Citrus seedling raising intelligent tidal irrigation method and system based on secondary fuzzy control

Abstract

The invention discloses an intelligent tidal irrigation method and system for citrus seedling culture based on secondary fuzzy control. The method comprises the following steps: acquiring real-time environment data of citrus seedling culture by using a plurality of sensors; processing the real-time environment data to obtain sensor time sequence data and sending the sensor time sequence data to the cloud; according to the sensor time sequence data, utilizing the trained citrus seedling transpiration prediction model returned by the cloud server to predict future transpiration; determining a transpiration grade according to the future transpiration amount; and outputting an irrigation strategy by using a fuzzy control technology according to the divided transpiration level and the real-time environment data. The system can realize daily timed automatic irrigation control, remarkably reduces human input and water resource waste, is suitable for nursery seedling culture scenes in main citrus producing areas such as hilly and mountainous areas, and has good popularization value and practical application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent agriculture and precise irrigation control of fruit tree seedlings, and in particular to a method, system, terminal device and computer-readable storage medium for intelligent tidal irrigation of citrus seedlings based on secondary fuzzy control. Background Art

[0002] Citrus cultivation is a crucial component of my country's fruit industry, and its nursery development has a profound impact on root growth, environmental adaptability, and subsequent yield quality. Tidal irrigation, an irrigation management method that combines periodic flooding with intermittent drainage, has been gradually adopted in fruit seedling cultivation, flower production, and cash crop production in recent years. This method, through alternating flooding and drainage, improves root zone aeration, promotes healthy root development, and helps enhance seedling survival and resilience.

[0003] However, with the expansion of citrus seedling production and the increasing complexity of nursery environmental variables, traditional tidal irrigation methods generally rely on timed, quantitative, or manual operation, lacking the ability to respond to multiple environmental factors. In actual production, factors such as external weather, soil properties, and seedling growth status are constantly changing. A single irrigation cycle and parameter settings often fail to adapt to the dynamic needs of seedlings for water and environmental conditions at different stages. This management model not only increases the intensity of operation and maintenance work, but also limits the further realization of the potential benefits of tidal irrigation.

[0004] Currently, the demand for intelligent and automated tidal irrigation management is becoming increasingly prominent. On the one hand, production practices urgently require real-time monitoring and integration of multi-source environmental information, such as temperature, humidity, light, and wind speed. On the other hand, scientifically adjusting irrigation strategies and optimizing control parameters based on the real-time status of the nursery and the growth stage of the seedlings has become a major bottleneck hindering the advancement of scientific and refined nursery irrigation. Therefore, the development of an intelligent tidal irrigation system that can comprehensively perceive multiple environmental variables, dynamically optimize the irrigation process, and improve operational efficiency and seedling quality has important theoretical and practical value. Summary of the Invention

[0005] In order to address the above-mentioned deficiencies in the prior art, the present invention provides a method, system, terminal device and computer-readable storage medium for intelligent tidal irrigation of citrus seedlings based on secondary fuzzy control.

[0006] The first object of the present invention is to provide an intelligent tidal irrigation method for citrus seedlings based on secondary fuzzy control.

[0007] The second object of the present invention is to provide an intelligent tidal irrigation system for citrus seedlings based on secondary fuzzy control.

[0008] The third object of the present invention is to provide a terminal device.

[0009] A fourth object of the present invention is to provide a computer-readable storage medium.

[0010] The first object of the present invention can be achieved by adopting the following technical solutions: A citrus seedling intelligent tidal irrigation method based on secondary fuzzy control, applied to the edge end, comprising: Use multiple sensors to collect real-time environmental data for citrus seedlings; Process real-time environmental data to obtain sensor time series data and send it to the cloud; Based on the sensor time series data, the trained citrus seedling transpiration prediction model returned by the cloud server is used to predict future transpiration; based on the future transpiration, the transpiration level is determined; According to the divided transpiration levels and real-time environmental data, fuzzy control technology is used to output irrigation strategies.

[0011] Furthermore, the sensor time series data is sent to the cloud server, including: The sensor time series data is packaged into standard MQTT messages and then published to the EMQX message broker server in the cloud through the MQTT client program. The EMQX message broker server uses the rule engine and data forwarding plug-in to store valid fields in the structured database in the cloud. Among them, the standard MQTT message includes payload and timestamp.

