In-situ paddy field moisture infiltration rate monitoring system and monitoring method

Abstract

The invention discloses an in-situ rice field moisture infiltration rate monitoring system and a detection method thereof. The monitoring system comprises an in-situ monitoring device deployed in a rice field, and a data processing terminal; wherein the in-situ monitoring device periodically and synchronously acquires the liquid level distance and meteorological parameters of the water surface of the rice field, preprocesses the liquid level distance and the meteorological parameters and then provides the preprocessed parameters to the data processing terminal; the data processing terminal automatically determines the daily decline amount of the liquid level and the daily evapotranspiration amount of crops, and the daily infiltration amount and the in-situ rice field moisture infiltration rate are obtained. The monitoring method comprises the following steps: collecting parameters; performing pretreatment; obtaining an effective infiltration time period list; obtaining the daily decline amount and the daily evapotranspiration amount of the liquid level; determining the daily infiltration amount; and determining the in-situ rice field moisture infiltration rate. According to the invention, non-contact and automatic monitoring is realized on the premise that the plow pan structure of the rice field is not damaged, infiltration characteristics under unstable conditions such as a dry-wet alternation period and the like can be effectively identified, and low-cost and high-time-resolution data support is provided for precise irrigation of the rice field and prevention and control of non-point source pollution.

Description

Technical Field

[0001] This invention belongs to the field of agricultural water conservancy and environmental monitoring technology, specifically a monitoring system and method for monitoring the in-situ water infiltration rate in paddy fields based on non-contact sensing and intelligent data processing. Background Technology

[0002] Paddy fields are a major water-intensive crop system globally. Accurately quantifying the amount of water infiltrating into paddy fields is a key foundation for implementing precision irrigation, improving water resource utilization, and controlling non-point source pollution.

[0003] In existing technologies, monitoring of water infiltration in paddy fields mainly employs physical capture methods (such as lysimeters) or simple liquid level monitoring. While the principle of physical capture methods is intuitive, it requires disrupting the original soil structure in the field, making it an invasive measurement. Furthermore, the equipment is expensive, making it difficult to deploy at low cost in large-scale farmland. Although simple liquid level monitoring can achieve non-contact measurement, it only obtains the total drop in surface water level and cannot decouple the two key processes of "water infiltration" and "surface evapotranspiration," resulting in monitoring data that cannot directly guide irrigation decisions.

[0004] Theoretically, the water balance method can be used to solve the decoupling problem. However, this method is currently mainly used as an offline research approach relying on manual intervention, lacking an integrated automated system. More importantly, in high-frequency in-situ monitoring of paddy fields, liquid level data inevitably includes drastic fluctuations and noise caused by rainfall, irrigation, and runoff. Existing technologies lack an intelligent algorithm capable of real-time analysis of high-time-series data and automatic identification and calibration of effective infiltration periods, preventing the water balance method from achieving automation and precision in complex field environments. Therefore, there is an urgent need to develop a system that integrates non-contact sensing technology and automated data processing algorithms to achieve low-cost, high-temporal-resolution infiltration rate monitoring without damaging the in-situ soil. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies, such as physical capture methods damaging in-situ soil structure, the inability of simple liquid level monitoring methods to decouple evapotranspiration and infiltration components, and the inability of traditional water balance methods to automatically handle high-frequency in-situ data interference. This invention provides an in-situ paddy field water infiltration rate monitoring system and method. This solution aims to achieve low-cost, automated, and high-temporal-resolution accurate monitoring of paddy field infiltration rate without disturbing the topsoil and plow pan layers through the deep integration of non-contact sensing technology and intelligent data processing algorithms.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] An in-situ paddy field water infiltration rate monitoring system, the monitoring system includes an in-situ monitoring device deployed in the paddy field and a data processing terminal;

[0008] The in-situ monitoring device includes a non-contact ranging sensor suspended vertically above the paddy field water surface and a meteorological parameter acquisition module. It synchronously collects the liquid level distance and meteorological parameters of the paddy field water surface on a periodic basis and preprocesses them on a periodic basis into minute-level liquid level distance time series data and minute-level environmental meteorological time series data, which are then provided to the data processing terminal.

[0009] The data processing terminal automatically identifies effective periods in the minute-level liquid level distance time series data where the distance reading of the non-contact ranging sensor steadily increases and marks them as effective infiltration periods, and determines the daily drop in liquid level based on the effective infiltration periods;

[0010] The data processing terminal automatically aggregates minute-level environmental meteorological time-series data to obtain the average value of hourly environmental meteorological time-series data, and calculates the hourly evapotranspiration of reference crops based on the average value of hourly environmental meteorological time-series data. The hourly evapotranspiration of reference crops is calculated hourly and accumulated to obtain the daily evapotranspiration of reference crops. Based on the daily evapotranspiration of reference crops and according to the crop coefficient method, the daily evapotranspiration of crops is obtained.

[0011] The data processing terminal obtains the daily infiltration rate based on the difference between the daily drop in liquid level and the daily evapotranspiration of crops, and obtains the in-situ water infiltration rate of paddy fields based on the daily infiltration rate.

[0012] The in-situ monitoring device includes a power supply unit, an in-situ data acquisition unit, a data processing unit, and a data storage and transmission unit. The power supply unit supplies power to the in-situ data acquisition unit, the data processing unit, and the data storage and transmission unit. The in-situ data acquisition unit collects the liquid level distance and meteorological parameters of the monitoring point in real time according to a cycle and transmits them to the data processing unit. The data processing unit determines the effective period and averages and preprocesses the liquid level distance and meteorological parameters within the effective period into minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data, which are then provided to the data processing terminal through the data storage and transmission unit.

[0013] The in-situ data acquisition unit includes a non-contact liquid level distance acquisition module, a meteorological parameter acquisition module, and a timestamp module. The liquid level distance acquisition module is used to collect the liquid level distance of the monitoring point in real time on a periodic basis and transmit it to the data processing unit. The meteorological parameter acquisition module is used to collect the wind speed, air temperature, relative humidity, and solar radiation of the monitoring point in real time on a periodic basis and transmit it to the data processing unit. The timestamp module provides a unified time reference for the real-time data collected synchronously by the liquid level distance acquisition module and the meteorological parameter acquisition module and supports the calculation of the day sequence number.

[0014] The in-situ data acquisition unit includes an interference monitoring module, which uses time-series matching to retain effective infiltration periods where rainfall is zero within the same time period.

[0015] The power supply unit includes a solar panel, a lithium battery, and a charging management circuit. The solar panel provides 12V DC power and supplies power to the in-situ data acquisition unit, data processing unit, and data storage and transmission unit through a step-down module. The lithium battery is used for energy storage and nighttime power supply and ensures stable operation of the in-situ monitoring device under no-light conditions. The solar panel is connected to the lithium battery through the charging management circuit.

[0016] The data processing unit includes a microcontroller and an OLED display. The microcontroller receives synchronously collected liquid level distance and meteorological parameters and preprocesses them into minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data. The OLED display is used to display the date, time, synchronously collected liquid level distance and meteorological parameters, and the preprocessed minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data.

[0017] The data storage and transmission unit includes a local storage module and a remote transmission module. The local storage module is used to store minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data transmitted from the data processing unit. The remote transmission module is used to upload the minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data transmitted from the data processing unit to the cloud server. The data processing terminal obtains the minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data by reading from the local storage module or downloading from the cloud.

