Crop growth simulation method and system

Abstract

The invention discloses a crop growth simulation method and system, and relates to the technical field of irrigation, and the method comprises the steps: building a physiological model according to the physiological change of crops, comprising the following steps: simulating a three-dimensional leaf model of a crop according to the leaf length, the leaf width, the leaf area, the leaf curl rate and the leaf distribution condition of the crop, simulating a three-dimensional physiological model of the crop according to the leaf area change degree of the crop, assimilation products of the crop under photosynthesis, respiration consumption of the crop under respiration, growth weight increment of the crop and organ dry weight of the crop; and jointly simulating the growth condition of the crop according to the three-dimensional leaf model of the crop and the three-dimensional physiological model of the crop. The device can truly simulate the crop growth process.

Description

Technical Field

[0001] This invention relates to the field of irrigation technology, and in particular to a method and system for simulating crop growth. Background Technology

[0002] Traditional agricultural production has many drawbacks, such as requiring a lot of human and material resources to pay attention to the environment, conditions and processes of crop growth. Crop models are a new research field for studying crop growth. They treat crops, environment and cultivation techniques as a whole system to simulate the physiological processes of crop growth and development, photosynthetic growth, organ formation and yield formation and their relationship with the environment.

[0003] Currently, crop growth simulation mainly employs three methods. Method one uses crop growth models, which simulate the crop growth process based on factors such as light, temperature, and moisture. Method two uses remote sensing technology, acquiring spectral information of crops via drones or satellites to provide large-scale and real-time monitoring, and simulating crop growth based on the monitored data. Method three uses data assimilation techniques, combining remote sensing data with crop growth models to improve the accuracy of the models; commonly used methods include Kalman filtering and Bayesian methods.

[0004] However, existing technologies for simulating crop growth processes do not fully consider the internal physiological conditions of crops, resulting in unrealistic simulations of crop growth processes. Summary of the Invention

[0005] This invention provides a method and system for simulating crop growth, which can solve the problem that in the prior art, the internal physiological conditions of crops are not fully considered when simulating crop growth, resulting in an unrealistic simulated crop growth process.

[0006] This invention provides a method for simulating crop growth, comprising the following steps: acquiring relevant data from a crop database, a meteorological database, and a soil database; simulating a three-dimensional leaf model of the crop based on its leaf length, leaf width, leaf area, leaf curling rate, and leaf distribution; simulating a three-dimensional physiological model of the crop based on changes in leaf area, assimilation products during photosynthesis, respiratory consumption during respiration, growth weight increment, and organ dry weight; and simulating crop growth using 3D Gaussian splashing technology based on the three-dimensional leaf model and the three-dimensional physiological model of the crop.

[0007] Furthermore, the three-dimensional leaf model of the crop is simulated based on the crop's leaf length, leaf width, leaf area, leaf curl rate, and leaf distribution. The ambient temperature is input into a leaf length model to describe the influence of ambient temperature on leaf growth, and the leaf length is obtained. The leaf length is input into a leaf width model to describe the relationship between leaf length and width, and the leaf width is obtained. The leaf area is simulated based on the upper and lower contour curve equations of the leaf. The leaf area when the leaf is extended and curled is input into a leaf curl rate model to describe the influence of leaf area on leaf curling, and the leaf curl rate is obtained. The leaf distribution is simulated based on the probability distribution.

[0008] Furthermore, the specific steps of simulating the three-dimensional physiological model of the crop based on the degree of change in crop leaf area, the assimilation products of the crop under photosynthesis, the respiratory consumption of the crop under respiration, the increase in crop growth weight, and the dry weight of crop organs include: Water and nitrogen are input into a leaf area dynamics model to describe the effects of water and nitrogen on leaf area changes, yielding the degree of leaf area change. Solar radiation, ambient temperature, water, and nitrogen are input into a photosynthesis model to describe the effects of multiple factors on photosynthesis, yielding assimilates under photosynthesis. Total respiratory consumption is obtained based on the respiratory consumption required for life maintenance and growth. Assimilates and respiratory consumption required for life maintenance are input into an incremental model to describe the influence of photosynthesis and respiration on growth weight increment, yielding growth weight increment. Assimilates are input into a dry weight model to describe the effect of assimilate distribution in various organs on organ dry weight, yielding organ dry weight.

[0009] Furthermore, the leaf curling rate of the crop is specifically obtained through the following steps: Leaf curling rate RR The water vapor conductance is obtained by comparing the water vapor conductance when the blade is curled with that when the blade is unfolded. The water vapor conductance represents the ability of the blade or stomata to conduct water vapor. The blade curl rate RR, The formula is: ; in, LA 0 Potential leaf area, i.e., the maximum leaf area without curling; LA The actual leaf area is given, and the leaf curl rate (RR) ranges from 0 to 1.

[0010] Furthermore, the simulation of blade distribution based on probability distribution specifically includes: Through probability distribution function f ( x ), to obtain the probability of the leaf being in the interval [-20°, 20°] and [-60°, 60°]. P ; The leaves are divided into multiple levels from low to high, based on probability. P Randomly draw blades of order 6 or 8 in the interval [-20°, 20°]; draw blades of order 6 or 8 or higher in the interval [-60°, 60°].

