Method for predicting soil moisture measurement and supplemental irrigation of field crops by fusing cosmic ray neutrons and remote sensing

Abstract

The application discloses a kind of field crop soil moisture measurement supplemental irrigation prediction methods of fusing cosmic ray neutron and remote sensing, comprising the following steps: determining crop species and growth stage;Using vehicle-mounted near-surface cosmic ray neutron instrument to obtain soil moisture measurement value, fusion remote sensing data is carried out spatial calibration and verification, and real-time value of soil moisture is obtained;Meteorological data, crop data and soil data are combined to build crop growth model, and the model is optimized using real-time value of soil moisture, and output soil moisture prediction value;Remote sensing information is obtained by unmanned aerial vehicle, and accumulated temperature inversion model is constructed, and real-time value of accumulated temperature is estimated;Real-time value of accumulated temperature is combined with meteorological data, and accumulated temperature prediction value is calculated, growth stage is predicted, difference analysis is carried out combining soil moisture prediction value and crop water requirement prediction value, and subsequent supplemental irrigation amount prediction value is obtained, and the application realizes the effective coupling of multiple data, provides reliable intelligent decision-making tool for implementing on-demand precision irrigation and improving agricultural water resource utilization efficiency.

Description

Technical Field

[0001] This invention relates to the field of agricultural irrigation technology, and more specifically, to a method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutrons and remote sensing. Background Technology

[0002] Field crops typically include major food crops such as wheat, corn, soybeans, and rice. Compared to small-scale family farming, field crops have a wide planting area, high yields, and a high level of mechanization, forming the foundation of agricultural production. Field crops play a vital role in the food industry, not only providing abundant food resources globally but also occupying an important position in economic and social development. Their production efficiency is directly related to national food security and farmers' livelihoods; therefore, improving the production efficiency and sustainable development of field crops has become an important direction for global agricultural science and technology research.

[0003] In the production management of field crops, water management is one of the key factors affecting crop growth and yield. Insufficient or excessive soil moisture can negatively impact crop growth. Therefore, rational soil moisture monitoring and supplemental irrigation technology is of great significance in field crop production. This technology involves monitoring soil moisture conditions and combining this with the crop's water requirements to adjust irrigation measures in a timely manner, thereby achieving precise and efficient water management. Traditional soil moisture monitoring and supplemental irrigation technologies mostly rely on manual measurement or simple soil moisture sensing tools. This method is not only labor-intensive and time-consuming, but also greatly affected by human factors, failing to accurately reflect changes in water availability during crop growth. Therefore, a technical solution is needed to address these issues. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art, achieve effective coupling of multi-source data, realize the transformation from static monitoring to dynamic prediction, and from single water information to synergistic perception of water and crop growth, provide a reliable intelligent decision-making tool for implementing on-demand precision irrigation and improving the efficiency of agricultural water resource utilization, and provide a method for predicting soil moisture supplementation irrigation for field crops that integrates cosmic ray neutrons and remote sensing.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] This invention discloses a method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing data, comprising the following steps:

[0007] S1. Determine the target crop species and their key water-required growth stages;

[0008] S2. Use a vehicle-mounted near-ground cosmic ray neutron instrument to obtain field soil moisture content measurements, and integrate UAV remote sensing data to perform spatial calibration and verification to obtain real-time soil moisture content values.

[0009] S3. Combine local meteorological data, crop growth monitoring data and soil physicochemical property data to construct a crop growth model, and use the real-time soil moisture content value in step S2 to assimilate and optimize the model so that the model has the ability to output predicted soil moisture content values.

[0010] S4. Acquire multi-dimensional remote sensing information of crops synchronously through the UAV platform, construct a crop accumulated temperature inversion model based on machine learning algorithm, and estimate the real-time value of crop accumulated temperature;

[0011] S5. Combine the real-time value of crop accumulated temperature with meteorological forecast data to calculate the predicted value of crop accumulated temperature, predict the crop growth stage, and perform difference analysis by combining the predicted value of soil moisture content in step S3 with the predicted value of crop water demand at each stage to obtain the predicted value of supplementary irrigation in the field for subsequent periods.

