Irrigation method and device for three-dimensional rice seedling raising factory

Abstract

The invention belongs to the technical field of seedling raising equipment, and relates to a rice three-dimensional seedling raising factory irrigation method and device. Performing pretreatment; extracting main environment characteristics; obtaining prediction data; obtaining an irrigation strategy; irrigation is performed after the irrigation strategy is adjusted. The method comprises the following steps: firstly, collecting environmental data and historical irrigation data in a three-dimensional rice seedling raising factory, then preprocessing the environmental data to obtain preprocessed environmental data, carrying out dimension reduction processing on the preprocessed environmental data, extracting main environmental characteristics, predicting the change trend of soil humidity and temperature according to the main environmental characteristics to obtain prediction data, and finally obtaining the prediction data. According to the method, prediction data and historical irrigation data are combined and optimized by using a particle swarm optimization algorithm to obtain an irrigation strategy, and finally irrigation is performed after the irrigation strategy is adjusted by using a depth Q network according to continuously updated environment data, so that the water demand of seedlings can be accurately adjusted in a complex environment, and water stress on the seedlings is avoided.

Description

Technical Field

[0001] The invention belongs to the technical field of rice seedling raising equipment and relates to an irrigation method and device for a vertical rice seedling raising factory. Background Art

[0002] Rice is one of my country's major grain crops. Improving the automation and intelligence of rice seedling production is crucial for ensuring high and stable rice yields. In the advancement of agricultural modernization, vertical rice seedling production plants are an innovative production method that utilizes vertical seedling growing platforms to provide a stable, effective, and controllable growing environment for rice seedlings.

[0003] However, due to the combined effects of multiple parameter coupling (such as temperature, humidity, light, CO2 concentration, soil moisture, etc.), uneven spatial distribution, rapid dynamic changes, and external interference factors in rice seedling factories, the environment in rice seedling factories is complex and changeable. The existing irrigation methods are difficult to meet the precise water requirements of seedlings in complex environments, which can easily cause water stress to the seedlings and seriously affect the health index of the seedlings. Summary of the Invention

[0004] The object of the present invention is to provide an irrigation method and device for a rice seedling plant, which can accurately adjust the water demand of the seedlings in a complex environment and avoid water stress on the seedlings.

[0005] To achieve the above objectives, the present invention provides the following technical solutions: A rice seedling plant irrigation method comprises the following steps: Collect environmental data within the rice seedling plant and obtain historical irrigation data within the rice seedling plant; preprocessing the environmental data to obtain preprocessed environmental data; Perform dimensionality reduction on the pre-processed environmental data to extract the main environmental features; Predict the changing trends of soil moisture and temperature based on the main environmental characteristics and obtain forecast data; The predicted data is combined with historical irrigation data and optimized using particle swarm optimization algorithm to obtain the irrigation strategy; Irrigation is carried out after adjusting the irrigation strategy based on the continuously updated environmental data using the deep Q network.

[0006] The present invention is also characterized in that: When preprocessing the environmental data, the environmental data are processed in sequence to remove outliers, fill missing values and normalize, and the environmental data are processed into values within the interval [0, 1] through the normalization formula.

[0007] When performing dimensionality reduction on the pre-processed environmental data, principal component analysis is used.

[0008] Among them, the long short-term memory network is used to predict the changing trends of soil moisture and temperature based on the main environmental characteristics.

[0009] A rice seedling plant irrigation device, comprising: Data acquisition module, used to collect environmental data in real time; Cisterns to store water for irrigation; Multiple groups of irrigation sprinklers, each group of irrigation sprinklers is connected to a water reservoir via a conveying assembly, and the conveying assembly is used to convey water in the water reservoir to a corresponding group of irrigation sprinklers; The control module is electrically connected to the multiple conveying components and the data acquisition module. The control module is used to receive environmental data fed back by the data acquisition module, preprocess the environmental data to obtain preprocessed environmental data, perform dimensionality reduction processing on the preprocessed environmental data to extract the main environmental features, and then predict the changing trends of soil moisture and temperature based on the main environmental features to obtain predicted data. The predicted data is then combined with historical irrigation data and optimized using a particle swarm optimization algorithm to obtain an irrigation strategy. Finally, the irrigation strategy is adjusted using a deep Q network based on the continuously updated environmental data to control the multiple conveying components to perform irrigation.