[0018] The in-situ monitoring device includes a pole, a support platform, a solar panel, and a sealed waterproof box. Two support platforms are mounted on the pole; one platform houses an air temperature and humidity sensor, a solar radiation sensor, and a wind speed sensor, while the other platform houses the solar panel. The air temperature and humidity sensor (for collecting air temperature and relative humidity), the solar radiation sensor (for collecting solar radiation), and the wind speed sensor (for collecting wind speed) constitute the meteorological parameter acquisition module in the in-situ data acquisition unit. A sealed waterproof box, suspended above the water surface of the paddy field, is located at the lower part of the pole. Inside the sealed waterproof box is a microcontroller and, via wiring, an SD card module, a timestamp module, an ultrasonic distance sensor, an OLED display, and a 4G DTU module, all connected to the microcontroller. The microcontroller and OLED display constitute the data processing unit in the in-situ data acquisition unit. The SD card module and 4G DTU module... The DTU module constitutes the data storage and transmission unit, and the ultrasonic distance sensor constitutes the liquid level distance acquisition module in the in-situ data acquisition unit and is installed at the bottom of the sealed waterproof box. The probe of the ultrasonic distance sensor passes through the bottom surface of the sealed waterproof box and is vertically downward aimed at the paddy field water surface. The distance between the probe of the ultrasonic distance sensor and the paddy field water surface is preset to 10-20cm.

[0019] The support platform is equipped with a tipping bucket rain gauge for collecting rainfall data.

[0020] A method for monitoring the infiltration rate of water in paddy fields in situ, the steps of which are as follows:

[0021] S1. Collect the water level distance and meteorological parameters of the paddy field water surface synchronously on a periodic basis;

[0022] S2. Determine the effective period, average the liquid level distance and meteorological parameters within the effective period, and obtain minute-level liquid level distance time series data and minute-level environmental meteorological time series data.

[0023] S3. Use the PELT change point detection algorithm or machine learning algorithm to determine the effective infiltration time period in the minute-level liquid level distance time series data in step S2, and form a list of effective infiltration time periods.

[0024] S4. Based on the list of effective infiltration periods, perform time weighting to obtain the average rate of change of liquid surface distance over time. Multiply the average rate of change of liquid surface distance over time by 24 hours to obtain the daily drop in liquid surface.

[0025] S5. Call the minute-level environmental meteorological time series data provided in step S2, and use the FAO-56 Penman-Monteith formula and crop coefficient method to obtain the daily evapotranspiration of crops.

[0026] S6. A simplified water balance calculation model based on the paddy field water balance equation: , This refers to the daily infiltration rate, expressed in mm. The daily drop in liquid level obtained in step S4, in mm. The daily evapotranspiration of the crop obtained in step S5 is expressed in mm.

[0027] S7. Obtain the in-situ water infiltration rate of paddy fields based on daily infiltration volume.

[0028] The effective period in step S2 refers to the period in which the number of effective samples exceeds 50% of the total number of samples. If the number of effective samples does not exceed 50% of the total number of samples, the period is skipped.

[0029] The specific steps for obtaining the list of effective infiltration time periods in step S3 using the PELT change point detection algorithm are as follows:

[0030] S311, Apply the minute-level liquid level distance time series data from step S2. The values ​​are smoothed using a rolling average to obtain smoothed and denoised minute-level liquid level distance time-series data. The calculation formula is as follows:

[0031]

[0032] in, For smoothing and noise reduction The liquid level distance value at any given time is in mm. For the data points in the minute-level liquid level distance time series data in step S2, To smooth the window size, used to eliminate high-frequency random noise interference;

[0033] S312. Take the derivative of the smoothed and denoised minute-level liquid level distance time series data to obtain the first-order rate of change. The calculation formula is as follows:

[0034]

[0035] in, for The rate of change of liquid level at any given time (i.e., the instantaneous infiltration rate), in mm / h. and These are the smoothed and denoised minute-level liquid level distance time-series data for the current and previous moments, respectively, in mm. The sampling period is in minutes, and the multiplier of 60 is used to convert the rate unit to a standard hourly rate.

[0036] S313. Use the PELT change point detection algorithm to analyze the first-order rate of change. The analysis was performed on the minute-level liquid level distance time series data. Dividing the data into several independent time periods with stable rates of change, the objective function of the PELT change point detection algorithm is as follows:

[0037]

[0038] in, For the set of identified change points, The number of change points, This is the piecewise cost function, used to measure the residual fit between this segment of data and the mean model. Penalty coefficients to prevent overfitting;

[0039] S314. Perform linear fitting on independent time periods with stable rates of change to obtain the precise slope and duration of each independent time period.

[0040] S315. Based on the preset physical constraints: 0 < Slope≤3mm / h and Duration≥5min, the precise slope and duration of all independent time periods with stable rates of change are screened, and the independent time periods with stable rates of change that meet the physical constraints are identified and marked as undetermined infiltration time periods.

[0041] S316. Set the time interval threshold for the undetermined infiltration period and calculate the time interval between adjacent undetermined infiltration periods. If the time interval between adjacent undetermined infiltration periods is not greater than the time interval threshold, merge the adjacent undetermined infiltration periods into the same continuous period and update the start and end times. Repeat the above comparison process until the time interval between all adjacent periods is greater than the time interval threshold. Then each period is a valid infiltration period. Finally, output a list of valid infiltration periods consisting of the start and end times.

[0042] The time interval threshold in step S316 is not less than 5 minutes.

[0043] The specific steps for obtaining the list of effective infiltration time periods in step S3 using a machine learning algorithm are as follows:

[0044] S321. Feature Engineering: Extract features from the minute-level liquid level distance time-series data in step S2, and set the sliding window size to... For each point in time Calculate the statistical feature vector within the sliding window: rolling mean. Rolling range Rolling standard deviation and the rolling slope, which reflects the instantaneous rate of change. The calculation formulas are as follows:

[0045]

[0046]

[0047]

[0048]

[0049] in, For the distance dataset within the sliding window, The rate of change of liquid level over time within the sliding window is calculated using the least squares linear regression method, and the unit is mm / h, reflecting the local instantaneous infiltration rate.

[0050] S322, Model Prediction and Label Mapping: Input the feature vector extracted in step S321 into the pre-trained random forest classification model to directly obtain the classification result of the random forest classification model, and convert the numerical result into the text label "Rising" or "Stable" according to the label mapping table.

[0051] S323, Peak Filtering: Perform peak detection on all data points initially marked as "Rising". If the slope of the current data point exceeds the peak threshold and the slope of the immediately following data point has fallen back to below the normal threshold, then the point is determined to be an interference signal and is removed.

[0052] S324, Time Period Merging: After aggregating the data points with the text label "Rising" remaining after peak filtering, merge them into continuous time periods, and remove time periods with a duration shorter than the set duration or a number of points lower than the set number of points to obtain the undetermined infiltration time period;

[0053] S325. Set the time interval threshold for the undetermined infiltration period and calculate the time interval between adjacent undetermined infiltration periods. If the time interval between adjacent undetermined infiltration periods is not greater than the time interval threshold, merge the adjacent undetermined infiltration periods into the same continuous period and update the start and end times. Repeat the above comparison process until the time interval between all adjacent periods is greater than the time interval threshold. Then each period is a valid infiltration period. Finally, output a list of valid infiltration periods consisting of the start and end times.

[0054] The set duration in step S324 is no less than 5 minutes and the set number of points is no less than 3.

[0055] The time interval threshold in step S325 is not less than 5 minutes.

[0056] The list of effective infiltration periods needs further screening. The screening method is as follows: the list of effective infiltration periods is matched with hourly rainfall data through time sequence matching, and only effective infiltration periods with zero rainfall in the same time period are retained to form the final list of effective infiltration periods. The final list of effective infiltration periods is used to replace the list of effective infiltration periods in step S4 for time weighting to obtain the daily drop in liquid level after screening.

[0057] The specific steps of the time-weighted method in step S4 are as follows:

[0058] S41. Compile a list of effective infiltration periods for the day and sum them up to obtain the total duration of the effective infiltration periods. and the total increase in liquid surface distance ;

[0059] S42. Calculate the average rate of change of time-weighted liquid surface distance. The calculation formula is: ,in, The time-weighted average rate of change of distance from the liquid surface is expressed in mm / h. This represents the total increase in liquid level distance for the effective infiltration periods of the day, in mm. The total duration of the list of effective infiltration periods for the day, in hours.

[0060] The process for obtaining the crop daily evapotranspiration in step S5 is as follows:

[0061] S51. Automatically aggregate the minute-level environmental meteorological time series data provided in step S2 by hour, and calculate the arithmetic mean of wind speed, air temperature, relative humidity and solar radiation within each hour to obtain the average value of hourly environmental meteorological time series data.