[0012] Furthermore, in step S2, the vehicle-mounted near-ground cosmic ray neutron spectrometer moves and observes along a pre-set transect route covering different soil textures and growth conditions once a week. Each observation acquires the volumetric water content of the surface soil within a 300-meter diameter area and a depth of approximately 0–30 cm. Simultaneously with each observation by the vehicle-mounted near-ground cosmic ray neutron spectrometer, a multi-rotor UAV platform is operated to conduct aerial photography over the observation transect to acquire canopy reflectivity and canopy temperature data with a spatial resolution better than 5 cm.

[0013] Furthermore, in step S2, the precise soil moisture content of multiple positioning points is simultaneously obtained using the time domain reflectometer method in the ground test plot for verification; the NDWI spatial distribution map generated by the UAV is spatially matched with the field soil moisture content measurement value; and a calibration model is established using the support vector machine regression algorithm, with the normalized differential moisture index NDWI, normalized vegetation index NDVI and canopy temperature as input features and the field soil moisture content measurement value as the target value.

[0014] Furthermore, by calibrating the model, high-resolution remote sensing information from UAVs is used as spatial weights to perform spatial downscaling calibration on the soil moisture content measurements at single points in the field, generating a real-time distribution map of soil moisture content with high spatial resolution for the entire field, which serves as the real-time value of soil moisture content.

[0015] Furthermore, step S3 includes:

[0016] S3.1 Basic Data and Model Preparation: Collect daily meteorological data for the past 5 years and real-time meteorological data for the current year for the target area; collect soil profile data for typical fields; select the AquaCrop crop growth model as the basic model, and initialize the model crop parameters according to the characteristics of crop varieties;

[0017] S3.2 Model Initialization and Assimilation: Taking the greening period as the starting point of the simulation, soil parameters and daily meteorological data are input, and the initial soil moisture content is set; the ensemble Kalman filter data assimilation algorithm is used to assimilate the real-time soil moisture content obtained through S2 into the soil water molecule module of the AquaCrop model as observed values.

[0018] S3.3 Prediction Output: After each assimilation is completed, the model is driven forward and combined with the meteorological forecast data for the next 7 days to output the predicted value of soil moisture content for each day of the next week.

[0019] Furthermore, the output levels of the predicted soil moisture content include: the volumetric moisture content of the 0-30cm surface layer and the average volumetric moisture content of the root zone, and simultaneously output the soil moisture deficit or the ratio of relative field capacity for water stress identification and irrigation decision threshold determination.

[0020] Furthermore, step S4 includes:

[0021] S4.1 Continuous Data Acquisition: Starting from the crop greening period, use UAV multi-spectral and thermal infrared sensors to conduct full-coverage aerial surveys every 7-10 days to continuously acquire the photochemical reflectance index PRI, which reflects the crop's water stress and photosynthetic efficiency.

[0022] S4.2 Ground surface temperature baseline: Automatic weather stations are set up in the field to record daily air temperatures and calculate the conventional effective accumulated temperature (GDD) based on air temperature. The effective accumulated temperature (GDD) can be calculated according to the following formula:

[0023]

[0024] in: The highest temperature of the day (°C). The lowest temperature of the day (°C). The baseline temperature for crops is set at 10℃.

[0025] S4.3 Model Construction: Using the time series and photochemical reflectance index PRI obtained by UAV as the main features, and the effective accumulated temperature GDD obtained by synchronous calculation as the baseline label, the random forest regression algorithm is used for training to construct the canopy effective accumulated temperature inversion model.

[0026] S4.4 Real-time accumulated temperature output: For each UAV observation, the processed canopy temperature distribution map and PRI distribution map are input into the trained canopy effective accumulated temperature inversion model, which can then invert the spatial distribution of the real-time crop accumulated temperature value of the entire field at the current moment.