[0010] Each conveying assembly includes: The water pump is arranged on the water reservoir, a water outlet is opened on the side of the water reservoir, the input end of the water pump is connected to the water outlet, and the output end of the water pump is connected to a corresponding group of irrigation nozzles through an irrigation pipe.

[0011] The control module includes: The controller is electrically connected to the multiple water pumps and the data acquisition module respectively, and is used to control the multiple water pumps to irrigate using the adjusted irrigation strategy.

[0012] The rice vertical seedling plant irrigation method and device of the present invention have the following advantages: First, the present invention first collects environmental data in a rice seedling raising factory and obtains historical irrigation data in the rice seedling raising factory, then preprocesses the environmental data to obtain preprocessed environmental data, performs dimensionality reduction processing on the preprocessed environmental data, extracts main environmental features, predicts the changing trends of soil moisture and temperature according to the main environmental features, obtains predicted data, combines the predicted data with historical irrigation data, optimizes it using a particle swarm optimization algorithm, obtains an irrigation strategy, and finally adjusts the irrigation strategy using a deep Q network according to the continuously updated environmental data and then irrigates, thereby being able to accurately adjust the water requirement of the seedlings in a complex environment, avoid water stress on the seedlings, and improve the health index of the seedlings.

[0013] Second, the present invention uses the PCA-LSTM-PSO-DQN algorithm to make intelligent irrigation decisions and adjust the irrigation amount, thereby achieving efficient water saving and precise irrigation, improving seedling cultivation efficiency, and having a positive impact on the sustainable development of agriculture. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic diagram of the overall process structure of the present invention; Figure 2 It is a schematic diagram of the overall structure of the present invention; Figure 3 This is a schematic diagram of the main structure of the present invention; Figure 4 It is a side structural schematic diagram of the present invention; Figure 5 It is a schematic diagram of the top structure of the present invention.

[0015] Reference numerals: 1. Vertical seedling rack; 2. Sprocket; 3. Irrigation nozzle; 4. Irrigation pipe; 5. Controller; 6. Drive motor; 7. Drive gear chain; 8. Float valve; 9. Water outlet; 10. Water pump; 11. Water reservoir. DETAILED DESCRIPTION

[0016] The technical solutions in the present invention will be described clearly and in detail below with reference to the accompanying drawings. In the description of the embodiments of the present invention, unless otherwise specified, " / " means or, for example, A / B can mean A or B: "and / or" in the text is only a description of the association relationship of associated objects, indicating that there can be three relationships, such as A and / or B, which can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present invention, "multiple" refers to two or more than two. The following terms "first" and "second" are used for descriptive purposes only and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features.

[0017] like Figure 1 As shown, the present invention provides an irrigation method for a rice seedling plant, comprising the following steps: Collect environmental data within the rice seedling plant and obtain historical irrigation data within the rice seedling plant; preprocessing the environmental data to obtain preprocessed environmental data; Perform dimensionality reduction on the pre-processed environmental data to extract the main environmental features; Predict the changing trends of soil moisture and temperature based on the main environmental characteristics and obtain forecast data; The predicted data is combined with historical irrigation data and optimized using particle swarm optimization algorithm to obtain the irrigation strategy; Irrigation is carried out after adjusting the irrigation strategy based on the continuously updated environmental data using the deep Q network.

[0018] In summary, the present invention first collects environmental data in a rice seedling raising factory and obtains historical irrigation data in the rice seedling raising factory, then preprocesses the environmental data to obtain preprocessed environmental data, performs dimensionality reduction on the preprocessed environmental data, extracts main environmental features, predicts the changing trends of soil moisture and temperature according to the main environmental features, obtains predicted data, combines the predicted data with historical irrigation data, optimizes them using a particle swarm optimization algorithm, obtains an irrigation strategy, and finally adjusts the irrigation strategy using a deep Q network according to the continuously updated environmental data before irrigation, thereby being able to accurately adjust the water requirement of the seedlings in a complex environment, avoid water stress on the seedlings, and improve the health index of the seedlings.

