Citrus seedling prediction irrigation method and device based on multiple sensors and storage medium

Abstract

The invention discloses a citrus seedling prediction irrigation method and device based on multiple sensors and a storage medium. The method comprises the steps that the multiple sensors are used for collecting real-time environment data of a citrus orchard; processing the real-time environment data to obtain sensor time sequence data; obtaining a training set according to the sensor time sequence data; training the irrigation prediction model by using the training set; inputting the actual water demand of the crops into the trained irrigation prediction model, and outputting the future water demand of the crops; according to the future water demand, the irrigation time, frequency and irrigation water quantity are determined. According to the data collected by the plurality of sensors and the irrigation prediction model, the future water demand of crops can be accurately predicted; the irrigation time, the irrigation frequency and the irrigation water amount are determined according to the water demand, and therefore accurate irrigation is achieved for citrus seedling culture.

Description

Technical Field

[0001] The present invention relates to the technical field of seedling cultivation, and in particular to a citrus seedling prediction irrigation method, device, terminal equipment and computer-readable storage medium based on multiple sensors. Background Art

[0002] my country's citrus planting area exceeds 2 million hm², with an output of 40 million tons. The citrus industry has become my country's largest fruit industry. Water is the basis for the growth and development of citrus and a necessary condition for its life activities. Whether the water management of citrus orchards is scientific and reasonable will directly affect the growth and development of citrus, the yield and quality of fruits. Especially in the critical period of water demand for citrus growth and development, it is necessary to ensure that there is sufficient water for soil irrigation. The total amount of water resources in the main citrus producing areas is relatively sufficient, with an average annual rainfall of 1200-1500mm, which can basically meet the water requirements of the citrus industry. However, there are problems such as unbalanced allocation of water and soil resources in the region and uneven temporal and spatial distribution of rainfall, which have led to frequent water and drought disasters (waterlogging and drought) in some main citrus producing areas. Since the citrus industry is mainly distributed in mountainous and hilly areas, the amount of water used for irrigation in citrus orchards is large, and the traditional citrus orchard irrigation method still mainly relies on manual experience, which makes the utilization rate of water resources low and causes the problem of low irrigation efficiency. The serious water shortage problem in the citrus industry restricts or hinders the healthy development of the citrus industry.

[0003] At present, the traditional method relies on manual irrigation, which can only maintain simple but laborious irrigation. It fails to take into account multiple parameters of the citrus seedling cultivation environment, as well as real-time monitoring data and real-time regulation of the irrigation system according to demand. Summary of the invention

[0004] In order to address the deficiencies of the above-mentioned prior art, the present invention provides a citrus seedling predictive irrigation method, device, terminal equipment and computer-readable storage medium based on multi-sensor. According to the real-time environmental data collected by multiple sensors, the XGBoost model is used to predict the future water requirement of crops, and the actual water application amount is determined according to the future water requirement, thereby achieving precise irrigation, improving water resource utilization, reducing agricultural production costs, and ultimately improving the yield and quality of citrus, which has important practical significance and application value.

[0005] The first object of the present invention is to provide a citrus seedling prediction irrigation method based on multiple sensors.

[0006] The second object of the present invention is to provide a citrus seedling prediction irrigation device based on multiple sensors.

[0007] The third object of the present invention is to provide a terminal device.

[0008] A fourth object of the present invention is to provide a computer-readable storage medium.

[0009] The first object of the present invention can be achieved by adopting the following technical solutions: A citrus seedling prediction irrigation method based on multiple sensors, the method comprising: Use multiple sensors to collect real-time environmental data from citrus orchards; Process the real-time environmental data to obtain sensor time series data; According to the sensor time series data, a training set is obtained; and the irrigation prediction model is trained using the training set; The actual crop water requirement is input into the trained irrigation prediction model to output the crop's future water requirement; based on the future water requirement, the irrigation time, frequency and amount of irrigation water are determined.

[0010] Furthermore, the irrigation prediction model is an XGBoost model.