[0027] Furthermore, step S5 includes:

[0028] S5.1 Accumulated Temperature and Growth Stage Prediction: Based on the real-time crop accumulated temperature value output by S4, combined with the temperature forecast for the next 7 days, the predicted crop accumulated temperature value for the next 7 days is calculated; based on the crop accumulated temperature and growth stage comparison table and the predicted crop accumulated temperature value, it is determined which specific growth stage the crop will enter or be in within the next week.

[0029] S5.2 Water Demand Forecast: Based on the future growth stage determined in S5.1, the crop water demand for the next 7 days is calculated using the Penman-Montes formula;

[0030] S5.3 Irrigation Quantity Calculation: Calculate the supplementary irrigation amount based on crop water requirement forecast and soil moisture content forecast;

[0031] S5.4 Output Irrigation Prescription Map: Based on the calculated supplementary irrigation amount, an irrigation prescription map is generated to guide field irrigation decisions; the irrigation prescription map will display the supplementary irrigation amount for each plot or zone, and the water demand of different areas can be variably adjusted according to the irrigation amount.

[0032] Furthermore, step S5.3 includes:

[0033] S5.3.1 Soil Water Storage Calculation: Calculate the predicted soil water content for the same period in the future, as output in step S3. Effective soil water storage capacity converted into root zone The calculation formula is:

[0034]

[0035] S5.3.2 Calculation of theoretical total water consumption: based on crop water requirements Calculate the theoretical total water consumption for the coming week;

[0036] S5.3.3 Irrigation Quantity Forecast: Calculate the supplementary irrigation quantity for the next week based on the following formula. :

[0037]

[0038] in, This is the forecast of effective rainfall for the next week. To improve irrigation water use efficiency, This represents the amount of water that can be consumed from the soil's effective water storage capacity within the next week, with max() taking the larger value.

[0039] The beneficial effects of this invention are:

[0040] This invention deeply integrates the advantages of vehicle-mounted cosmic ray neutron method for large-scale soil moisture detection, the fine spatial information acquisition capability of UAV remote sensing, and the mechanistic simulation function of crop growth model. Through data assimilation and machine learning technology, it achieves effective coupling of multi-dimensional data, realizing a leap from static monitoring to dynamic prediction, and from single moisture information to the synergistic perception of water and crop growth. It provides an advanced and reliable intelligent decision-making tool for implementing on-demand precision irrigation and improving the efficiency of agricultural water resource utilization. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of one embodiment. Detailed Implementation

[0042] The technical solutions in this embodiment will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] This embodiment takes a large-scale winter wheat planting area in the North China Plain as an example to illustrate the implementation process of the method of the present invention from the greening stage to the late grain filling stage.

[0044] S1: Determine the crop type and its main water-required growth stage.

[0045] The target crop in this embodiment is winter wheat grown in the North China Plain. Its main water-required growth stages are, in order: greening stage, jointing stage, heading and flowering stage, early grain filling stage, and late grain filling stage. Among them, the jointing stage, heading and flowering stage, and early grain filling stage are considered water-sensitive critical water-required stages, which require key monitoring and prediction. The typical time windows corresponding to each stage are as follows: the greening stage is from late February to early March (approximately 55-70 days after sowing); the jointing stage is from mid-March (approximately 70-85 days after sowing); the jointing stage is from late March to early April (approximately 85-100 days after sowing); the booting stage is from mid-April (approximately 100-110 days after sowing); the heading and flowering stage is from late April to early May (approximately 110-125 days after sowing); the early grain-filling stage is from early May to mid-May (approximately 125-140 days after sowing); and the late grain-filling stage is from late May to early June (approximately 140-155 days after sowing).

[0046] S2: The field soil moisture content was measured using a vehicle-mounted near-ground cosmic ray neutron instrument, and spatially calibrated and verified by integrating UAV remote sensing data to obtain real-time soil moisture content values.

[0047] S2.1: Data Acquisition

[0048] After the winter wheat turns green, a vehicle-mounted cosmic ray neutron spectrometer (CRNP) is used to conduct mobile observations approximately once a week along a pre-designed transect route covering different soil textures and growth conditions. Each observation acquires measurements of the volumetric water content of the surface soil within a diameter of approximately 300 meters and a depth of approximately 0–30 cm. .