[0012] Furthermore, the sensor time series data is obtained through the following process: Packaging real-time environmental data into MQTT protocol data packets; Cleaning the data in the data package; the cleaning includes processing missing values, duplicate values ​​and outliers, and performing consistency checks to reduce data noise and errors; Extract the payload and the time when the data was acquired based on the cleaned data; The cleaned data and the time of data acquisition are matched to obtain sensor time series data.

[0013] Furthermore, the citrus seedling transpiration prediction model is a hybrid prediction model that combines a convolutional neural network, a gated recurrent unit, and a feedforward neural network; The trained citrus seedling transpiration prediction model is obtained through the following process: Using sensor time series data, we construct supervised learning sample pairs, including: The sensor time series data within a fixed time window in the past is used as the input data of the transpiration prediction model of citrus seedlings to represent the changing trend of the environment in which the citrus seedlings are located. The time series of crop transpiration corresponding to a fixed time length in the future is used as the supervision label of the citrus seedling transpiration prediction model; the time series of crop transpiration is the measured transpiration data; The supervised learning samples were used to train the transpiration prediction model for citrus seedlings, including: The input data is fed into the fused convolutional neural network and gated recurrent unit respectively to extract local spatial features and learn long-term dependencies; The outputs of the fused convolutional neural network and the gated recurrent unit are input into the feedforward neural network to integrate spatiotemporal features; The output data of the feedforward neural network is connected through residual connection to obtain the future transpiration.

[0014] Furthermore, the transpiration level is determined based on the ratio of the future transpiration amount to the standard transpiration amount, where the standard transpiration amount is calculated by the Penman formula based on the sensor time series data.

[0015] Furthermore, the method of outputting an irrigation strategy using fuzzy control technology based on the divided transpiration levels and real-time environmental data includes: The planting time of citrus seedlings is input into the first-level fuzzy controller, which outputs the target air humidity temperature and target soil temperature and humidity required at the current stage; Calculate the deviation between real-time environmental data and target air humidity and temperature and target soil temperature and humidity; The transpiration level and deviation value determined based on real-time environmental data are input into the secondary fuzzy controller to obtain the adjustment of tidal water level, irrigation duration and irrigation start time.

[0016] Furthermore, the use of multiple sensors to collect real-time environmental data of citrus seedlings includes: Connect the Raspberry Pi controller to multiple sensors; Using the Raspberry Pi host to run the lightweight system of Raspberry Pi OS, by setting up scheduled tasks and Python scripts, multiple sensors can collect real-time environmental data.

[0017] Furthermore, the real-time environmental data includes air temperature, air humidity, soil temperature, soil humidity and soil conductivity.

[0018] The second object of the present invention can be achieved by adopting the following technical solutions: A citrus seedling intelligent tidal irrigation system based on two-level fuzzy control, the system includes an edge end and a cloud end, wherein: The edge end includes: A data acquisition module is used to collect real-time environmental data of citrus seedlings using multiple sensors; The data processing module is used to process real-time environmental data, obtain sensor time series data and send it to the cloud; The transpiration prediction module is used to predict future transpiration based on sensor time series data using the trained citrus seedling transpiration prediction model returned by the cloud server; and to determine the transpiration level based on the future transpiration; Irrigation strategy adjustment module, used to output irrigation strategy based on the divided transpiration levels and real-time environmental data using fuzzy control technology; The cloud end is used to receive real-time environmental data and sensor time series data sent by the edge end, and use the sensor time series data to train the citrus seedling transpiration prediction model, and return the trained citrus seedling transpiration prediction model to the edge end.

[0019] The third object of the present invention can be achieved by adopting the following technical solutions: A terminal device includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the above-mentioned intelligent tidal irrigation method for citrus seedlings based on secondary fuzzy control is implemented.

[0020] The fourth object of the present invention can be achieved by adopting the following technical solutions: A computer-readable storage medium stores a program, which, when executed by a processor, implements the above-mentioned intelligent tidal irrigation method for citrus seedlings based on secondary fuzzy control.