[0011] Furthermore, the samples in the training set include crop water requirements and the time to obtain processed real-time environmental data; The method of training the irrigation prediction model using the training set includes: The current or past crop water requirement is used as input data, and the crop water requirement in the future time is used as output data; Using irrigation prediction models to learn the relationship between input and output data; The learned irrigation prediction model is verified using the test set. If the evaluation index meets the set requirements, the training is completed.

[0012] Further, the sensor time series data includes processed real-time environmental data and the time when the processed real-time environmental data is acquired; The training set is obtained according to the sensor time series data, including: Calculate crop evapotranspiration based on processed real-time environmental data from sensor time series data; Calculate crop water requirements based on crop evapotranspiration; The crop water requirement and the time of obtaining processed real-time environmental data are used as samples in the training set.

[0013] Furthermore, the crop evapotranspiration is calculated using the Penman formula based on the processed real-time environmental data from the sensor time series data.

[0014] Furthermore, the real-time environmental data is processed to obtain sensor time series data, including: Pack the real-time environment data into MQTT protocol data packets; Cleaning the data in the data packet; the cleaning includes processing missing values, duplicate values ​​and abnormal values, and performing consistency checks to reduce noise and errors in the data; Based on the cleaned data, extract the payload and the time when the data was obtained; The cleaned data and the time of acquiring the data are matched to obtain the sensor time series data.

[0015] Furthermore, the use of multiple sensors to collect real-time environmental data of the citrus orchard includes: The real-time environmental data of the citrus orchard is obtained through a collection system; the collection system includes a local collection system and a cloud service platform, the local collection system is used to collect the real-time environmental data of the citrus orchard using multiple sensors, and the cloud service platform is used to access and query the real-time environmental data of the citrus orchard; the real-time environmental data includes air temperature and humidity, net solar radiation, atmospheric pressure and wind speed data.

[0016] The second object of the present invention can be achieved by adopting the following technical solutions: A citrus seedling prediction irrigation device based on multiple sensors, the device comprising: A data acquisition module for collecting real-time environmental data of the citrus orchard using multiple sensors; A data processing module is used to process real-time environmental data to obtain sensor time series data; The model training module is used to obtain a training set based on the sensor time series data; and to train the irrigation prediction model using the training set; The water demand prediction module is used to input the actual crop water demand into the trained irrigation prediction model and output the future water demand of the crop; based on the future water demand, the irrigation time, frequency and irrigation water volume are determined.

[0017] 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 citrus seedling prediction irrigation method based on multiple sensors is implemented.

[0018] The fourth object of the present invention can be achieved by adopting the following technical solutions: A computer-readable storage medium stores a program, and when the program is executed by a processor, the above-mentioned citrus seedling prediction irrigation method based on multiple sensors is implemented.

[0019] The present invention has the following beneficial effects compared with the prior art: 1. The present invention integrates multiple sensor data, data analysis and predictive irrigation models to effectively predict the irrigation water application amount of citrus seedlings to achieve precise irrigation control, avoid excessive or insufficient irrigation, and ensure that citrus seedlings are adequately supplied with water under optimal growth conditions.

[0020] 2. Compared with traditional irrigation methods, the XGBoost model used in the present invention can accurately predict the amount of irrigation water for citrus seedlings, providing farmers and managers with a scientific decision-making basis, optimizing irrigation plans, reasonably arranging irrigation time and frequency, improving water resource utilization efficiency, reducing production costs, and ultimately improving crop yield and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0022] Figure 1 This is a flow chart of a citrus seedling prediction irrigation method based on multiple sensors according to Example 1 of the present invention; Figure 2 This is a schematic diagram of the structure of a citrus seedling prediction irrigation system according to Example 1 of the present invention; Figure 3 This is a schematic diagram of an MQTTx client according to Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the EMQX linkage MySQL function of Example 1 of the present invention; Figure 5 A scatter plot of crop evapotranspiration distributed by hour in Example 1 of the present invention; Figure 6 A scatter plot of crop evapotranspiration distributed by date and time according to Example 1 of the present invention; Figure 7 This is a schematic diagram of real-time sensor data for predictive irrigation of citrus seedlings according to Example 1 of the present invention; Figure 8 This is a schematic diagram of a monitoring system on a citrus seedling prediction irrigation Internet of Things platform according to Example 1 of the present invention; Fig. 9 This is a structural block diagram of a citrus seedling prediction irrigation device based on multiple sensors according to Example 2 of the present invention; Fig.10 This is a structural block diagram of the terminal device of embodiment 3 of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. It should be understood that the specific embodiments described are only used to explain the present application and are not used to limit the present application.