[0049] The specific parameters are as follows:

[0050]

[0051] S2.2: Synchronous Remote Sensing Monitoring

[0052] During each vehicle-mounted CRNP observation, a multi-rotor drone platform (such as the DJI M300RTK) equipped with a multi / hyperspectral camera (such as the RedEdge-MX) and a thermal infrared camera (such as the ALTUM-PT) is operated to take aerial photos over the observation transect to obtain canopy multi / hyperspectral reflectance (for calculating normalized differential moisture index NDWI, normalized differential vegetation index NDVI, etc.) and canopy temperature data with a spatial resolution better than 5 cm.

[0053]

[0054] S2.3: Data Calibration and Fusion

[0055] Accurate soil moisture content at multiple locations was simultaneously acquired using Time Domain Reflectometry (TDR) in the ground test plot for validation. The NDWI spatial distribution map generated by the UAV was spatially matched with the concurrent CRNP measurements. A calibration model was established using Support Vector Machine Regression (SVR) algorithm, with NDWI, NDVI, and canopy temperature as input features and CRNP measurements as the target value. This model, using high-resolution UAV remote sensing information as spatial weights, spatially downscaled and calibrated the single-point CRNP measurements, generating a high spatial resolution (5-meter grid) real-time soil moisture distribution map for the entire field (approximately 500 acres), serving as the real-time soil moisture value. The specific steps are broken down as follows:

[0056] S2.3.1: Ground calibration of volumetric water content: Within a 100-acre field, 3-5 TDR (Temperature Reduction) points are set up according to soil texture / topography / growth status. At each point, the volumetric water content of three layers (0-10cm, 10-20cm, and 20-30cm) is measured, and the weighted average is obtained. Simultaneously collect a small number of drying samples (6-10 samples each time, with increased density during critical periods) for TDR and volumetric moisture content consistency verification. The time difference between TDR measurement and vehicle-mounted CRNP should be controlled within 2 hours as much as possible.

[0057] S2.3.2: Spatial matching of UAV and cosmic neutron ray measurement: Using each CRNP observation point as the center, the UAV is used to measure the NDVI, NDWI, and canopy temperature Tcanopy in a region with a radius of 150m.

[0058] S2.3.3: CRNP Measurement Calibration: Using the SVR algorithm with NDVI, NDWI, and canopy temperature (Tcanopy) as input features, 120-200 samples were selected and trained through grid search and 5-fold cross-validation to construct a calibration model. The required values ​​are R² ≥ 0.70 and RMSE ≤ 0.02–0.03m. 3 ·m -3 (Soil volumetric moisture content)

[0059] S2.3.4: Calibration Model Output: Using the high-resolution information from the UAV as spatial weights, spatial downscaling calibration is performed on the CRNP points to generate a 5m×5m grid soil moisture distribution map of the field. This map is defined as the real-time soil moisture value. .

[0060] S3: Combine local meteorological data, crop growth monitoring data and soil physicochemical property data to construct a crop growth model, and use the real-time soil moisture content value in step S2 to assimilate and optimize the model so that the model has the ability to output soil moisture content prediction values, so as to realize the crop growth model to output soil moisture content prediction values.

[0061] S3.1: Basic Data and Model Preparation

[0062] Daily meteorological data for the past 5 years and real-time meteorological data for the current year were collected for the target area. The details are shown in the table below:

[0063]

[0064] Soil profile data were collected from typical fields. Layered sampling was conducted at depths of 0–20 cm, 20–40 cm, 40–60 cm, 60–90 cm, and 90–120 cm. Parameters such as soil texture, bulk density, field water holding capacity, and wilting point were measured. Recommended ranges for key soil parameters were used for model parameter initialization, quality control, and outlier removal. Details are shown in the table below.

