environmental parameters and seedling cultivation effects;

[0063] S23. Model training: Divide historical data into a training set and a test set. The training set is used to train the model, and the test set is used to evaluate the model's performance. The selected machine learning algorithm is trained using the training set data to obtain a prediction model. During the training process, the model parameters (such as learning rate, number of iterations, neural network structure, etc.) are continuously adjusted to improve the model's prediction accuracy.

[0064] S24. Model Evaluation and Optimization: Use the test set data to evaluate the trained model, calculate indicators such as prediction accuracy, recall rate, and F1 score to assess the model's performance, and optimize the model based on the evaluation results. Optimization methods include adjusting model parameters, selecting a more appropriate machine learning algorithm, and increasing training data. Repeat the model training, evaluation, and optimization process until the model's performance meets the requirements.

[0065] S25. Model deployment: Deploy the trained model to the environmental control system of the seedling facility to achieve real-time prediction and intelligent control.

[0066] S3. Seedling Simulation and Program Development:

[0067] Different combinations of environmental parameters are input into the prediction model to simulate the seedling cultivation effect. Based on the simulation results, the optimal seedling cultivation environment control plan is formulated. In this step, a simulation interface or software is developed to allow users to input different parameter combinations for simulation and use optimization algorithms (such as genetic algorithms and particle swarm algorithms) to find the optimal solution. The specific operations are as follows:

[0068] Select an optimization algorithm: Based on the complexity and requirements of the problem, choose an appropriate optimization algorithm, such as genetic algorithm or particle swarm optimization. These algorithms can efficiently search the parameter space and find a solution close to the global optimal solution.

[0069] Define the objective function: Based on the simulation results, define an objective function to evaluate the seedling cultivation effect under different environmental parameter combinations. The objective function is the weighted sum of the seedling growth rate, the product of the survival rate, or other comprehensive indicators.

[0070] Set algorithm parameters: According to the requirements of the optimization algorithm, set necessary parameters, such as population size, number of iterations, mutation rate, inertia weight, etc.

[0071] Run the optimization algorithm: Integrate the optimization algorithm in the simulation interface or software and allow the user to start the algorithm to search for the optimal solution. The algorithm automatically iterates the search until the preset stopping condition is reached (such as reaching the maximum number of iterations, objective function convergence, etc.);

[0072] Output the optimal solution: After the algorithm is finished running, the optimal combination of environmental parameters and its corresponding seedling effect prediction value are output. The user can formulate an actual seedling environment control plan based on this information. The seedling environment control plan includes the set values ​​of environmental conditions such as temperature, humidity, light intensity, CO2 concentration, and the control strategy.

[0073] S4. Real-time environmental control:

[0074] According to the established control scheme, the environmental parameters in the seedling raising facility are adjusted in real time, and the actual seedling raising effect is monitored and compared with the predicted results to adjust the control scheme. In this embodiment, an automatic control system (such as a PLC or intelligent controller) is connected to an environmental parameter adjustment device (such as a heater, a humidifier, a lighting device, etc.) to achieve automatic control, and a monitoring interface is developed to display the environmental parameters and seedling raising status in real time, facilitating manual intervention.

[0075] S5. Data Feedback and Model Optimization:

[0076] Collect data from the actual seedling cultivation process as new training samples, establish a data feedback mechanism to ensure that new data can be collected in a timely manner and used for model updates, specifically by setting regular data upload and update cycles, or uploading new data in real time according to actual needs. When the model performance deteriorates, use the new data to update and optimize the prediction model, and use online learning or incremental learning technology to achieve dynamic updates of the model. These technologies allow the model to adjust parameters and optimize the model in real time when receiving new data without retraining the entire model. Online learning technology uses a rolling window method to update the model, using only a subset of data to update the model parameters each time, thereby achieving online learning and incremental learning.

[0077] The embodiments of the present invention are presented for purposes of illustration and description and are not intended to be exhaustive or to limit the invention to the disclosed forms. Many modifications and variations will be apparent to those skilled in the art. The embodiments are chosen and described in order to better illustrate the principles of the invention and its practical application and to enable those skilled in the art to understand the invention and design various embodiments with various modifications as suited for specific applications.

Claims

1. A method for controlling the environment of seedling facilities based on AI algorithm, characterized in that: The following steps are involved: S1. Data collection and preprocessing: Collect environmental parameter data within the nursery facility and external meteorological data, and clean, standardize or normalize the collected data; S2. Establish an AI prediction model: Based on historical data, use machine learning algorithms to establish a prediction model between environmental parameters and seedling cultivation results; S3. Seedling Simulation and Program Development: Different environmental parameter combinations are input into the prediction model to simulate the seedling effect. Based on the simulation results, the optimal seedling environment control program is developed; S4. Real-time environmental control: According to the established control plan, the environmental parameters in the seedling raising facility are adjusted in real time, and the actual seedling raising effect is monitored and compared with the predicted results to adjust the control plan; S5. Data feedback and model optimization: Collect data from the actual seedling cultivation process as new training samples. When the model performance declines, use the new data to update and optimize the prediction model.