[0021] The present invention has the following beneficial effects compared to the prior art: The present invention uses multiple sensors combined with edge terminals to achieve real-time monitoring and stable data upload of the entire citrus seedling cultivation process, providing high-quality data support for irrigation prediction and control. By combining a citrus seedling transpiration prediction model that combines a convolutional neural network, a gated recurrent unit, and a feedforward neural network with fuzzy control technology, it can dynamically determine the irrigation water level, duration, and start time, thereby achieving data-driven precise dynamic irrigation, improving water resource utilization efficiency, reducing manual intervention, and improving citrus seedling yield and seedling quality. The present invention can achieve daily scheduled automatic irrigation control, significantly reducing manpower input and water resource waste. It is suitable for nursery seedling cultivation scenarios in major citrus producing areas such as hilly and mountainous areas, and has good promotion value and practical application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings described below are only some of the embodiments of the present application, and all other drawings obtained by those of ordinary skill in the art without creative labor based on these drawings are within the scope of protection of the present application. It should be understood that the specific embodiments described are only used to explain the present application and not to limit the present application.

[0023] Figure 1 Part structure diagram of the citrus seedling intelligent tidal irrigation system based on two-stage fuzzy control of the embodiment 1 of the present application; Figure 2 Flow chart of the citrus seedling intelligent tidal irrigation method based on two-stage fuzzy control of the embodiment 1 of the present application; Figure 3 Schematic diagram of the EMQX linkage MySQL function of the embodiment 1 of the present application; Figure 4 The citrus seedling transpiration prediction model architecture diagram of the embodiment 1 of the present application; Figure 5 The fitting diagram of the citrus seedling transpiration prediction model used in the embodiment 1 of the present application and the actual transpiration; Figure 6 The Internet of Things platform main interface schematic diagram of the citrus seedling intelligent tidal irrigation system of the embodiment 1 of the present application; Figure 7 The auxiliary decision-making and analysis interface schematic diagram of the citrus seedling intelligent tidal irrigation system of the embodiment 1 of the present application; Figure 8 Structure block diagram of the citrus seedling intelligent tidal irrigation system based on two-stage fuzzy control of the embodiment 2 of the present application; Figure 9 Structure block diagram of the terminal device of the embodiment 3 of the present application. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be described clearly and completely below with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, and not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of protection of the present application. It should be understood that the specific embodiments described are only used to explain the present application and not to limit the present application.

[0025] Embodiment 1: As Figure 1 , 2As shown, this embodiment provides a citrus seedling intelligent tidal irrigation method based on secondary fuzzy control, and the specific implementation process includes the following steps: S201. Utilize multiple sensors to obtain real-time environmental data of citrus orchard seedling cultivation.

[0026] The local acquisition system includes a Raspberry Pi controller, Raspberry Pi OS system environment, multiple sensors, and an MQTT client program.

[0027] The Raspberry Pi controller is connected to multiple sensors via an RS485 adapter module. One end of the RS485 adapter module is connected to the Raspberry Pi's serial communication port (USB-to-serial port chip), and the other end is connected to the sensor's RS485 port. The Raspberry Pi runs the lightweight Raspberry Pi OS, which uses scheduled tasks and Python data collection scripts to automatically schedule data collection and upload.

[0028] The Raspberry Pi obtains real-time data on the citrus seedling environment through various sensors connected to it, including but not limited to: air temperature, air humidity, soil temperature, soil moisture, soil conductivity, light intensity, atmospheric pressure, net solar radiation, wind speed and other environmental factors.

[0029] The Raspberry Pi writes the collected sensor data into a locally deployed MySQL database for use by the local front-end page and subsequent prediction algorithms and fuzzy control modules. It also packages the data into data frames in the MQTT protocol format and publishes them to the remotely deployed EMQX platform through the built-in MQTT client program.

[0030] The MQTT protocol is a lightweight, low-power IoT messaging protocol that uses a publish-subscribe model, making it ideal for multi-device communication in agricultural IoT environments. The communication process is as follows: a Raspberry Pi, acting as an MQTT publisher, publishes messages to an EMQX server in the cloud. The EMQX server receives the messages and forwards them to a structured database (such as MySQL) or other processing modules through a rules engine.

[0031] Throughout the embodiment, all sensors are standard industrial-grade sensors connected via RS485 interfaces.