[0065]

[0066] The AquaCrop crop growth model was selected as the base model, and the crop parameters of the model were initialized according to the characteristics of winter wheat varieties in North China. At the same time, the jointing stage, heading and flowering stage and the early grain filling stage were set as water stress sensitive stages to enhance the response to water shortage during critical periods.

[0067] The details are shown in the table below:

[0068]

[0069] S3.2: Model Initialization and Assimilation

[0070] Using the greening period as the starting point for the simulation (preferably late February to early March), soil parameters and daily meteorological data are input, and the initial soil moisture content is set. The initial soil moisture content is preferably generated by S2. Provided (0~30cm), and converted into the top 0~30cm soil water storage S0~30:

[0071]

[0072] For the initial moisture content of the soil layer below 0-30cm (30-120cm), it is preferably set at 0.65-0.80 times the field capacity, i.e.:

[0073]

[0074] The Ensemble Kalman Filter (EnKF) data assimilation algorithm is used to convert the real-time soil moisture values ​​obtained through S2. As observed values, the soil water molecule module of the AquaCrop model was assimilated; the root zone water storage Wr (mm) and the water content of each soil layer were also assimilated. As state variables, by continuously adjusting the state variables, the model output is... Compared with observed values Minimize the error between them, thereby dynamically correcting and optimizing the model. To ensure assimilation consistency, The optimal method is to calculate the water content of each soil layer in the model by weighting, and to use an output diameter of 0-30cm consistent with S2.

[0075] To improve spatial representativeness, observation values Input assimilation method by zone: Divide into 3-5 management zones according to soil texture / topography / growth status, and calculate the average for each zone. They are assimilated separately to improve the ability to depict spatial heterogeneity.

[0076] The assimilation settings are detailed in the table below:

[0077]

[0078] The assimilation error and process perturbation are preferably set according to the table below to describe the observation uncertainty and the model process uncertainty, and to ensure the stability of the EnKF update.

[0079]

[0080] S3.3: Predicted Output

[0081] The optimized model can more accurately reflect the actual water dynamics in the field. After each assimilation, the model is driven forward and, combined with the meteorological forecast data for the next 7 days, outputs the predicted daily soil moisture content for the next week. The 7-day weather forecast data shall include at least Tmax, Tmin, P, Rs, U2, and RH, and preferably adopt a direct-driven model with daily time steps.

[0082] The preferred output layer includes: surface volume water content of 0-30cm. and the average volumetric water content of the root zone Simultaneously, it outputs soil moisture deficit D (mm) or relative field capacity ratio θ / θFC for water stress identification and irrigation decision threshold determination. The specific output content (rolling over 7 days after assimilation) is shown in the table below:

[0083]

[0084] Spatial trend maps are preferably generated using a zoning and grid mapping method: first, the fields are divided into 3 to 5 management zones, and AquaCrop is run independently on each zone to obtain the zone scale. Then, using the gridding of the assimilation day. Using the spatial basis, the relative deviation coefficient K(x) of the grid is calculated and mapped to the zonal prediction results for the next 7 days, thus obtaining the spatialized prediction of the 5m grid.

[0085] The preferred calculation rules are:

[0086]

[0087] Meanwhile, to enhance prediction stability, a convergence / decay mechanism is preferably set for K(x): K(x) remains unchanged for the next 7 days, or converges to 1.0 daily with a decay coefficient in the range of 0.90 to 1.10, so as to reduce the spatial amplification effect caused by single remote sensing errors.

[0088] In addition, the key parameter settings for spatialization prediction (partitioning and grid mapping) are shown in the table below:

[0089]

[0090] Through the above steps, the optimized AquaCrop model can output daily soil moisture predictions for the next week after each assimilation. It also includes a 5m spatial trend map, thereby enabling dynamic prediction and early warning of water demand during key water requirement stages of winter wheat.

[0091] S4: Acquire multi-dimensional remote sensing information of crops synchronously through the UAV platform. Use UAVs to acquire multi-dimensional remote sensing data such as canopy temperature and spectral reflectance of crops in the field. Combine the above data with machine learning algorithms to construct an accumulated temperature inversion model and realize the output of real-time value of crop accumulated temperature.