2. A method for controlling the environment of a seedling facility based on an AI algorithm according to claim 1, characterized in that: In step S1, environmental parameter data are collected in real time through a sensor network deployed in the seedling facility, including temperature and humidity, light intensity, CO2 concentration, soil moisture, soil pH, and culture medium nutrient content; the external meteorological data directly calls the monitoring data of the meteorological observation station.

3. The method for controlling the environment of a seedling facility based on an AI algorithm according to claim 1, wherein: In step S1, Python's Pandas and NumPy libraries are used for cleaning, and MinMaxScaler or StandardScaler is used for standardization or normalization. The specific operations are as follows: 1) Data cleaning Remove outliers: Draw a time series graph or scatter plot of the data to check and identify outliers that significantly deviate from the normal data range, or use statistical methods to define the range of normal data. Any data point outside this range is considered an outlier. For identified outliers, choose to delete them or replace them with other reasonable values. Filling missing values: For a small number of or occasional missing values, we can use the previous valid value, the next valid value, the mean, or the median for simple filling. For time series data with a large number of missing values, we can use linear interpolation, polynomial interpolation, or spline interpolation to fill them. 2) Standardization: Convert the data to a distribution with a mean of 0 and a standard deviation of 1 to make data of different dimensions comparable. The formula is: z = (x-μ) / σ, where x is the original data, μ is the mean of the data, and σ is the standard deviation of the data. 3) Normalization processing: Scale the data to a specific range to eliminate the dimensionality of the data. The formula is: x' = (x-x_min) / (x_max-x_min), where x is the original data, x_min is the minimum value of the data, and x_max is the maximum value of the data.

4. The method for controlling the environment of a seedling facility based on an AI algorithm according to claim 1, wherein: The step S2 specifically includes the following sub-steps: S21. Data preparation: Collect environmental parameter data within the nursery facility and corresponding nursery effect data from historical records; S22. Select a machine learning algorithm: Select an appropriate machine learning algorithm based on the requirements of the nursery facility environmental control and the characteristics of the data; S23. Model training: Divide historical data into training and test sets, and use the training set data to train the selected machine learning algorithm to obtain a prediction model; S24. Model evaluation and optimization: Use the test set data to evaluate the trained model, calculate indicators to assess the performance of the model, and optimize the model based on the evaluation results until the model performance meets the requirements; S25. Model deployment: Deploy the trained model to the environmental control system of the seedling facility to achieve real-time prediction and intelligent control.

5. The method for controlling the environment of a seedling facility based on an AI algorithm according to claim 4, characterized in that: In step S22, the machine learning algorithm includes but is not limited to regression analysis, decision tree, and neural network, as follows: Regression analysis: Applicable to scenarios where continuous values ​​are predicted, and can establish linear or nonlinear relationships between environmental parameters and seedling cultivation effects; Decision tree: Suitable for classification and regression tasks. It uses a tree structure to represent the decision-making process and can intuitively demonstrate the impact of environmental parameters on seedling production. Neural network: Suitable for modeling complex nonlinear relationships, and uses multi-layer neural networks to approximate the functional relationship between environmental parameters and seedling cultivation effects.

6. The method for controlling the environment of a seedling facility based on an AI algorithm according to claim 1, wherein: In step S3, a simulation interface or software is developed to allow the user to input different parameter combinations for simulation, and an optimization algorithm is used to find the optimal solution. The specific operations are as follows: Select an optimization algorithm: Based on the complexity and requirements of the problem, select an optimization algorithm to search the parameter space and find a solution close to the global optimal solution; Define an objective function: Based on the simulation results, define an objective function to evaluate the seedling cultivation effect under different environmental parameter combinations. The objective function is the weighted sum of the seedling growth rate, the product of the survival rate, or other comprehensive indicators; Set algorithm parameters: Set necessary parameters according to the requirements of the optimization algorithm; Run the optimization algorithm: Integrate the optimization algorithm in the simulation interface or software and allow the user to start the algorithm to search for the optimal solution. The algorithm automatically iterates the search until the preset stopping condition is reached. Output the optimal solution: After the algorithm runs, it outputs the optimal combination of environmental parameters and its corresponding predicted value of seedling cultivation effect. Users can formulate actual seedling cultivation environment control plan based on this information.

7. A method for controlling the environment of a seedling facility based on an AI algorithm according to claim 6, characterized in that: The seedling raising environment control program includes set values ​​and control strategies for environmental conditions such as temperature, humidity, light intensity, and CO2 concentration.

8. The method for controlling the environment of a seedling facility based on an AI algorithm according to claim 1, wherein: In step S4, an automatic control system is connected to the environmental parameter adjustment device to achieve automatic regulation, and a monitoring interface is developed to display the environmental parameters and seedling status in real time to facilitate manual intervention.

9. The method for controlling the environment of a seedling facility based on an AI algorithm according to claim 1, wherein: In step S5, a data feedback mechanism is established to ensure that new data can be collected in a timely manner and used for model updating, specifically by setting regular data upload and update cycles, or uploading new data in real time according to actual needs.

10. The method for controlling the environment of a seedling facility based on an AI algorithm according to claim 1, characterized in that: In step S5, online learning or incremental learning technology is used to achieve dynamic updating of the model. The online learning technology uses a rolling window method to update the model, and only uses a subset of data to update the model parameters each time, thereby achieving online learning and incremental learning.

Images