[0024] Embodiment 1: like Figure 1 , 2 As shown, this embodiment provides a citrus seedling prediction irrigation method based on multiple sensors, and the specific implementation process includes the following steps: S101. Use multiple sensors to obtain real-time data of a citrus orchard.

[0025] In this step, real-time data of the citrus orchard is obtained through the collection system, where the collection system consists of two parts: a local collection system and a cloud service platform.

[0026] Furthermore, step S101 includes: (1) Collect real-time data using multiple sensors in the local acquisition system.

[0027] The local acquisition system includes STM32 master control, RTOS real-time operating system, sensors and MQTTx client.

[0028] The STM32 master is connected to multiple sensors through the RS485 isolation module, where one end of the RS485 isolation module is connected to the serial communication interface (USART / UART) of the STM32 master, and the other end is connected to the RS485 communication interface of the sensor. The RTOS real-time operating system is used to initialize the STM32 master, enable the 485 isolation module, configure the network program, start the data acquisition program, and the data upload program.

[0029] The STM32 master controller uses multiple connected sensors to obtain real-time data of the citrus orchard, including air temperature and humidity, soil temperature and humidity, light intensity, atmospheric pressure, net solar radiation, wind speed and other data.

[0030] like Figure 3, the STM32 master packaged the data collected by the sensor into MQTT protocol data packets and transmitted them to the MQTTx platform. The MQTT protocol is a lightweight, low-bandwidth message transmission protocol designed to achieve reliable communication between devices in IoT applications; it uses a publish-subscribe model, that is, communication between an MQTT server and multiple MQTT clients, which is suitable for data communication in the citrus seedling prediction irrigation system.

[0031] In one embodiment, the plurality of sensors are RS485 sensors.

[0032] MQTTx is an open source cross-platform MQTT desktop client developed by EMQ, which is compatible with Linux, Windows and MacOS systems. The user interface UI of MQTTx adopts a chat-style design, making the operation logic more concise and intuitive. It supports users to quickly create and save multiple MQTT connections, facilitate testing MQTT connections, and subscribe and publish MQTT messages. Use MQTTx as a client for sending and receiving MQTT messages, that is, the client interface for sending and receiving sensor data of the local acquisition system. Every time MQTTx receives data, it will log it into a .log document, which contains the time when the message was received and the data of each sensor at the current time. The data will also be synchronized to the cloud server.

[0033] (2) Use the cloud server platform to access and query the collected real-time data.

[0034] like Figure 4 In this embodiment, the EMQX IoT platform and MySQL are used on the Alibaba Cloud server to store and read and write data of the MQTT protocol. As an IoT platform, EMQX is responsible for accessing devices and transmitting messages; as a data storage platform, MySQL is responsible for storing device status and original data, as well as message data storage and data analysis. MySQL data integration is a built-in function in EMQX, which can achieve complex business development through simple configuration. After querying the sensor data through the MySQL export function, the query results of the sensor data are exported as CSV files, which are saved in the cloud and locally as the acquired sensor data for use in subsequent steps.

[0035] In this embodiment, EMQX is used to integrate the received MQTT data. EMQX forwards device events and data to MySQL through the rule engine and Sink. By reading the data in MySQL, the device status can be sensed, the device online and offline records can be obtained, and the device data can be analyzed. The specific workflow is as follows: (1) First, connect the device to EMQX. After the IoT device is successfully connected through the MQTT protocol, an online event will be triggered. The event contains information such as the device ID, source IP address, and other attributes.