[0092] S4.1: Continuous Data Acquisition

[0093] Starting from the greening period, using multi / hyperspectral and thermal infrared sensors from drones, a full-coverage aerial survey will be conducted every 7-10 days to continuously acquire canopy temperature data. The hyperspectral index (photochemical reflectance index PRI) reflects crop water stress and photosynthetic efficiency.

[0094] The specific data acquisition parameters are as follows:

[0095]

[0096] S4.2: Ground surface temperature benchmark

[0097] Automatic weather stations are deployed in the fields to record daily temperatures, and the general effective accumulated temperature (GDD) based on these temperatures is calculated. The effective accumulated temperature (GDD) can be calculated using the following formula:

[0098]

[0099] in: The highest temperature of the day (°C). The lowest temperature of the day (°C). This is the base temperature for crops, typically 10°C.

[0100] S4.3: Model Building

[0101] Timing acquired by drones Using PRI data as the main feature and synchronously calculated GDD as the baseline label, a random forest regression algorithm was used to train and construct a canopy effective accumulated temperature inversion model. This model establishes a nonlinear relationship between canopy microclimate information (remote sensing inversion) and the actual thermal effect (accumulated temperature) of crops. When using the random forest algorithm, the data is first preprocessed, i.e.... PRI data was converted into feature vectors for each plot, with GDD used as the target label to ensure temporal synchronization with remote sensing data; subsequently, known meteorological data GDD was used in conjunction with data acquired by the UAV. A training set was established using PRI data. Finally, a random forest regression algorithm was used to train the model, determining the optimal number of decision trees, tree depth, and other parameters. Cross-validation was then used to evaluate the model's accuracy, ensuring it effectively reflects changes in crop accumulated temperature. The specific training parameters for the random forest model are shown in the table below.

[0102]

[0103] S4.4: Real-time accumulated temperature output

[0104] For each drone observation, the processed canopy temperature distribution map and PRI distribution map are input into a trained random forest model, which can then be used to derive the real-time crop accumulated temperature value for the entire field at that moment. The spatial distribution of this value better reflects the actual heat accumulation in the field caused by water stress and differences in growth.

[0105] S5: Combine the real-time value of crop accumulated temperature with meteorological forecast data to calculate the predicted value of crop accumulated temperature, predict the crop growth stage, determine the crop growth stage based on the predicted value of crop accumulated temperature, and then obtain the predicted value of crop water requirement for that growth stage. Combine the predicted value of soil moisture content in step S3 with the predicted value of crop water requirement for each stage to perform difference analysis and obtain the predicted value of supplementary irrigation in the field for subsequent periods.

[0106] S5.1: Accumulated Temperature and Fertility Stage Prediction

[0107] Based on the latest real-time crop accumulated temperature value output in step S4 Based on the 7-day temperature forecast, the predicted accumulated temperature value for crops up to the 7th day is calculated. .

[0108]

[0109] Based on the table of accumulated temperature and growth period for local winter wheat varieties (shown above), and according to predictions... This determines which specific growth stage the crop will enter or be in within the next week. For example, based on the predicted accumulated temperature, it might be determined that the crop will enter the peak heading stage in 3 days, as shown in the example table below:

[0110]

[0111] S5.2: Water Demand Forecast

[0112] Based on the future growth stage determined in step S5.1, the crop water requirement for the next 7 days is calculated using the Penman-Monteith formula. The calculation formula is as follows:

[0113]

[0114] in, For crop coefficients, This serves as a reference for crop evapotranspiration over the next seven days. This is a forecast of crop water requirements for the coming week.

[0115] The following is an example of crop water requirement prediction calculation:

[0116]

[0117] S5.3: Irrigation Quantity Decision Calculation

[0118] The supplementary irrigation amount is calculated based on crop water requirement forecasts and soil moisture content forecasts. The steps are explained below:

[0119] S5.3.1: Soil Water Storage Calculation: Calculate the predicted soil moisture content for the same period in the future, as output from S3. Converted into effective soil water storage in the root zone (0-60cm) .