[0062] S52. Substitute the average hourly environmental meteorological time-series data into the FAO-56 Penman-Monteith formula to calculate the reference crop hourly evapotranspiration. The FAO-56 Penman-Monteith formula is:

[0063]

[0064] in, For reference crop hourly evapotranspiration rate, the unit is mm / h. The slope of the saturated water vapor pressure curve is expressed in kPa / °C. Net radiation, measured in MJ / m²·h, is calculated based on solar radiation. Soil heat flux, expressed in MJ / m²·h. The humidity constant is expressed in kPa / °C. Air temperature, in °C. Wind speed at a height of 2m, in m / s. and These are the saturated vapor pressure and the actual vapor pressure, respectively, in kPa. They are calculated based on measured air temperature and relative humidity.

[0065] S53, Refer to the crop hourly evapotranspiration rate Multiply by the unit time to convert to reference crop hourly evapotranspiration, calculate the reference crop hourly evapotranspiration hourly, and sum them up to obtain the reference crop daily evapotranspiration. ;

[0066] S54. Obtain daily crop evapotranspiration using the crop coefficient method. :

[0067]

[0068] in, This refers to the daily evapotranspiration of crops, expressed in mm. For reference, the daily evapotranspiration of crops is expressed in mm. The crop coefficient is a dynamic value that is combined with the actual growth period of the monitored rice.

[0069] The method for obtaining the in-situ paddy field water infiltration rate in step S7 is as follows: divide the daily infiltration amount obtained in step S6 by 1 day and perform dimension conversion to obtain the in-situ paddy field water infiltration rate in mm / d; the in-situ paddy field water infiltration rate is numerically equal to the daily infiltration amount.

[0070] The present invention has the following advantages over the prior art:

[0071] This invention integrates non-contact sensing technology and intelligent data processing algorithms to achieve automated, high temporal resolution, and low-cost monitoring of paddy field infiltration rates without damaging the in-situ soil structure. Specifically, compared to traditional excavated lysimeters, the suspended ultrasonic ranging technology avoids physical disturbance to the plow pan of the paddy field, ensuring that the monitoring data can truly reflect the natural hydrological characteristics of the field in situ, and solving the problems of boundary effects and poor representativeness of existing invasive measurements. At the same time, addressing the deficiency of simple liquid level monitoring in separating evaporation and infiltration components, the invention uses a co-located integrated micro-weather station and an embedded FAO-56 calculation model to decouple water evaporation loss in real time, eliminating monitoring uncertainties.

[0072] The monitoring method of this invention overcomes the bottleneck of existing technologies that struggle to accurately capture transient hydrological changes in complex field environments. Thanks to the system's high-frequency sampling architecture and unique effective infiltration period identification algorithm (PELT / machine learning), it can keenly identify and quantify the non-steady-state infiltration process caused by agronomic management (such as drying and re-flooding) from raw time-series data containing rainfall, irrigation, and wind and wave noise. Through field measurements in multiple paddy fields during the later tillering stage, the monitoring system successfully captured the preferential flow phenomenon caused by soil shrinkage cracks in the early stage of drying and re-flooding. The average infiltration rate after re-flooding for 5 days ranged from 1.63 to 5.30 cm / day, with an average of about 3.0 cm / day. This experimental result is consistent with the actual situation after paddy field re-flooding and reveals the instantaneous high-flux leakage characteristics that are easily overlooked by traditional low-frequency monitoring.

[0073] The in-situ paddy field water infiltration rate monitoring system provided by this invention is based on a hardware architecture of general-purpose microcontrollers and standard industrial sensors, which significantly reduces equipment costs and enables it to be deployed on a large scale in vast farmlands. This provides strong technical support for the efficient utilization of regional agricultural water resources and the prevention and control of non-point source pollution. Attached Figure Description

[0074] Appendix Figure 1 This is a block diagram of the architecture of the in-situ paddy field infiltration rate monitoring system provided by the present invention;

[0075] Appendix Figure 2 This is a schematic diagram of the paddy field deployment portion of the in-situ paddy field infiltration rate monitoring system provided by the present invention.

[0076] Appendix Figure 3 A schematic diagram of the internal core components of the sealed waterproof box provided by the present invention;

[0077] Appendix Figure 4 This is an overall flowchart of the in-situ paddy field infiltration rate monitoring method provided by the present invention.

[0078] Appendix Figure 5 A flowchart for obtaining a list of effective infiltration periods using the PELT change point detection algorithm provided by the present invention;

[0079] Appendix Figure 6 This is a schematic diagram of the original liquid level distance data collected by the in-situ paddy field infiltration rate monitoring system provided by the present invention.

[0080] Appendix Figure 7 The method of the present invention is for Figure 6 This is a schematic diagram showing the effect of marking the effective infiltration period (highlighted area) after the data is automatically identified.

[0081] The components are as follows: 1—Air temperature and humidity sensor; 2—Solar radiation sensor; 3—Wind speed sensor; 4—Support platform; 5—Tipping bucket rain gauge; 6—Solar panel; 7—Upright pole; 8—Sealed waterproof box; 9—Microcontroller; 10—SD card module; 11—Time stamp module; 12—Ultrasonic distance sensor; 13—OLED display screen; 14—4G DTU module; 100—Power supply unit; 200—In-situ data acquisition unit; 300—Data processing unit; 400—Data storage and transmission unit. Detailed Implementation

[0082] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided so that the invention will be thorough and complete, and the concept of the exemplary embodiments will be fully conveyed to those skilled in the art. The same reference numerals in the drawings denote the same or similar structures, and therefore their detailed description will be omitted.

[0083] The terms “a,” “one,” “the,” and “the” are used to indicate the existence of one or more elements / components / etc.; the terms “including” and “having” are used to indicate an open-ended meaning of inclusion and that other elements / components / etc. may exist in addition to the listed elements / components / etc.

[0084] like Figure 1 As shown: This invention provides an in-situ paddy field water infiltration rate monitoring system, which consists of two parts in terms of physical architecture: an in-situ monitoring device deployed in the field and a remotely deployed data processing terminal.

[0085] The in-situ monitoring device includes a non-contact ranging sensor suspended vertically above the paddy field water surface and a meteorological parameter acquisition module. It periodically collects water level distance and meteorological parameters from the paddy field water surface and preprocesses them into minute-level water level distance time-series data and minute-level environmental meteorological time-series data, providing this data to a data processing terminal. The data processing terminal automatically identifies effective periods in the minute-level water level distance time-series data where the distance reading of the non-contact ranging sensor steadily increases and marks them as effective infiltration periods. Based on these effective infiltration periods, it determines the daily drop in water level. The data processing terminal automatically aggregates the minute-level environmental meteorological time-series data to obtain the average hourly environmental meteorological time-series data. Based on this average, it calculates the hourly evapotranspiration of a reference crop, calculates the hourly evapotranspiration of the reference crop hourly, and accumulates it to obtain the daily evapotranspiration of the reference crop. Based on the daily evapotranspiration of the reference crop and using the crop coefficient method, it obtains the daily infiltration rate of the crop. The data processing terminal obtains the daily infiltration rate based on the difference between the daily drop in water level and the daily evapotranspiration of the crop, and then obtains the in-situ paddy field water infiltration rate based on the daily infiltration rate.

[0086] like Figure 1 , Figure 2 and Figure 3 As shown, the in-situ monitoring device includes a power supply unit 100, an in-situ data acquisition unit 200, a data processing unit 300, and a data storage and transmission unit 400.

[0087] The power supply unit 100 is used to supply power to the in-situ data acquisition unit 200, the data processing unit 300 and the data storage and transmission unit 400. The power supply unit 100 is preferably composed of a solar panel 6, a lithium battery and a charging management circuit. The solar panel 6 provides 12V DC power and supplies power to the in-situ data acquisition unit 200, the data processing unit 300 and the data storage and transmission unit 400 through a step-down module. The lithium battery is used for energy storage and nighttime power supply and to ensure that the in-situ monitoring device operates stably under no-light conditions. The solar panel 6 is connected to the lithium battery through the charging management circuit.