[0032] The MQTT client, developed using lightweight libraries like paho-mqtt, is compatible with Linux systems and supports features like disconnection reconnection and QoS settings. After each data collection session, the client automatically publishes a JSON-formatted payload containing the current timestamp and various environmental values ​​to a designated topic, enabling real-time subscription and data processing in the cloud.

[0033] Through this hybrid mode of "Raspberry Pi edge acquisition + MQTT remote reporting + local database caching", the system can not only support stable operation in an offline state, but also has cloud integration and remote expansion capabilities, meeting the dual requirements of the citrus seedling prediction irrigation system for data real-time and system stability.

[0034] S202: Process the real-time environmental data to obtain sensor time series data.

[0035] In this example, the local data collection system is deployed on a Raspberry Pi edge node. All raw environmental data collected by sensors (such as air temperature and humidity, soil temperature and humidity, light intensity, atmospheric pressure, wind speed, and electrical conductivity) is first cleaned and preprocessed locally. Data preprocessing is performed by a Python program running locally on the Raspberry Pi. The process includes imputing missing values ​​(using forward filling or linear interpolation), removing outliers (such as using a sliding median or standard deviation threshold), unifying data formats, normalizing values, and aligning timestamps. This cleaning process ensures the integrity, consistency, and continuity of the generated environmental data for use in time series modeling.

[0036] After data cleaning is completed, it will be distributed synchronously to two storage terminals: first, it will be written into the locally deployed MySQL database for local prediction model calls and real-time display in the front-end system; second, Figure 3 As shown, the message is packaged into a standard MQTT message (including payload and timestamp) and published to the EMQX message broker in the cloud via the MQTT client. Upon receiving the message, the EMQX platform uses its rules engine and data forwarding plugin (Sink) to store the valid fields in a structured database (Alibaba Cloud MySQL) in the cloud for subsequent large-scale model training and historical trend analysis.

[0037] During the model input preparation phase, the system uses a sliding window approach to generate continuous sample segments from the processed local time series. A single input sample sequence is constructed from 144 consecutive entries (corresponding to sampling every 10 minutes over the past 24 hours). Multiple datasets are then generated using sliding windows for each entry. This format is suitable for both real-time inference of local models and batch training of cloud-based models.

[0038] By completing all data cleaning and preprocessing locally and implementing "one-time processing, two-way writing", the system ensures data processing efficiency, real-time prediction and consistency of training data, and builds a unified data flow architecture of "edge intelligence + cloud convergence", providing a high-quality input foundation for subsequent evaporation prediction and fuzzy control.

[0039] S203. Utilize the sensor time series data to construct a training set, and utilize the training set to train a citrus seedling transpiration prediction model.

[0040] This step is completed by the cloud server and includes two stages: training data preparation and deep learning model training: (1) Construct training data.

[0041] Based on the historical time series data collected by multiple sensors, supervised learning sample pairs are constructed. Each sample consists of the following two parts: Input: It is composed of a time series of multidimensional environmental variables (air temperature, humidity, soil temperature and humidity, soil EC value) within a fixed time window in the past (24 hours), which represents the changing trend of the environment in which the citrus seedlings are located; Output part: The time series of crop transpiration corresponding to a fixed time length (24 hours) in the future serves as a supervision label.

[0042] The environmental data comes from the sensor historical records stored in the cloud platform, which are arranged in 10-minute intervals and form the input sequence after data cleaning and normalization; the output part is the measured transpiration data.

[0043] (2) Training a citrus seedling transpiration prediction model.

[0044] like Figure 4 As shown in the figure, based on the above sample pairs, a deep learning model combining a convolutional neural network (CNN), a gated recurrent unit (GRU), and a feedforward neural network (FFN) is trained. This model has the following structural features: the CNN is used to extract local variation features in the input sequence, the GRU is used to model long-term dependencies and temporal trends, and the FFN is used to integrate global semantic features and enhance nonlinear expression capabilities.

[0045] In this example, in order to find the most suitable network structure for citrus seedling transpiration prediction task, multiple comparative experiments were conducted, including mainstream models such as LSTM, pure CNN, Transformer and Multi-layer Perceptron (MLP). Figure 5 As shown in the figure, by comparing multiple indicators such as mean absolute error (MAE), mean square error (MSE) and coefficient of determination (R²), it was finally determined that the fusion structure performed best in terms of comprehensive accuracy, stability and training convergence efficiency, and had stronger feature extraction and sequence modeling capabilities.