[0036] (2) Device message publishing and reception: The device publishes telemetry and status data through specific topics. After receiving the message, EMQX will compare it in the rule engine.

[0037] (3) Write data to MySQL. The rule triggers the operation of writing messages to MySQL.

[0038] (4) With the help of SQL templates, users can extract data from the rule processing results and construct SQL to send to MySQL for execution, thereby writing or updating specific fields of the message into the corresponding tables and columns of the database.

[0039] In this embodiment, the MySQL scheduled dump function is used to export a CSV format data set at a certain time point every day. The steps are as follows: (1) Write SQL statements: Write SQL query statements to export table data; (2) Test the export command: Test the SQL export query in the command line; (3) Create a script: Write the export command into an executable script; (4) Set up scheduled tasks: Use cron to set up scheduled tasks.

[0040] The configuration of the cloud server platform in this step is the Ubuntu 20.04 64-bit system on the Alibaba Cloud server, a 2-core CPU, and 2GB of memory.

[0041] S102: Process the real-time data to obtain sensor time series data.

[0042] This step includes data cleaning, preprocessing, data analysis and visualization of multi-sensor data to obtain the training set required for training the prediction model.

[0043] In this embodiment, all the sensor data collected include air temperature and air humidity, soil temperature and humidity, solar net radiation, atmospheric pressure, wind speed and light intensity, and are sent to the Internet of Things platform in real time through wireless transmission technology to keep track of the meteorological environment and moisture dynamics of the crops at any time. Among them, the five data of air temperature, air humidity, solar net radiation, atmospheric pressure and wind speed are used as real-time environmental data for crop evapotranspiration ( ET 0 )calculate.

[0044] Data cleaning includes using fillna(), dropna(), drop_duplicates() and other functions to process missing values, duplicate values ​​and outliers, converting the format of MQTT data, and performing consistency checks to reduce data noise and errors.

[0045] Preprocessing is to extract the payload (message body of MQTT data) and the time of acquiring data from the MQTT data, match the values ​​of each sensor with the time of acquiring data, and form a set of sensor time series data, including date and time and the values ​​collected by each sensor at the corresponding date and time, for use in subsequent model building.

[0046] Data analysis and visualization uses frequency distribution, percentages, etc. to summarize the characteristics of sensor data, and draws scatter plots, line graphs, and bar graphs of the data of each sensor at the current time, which can better understand and display the data; through data line graphs, etc., the trend of data changes over time can be discovered, such as seasonal changes in crop water demand.

[0047] The above operations are completed using Python's Csv, Json, Pandas, Matplotlib, and Seaborn libraries.

[0048] S103. Obtain a training set based on the sensor time series data; and use the training set to train the irrigation prediction model.

[0049] Further, step S103 includes: (1) Obtain a training set based on sensor time series data.

[0050] This step uses the sensor time series data obtained in step S102 to calculate crop evapotranspiration ( ET 0 ), calculated using the Penman-Monteith equation.

[0051] The Penman formula is as follows (in mm / day):

[0052] in: Δ is the slope of the saturated vapor pressure curve (kPa / °C), calculated from 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 is approximately 0 on a daily scale; γ is the hygrometer constant (kPa / °C), calculated from air temperature and atmospheric pressure; T is the average temperature (°C), calculated from the air temperature; u 2 is the wind speed (m / s); e s is the saturated vapor pressure (kPa), calculated from the air temperature and air humidity; e a is the actual vapor pressure (kPa), calculated from the air temperature and air humidity.

[0053] This example uses Python to calculate crop evapotranspiration.

[0054] like Figure 5 , the calculation results of crop evapotranspiration distributed by hour are consistent with the variation of crop evapotranspiration of citrus seedlings in a day; Figure 6 The calculation results of crop evapotranspiration distributed by date and time are consistent with the daily changes in crop evapotranspiration of citrus seedlings in a season.