[0088] The in-situ data acquisition unit 200 is used to periodically and synchronously acquire the liquid level distance and meteorological parameters of the monitoring points and transmit them to the data processing unit 300. Specifically, the in-situ data acquisition unit 200 includes a non-contact liquid level distance acquisition module, a meteorological parameter acquisition module, a timestamp module 11, and an interference monitoring module. Among them, the liquid level distance acquisition module is used to periodically and synchronously acquire the liquid level distance of the monitoring points and transmit it to the data processing unit 300. The liquid level distance acquisition module is preferably an intelligent ultrasonic distance sensor 12 (HC-SR04, US-100) with integrated temperature compensation algorithm. It is connected to the ESP32 microcontroller 9 through the GPIO digital interface to output calibrated liquid level distance data to achieve accurate non-contact measurement. The meteorological parameter acquisition module is used to periodically collect real-time data on wind speed, air temperature, relative humidity, and solar radiation from monitoring points and transmit it to the data processing unit 300. The meteorological parameter acquisition module includes an air temperature and humidity sensor 1, a solar radiation sensor 2, and a wind speed sensor 3. The solar radiation sensor 3 is connected to the ESP32 microcontroller 9 via an RS485 communication converter to a UART interface. The air temperature and humidity sensor 1 is connected to the ESP32 microcontroller 9 via an I2C bus interface. The wind speed sensor 3 is connected to the ESP32 microcontroller 9 via a GPIO digital interface, thus ensuring synchronous digital acquisition of multi-source meteorological data. The timestamp module 11 provides a unified time reference for the real-time data synchronously collected by the liquid level distance acquisition module and the meteorological parameter acquisition module and supports the calculation of day numbers. The timestamp module 11 is preferably a high-precision RTC clock module, connected to the ESP32 microcontroller 9 via an I2C bus interface. The interference monitoring module uses time-series matching to retain the effective infiltration period when the rainfall is zero within the same time period. The interference monitoring module is preferably a tipping bucket rain gauge 5, which is connected to the ESP32 microcontroller 9 through the GPIO digital interface to output pulse signals. The data processing unit 300 calculates the rainfall event and rainfall intensity by counting the number of pulses.

[0089] The data processing unit 300 determines the effective period and averages the liquid level distance and meteorological parameters within the effective period, preprocessing them into minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data, which are then provided to the data processing terminal via the data storage and transmission unit 400. The data processing unit 300 is the control core of the in-situ monitoring device, preferably an ESP32 microcontroller 9 with integrated wireless communication functionality. The microcontroller 9 receives synchronously collected liquid level distance and meteorological parameters and preprocesses them into minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data. The data processing unit 300 also includes an OLED display screen 13, which displays the date, time, synchronously collected liquid level distance and meteorological parameters, and the preprocessed minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data.

[0090] The data storage and transmission unit 400 includes a local storage module and a remote transmission module. The local storage module is used to store minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data transmitted from the data processing unit 300. The local storage module is preferably an SD card module 11, which is connected to the ESP32 microcontroller 9 in the data processing unit 300 via an SPI high-speed interface for local storage of minute-level and hourly average data. The remote transmission module is used to upload the minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data transmitted from the data processing unit 300 to a cloud server. The remote transmission module is preferably a 4G DTU module 14, which is connected to the ESP32 microcontroller 9 in the data processing unit 300 via a UART serial interface for uploading data to the cloud server to achieve remote monitoring and data backup. The data processing terminal obtains the minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data by reading from the local storage module or downloading from the cloud.

[0091] like Figure 2 and Figure 3 As shown, the in-situ monitoring device includes a pole 7, a support platform 4, a solar panel 6, and a sealed waterproof box 8. The pole 7 is a metal pole with a diameter of 76mm and a length of 2.5m. Two support platforms 4 are arranged on the pole 7. One support platform 4, about 2m above the ground, is used to install an air temperature and humidity sensor 1, a solar radiation sensor 2, a wind speed sensor 3, and a tipping bucket rain gauge 5 for collecting rainfall. The other support platform 4, located at the top of the pole 7, is used to install the solar panel 6. The air temperature and humidity sensor 1 for collecting air temperature and relative humidity, the solar radiation sensor 2 for collecting solar radiation, and the wind speed sensor 3 for collecting wind speed constitute the meteorological parameter acquisition module in the in-situ data acquisition unit 200. A sealed waterproof box 8 is arranged at the lower part of the pole 7, suspended above the water surface of the paddy field. This sealed waterproof box 8 is the core control and data processing center of the system. The sealed waterproof box 8 houses a microcontroller 9 and an SD card module 10, a timestamp module 11, an ultrasonic distance sensor 12, an OLED display screen 13, and a 4G network, all connected to the microcontroller 9 via wiring. The DTU module 14, microcontroller 9, and OLED display 13 constitute the data processing unit 300 in the in-situ data acquisition unit 200. The SD card module 10 and 4G DTU module 14 constitute the data storage and transmission unit 400. The ultrasonic distance sensor 12 constitutes the liquid level distance acquisition module in the in-situ data acquisition unit 200.

[0092] The key deployment feature of this invention lies in the installation method of the liquid level distance acquisition module (ultrasonic distance sensor 12), such as... Figure 3As shown, the ultrasonic distance sensor 12 is installed at the bottom of the sealed waterproof box 8. Its probe passes through the bottom surface of the box and points vertically downwards, directly aiming at the water surface of the paddy field. The distance between the sensor and the water surface is preset to 10-20cm, so as to realize non-contact high-frequency monitoring of the distance to the water surface.

[0093] In addition, this system includes a data processing terminal, preferably a general-purpose computer, as a backend analysis platform. This terminal is configured to acquire in-situ monitoring time-series data by reading from an SD card or downloading from the cloud, and to run the following core computational logic:

[0094] Automatically identify effective periods in minute-level liquid level distance time series data where the distance reading of the non-contact ranging sensor steadily increases and mark them as effective infiltration periods, and determine the daily drop in liquid level based on the effective infiltration periods;

[0095] Automatically aggregate minute-level environmental meteorological time-series data to obtain hourly-level environmental meteorological time-series data average values, and calculate reference crop hourly evapotranspiration based on the hourly-level environmental meteorological time-series data average values. Calculate reference crop hourly evapotranspiration hourly and accumulate it to obtain reference crop daily evapotranspiration. Based on the reference crop daily evapotranspiration and according to the crop coefficient method, obtain crop daily evapotranspiration.

[0096] The daily infiltration rate is obtained by calculating the difference between the daily drop in water level and the daily evapotranspiration of the crop, and the in-situ water infiltration rate of the paddy field is obtained based on the daily infiltration rate.

[0097] See Figure 4 This is an overall flowchart of the in-situ paddy field water infiltration rate monitoring method provided by the present invention. The method employs a hybrid processing mode of in-situ data acquisition and offline calculation. The on-site data acquisition and preprocessing stages are automatically executed by the in-situ monitoring device; the core algorithm and infiltration calculation stages are executed offline by the data processing terminal after acquiring the preprocessed data provided by the in-situ monitoring device.

[0098] To improve data quality, the system executes strict preprocessing logic:

[0099] (1) The raw data all have precise timestamps provided by the RTC clock module;

[0100] (2) The system performs physical validity verification and filtering on the raw data, removes outliers that exceed the preset range (e.g., distance 0mm to 5000mm, temperature -40℃ to 80℃, humidity 0 to 100%), and performs shake-removal processing on rain pulses with an interval of less than 100ms.

[0101] (3) The system sets up a fault tolerance mechanism for the acquisition process (for example, a 30ms timeout limit for ultrasonic ranging and a 500ms timeout protection for Modbus reading).

[0102] (4) When the number of valid samples in a period exceeds 50% of the total number of samples, the system calculates the minute average of each parameter and stores it in the local storage module; if there are not enough valid samples, the period is skipped.