[0046] Model training is performed on cloud servers, supporting large-scale data batch processing, GPU parallel training, and remote parameter optimization. After training is complete, the final model file is automatically downloaded to a Raspberry Pi node deployed at the edge, enabling local model inference and daily automated forecasting.

[0047] S204: Use the trained model to predict future transpiration, and calculate the transpiration level based on the future transpiration.

[0048] This step is performed locally by the edge computing node (Raspberry Pi) and is used for daily scheduling to predict crop transpiration levels for the next 24 hours and classify them into different levels: (1) Local data collection and input preparation.

[0049] The Raspberry Pi reads environmental data collected by sensors over the past 24 hours in real time and formats it into the standard input format required by the model. The input dimensions remain consistent with those used during training. The processing includes normalization, missing value filling, and time alignment.

[0050] (2) Local model reasoning.

[0051] The input data is fed into the deployed deep learning model to obtain a series of evaporation predictions for the next 24 hours with a step size of 10 minutes. The prediction results are the ET series, with the unit of mm / 10min. These predictions are then added together to obtain the total evaporation for the next 24 hours, with the unit of mm.

[0052] (3) Calculation of standard transpiration rate.

[0053] This step uses the sensor time series data obtained in step S202 to calculate the standard transpiration of the citrus seedlings per day (according to the sunshine cycle, i.e., from 6 am to 6 am the next day) ( ET 0), calculated using the Penman-Monteithequation.

[0054] The Penman formula is as follows (unit: mm / day):

[0055] in: Δ is the slope of the saturated vapor pressure curve (kPa / °C), which is calculated from the air temperature and air humidity; R n is the net solar radiation (MJ / m² / day); G is the soil heat flux (MJ / m² / day), which can usually be ignored and approximated to 0 on a daily scale; γ is the hygrometer constant (kPa / °C), calculated from air temperature and atmospheric pressure; T is the average air temperature (°C), calculated from the air temperature; u 2 is wind speed (m / s); es is the saturated vapor pressure (kPa), calculated from the air temperature and humidity; e a is the actual vapor pressure (kPa), calculated from the air temperature and humidity.

[0056] This example uses Python to calculate crop evapotranspiration.

[0057] (4) Classification of transpiration levels.

[0058] The transpiration intensity level is divided according to the ratio of the total transpiration predicted in the next 24 hours to the standard transpiration ET0: Ratio < 0.5: defined as “low transpiration”; 0.5 ≤ ratio ≤ 1.5: defined as “medium transpiration”; 0.6 Ratio >1.5: defined as “high transpiration”; This level is used as one of the fuzzy control input parameters to indicate the transpiration status of crops on that day.

[0059] S205: Call the secondary fuzzy control system based on the transpiration level and output the optimal irrigation strategy.

[0060] This step is executed locally on the Raspberry Pi device at the edge. The fuzzy control program is called daily to output the real-time irrigation strategy parameters for the current cycle based on the growth stage of the citrus seedlings, environmental deviation, and transpiration prediction level. The system uses a two-level fuzzy control architecture, mainly consisting of two fuzzy controllers connected in series: First, a first-level fuzzy controller uses the planting date of the citrus seedlings as input to determine the current growth stage (germination, seedling, flowering, fruit expansion, or maturity). It then outputs the target air humidity, target air temperature, target soil temperature, target soil moisture, target water level, and ideal irrigation duration for that stage. These target values ​​serve as the baseline parameters for the tidal irrigation control system.

[0061] Next, the second-level fuzzy controller receives three types of input: the deviation between the real-time air temperature and humidity and the set target (target air temperature and humidity); the deviation between the soil temperature and humidity and the target soil temperature and humidity; and the transpiration intensity level (low, medium, or high). Based on the established fuzzy rule base, this controller performs fuzzy reasoning and defuzzification, outputting three adjustment parameters: the adjustment amount for the tidal water level (in units of ±cm), the adjustment amount for the irrigation duration (in units of ±minutes), and the adjustment amount for the irrigation start time (in units of ±hours, which can be advanced or delayed).