[0055] The crop water requirement is obtained by the following formula ( ET C ), in mm / day:

[0056] in, K is the crop coefficient.

[0057] The crop coefficient can be obtained through experimental data or by referring to the data released by the agricultural and rural departments, and is determined by the growth stage of the crop. The paper verifies the crop coefficient of citrus seedlings ( K ), generally takes a value of 0.5.

[0058] Calculate the crop water requirement for citrus nursery systems ( ET C ), the crop water requirement and the corresponding time are divided into training set and test set in a ratio of 7:3 through the train_test_split function in the sklearn library in the python environment.

[0059] (2) Use the training set to train the irrigation prediction model.

[0060] Train the XGBoost prediction model on the training set to let the model learn the relationship between input data (the input crop water requirement can be current or past data) and output data (future crop water requirement). Subsequently, use the pre-divided test set to verify the model performance. The evaluation indicators may include mean square error (MSE), mean absolute error (MAE), etc. It is used to evaluate the deviation between the model prediction results and the true value. After the training is completed, the generated prediction model is used for actual prediction. The specific steps are as follows: (1) Load the trained prediction model: Use the loading function provided by the XGBoost library to load the previously trained model into memory. In XGBoost, you can load the model through the pickle module or the load_model method of the xgboost.Booster object.

[0061] (2) Data format adjustment: Convert the preprocessed prediction data into the input format required by the XGBoost model.

[0062] (3) Perform prediction operations: Use the loaded and configured model to predict the formatted input data. In XGBoost, predictions can be made using the model.predict() function, where model is the loaded model and dtest is the prediction data converted to a suitable format. The format and content of the prediction results depend on the output settings of the model. For example, predicting the crop water requirement of citrus seedlings over a certain period of time in the future ET C The result is time series data for a specific period in the future, which represents the ET C Predicted value.

[0063] S104. Predicting future crop water requirements based on actual crop water requirements using the trained irrigation prediction model; determining irrigation time, frequency, and irrigation water volume based on future crop water requirements.

[0064] The actual ET C The forecast value is used to predict the future crop water requirement at a specific point in the future, so as to accurately determine the time, frequency and amount of irrigation. The following are the parameters for forecasting irrigation: (1) Irrigation time: Select the best irrigation period based on weather forecasts and historical data (such as avoiding high temperatures at noon).

[0065] (2) Irrigation frequency: Set a reasonable irrigation interval (e.g., once every 3 days) based on soil moisture and predicted future crop water requirements.

[0066] (3) Irrigation size: Calculate the amount of water for each irrigation based on the predicted future crop water demand (e.g. 40 mm of irrigation water per day) to ensure that crop needs are met.

[0067] After completing the above steps, the irrigation system is precisely controlled to achieve a relatively balanced growth of the citrus seedlings.

[0068] This step uses the decision tree prediction model XGBoost, which has significant benefits in predicting citrus seedling irrigation systems. By integrating the higher prediction accuracy of multiple decision trees, the XGBoost model is able to capture complex data patterns and nonlinear relationships, thereby providing high-precision irrigation water application predictions to avoid excessive or insufficient irrigation. The model can effectively process a variety of feature data, including air temperature, air humidity, wind speed, net solar radiation, atmospheric pressure, etc., comprehensively consider the impact of each variable, and provide comprehensive prediction results. Due to its efficient data processing and parallel computing capabilities, the XGBoost model can handle large-scale data sets and quickly respond and adapt to real-time changes. Its robustness to missing values ​​and outliers and its flexible parameter adjustment capabilities enable it to maintain stable prediction performance in the face of environmental changes and data fluctuations.

[0069] This embodiment also designs an Internet of Things platform, which is mainly composed of an online management module, a real-time monitoring system and example applications.