[0103] The steps of this in-situ paddy field water infiltration rate monitoring method are as follows:

[0104] S1. Based on the periodic sampling and multi-sensor parallel acquisition mechanism, the data processing unit 300 records in a one-minute cycle. In the first 20 seconds of each cycle, it synchronously collects the water level distance and meteorological parameters of the paddy field at a frequency of 1Hz, and then goes into sleep mode for 40 seconds.

[0105] S2. When the number of valid samples in a period exceeds 50% of the total number of samples, it is determined as a valid period. The liquid level distance and meteorological parameters in the valid period are averaged to obtain minute-level liquid level distance time series data and minute-level environmental meteorological time series data. When the number of valid samples does not exceed 50% of the total number of samples, the period is skipped.

[0106] S3. Use the PELT change point detection algorithm or machine learning algorithm to determine the effective infiltration time period in the minute-level liquid level distance time series data in step S2, and form a list of effective infiltration time periods.

[0107] S4. Based on the list of effective infiltration periods, perform time-weighted calculation to obtain the average rate of change of liquid level distance over time. Multiply the average rate of change of liquid level distance over time by 24 hours to obtain the daily drop in liquid level. The specific steps of the time-weighted method are as follows:

[0108] S41. Compile a list of effective infiltration periods for the day and sum them up to obtain the total duration of the effective infiltration periods. and the total increase in liquid surface distance ;

[0109] S42. Calculate the average rate of change of time-weighted liquid surface distance. The calculation formula is: ,in, The time-weighted average rate of change of distance from the liquid surface is expressed in mm / h. This represents the total increase in liquid level distance for the effective infiltration periods of the day, in mm. The total duration of the list of effective infiltration periods for the day, in hours;

[0110] S5. Call the minute-level environmental meteorological time series data provided in step S2, and use the FAO-56 Penman-Monteith formula and crop coefficient method to obtain the daily evapotranspiration of crops.

[0111] S6. The paddy field water balance equation can be established as follows: ,in, The daily drop in liquid level obtained in step S4, in mm. The daily evapotranspiration of the crop obtained in step S5 is expressed in mm. This refers to the daily infiltration rate, expressed in mm. Surface runoff, in mm. Rainfall amount, in mm. Irrigation volume, unit: mm;

[0112] Under the condition of "no rainfall" = 0), no irrigation ( = 0), no runoff ( Under the condition of "= 0"), a simplified water balance calculation model based on the paddy field water balance equation is obtained: It should be noted that, since the ultrasonic ranging sensor 12 is installed above the water surface, the drop in liquid level is directly reflected as an increase in the distance between the ultrasonic ranging sensor 12 and the water surface. Therefore, the daily drop in liquid level is numerically equal to the increase in the ranging distance of the ultrasonic ranging sensor 12 during the same period.

[0113] S7. Divide the daily infiltration amount obtained in step S6 by 1 day and perform dimension conversion to obtain the in-situ paddy field water infiltration rate in mm / d. That is, the in-situ paddy field water infiltration rate is numerically equal to the daily infiltration amount.

[0114] Given Figure 6 The minute-average liquid level distance time-series data inevitably contains high-frequency noise fluctuations and interfering periods characterized by decreasing distance readings. To accurately extract valid data representing the natural infiltration process from this complex data background, an interference discrimination algorithm must be performed on the minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data uploaded in step S2 before performing infiltration calculations. The goal of the interference discrimination algorithm is to automatically analyze the time-series data, identify, and label all valid infiltration periods that meet the simplified conditions (no rainfall, irrigation, or runoff). This valid infiltration period is defined as the period in which the liquid level distance S steadily increases, corresponding to... Figure 7 The blue area represents periods of stable liquid level decline caused by evapotranspiration and infiltration; disturbance periods are defined as periods where the liquid level distance S decreases or where the data fluctuates drastically. Figure 7 The white area in the middle.

[0115] like Figure 5 As shown, in the above method, the specific steps for obtaining the list of effective infiltration time periods in step S3 using the PELT change point detection algorithm are as follows:

[0116] S311, Smoothing and Noise Reduction: The minute-level liquid level distance time-series data from step S2 is processed using... The values ​​are smoothed using a rolling average to obtain smoothed and denoised minute-level liquid level distance time-series data. The calculation formula is as follows:

[0117]

[0118] in, For smoothing and noise reduction The liquid level distance value at any given time is in mm. For the data points in the minute-level liquid level distance time series data in step S2, The window size is smooth, unitless, and preferably odd, to eliminate high-frequency random noise interference;

[0119] S312. Calculate the rate of change: Take the derivative of the smoothed and denoised minute-level liquid level distance time series data to obtain the first-order rate of change. The calculation formula is as follows:

[0120]

[0121] in, for The rate of change of liquid level at any given time (i.e., the instantaneous infiltration rate), in mm / h. and These are the smoothed and denoised minute-level liquid level distance time-series data for the current and previous moments, respectively, in mm. The sampling period (1 min can be selected) is in min, and the multiplier 60 is used to convert the rate unit to a standard hourly rate.

[0122] S313, PELT (Pruned Exact Linear Time) segmentation: The PELT change point detection algorithm (which automatically identifies the moment when the mean of the rate of change changes abruptly by minimizing the penalty cost function) is used to segment the first-order rate of change. The analysis was performed on the minute-level liquid level distance time series data. Dividing the data into several independent time periods with stable rates of change, the objective function of the PELT change point detection algorithm is as follows:

[0123]

[0124] in, For the set of identified change points, The number of change points, This is the piecewise cost function, used to measure the residual fit between this segment of data and the mean model. Penalty coefficients to prevent overfitting;

[0125] S314. Time period quantification: Perform linear fitting on independent time periods with a stable change rate to obtain the precise slope Slope and duration Duration of each independent time period. The calculation formulas are as follows:

[0126]

[0127]

[0128] Among them, is the precise slope of the liquid level distance within the independent time period, with the unit of mm / h, is the number of data points within this independent time period, without a unit, is the time of the th data point, with the unit of min, is the th data point corresponding to the smoothed and noise-reduced liquid level distance value, with the unit of mm. The multiplier 60 is used to convert the rate unit from mm / min to mm / h, is the duration of the independent time period, with the unit of min, and are respectively the start time and end time of this independent time period, with the unit of min;

[0129] S315. Physical constraint screening: According to the preset physical constraint conditions: 0 < Slope ≤ 3 mm / h and Duration ≥ 5 min, screen the precise slope Slope and duration Duration of all independent time periods with a stable change rate, identify the independent time periods with a stable change rate that meet the physical constraint conditions, and mark them as pending infiltration time periods;

[0130] S316. Time period merging: Set the time gap threshold for the pending infiltration time periods and calculate the time interval between adjacent pending infiltration time periods. If the time interval between adjacent pending infiltration time periods is not greater than the time gap threshold, merge the adjacent pending infiltration time periods into the same continuous time period and update the start and end times. Repeat the above comparison process until the time intervals of all adjacent time periods are greater than the time gap threshold. Then each time period is an effective infiltration time period, and finally output an effective infiltration time period list composed of the start and end times of the effective infiltration time periods;

[0131] S317. Impurity removal screening: Match the effective infiltration time period list with the hourly rainfall data in time series, and only retain the effective infiltration time periods with zero rainfall within the same time period to form the final effective infiltration time period list, and use the final effective infiltration time period list to replace the effective infiltration time period list in step S4 for time weighting.

[0132] In the above method, the list of effective infiltration periods in step S3 can also be obtained using a machine learning algorithm. This machine learning algorithm is also executed on the data processing terminal and employs a data-driven classification method, specifically including the following steps:

[0133] S321. Feature Engineering: Extract features from the minute-level liquid level distance time-series data in step S2, and set the sliding window size to... (For example =3), for each time point Calculate the statistical feature vector within the sliding window: rolling mean. Rolling range Rolling standard deviation and the rolling slope, which reflects the instantaneous rate of change. The calculation formulas are as follows:

[0134]

[0135]

[0136]

[0137]

[0138] in, For the distance dataset within the sliding window, The rate of change of liquid level over time within the sliding window is calculated using the least squares linear regression method, and the unit is mm / h, reflecting the local instantaneous infiltration rate.