[0062] This control logic is executed once a day to generate a new round of irrigation adjustment parameters, which are used to drive subsequent irrigation control modules to achieve linked responsive management of crop water demand dynamics and meteorological changes.

[0063] After the control parameters are output, the Raspberry Pi automatically triggers the underlying relay, water pump, or solenoid valve control circuit to perform tidal irrigation regulation tasks, including: Control the pumping height and set the tidal water level; Set irrigation duration (via the timer module); Modify the irrigation start time point according to the "advance / delay" time parameters; The system executes the entire process once daily at 6:00 AM by default. Users can also remotely monitor and manually intervene via the web or app platforms. Environmental data and execution logs during execution are transmitted back to the cloud database in real time and visualized on the management platform, facilitating subsequent data analysis and system optimization.

[0064] The IoT front-end system provided by the present invention is deployed in edge nodes (such as Raspberry Pi), built based on the Vue framework and Node.js, and has multi-module integration, integrated control and human-computer interaction functions, forming an information bridge between the system user layer and the perception control layer.

[0065] Figure 6 This is a schematic diagram of the main interface of the Internet of Things front-end system provided in this embodiment. The interface centrally displays multiple core functional modules of the system, including a real-time data monitoring module, a real-time video detection module, an alarm information module, and an automatic irrigation control module.

[0066] The real-time data monitoring module simultaneously displays the values ​​collected by various sensors in the current environment, including air temperature, air humidity, soil temperature and humidity, light intensity, wind speed, and atmospheric pressure. This data is dynamically updated through charts and can be linked to historical data displays, allowing users to monitor the seedling environment at all times. The real-time video detection module accesses on-site camera images and provides on-site visual monitoring through a video stream playback window, helping users remotely assess crop growth and operational safety.

[0067] The alarm information module features threshold warnings and status notifications. When environmental variables exceed set ranges, sensors experience anomalies, data interruptions occur, or model outputs experience abnormal levels, the system will issue a front-end warning, enhancing system security and controllability. The automatic irrigation control module displays daily irrigation adjustment parameters generated by the system, including recommended irrigation times, water levels, and duration. It also supports manual parameter modification and single-click irrigation command triggering, creating a flexible control method that combines automatic and manual regulation.

[0068] Figure 7The schematic diagram of the intelligent tidal irrigation system auxiliary decision-making and analysis interface provided in this embodiment mainly integrates functions such as the transpiration prediction visualization module, the citrus growth analysis module and the citrus seedling expert system module.

[0069] The transpiration forecast module graphically displays the model's predicted transpiration trend curve for the next 24 hours, helping users assess the future water demand of crops. The system also supports comparing historical actual transpiration data with predicted results to analyze model deviations and environmental trends.

[0070] The Citrus Growth Analysis module comprehensively assesses crop growth based on environmental data, historical irrigation records, and image information. It provides staged growth scores, temperature and humidity adaptability analysis, and potential pest and disease warnings, enhancing intelligent management. The Citrus Seedling Expert System module features a built-in knowledge base, allowing users to query recommended irrigation standards, fertilizer and water management plans, and pest and disease control recommendations based on the seedling stage or abnormal environmental conditions. This intelligent question-and-answer guidance aids agricultural management decision-making.

[0071] Those skilled in the art will appreciate that all or part of the steps in the method for implementing the above embodiments may be completed by instructing related hardware through a program, and the corresponding program may be stored in a computer-readable storage medium.

[0072] It should be noted that although the method operations of the above embodiments are described in a particular order in the accompanying drawings, this does not require or imply that the operations must be performed in this particular order, or that all of the illustrated operations must be performed to achieve the desired results. Rather, the depicted steps may be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into a single step, and / or a single step may be broken down into multiple steps.

[0073] Example 2: like Figure 8 As shown, this embodiment provides a citrus seedling intelligent tidal irrigation system based on two-level fuzzy control, which includes an edge terminal 801 and a cloud terminal 802, wherein: The edge terminal 801 includes: A data acquisition module is used to collect real-time environmental data of citrus seedlings using multiple sensors; The data processing module is used to process real-time environmental data, obtain sensor time series data and send it to the cloud; The transpiration prediction module is used to predict future transpiration based on sensor time series data using the trained citrus seedling transpiration prediction model returned by the cloud server; and to determine the transpiration level based on the future transpiration; Irrigation strategy adjustment module, used to output irrigation strategy based on the divided transpiration levels and real-time environmental data using fuzzy control technology; The cloud 802 is used to receive real-time environmental data and sensor time series data sent by the edge end, and use the sensor time series data to train the citrus seedling transpiration prediction model, and return the trained citrus seedling transpiration prediction model to the edge end.