[0070] The online management module is used to set different operating permissions for different users to realize the most basic functions of the IoT platform. It mainly includes three types of users with different permissions: managers, ordinary users, and guests. Among them, managers are allowed to have the highest permissions and can fully control system configuration, user permission allocation, data analysis, report generation and other advanced operations; ordinary users have limited permissions and can access and operate functional modules related to their permissions, such as monitoring various sensor data and controlling irrigation systems; guest users have the lowest permissions and can only view specific public information without making any changes. Through this permission setting, the online management module allows different users to perform corresponding operations according to their respective permissions.

[0071] like Figure 7 As shown in the figure, the real-time monitoring system, as the core of the IoT platform, is mainly used to monitor the values ​​collected in real time by multiple sensors, such as air temperature and humidity, soil temperature and humidity, light intensity, atmospheric pressure, and solar net radiation. The real-time monitoring system can dynamically grasp the changes in the growth environment of citrus seedlings through the real-time collection and analysis of these sensor data. Each sensor in the system is calibrated to ensure the accuracy and reliability of the data. The monitoring system adopts multi-layer data verification and redundancy mechanisms to prevent data loss and false alarms.

[0072] like Figure 8 As shown, the example application allows users to set and adjust specific irrigation parameters according to different seedling requirements. The irrigation system automatically adjusts the irrigation strategy based on these parameters and real-time monitoring data to ensure the best seedling environment.

[0073] Those skilled in the art will appreciate that all or part of the steps in the system 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.

[0074] It should be noted that although the system operations of the above embodiments are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. On the contrary, the steps depicted can change the order of execution. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step, and / or one step can be decomposed into multiple steps.

[0075] Embodiment 2: like Fig. 9 As shown, this embodiment provides a citrus seedling prediction irrigation device based on multiple sensors, which includes a data acquisition module 901, a data processing module 902, a model training module 903 and a water demand prediction module 904. The specific functions of each module are as follows: The data collection module 901 is used to collect real-time environmental data of the citrus orchard using multiple sensors; The data processing module 902 is used to process the real-time environmental data to obtain sensor time series data; The model training module 903 is used to obtain a training set according to the sensor time series data; and train the irrigation prediction model using the training set; The water demand prediction module 904 is used to input the actual crop water demand into the trained irrigation prediction model and output the future water demand of the crop; and determine the irrigation time, frequency and irrigation water volume according to the future water demand.

[0076] Embodiment 3: like Fig.10As shown, this embodiment provides a terminal device, which includes a processor 1002, a memory, an input device 1003, a display device 1004 and a network interface 1005 connected through a system bus 901. The processor 1002 is used to provide computing and control capabilities, and the memory includes a non-volatile storage medium 1006 and an internal memory 1007. The non-volatile storage medium 1006 stores an operating system, a computer program and a database. The internal memory 1007 provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium 1006. When the computer program is executed by the processor 1002, the method for predicting irrigation for citrus seedlings in the above-mentioned embodiment 1 is implemented as follows: Use multiple sensors to collect real-time environmental data from citrus orchards; Process the real-time environmental data to obtain sensor time series data; According to the sensor time series data, a training set is obtained; and the irrigation prediction model is trained using the training set; The actual crop water requirement is input into the trained irrigation prediction model to output the crop's future water requirement; based on the future water requirement, the irrigation time, frequency and amount of irrigation water are determined.

[0077] Embodiment 4: This embodiment provides a storage medium, which is a computer-readable storage medium and stores a computer program. When the computer program is executed by a processor, the method for predicting irrigation for citrus seedling cultivation of the above embodiment 1 is implemented as follows: Use multiple sensors to collect real-time environmental data from citrus orchards; Process the real-time environmental data to obtain sensor time series data; According to the sensor time series data, a training set is obtained; and the irrigation prediction model is trained using the training set; The actual crop water requirement is input into the trained irrigation prediction model to output the crop's future water requirement; based on the future water requirement, the irrigation time, frequency and amount of irrigation water are determined.

[0078] It should be noted that the computer-readable storage medium of the present embodiment may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with 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 of the above.