[0139] S322, Model Prediction and Label Mapping: Input the feature vector extracted in step S321 into the pre-trained random forest classification model to directly obtain the classification result of the random forest classification model, and convert the numerical result into the text label "Rising" or "Stable" according to the label mapping table.

[0140] S323, Spike Filtering: In order to eliminate false alarms, spike detection is performed on all data points initially marked as "Rising". If the slope of the current data point exceeds the spike threshold (preferably 7.0 mm / h) and the slope of the immediately following data point has fallen back to below the normal threshold (preferably 3.0 mm / h), then the point is determined to be an interference signal (marked as "Filtered Spike") and is removed.

[0141] S324, Time Period Merging: After aggregating the peak filtering, the remaining data points with the text label "Rising" are merged into continuous time periods, and time periods with a duration shorter than the set duration (5 minutes) or containing fewer than the set number of points (3) are removed to obtain the undetermined infiltration time periods;

[0142] S325. Set the time interval threshold (5min) for the undetermined infiltration period and calculate the time interval between adjacent undetermined infiltration periods. If the time interval between adjacent undetermined infiltration periods is not greater than the time interval threshold, merge the adjacent undetermined infiltration periods into the same continuous period and update the start and end times. Repeat the above comparison process until the time interval between all adjacent periods is greater than the time interval threshold. Then each period is a valid infiltration period. Finally, output a list of valid infiltration periods consisting of the start and end times.

[0143] S326. The effective infiltration period list is matched with the hourly rainfall data through time sequence matching, and only the effective infiltration periods with zero rainfall in the same period are retained to form the final effective infiltration period list. The final effective infiltration period list is used to replace the effective infiltration period list in step S4 for time weighting.

[0144] In the above method, the process of obtaining the daily evapotranspiration of crops is as follows:

[0145] S51. Automatically aggregate the minute-level environmental meteorological time series data provided in step S2 by hour, and calculate the arithmetic mean of wind speed, air temperature, relative humidity and solar radiation within each hour to obtain the average value of hourly environmental meteorological time series data.

[0146] S52. Substitute the average hourly environmental meteorological time-series data into the FAO-56 Penman-Monteith formula to calculate the reference crop hourly evapotranspiration. The FAO-56 Penman-Monteith formula is:

[0147]

[0148] in, For reference crop hourly evapotranspiration rate, the unit is mm / h. The slope of the saturated water vapor pressure curve is expressed in kPa / °C. Net radiation, measured in MJ / m²·h, is calculated based on solar radiation. Soil heat flux, expressed in MJ / m²·h. The humidity constant is expressed in kPa / °C. Air temperature, in °C. Wind speed at a height of 2m, in m / s. and These are the saturated vapor pressure and the actual vapor pressure, respectively, in kPa. They are calculated based on measured air temperature and relative humidity.

[0149] S53, Refer to the crop hourly evapotranspiration rate Multiply by the unit time to convert to reference crop hourly evapotranspiration, calculate the reference crop hourly evapotranspiration hourly, and sum them up to obtain the reference crop daily evapotranspiration. ;

[0150] S54. Obtain daily crop evapotranspiration using the crop coefficient method. :

[0151]

[0152] in, This refers to the daily evapotranspiration of crops, expressed in mm. For reference, the daily evapotranspiration of crops is expressed in mm. This is a crop coefficient, without units. The determination of the value is a well-known technique in the field. It is preferable to dynamically select the value based on the standard of FAO-56 document issued by the Food and Agriculture Organization of the United Nations (FAO) and in combination with the actual growth stage of the monitored rice (tillering stage, booting stage, maturity stage, etc.).

[0153] Based on the data provided in step S51, the calculation methods for the required parameters in step S52 are as follows.

[0154] The slope of the saturated vapor pressure curve was calculated based on the measured hourly average air temperature. The calculation formula is:

[0155]

[0156] in, The slope of the saturated water vapor pressure curve is expressed in kPa / ℃. The average air temperature is expressed in °C (°C) for one hour.

[0157] Calculation of ecliptic constants based on station altitude The calculation formula is:

[0158]

[0159]

[0160] in, This is the wet / dry constant, with units of kPa / ℃. The pressure at the location of the device is atmospheric pressure, expressed in kPa. The altitude of the device's location is in meters (m).

[0161] Net radiation calculated based on measured solar radiation meteorological data The calculation formula is:

[0162]

[0163] in, Net radiation, in MJ / m²·h The canopy albedo is dimensionless and has a value of 0.23. Total solar radiation, expressed in MJ / m²·h. The Stefan-Boltzmann constant is expressed in MJ·K. -4 ·m -2 ·h -1 The value is 2.043 × 10 -10 , The hourly average absolute temperature, expressed in Kelvin (K). This is the actual water vapor pressure, in kPa. Clear-sky radiation, measured in MJ / m²·h.

[0164] Calculation of clear-sky radiation based on device altitude and astronomical radiation The calculation formula is:

[0165]

[0166] in, Clear-sky radiation, measured in MJ / m²·h. The altitude of the device's location is in meters (m). It represents astronomical radiation, measured in MJ / m²·h.

[0167] Calculation of astronomical radiation for an hourly period based on the solar constant and Earth-Sun position. The calculation formula is:

[0168]

[0169] in, Astronomical radiation, measured in MJ / m²·h. The solar constant, expressed in MJ / m²·min, has a value of 0.082. It is inversely proportional to the Earth-Sun distance and is dimensionless. Solar declination, measured in rad. The latitude of the device's location is expressed in rad. and These are the solar hour angles at the beginning and end of the hourly time period, respectively, in rad.

[0170] Calculate the inverse ratio of Earth-Sun distance based on accumulated days over years. Solar declination and solar hour angle The calculation formula is:

[0171]

[0172]

[0173]

[0174] in, It is inversely proportional to the Earth-Sun distance and is dimensionless. Solar declination, measured in rad. It refers to the day of the year, ranging from 1 to 366, and is dimensionless. Solar hour angle, unit: rad. To calculate the midpoint time of the time period, in hours. The longitude of the device's location is in degrees. The central longitude of the time zone is given in degrees.

[0175] Calculation of soil heat flux based on the positive and negative states of net radiation The calculation formula is:

[0176]

[0177] in, Soil heat flux, expressed in MJ / m²·h. Net radiation, in MJ / m²·h Soil heat flux coefficient, dimensionless;

[0178] Soil heat flux coefficient The rule for determining the value is: when When ≥0, it corresponds to the daytime period. The value is 0.1; when When the value is less than 0, it corresponds to the nighttime period. The value is 0.5.

[0179] Calculation of saturated vapor pressure based on measured air temperature The calculation formula is:

[0180]

[0181] in, The pressure is the saturated vapor pressure, expressed in kPa. Air temperature, unit: °C;

[0182] Calculate the actual water vapor pressure based on measured relative humidity and saturated water vapor pressure. The calculation formula is:

[0183]

[0184] in, This is the actual water vapor pressure, in kPa. The pressure is the saturated vapor pressure, expressed in kPa. Relative humidity, in percent.

[0185] The monitoring method of this invention overcomes the bottleneck of existing technologies that struggle to accurately capture transient hydrological changes in complex field environments. Thanks to the system's high-frequency sampling architecture and unique effective infiltration period identification algorithm (PELT / machine learning), it can keenly identify and quantify the non-steady-state infiltration process caused by agronomic management (such as drying and re-flooding) from raw time-series data containing rainfall, irrigation, and wind and wave noise. Through field measurements in multiple paddy fields during the later tillering stage, the monitoring system successfully captured the preferential flow phenomenon caused by soil shrinkage cracks in the early stage of drying and re-flooding. The average infiltration rate after re-flooding for 5 days ranged from 1.63 to 5.30 cm / day, with an average of about 3.0 cm / day. This experimental result is consistent with the actual situation after paddy field re-flooding and reveals the instantaneous high-flux leakage characteristics that are easily overlooked by traditional low-frequency monitoring.