[0120]

[0121] S5.3.2: Calculation of theoretical total water consumption: based on crop water requirements Calculate the theoretical total water consumption for the coming week.

[0122] S5.3.3: Irrigation Quantity Forecast: Calculate the supplementary irrigation quantity for the coming week based on the following formula. :

[0123]

[0124] in, This is the forecast of effective rainfall for the next week. The irrigation water use efficiency is taken as 0.85. The consumable amount of effective soil water storage in the next week (and) (This is related to the lower limit of water consumption allowed by crops), max() takes the larger value.

[0125] The following is an example of irrigation volume prediction calculation:

[0126]

[0127] S5.4: Output irrigation prescription diagram

[0128] Based on the calculated supplementary irrigation amount This generates irrigation prescription maps to guide field irrigation decisions. The irrigation prescription map displays the supplemental irrigation amount for each plot or zone and allows for variable adjustments to water requirements in different areas based on the irrigation amount. An example of irrigation prescription map output data is shown below:

[0129]

[0130] The above content details the specific application of this invention during the critical growth period of winter wheat in North China. Through the cyclical execution of S2 to S4, continuous tracking and prediction of soil moisture and crop progress are achieved. Finally, through S5, forward-looking and precise irrigation decisions are made, effectively improving water resource utilization efficiency.

[0131] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing, characterized in that... Includes the following steps: S1. Determine the target crop species and their key water-required growth stages; S2. Use a vehicle-mounted near-ground cosmic ray neutron instrument to obtain field soil moisture content measurements, and integrate UAV remote sensing data to perform spatial calibration and verification to obtain real-time soil moisture content values. S3. Combine local meteorological data, crop growth monitoring data and soil physicochemical property data to construct a crop growth model, and use the real-time soil moisture content value in step S2 to assimilate and optimize the model so that the model has the ability to output predicted soil moisture content values. S4. Acquire multi-dimensional remote sensing information of crops synchronously through the UAV platform, construct a crop accumulated temperature inversion model based on machine learning algorithm, and estimate the real-time value of crop accumulated temperature; S5. Combine the real-time value of crop accumulated temperature with meteorological forecast data to calculate the predicted value of crop accumulated temperature, predict the crop growth stage, and perform difference analysis by combining the predicted value of soil moisture content in step S3 with the predicted value of crop water demand at each stage to obtain the predicted value of supplementary irrigation in the field for subsequent periods.

2. The method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing as described in claim 1, characterized in that, In step S2, the vehicle-mounted near-ground cosmic ray neutron spectrometer moves along a preset route to conduct observations once a week. Each observation acquires the volumetric water content of the surface soil. Simultaneously with each observation by the vehicle-mounted near-ground cosmic ray neutron spectrometer, an unmanned aerial vehicle platform is operated to take aerial photos over the observation strip to acquire data on canopy reflectivity and canopy temperature.

3. The method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing as described in claim 1, characterized in that... In step S2, the precise soil moisture content of multiple positioning points in the ground test area is obtained simultaneously using the time domain reflectometer method for verification; the spatial distribution map of NDWI generated by the UAV is spatially matched with the field soil moisture content measurement value; and a calibration model is established using the support vector machine regression algorithm, with the normalized differential moisture index NDWI, normalized vegetation index NDVI and canopy temperature as input features, and the field soil moisture content measurement value as the target value.

4. The method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing as described in claim 3, is characterized in that... By calibrating the model, high-resolution remote sensing information from UAVs is used as spatial weights to perform spatial downscaling calibration on the soil moisture content measurements at single points in the field, generating a real-time distribution map of soil moisture content with high spatial resolution for the entire field, which serves as the real-time value of soil moisture content.