[0019] Among them, environmental data includes temperature, air humidity, soil humidity, light intensity and wind speed.

[0020] Among them, when preprocessing the environmental data, the environmental data is processed by removing outliers, filling missing values and normalizing in sequence. The present invention processes the environmental data into a numerical value in the interval [0, 1] by removing outliers, filling missing values and normalizing in sequence, and processes the environmental data through a normalization formula, thereby ensuring the quality and consistency of the environmental data and providing accurate input for subsequent feature extraction and decision making.

[0021] When performing dimensionality reduction on pre-processed environmental data, principal component analysis (PCA) is used to convert multidimensional data into linearly uncorrelated variables, called principal components, through linear transformation. PCA is used to process environmental data, extract the main environmental features that are most useful for predicting soil moisture and temperature, remove redundant information, and reduce computational complexity. The specific formula of PCA is as follows: .

[0022] Where, X To preprocess the environment data matrix, W is the eigenvector matrix, Z is the environmental data matrix after dimensionality reduction.

[0023] To predict soil moisture and temperature trends based on key environmental characteristics, a long short-term memory network (LSTM) model is used. Key environmental characteristics serve as inputs to the LSTM model, improving irrigation accuracy and efficiency. The LSTM model uses key environmental characteristics to capture long-term dependencies in time series and predict future trends in soil moisture and temperature. The core of the LSTM model is to control the flow of information through forget gates, input gates, and output gates, capturing the impact of past environmental conditions on current and future environments. The specific formula for the LSTM model is as follows:

[0024] Where, is the output of the forget gate (a value between 0 and 1), is the Sigmoid activation function, is the weight matrix of the forget gate, is the hidden state at the previous moment, is the input at the current moment, is the bias term of the forget gate, is the output of the input gate (a value between 0 and 1), is the weight matrix of the input gate, is the bias term of the input gate, is the value of the candidate memory cell, is the hyperbolic tangent activation function, is the weight matrix of the candidate memory unit, is the bias term of the candidate memory unit, is the memory unit state at the current moment, is the output of the forget gate, is the memory unit state at the previous moment, is the output of the input gate, is the output of the output gate (a value between 0 and 1), is the weight matrix of the output gate, is the bias term of the output gate, is the hidden state at the current moment.

[0025] The predicted data is combined with the historical irrigation data and optimized using the particle swarm optimization algorithm (PSO). The predicted data and the historical irrigation data are used as the input of the particle swarm optimization algorithm. The specific optimization objectives are: 1) Maximize water resource utilization efficiency: Combined with the transpiration of water at various locations on the vertical seedling racks, the irrigation volume is precisely controlled to reduce waste and ensure efficient use of water resources.

[0026] 2) Minimize irrigation costs: Reduce irrigation costs by optimizing irrigation time, frequency and water volume.

[0027] 3) Ensure a stable growth environment for seedlings: Ensure a stable growth environment for seedlings by maintaining appropriate soil moisture and temperature.

[0028] 4) Improve the quality of rice seedling growth: Improve the quality of rice seedling growth by optimizing irrigation strategies. Reduce environmental impact: Reduce negative impacts on the environment by optimizing irrigation strategies.

[0029] 5) Improve system response speed: The optimization model can quickly respond to environmental changes and adjust irrigation strategies in a timely manner. The specific calculation process is as follows:

[0030] The particle velocity update formula is as follows: .

[0031] The particle position update formula is as follows: .

[0032] Where, For particles i In time t +1 speed, For particles i exist t The speed of time, is the inertia weight, which controls the inertia of the particle velocity. and is the acceleration constant, which controls the speed at which particles move to the individual optimal position and the global optimal position. and is a random number in the range [0, 1] to increase the randomness of the search, For particles i The historical optimal position (individual optimal), is the historical optimal position of the entire group (global optimal), is the position of particle i at time t, For particles i In time t +1 position, For particles i exist t The location at the moment.

[0033] Among them, the optimized irrigation strategy includes irrigation amount and irrigation frequency.