[0186] In this embodiment of the invention, the term "multiple" refers to two or more, unless otherwise explicitly defined. The terms "install," "connect," and "fix" should be interpreted broadly. For example, "connect" can mean a fixed connection, a detachable connection, or an integral connection. Those skilled in the art can understand the specific meaning of the above terms in this embodiment of the invention based on the specific circumstances.

[0187] In the description of the embodiments of the present invention, it should be understood that the terms "upper" and "lower" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or unit referred to must have a specific orientation or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of the present invention.

[0188] In the description of this specification, the terms "an embodiment," "a preferred embodiment," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0189] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention. Technologies not covered in this invention can be implemented using existing technologies.

Claims

1. An in-situ paddy field water infiltration rate monitoring system, characterized in that: The monitoring system includes in-situ monitoring devices deployed in the paddy fields, as well as data processing terminals; The in-situ monitoring device includes a non-contact ranging sensor suspended vertically above the paddy field water surface and a meteorological parameter acquisition module. It synchronously collects the liquid level distance and meteorological parameters of the paddy field water surface on a periodic basis and preprocesses them on a periodic basis into minute-level liquid level distance time series data and minute-level environmental meteorological time series data, which are then provided to the data processing terminal. The data processing terminal automatically identifies effective periods in the minute-level liquid level distance time series data where the distance reading of the non-contact ranging sensor steadily increases and marks them as effective infiltration periods, and determines the daily drop in liquid level based on the effective infiltration periods; The data processing terminal automatically aggregates minute-level environmental meteorological time-series data to obtain the average value of hourly environmental meteorological time-series data, and calculates the hourly evapotranspiration of reference crops based on the average value of hourly environmental meteorological time-series data. The hourly evapotranspiration of reference crops is calculated hourly and accumulated to obtain the daily evapotranspiration of reference crops. Based on the daily evapotranspiration of reference crops and according to the crop coefficient method, the daily evapotranspiration of crops is obtained. The data processing terminal obtains the daily infiltration rate based on the difference between the daily drop in liquid level and the daily evapotranspiration of crops, and obtains the in-situ water infiltration rate of paddy fields based on the daily infiltration rate.

2. The in-situ paddy field water infiltration rate monitoring system according to claim 1, characterized in that: The in-situ monitoring device includes a power supply unit (100), an in-situ data acquisition unit (200), a data processing unit (300), and a data storage and transmission unit (400). The power supply unit (100) is used to supply power to the in-situ data acquisition unit (200), the data processing unit (300), and the data storage and transmission unit (400). The in-situ data acquisition unit (200) is used to collect the liquid level distance and meteorological parameters of the monitoring point in real time according to the cycle and transmit them to the data processing unit (300). The data processing unit (300) determines the effective period and averages and preprocesses the liquid level distance and meteorological parameters within the effective period into minute-level liquid level distance time series data and minute-level environmental meteorological time series data, and provides them to the data processing terminal through the data storage and transmission unit (400).

3. The in-situ paddy field water infiltration rate monitoring system according to claim 2, characterized in that: The in-situ data acquisition unit (200) includes a non-contact liquid level distance acquisition module, a meteorological parameter acquisition module, and a timestamp module (11). The liquid level distance acquisition module is used to collect the liquid level distance of the monitoring point in real time according to the cycle and transmit it to the data processing unit (300). The meteorological parameter acquisition module is used to collect the wind speed, air temperature, relative humidity and solar radiation of the monitoring point in real time according to the cycle and transmit it to the data processing unit (300). The timestamp module (11) provides a unified time reference for the real-time data collected synchronously by the liquid level distance acquisition module and the meteorological parameter acquisition module and supports the calculation of the day sequence number.

4. The in-situ paddy field water infiltration rate monitoring system according to claim 2, characterized in that: The in-situ data acquisition unit (200) includes an interference monitoring module, which uses time-series matching to retain effective infiltration periods when rainfall is zero within the same time period.

5. The in-situ paddy field water infiltration rate monitoring system according to claim 2, characterized in that: The power supply unit (100) includes a solar panel (6), a lithium battery and a charging management circuit. The solar panel (6) provides 12V DC power and supplies power to the in-situ data acquisition unit (200), data processing unit (300) and data storage and transmission unit (400) through a step-down module. The lithium battery is used for energy storage and nighttime power supply and ensures stable operation of the in-situ monitoring device under no-light conditions. The solar panel (6) is connected to the lithium battery through the charging management circuit. The data processing unit (300) includes a microcontroller (9) and an OLED display (13). The microcontroller (9) is used to receive synchronously collected liquid level distance and meteorological parameters and preprocess them into minute-level liquid level distance time series data and minute-level environmental meteorological time series data. The OLED display (13) is used to display the date, time, synchronously collected liquid level distance and meteorological parameters, preprocessed minute-level liquid level distance time series data and minute-level environmental meteorological time series data. The data storage and transmission unit (400) includes a local storage module and a remote transmission module. The local storage module is used to store minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data transmitted from the data processing unit (300). The remote transmission module is used to upload minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data transmitted from the data processing unit (300) to the cloud server. The data processing terminal obtains minute-level liquid level distance time-series data and minute-level environmental meteorological time-series data by reading the local storage module or downloading from the cloud.

6. The in-situ paddy field water infiltration rate monitoring system according to any one of claims 1-5, characterized in that: The in-situ monitoring device includes a pole (7), a support platform (4), a solar panel (6), and a sealed waterproof box (8). Two support platforms (4) are arranged on the pole (7). One support platform (4) is used to install an air temperature and humidity sensor (1), a solar radiation sensor (2), and a wind speed sensor (3), while the other support platform (4) is used to install a solar panel (6). The air temperature and humidity sensor (1) for collecting air temperature and relative humidity, the solar radiation sensor (2) for collecting solar radiation, and the wind speed sensor (3) for collecting wind speed constitute the meteorological parameter acquisition module in the in-situ data acquisition unit (200). A sealed waterproof box (8) is arranged at the lower part of the pole (7) and suspended above the water surface of the paddy field. The sealed waterproof box (8) contains a microcontroller (9) and an SD card module (10), a timestamp module (11), an ultrasonic distance sensor (12), an OLED display (13), and a 4G network connected to the microcontroller (9) via wires. The DTU module (14), microcontroller (9) and OLED display (13) constitute the data processing unit (300) in the in-situ data acquisition unit (200). The SD card module (10) and 4G DTU module (14) constitute the data storage and transmission unit (400). The ultrasonic distance sensor (12) constitutes the liquid level distance acquisition module in the in-situ data acquisition unit (200) and is installed at the bottom of the sealed waterproof box (8). The probe of the ultrasonic distance sensor (12) passes through the bottom surface of the sealed waterproof box (8) and is vertically downward aligned with the paddy field water surface. The distance between the probe of the ultrasonic distance sensor (12) and the paddy field water surface is preset to 10-20cm.

7. A method for monitoring the infiltration rate of water in paddy fields in situ, characterized in that: The steps of this monitoring method are as follows: S1. Collect the water level distance and meteorological parameters of the paddy field water surface synchronously on a periodic basis; S2. Determine the effective period, average the liquid level distance and meteorological parameters within the effective period, and obtain minute-level liquid level distance time series data and minute-level environmental meteorological time series data. S3. Use the PELT change point detection algorithm or machine learning algorithm to determine the effective infiltration time period in the minute-level liquid level distance time series data in step S2, and form a list of effective infiltration time periods. S4. Based on the list of effective infiltration periods, perform time weighting to obtain the average rate of change of liquid surface distance over time. Multiply the average rate of change of liquid surface distance over time by 24 hours to obtain the daily drop in liquid surface. S5. Call the minute-level environmental meteorological time series data provided in step S2, and use the FAO-56 Penman-Monteith formula and crop coefficient method to obtain the daily evapotranspiration of crops. S6. A simplified water balance calculation model based on the paddy field water balance equation: , This refers to the daily infiltration rate, expressed in mm. The daily drop in liquid level obtained in step S4, in mm. The daily evapotranspiration of the crop obtained in step S5 is expressed in mm. S7. Obtain the in-situ water infiltration rate of paddy fields based on daily infiltration volume.