[0074] The specific implementation of each module in this embodiment can be found in the above-mentioned embodiment 1, and will not be described one by one here; it should be noted that the system provided in this embodiment is only illustrated by the division of the above-mentioned functional modules. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure can be divided into different functional modules to complete all or part of the functions described above.

[0075] Example 3: This embodiment provides a terminal device, which can be a computer, such as Figure 9 As shown, it comprises a processor 902, a memory, an input device 903, a display 904 and a network interface 905 connected via a system bus 901. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium 906 and an internal memory 907. The non-volatile storage medium 906 stores an operating system, a computer program and a database. The internal memory 907 provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. When the processor 902 executes the computer program stored in the memory, the intelligent tidal irrigation method for citrus seedlings based on secondary fuzzy control of the above-mentioned embodiment 1 is implemented.

[0076] Example 4: This embodiment provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the intelligent tidal irrigation method for citrus seedlings based on secondary fuzzy control of the above-mentioned embodiment 1 is implemented.

[0077] In summary, the present invention discloses a method, system, terminal device, and computer-readable storage medium for intelligent tidal irrigation of citrus seedlings based on two-level fuzzy control. First, multiple types of sensors (air temperature and humidity, soil temperature and humidity, conductivity, light, wind speed, and ultraviolet intensity) and edge computing devices are used to achieve high-frequency and continuous monitoring of seedling environment data, and the collected data is uploaded to the cloud platform in real time through the Internet of Things protocol. Second, The collected environmental data are cleaned, normalized and time series constructed to ensure data integrity and consistency. The constructed time series data are then used to train a deep learning model based on a fusion of convolutional neural networks (CNNs), gated recurrent units (GRUs) and feedforward neural networks (FFNs). The trained model is used to fully exploit the spatiotemporal characteristics of multi-source environmental data to achieve dynamic prediction of citrus seedling transpiration in the next 24 hours. The predicted results are compared with the standard transpiration calculated using the Penman formula to automatically complete the transpiration level classification. Finally, the classified transpiration level is input into a two-level fuzzy controller together with parameters such as the current citrus growth stage and environmental deviation: the ideal environmental target corresponding to the growth stage is automatically generated through the first-level fuzzy control, and the second-level fuzzy control integrates the predicted transpiration level with the actual environmental deviation, dynamically infers and outputs parameters such as the specific water level, duration and start-up timing of tidal irrigation, thereby automatically linking water pumps, solenoid valves and other equipment to accurately control the irrigation process, achieving closed-loop feedback and autonomous adjustment throughout the entire process. The system supports daily scheduled automatic operation and can be remotely monitored and manually intervened through the front-end platform, improving the intelligence, refinement and response speed of irrigation, significantly improving water utilization efficiency and seedling cultivation results, and is particularly suitable for the intelligent management and promotion of citrus seedling cultivation in complex environments such as hilly and mountainous areas.

[0078] It should be noted that the computer-readable storage medium of this embodiment may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0079] The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes based on the technical solution and inventive concept of the present invention within the scope disclosed by the present invention, which falls within the scope of protection of the present invention.

Claims

1. A citrus seedling intelligent tidal irrigation method based on secondary fuzzy control, applied to the edge end, characterized in that: The method comprises: Use multiple sensors to collect real-time environmental data for citrus seedlings; Process real-time environmental data to obtain sensor time series data and send it to the cloud; Based on the sensor time series data, the trained citrus seedling transpiration prediction model returned by the cloud server is used to predict future transpiration; based on the future transpiration, the transpiration level is determined; According to the divided transpiration levels and real-time environmental data, fuzzy control technology is used to output irrigation strategies.

2. The intelligent tidal irrigation method for citrus seedling cultivation according to claim 1, characterized in that: Send sensor time series data to the cloud server, including: The sensor time series data is packaged into standard MQTT messages and then published to the EMQX message broker server in the cloud through the MQTT client program. The EMQX message broker server uses the rule engine and data forwarding plug-in to store valid fields in the structured database in the cloud. Among them, the standard MQTT message includes payload and timestamp.