[0079] The computer readable storage medium may be written in one or more programming languages ​​or a combination thereof to execute the computer program of the present embodiment, and the programming language includes an object-oriented programming language, such as Java, Python, C++, and a conventional procedural programming language, such as C or a similar programming language. The program may be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect through the Internet).

[0080] The above is only a preferred embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solution and inventive concept of the present invention within the scope disclosed by the present invention, which shall fall within the protection scope of the present invention.

Claims

1. A multi-sensor based citrus seedling prediction irrigation method, characterized in that: The method comprises: Use multiple sensors to collect real-time environmental data from citrus orchards; Process the real-time environmental data to obtain sensor time series data; According to the sensor time series data, a training set is obtained; and the irrigation prediction model is trained using the training set; The actual crop water requirement is input into the trained irrigation prediction model to output the crop's future water requirement; based on the future water requirement, the irrigation time, frequency and amount of irrigation water are determined.

2. The citrus seedling prediction irrigation method according to claim 1, characterized in that: The irrigation prediction model is an XGBoost model.

3. The citrus seedling prediction irrigation method according to any one of claims 1 and 2, characterized in that: The samples in the training set include crop water requirements and the time to obtain processed real-time environmental data; The method of training the irrigation prediction model using the training set includes: The current or past crop water requirement is used as input data, and the crop water requirement in the future time is used as output data; Using irrigation prediction models to learn the relationship between input and output data; The learned irrigation prediction model is verified using the test set. If the evaluation index meets the set requirements, the training is completed.

4. The citrus seedling prediction irrigation method according to any one of claims 1 and 2, characterized in that: The sensor time series data includes processed real-time environmental data and the time when the processed real-time environmental data is obtained; The training set is obtained according to the sensor time series data, including: Calculate crop evapotranspiration based on processed real-time environmental data from sensor time series data; Calculate crop water requirements based on crop evapotranspiration; The crop water requirement and the time of obtaining processed real-time environmental data are used as samples in the training set.

5. The citrus seedling prediction irrigation method according to claim 4, characterized in that: The crop evapotranspiration is calculated using the Penman formula based on the processed real-time environmental data from the sensor time series data.

6. The citrus seedling prediction irrigation method according to any one of claims 1 and 2, characterized in that: Process the real-time environmental data to obtain sensor time series data, including: Pack the real-time environment data into MQTT protocol data packets; Cleaning the data in the data packet; the cleaning includes processing missing values, duplicate values ​​and abnormal values, and performing consistency checks to reduce noise and errors in the data; Based on the cleaned data, extract the payload and the time when the data was obtained; The cleaned data and the time of acquiring the data are matched to obtain the sensor time series data.

7. The citrus seedling prediction irrigation method according to any one of claims 1 and 2, characterized in that: The method of collecting real-time environmental data of the citrus orchard using multiple sensors includes: The real-time environmental data of the citrus orchard is obtained through a collection system; the collection system includes a local collection system and a cloud service platform, the local collection system is used to collect the real-time environmental data of the citrus orchard using multiple sensors, and the cloud service platform is used to save and query the real-time environmental data of the citrus orchard; the real-time environmental data includes air temperature and humidity, net solar radiation, atmospheric pressure and wind speed data.

8. A citrus seedling prediction irrigation device based on multiple sensors, characterized in that: The device comprises: A data acquisition module for collecting real-time environmental data of the citrus orchard using multiple sensors; A data processing module is used to process real-time environmental data to obtain sensor time series data; The model training module is used to obtain a training set based on the sensor time series data; and to train the irrigation prediction model using the training set; The water demand prediction module is used to input the actual crop water demand into the trained irrigation prediction model and output the future water demand of the crop; based on the future water demand, the irrigation time, frequency and irrigation water volume are determined.

9. A terminal device, comprising a processor and a memory for storing a program executable by the processor, characterized in that: When the processor executes the program stored in the memory, the citrus seedling prediction irrigation method described in any one of claims 1 to 7 is implemented.

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 citrus seedling prediction irrigation method according to any one of claims 1 to 7 is implemented.

Images