[0034] Among them, when adjusting the irrigation strategy using the Deep Q Network (DQN) based on continuously updated environmental data, different irrigation strategies are adopted for seedlings at different growth stages, making full use of the two rows of sprinklers to implement the corresponding strategies. The specific formula of the Deep Q Network is as follows: 。

[0035] Where, is the state-action value function, is the learning rate, r It's a reward. is the discount factor, s and a They are state and action respectively.

[0036] The input of the deep Q network is the irrigation strategy and updated environmental data. The irrigation strategy is adjusted according to the updated environmental data, that is, the irrigation amount and irrigation frequency are adjusted to ensure that the irrigation amount is adjusted in time according to the current environmental conditions and predicted trends, thereby ensuring optimal water management.

[0037] like Figure 2 、 Figure 3 、 Figure 4 、 Figure 5 As shown, the present invention also provides an irrigation device for a rice vertical seedling plant, including a data acquisition module, a water reservoir 11, multiple groups of irrigation nozzles 3, multiple conveying components and a control module. The water reservoir 11 is arranged at one end of the vertical seedling plant frame 1. The water reservoir 11 is used to store water required for irrigation. Multiple groups of irrigation nozzles 3 are arranged in the vertical seedling plant frame 1. The multiple groups of irrigation nozzles 3 are evenly arranged along the length direction of the vertical seedling plant frame 1. Each group of irrigation nozzles 3 is connected to the water reservoir 11 through a conveying component. The conveying component is used to convey the water in the water reservoir 11 to the corresponding group of irrigation nozzles 3. The control module is electrically connected to the multiple conveying components and the data acquisition module. The control module is used to receive feedback from the data acquisition module. Environmental data, preprocessing the environmental data to obtain preprocessed environmental data, then performing dimensionality reduction processing on the preprocessed environmental data to extract the main environmental features, and then predicting the changing trends of soil moisture and temperature based on the main environmental features to obtain predicted data, and then combining the predicted data with historical irrigation data, and optimizing it using the particle swarm optimization algorithm to obtain an irrigation strategy, and finally adjusting the irrigation strategy using the deep Q network according to the continuously updated environmental data to control multiple conveying components for irrigation, at the same time, the irrigation sprinklers 3 are preferably two groups, and the two groups of irrigation sprinklers 3 are arranged along the length direction of the vertical seedling rack 1, and each group of irrigation sprinklers 3 has multiple, and the multiple irrigation sprinklers 3 are evenly arranged along the height direction of the vertical seedling rack 1.

[0038] like Figure 2 、 Figure 5 As shown, the conveying assembly includes a water pump 10, which is arranged on a water reservoir 11. A water outlet 9 is opened on the side of the water reservoir 11. The input end of the water pump 10 is connected to the water outlet 9, and the output end of the water pump 10 is connected to a corresponding group of irrigation nozzles 3 through an irrigation pipe 4.

[0039] like Figure 2As shown, each control module includes a controller 5, a sensor module is arranged on the vertical rice seedling rack 1, and the controller 5 is arranged on the side of the water reservoir 11. The controller 5 is electrically connected to multiple water pumps 10 and a data acquisition module respectively. The deep Q network is arranged in the controller 5. The controller 5 is used to control the multiple water pumps 10 to irrigate the rice seedlings in the vertical rice seedling rack 1 using the adjusted irrigation strategy.

[0040] Among them, the data acquisition module is a sensor module, which detects environmental data including temperature, air humidity, soil humidity, light intensity and wind speed through the sensor module.

[0041] like Figure 2 、 Figure 3 As shown, multiple toothed discs 2 are respectively provided on both sides of the vertical seedling growing frame 1, and the multiple toothed discs 2 are arranged along the side edges of the vertical seedling growing frame 1. A transmission gear chain 7 is provided on the multiple toothed discs 2, and multiple seedling raising troughs are provided between the two transmission gear chains 7. Multiple seedling raising troughs are arranged along the two transmission gear chains 7. Each seedling raising trough is used to plant seedlings, and the two ends of each seedling raising trough are respectively connected to the two transmission gear chains 7. Drive motors 6 are respectively provided on both sides of the vertical seedling growing frame 1 near the two transmission gear chains 7. The output end of the drive motor 6 is provided with a gear, and the gear is meshed with the corresponding transmission gear chain 7. When the drive motor 6 is started, the drive motor 6 drives the transmission gear chain 7 to rotate through the gear. The two transmission gear chains 7 drive the multiple seedling raising troughs to circulate in the vertical seedling growing frame 1, so as to make the irrigation more uniform.