3. The intelligent tidal irrigation method for citrus seedling cultivation according to any one of claims 1 and 2, characterized in that: The sensor time series data is obtained through the following process: Packaging real-time environmental data into MQTT protocol data packets; Cleaning the data in the data package; the cleaning includes processing missing values, duplicate values ​​and outliers, and performing consistency checks to reduce data noise and errors; Extract the payload and the time when the data was acquired based on the cleaned data; The cleaned data and the time of data acquisition are matched to obtain sensor time series data.

4. The intelligent tidal irrigation method for citrus seedling cultivation according to claim 1, characterized in that: The citrus seedling transpiration prediction model is a hybrid prediction model that combines a convolutional neural network, a gated recurrent unit, and a feedforward neural network; The trained citrus seedling transpiration prediction model is obtained through the following process: Using sensor time series data, we construct supervised learning sample pairs, including: The sensor time series data within a fixed time window in the past is used as the input data of the transpiration prediction model of citrus seedlings to represent the changing trend of the environment in which the citrus seedlings are located. The time series of crop transpiration corresponding to a fixed time length in the future is used as the supervision label of the citrus seedling transpiration prediction model; the time series of crop transpiration is the measured transpiration data; The supervised learning samples were used to train the transpiration prediction model for citrus seedlings, including: The input data is fed into the fused convolutional neural network and gated recurrent unit, which are used to extract local spatial features and learn long-term dependencies respectively; The outputs of the fused convolutional neural network and the gated recurrent unit are input into the feedforward neural network to integrate spatiotemporal features; The output data of the feedforward neural network is connected through residual connection to obtain the future transpiration.

5. The intelligent tidal irrigation method for citrus seedling cultivation according to claim 1, characterized in that: The transpiration level is determined based on the ratio of the future transpiration to the standard transpiration, which is calculated by the Penman formula based on the sensor time series data.

6. The intelligent tidal irrigation method for citrus seedling cultivation according to claim 1, characterized in that: The method of outputting an irrigation strategy using fuzzy control technology based on the divided transpiration levels and real-time environmental data includes: The planting time of citrus seedlings is input into the first-level fuzzy controller, which outputs the target air humidity temperature and target soil temperature and humidity required at the current stage; Calculate the deviation between real-time environmental data and target air humidity and temperature and target soil temperature and humidity; The transpiration level and deviation value determined based on real-time environmental data are input into the secondary fuzzy controller to obtain the adjustment of tidal water level, irrigation duration and irrigation start time.

7. The intelligent tidal irrigation method for citrus seedling cultivation according to claim 1, characterized in that: The method of collecting real-time environmental data of citrus seedlings using multiple sensors includes: Connect the Raspberry Pi controller to multiple sensors; Using the Raspberry Pi host to run the lightweight system of Raspberry Pi OS, by setting up scheduled tasks and Python scripts, multiple sensors can collect real-time environmental data.

8. The intelligent tidal irrigation method for citrus seedling cultivation according to any one of claims 1 to 2 and 4 to 7, characterized in that: The real-time environmental data includes air temperature, air humidity, soil temperature, soil humidity and soil conductivity.

9. A citrus seedling intelligent tidal irrigation system based on secondary fuzzy control, the system comprising an edge terminal and a cloud terminal, wherein: The edge end includes: A data acquisition module is used to collect real-time environmental data of citrus seedlings using multiple sensors; The data processing module is used to process real-time environmental data, obtain sensor time series data and send it to the cloud; The transpiration prediction module is used to predict future transpiration based on sensor time series data using the trained citrus seedling transpiration prediction model returned by the cloud server; and to determine the transpiration level based on the future transpiration; Irrigation strategy adjustment module, which is used to output irrigation strategies using fuzzy control technology based on the divided transpiration levels and real-time environmental data; The cloud end is used to receive real-time environmental data and sensor time series data sent by the edge end, and use the sensor time series data to train the citrus seedling transpiration prediction model, and return the trained citrus seedling transpiration prediction model to the edge end.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting tidal irrigation for citrus seedlings according to any one of claims 1 to 8 is implemented.

Images