8. The method for monitoring the in-situ water infiltration rate in paddy fields according to claim 7, characterized in that: The effective period in step S2 refers to the period in which the number of effective samples exceeds 50% of the total number of samples. If the number of effective samples does not exceed 50% of the total number of samples, the period is skipped.

9. The method for monitoring the in-situ water infiltration rate in paddy fields according to claim 7, characterized in that: The specific steps for obtaining the list of effective infiltration time periods in step S3 using the PELT change point detection algorithm are as follows: S311, Apply the minute-level liquid level distance time series data from step S2. The values ​​are smoothed using a rolling average to obtain smoothed and denoised minute-level liquid level distance time-series data. The calculation formula is as follows: in, For smoothing and noise reduction The liquid level distance value at any given time is in mm. For the data points in the minute-level liquid level distance time series data in step S2, To smooth the window size, used to eliminate high-frequency random noise interference; S312. Take the derivative of the smoothed and denoised minute-level liquid level distance time series data to obtain the first-order rate of change. The calculation formula is as follows: in, for The rate of change of liquid level at any given time (i.e., the instantaneous infiltration rate), in mm / h. and These are the smoothed and denoised minute-level liquid level distance time-series data for the current and previous moments, respectively, in mm. The sampling period is in minutes, and the multiplier of 60 is used to convert the rate unit to a standard hourly rate. S313. Use the PELT change point detection algorithm to analyze the first-order rate of change. The analysis was performed on the minute-level liquid level distance time series data. Dividing the data into several independent time periods with stable rates of change, the objective function of the PELT change point detection algorithm is as follows: in, For the set of identified change points, The number of change points, This is the piecewise cost function, used to measure the residual fit between this segment of data and the mean model. Penalty coefficients to prevent overfitting; S314. Perform linear fitting on independent time periods with stable rates of change to obtain the accurate slope and duration of each independent time period. S315. Based on the preset physical constraints: 0 < Slope≤3mm / h and Duration≥5min, the precise slope and duration of all independent time periods with stable rates of change are screened, and the independent time periods with stable rates of change that meet the physical constraints are identified and marked as undetermined infiltration time periods. S316. Set the time interval threshold for the undetermined infiltration period and calculate the time interval between adjacent undetermined infiltration periods. If the time interval between adjacent undetermined infiltration periods is not greater than the time interval threshold, merge the adjacent undetermined infiltration periods into the same continuous period and update the start and end times. Repeat the above comparison process until the time interval between all adjacent periods is greater than the time interval threshold. Then each period is a valid infiltration period. Finally, output a list of valid infiltration periods consisting of the start and end times.

10. The method for monitoring the in-situ water infiltration rate in paddy fields according to claim 7, characterized in that: The specific steps for obtaining the list of effective infiltration time periods in step S3 using a machine learning algorithm are as follows: S321. Feature Engineering: Extract features from the minute-level liquid level distance time-series data in step S2, and set the sliding window size to... For each point in time Calculate the statistical feature vector within the sliding window: rolling mean. Rolling range Rolling standard deviation and the rolling slope, which reflects the instantaneous rate of change. The calculation formulas are as follows: in, For the distance dataset within the sliding window, The rate of change of liquid level over time within the sliding window is calculated using the least squares linear regression method, and the unit is mm / h, reflecting the local instantaneous infiltration rate. S322, Model Prediction and Label Mapping: Input the feature vector extracted in step S321 into the pre-trained random forest classification model to directly obtain the classification result of the random forest classification model, and convert the numerical result into the text label "Rising" or "Stable" according to the label mapping table. S323, Peak Filtering: Perform peak detection on all data points initially marked as "Rising". If the slope of the current data point exceeds the peak threshold and the slope of the immediately following data point has fallen back to below the normal threshold, then the point is determined to be an interference signal and is removed. S324, Time Period Merging: After aggregating the data points with the text label "Rising" remaining after peak filtering, merge them into continuous time periods, and remove time periods with a duration shorter than the set duration or a number of points lower than the set number of points to obtain the undetermined infiltration time period; S325. Set the time interval threshold for the undetermined infiltration period and calculate the time interval between adjacent undetermined infiltration periods. If the time interval between adjacent undetermined infiltration periods is not greater than the time interval threshold, merge the adjacent undetermined infiltration periods into the same continuous period and update the start and end times. Repeat the above comparison process until the time interval between all adjacent periods is greater than the time interval threshold. Then each period is a valid infiltration period. Finally, output a list of valid infiltration periods consisting of the start and end times.

11. The method for monitoring the in-situ water infiltration rate in paddy fields according to any one of claims 7, 9, and 10, characterized in that: The list of effective infiltration periods needs further screening. The screening method is as follows: the list of effective infiltration periods is matched with hourly rainfall data through time sequence matching, and only effective infiltration periods with zero rainfall in the same time period are retained to form the final list of effective infiltration periods. The final list of effective infiltration periods is used to replace the list of effective infiltration periods in step S4 for time weighting to obtain the daily drop in liquid level after screening.

12. The method for monitoring the in-situ water infiltration rate in paddy fields according to claim 7, characterized in that: The specific steps of the time-weighted method in step S4 are as follows: S41. Compile a list of effective infiltration periods for the day and sum them up to obtain the total duration of the effective infiltration periods. and the total increase in liquid surface distance ; S42. Calculate the average rate of change of time-weighted liquid surface distance. The calculation formula is: ,in, The time-weighted average rate of change of distance from the liquid surface is expressed in mm / h. This represents the total increase in liquid level distance for the effective infiltration periods of the day, in mm. The total duration of the list of effective infiltration periods for the day, in hours.

13. The method for monitoring the in-situ water infiltration rate in paddy fields according to claim 7, characterized in that: The process for obtaining the crop daily evapotranspiration in step S5 is as follows: S51. Automatically aggregate the minute-level environmental meteorological time series data provided in step S2 by hour, and calculate the arithmetic mean of wind speed, air temperature, relative humidity and solar radiation within each hour to obtain the average value of hourly environmental meteorological time series data. S52. Substitute the average hourly environmental meteorological time-series data into the FAO-56 Penman-Monteith formula to calculate the reference crop hourly evapotranspiration. The FAO-56 Penman-Monteith formula is: in, For reference crop hourly evapotranspiration rate, the unit is mm / h. The slope of the saturated water vapor pressure curve is expressed in kPa / °C. Net radiation, measured in MJ / m²·h, is calculated based on solar radiation. Soil heat flux, expressed in MJ / m²·h. The humidity constant is expressed in kPa / °C. Air temperature, in °C. Wind speed at a height of 2m, in m / s. and These are the saturated vapor pressure and the actual vapor pressure, respectively, in kPa. They are calculated based on measured air temperature and relative humidity. S53, Refer to the crop hourly evapotranspiration rate Multiply by the unit time to convert to the hourly evapotranspiration of the reference crop, calculate the hourly evapotranspiration of the reference crop hourly, and sum them up to obtain the daily evapotranspiration of the reference crop. ; S54. Obtain daily crop evapotranspiration using the crop coefficient method. : in, This refers to the daily evapotranspiration of crops, expressed in mm. For reference, the daily evapotranspiration of crops is expressed in mm. The crop coefficient is a dynamic value that is combined with the actual growth period of the monitored rice.

14. The method for monitoring the in-situ water infiltration rate in paddy fields according to claim 7, characterized in that: The method for obtaining the in-situ paddy field water infiltration rate in step S7 is as follows: divide the daily infiltration amount obtained in step S6 by 1 day and perform dimension conversion to obtain the in-situ paddy field water infiltration rate in mm / d; the in-situ paddy field water infiltration rate is numerically equal to the daily infiltration amount.

Images