[0042] like Figure 2 As shown, a float valve 8 is provided in the water reservoir 11 , and the water level in the water reservoir 11 is controlled by the float valve 8 .

[0043] Irrigation pipes 4 are equipped with solenoid valves and flowmeters. The solenoid valves act as switches for water flow and are electrically connected to a controller 5. These valves are opened or closed based on the irrigation strategy. The rapid response of the solenoid valves ensures accurate and timely irrigation, avoiding water waste. The flowmeters measure the flow of water through the irrigation pipes 4 and are electrically connected to the controller 5, providing real-time flow data for irrigation strategies. The precise measurements from the flowmeters help the controller 5 adjust the operating status of the water pump 10, ensuring accurate irrigation water flow.

[0044] It will be understood that the present invention is described by way of some embodiments, and it will be appreciated by those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of the present invention are intended to be protected by the present invention.

Claims

1. A rice seedling plant irrigation method, characterized in that: The following steps are involved: Collect environmental data within the rice seedling plant and obtain historical irrigation data within the rice seedling plant; preprocessing the environmental data to obtain preprocessed environmental data; Perform dimensionality reduction on the pre-processed environmental data to extract the main environmental features; Predict the changing trends of soil moisture and temperature based on the main environmental characteristics and obtain forecast data; The predicted data is combined with historical irrigation data and optimized using particle swarm optimization algorithm to obtain the irrigation strategy; Irrigation is carried out after adjusting the irrigation strategy based on the continuously updated environmental data using the deep Q network.

2. The rice seedling plant irrigation method according to claim 1, characterized in that: When preprocessing environmental data, the environmental data are processed in sequence to remove outliers, fill missing values and normalize, and the environmental data are processed into values within the interval [0, 1] through the normalization formula.

3. The rice seedling plant irrigation method according to claim 1, characterized in that: Principal component analysis is used to reduce the dimensionality of preprocessed environmental data.

4. The rice seedling plant irrigation method according to claim 1, characterized in that: Long short-term memory networks are used to predict the changing trends of soil moisture and temperature based on the main environmental characteristics.

5. A rice seedling plant irrigation device, characterized in that: The method according to claim 1, comprising: Data acquisition module, used to collect environmental data in real time; a reservoir (11) for storing water required for irrigation; A plurality of groups of irrigation nozzles (3), each group of irrigation nozzles (3) is connected to a water reservoir (11) via a conveying assembly, the conveying assembly being used to convey water in the water reservoir (11) to a corresponding group of irrigation nozzles (3); The control module is electrically connected to the multiple conveying components and the data acquisition module. The control module is used to receive environmental data fed back by the data acquisition module, preprocess the environmental data to obtain preprocessed environmental data, perform dimensionality reduction processing on the preprocessed environmental data to extract main environmental features, predict the changing trends of soil moisture and temperature based on the main environmental features to obtain predicted data, combine the predicted data with historical irrigation data, optimize it using a particle swarm optimization algorithm to obtain an irrigation strategy, and finally adjust the irrigation strategy using a deep Q network based on the continuously updated environmental data to control the multiple conveying components to perform irrigation.

6. The rice vertical seedling plant irrigation device according to claim 5, characterized in that: Each of the conveying assemblies comprises: A water pump (10) is arranged on a water reservoir (11). A water outlet (9) is provided on the side of the water reservoir (11). The input end of the water pump (10) is connected to the water outlet (9), and the output end of the water pump (10) is connected to a corresponding group of irrigation nozzles (3) through an irrigation pipe (4).

7. The rice vertical seedling plant irrigation device according to claim 6, characterized in that: Each of the control modules comprises: The controller (5) is electrically connected to the plurality of water pumps (10) and the data acquisition module respectively, and is used to control the plurality of water pumps (10) to perform irrigation using the adjusted irrigation strategy.

Images