5. The method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing as described in claim 1, characterized in that, Step S3 includes: S3.1 Basic Data and Model Preparation: Collect daily meteorological data for the past 5 years and real-time meteorological data for the current year for the target area; collect soil profile data for typical fields; select the AquaCrop crop growth model as the basic model, and initialize the model crop parameters according to the characteristics of crop varieties; S3.2 Model Initialization and Assimilation: Taking the greening period as the starting point of the simulation, soil parameters and daily meteorological data are input, and the initial soil moisture content is set; the ensemble Kalman filter data assimilation algorithm is used to assimilate the real-time soil moisture content obtained through S2 into the soil water molecule module of the AquaCrop model as observed values. S3.3 Prediction Output: After each assimilation is completed, the model is driven forward and combined with the meteorological forecast data for the next 7 days to output the predicted value of soil moisture content for each day of the next week.

6. The method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing as described in claim 5, is characterized in that... The output levels of the predicted soil moisture content include: the volumetric moisture content of the 0-30cm surface layer and the average volumetric moisture content of the root zone. The soil water deficit or relative field capacity ratio is also output simultaneously for water stress identification and irrigation decision threshold determination.

7. The method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing as described in claim 1, characterized in that, Step S4 includes: S4.1 Continuous Data Acquisition: Starting from the crop greening period, use UAV multi-spectral and thermal infrared sensors to conduct full-coverage aerial surveys every 7-10 days to continuously acquire the photochemical reflectance index PRI, which reflects the crop's water stress and photosynthetic efficiency. S4.2 Ground surface temperature baseline: Automatic weather stations are set up in the field to record daily air temperatures and calculate the conventional effective accumulated temperature (GDD) based on air temperature. The effective accumulated temperature (GDD) can be calculated according to the following formula: in: The highest temperature of the day (°C). The lowest temperature of the day (°C). The baseline temperature for crops is set at 10℃. S4.3 Model Construction: Using the time series and photochemical reflectance index PRI obtained by UAV as the main features, and the effective accumulated temperature GDD obtained by synchronous calculation as the baseline label, the random forest regression algorithm is used for training to construct the canopy effective accumulated temperature inversion model. S4.4 Real-time accumulated temperature output: For each UAV observation, the processed canopy temperature distribution map and PRI distribution map are input into the trained canopy effective accumulated temperature inversion model, which can then invert the spatial distribution of the real-time crop accumulated temperature value of the entire field at the current moment.

8. The method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing as described in claim 1, characterized in that, Step S5 includes: S5.1 Accumulated Temperature and Growth Stage Prediction: Based on the real-time crop accumulated temperature value output by S4, combined with the temperature forecast for the next 7 days, the predicted crop accumulated temperature value for the next 7 days is calculated; based on the crop accumulated temperature and growth stage comparison table and the predicted crop accumulated temperature value, it is determined which specific growth stage the crop will enter or be in within the next week. S5.2 Water requirement forecast: Based on the future growth stage determined in S5.1, the crop water requirement for the next 7 days is calculated using the Penman-Montes formula; S5.3 Irrigation Quantity Calculation: Calculate the supplementary irrigation amount based on crop water requirement forecast and soil moisture content forecast; S5.4 Output Irrigation Prescription Map: Based on the calculated supplementary irrigation amount, an irrigation prescription map is generated to guide field irrigation decisions; the irrigation prescription map will display the supplementary irrigation amount for each plot or zone, and the water demand of different areas can be variably adjusted according to the irrigation amount.

9. The method for predicting soil moisture and supplemental irrigation for field crops by integrating cosmic ray neutron and remote sensing as described in claim 8, characterized in that, Step S5.3 includes: S5.3.1 Soil Water Storage Calculation: Calculate the predicted soil water content for the same period in the future, as output in step S3. Effective soil water storage capacity converted into root zone The calculation formula is: S5.3.2 Calculation of theoretical total water consumption: based on crop water requirements Calculate the theoretical total water consumption for the coming week; S5.3.3 Irrigation Quantity Forecast: Calculate the supplementary irrigation quantity for the next week based on the following formula. : in, This is the forecast of effective rainfall for the next week. To improve irrigation water use efficiency, This represents the amount of water that can be consumed from the soil's effective water storage capacity within the next week, with max() taking the larger value.

Images