[0011] Furthermore, the degree of change in the leaf area of ​​the crop is obtained through the following specific steps: The degree of change in leaf area is obtained from the leaf area index; Leaf area index under suitable water and nitrogen nutrition conditions LAI The formula for obtaining it is: in, L max The maximum green leaf area of ​​corn under the required meteorological conditions. FD The percentage of yellow leaves. DR Yellowing rate C This is the conversion factor for the yellowing rate. DVS For crop biology students for a long time, a Both b and are coefficients; Leaf area index under water and nitrogen nutrient stress conditions RLAI The formula for obtaining it is: in, RLAI(i) , RLAI(i - 1) For actual water and nitrogen nutrient stress i Time and i-1 Leaf area index at that time; LAI(i) , LAI(i - 1) Under suitable water and fertilizer conditions i Time and i-1 Leaf area index at that time; DLAI This represents the increase in leaf area index; SWF and SNF1 These are the correction factors for water and amino nutrient stress, respectively.

[0012] Furthermore, the specific steps for simulating crop growth using 3D Gaussian splashing technology include: Leaf simulation data is provided through a 3D leaf model of the crop, and morphological construction parameters for crop growth and dry weight of crop organs are provided through a 3D physiological model of the crop. The temporal binding of crop leaf simulation data and morphological construction parameters is achieved using 3D Gaussian splashing technology.

[0013] This invention provides a crop growth simulation system, comprising: The data acquisition module is used to acquire relevant data from crop databases, meteorological databases, and soil databases; The model simulation module is used to simulate a three-dimensional leaf model of a crop based on its leaf length, leaf width, leaf area, leaf curling rate, and leaf distribution; and to simulate a three-dimensional physiological model of a crop based on the degree of change in leaf area, assimilation products of the crop under photosynthesis, respiratory consumption of the crop under respiration, growth weight increment of the crop, and organ dry weight of the crop. The crop growth simulation module is used to simulate crop growth using 3D Gaussian splashing technology, based on the crop's three-dimensional leaf model and three-dimensional physiological model.

[0014] This invention provides a method and system for simulating crop growth, which has the following advantages compared with the prior art: The crop growth process is simulated using a three-dimensional leaf model and a three-dimensional physiological model. The three-dimensional leaf model simulates the shape of the leaves by measuring leaf length, width, area, curling rate, and distribution. The three-dimensional physiological model simulates the physiological state of the crop by measuring changes in leaf area, assimilation products during photosynthesis, respiratory consumption during respiration, weight gain, and organ dry weight. Finally, the three-dimensional leaf model and the three-dimensional physiological model are combined using 3D Gaussian sputtering technology to take into account the shape and physiological changes that occur during crop growth and reflect real crop growth. Attached Figure Description

[0015] Figure 1 A flowchart illustrating a crop growth simulation method provided in an embodiment of the present invention; Figure 2 An internal organizational diagram of the intelligent decision-making system provided in this embodiment of the invention; Figure 3 This is a schematic diagram of corn plant modeling provided in an embodiment of the present invention; Figure 4 A schematic diagram of the arrangement of a field thermal infrared imager and a hyperspectral camera is provided for an embodiment of the present invention; Figure 5 This is a partial schematic diagram provided for an embodiment of the present invention.

[0016] Reference numerals: 1-thermal infrared equipment, 2-hyperspectral imaging equipment, 3-laser rangefinder, 4-electric telescopic pole. Detailed Implementation

[0017] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0018] See Figure 1 This invention provides a method for simulating crop growth, comprising the following steps: Step 1: Obtain relevant data from crop databases, meteorological databases, and soil databases.

[0019] Step 2: Simulate a three-dimensional leaf model of the crop based on the leaf length, leaf width, leaf area, leaf curling rate, and leaf distribution; simulate a three-dimensional physiological model of the crop based on the degree of change in leaf area, assimilation products of the crop under photosynthesis, respiratory consumption of the crop under respiration, growth weight increment of the crop, and organ dry weight of the crop.

[0020] Step 3: Based on the three-dimensional leaf model and the three-dimensional physiological model of the crop, use 3D Gaussian splashing technology to simulate crop growth.

[0021] Specifically, the ambient temperature is input to a leaf length model describing the influence of ambient temperature on leaf growth, and the leaf length is obtained; the leaf length is input to a leaf width model describing the relationship between leaf length and width, and the leaf width is obtained; the leaf area is simulated based on the upper and lower contour curve equations of the leaf; the leaf area when the leaf is extended and curled is input to a leaf curl rate model describing the influence of leaf area on leaf curling, and the leaf curl rate is obtained; the leaf distribution is simulated based on the probability distribution.

[0022] Step 4: Crop physiological data include: the degree of change in crop leaf area; crop assimilation products under photosynthesis; crop respiration consumption under respiration; crop growth weight gain; crop organ dry weight; and crop yield.

[0023] Specifically, water and nitrogen are input into a leaf area dynamics model to describe the effects of water and nitrogen on leaf area changes, obtaining the degree of leaf area change; solar radiation, ambient temperature, water, and nitrogen are input into a photosynthesis model to describe the effects of multiple factors on photosynthesis, obtaining assimilates under photosynthesis; total respiratory consumption is obtained based on the respiratory consumption required for life maintenance and growth; assimilates and respiratory consumption required for life maintenance are input into an incremental model to describe the influence of photosynthesis and respiration on growth weight increment, obtaining growth weight increment; and assimilates are input into a dry weight model to describe the effect of the distribution of assimilates in various organs on organ dry weight, obtaining organ dry weight.

[0024] I. Three major databases.

[0025] (a) Crop Database.

[0026] The crop database consists of two parts: an existing database and a measured database. It mainly includes relevant data affecting crop growth and development, such as plant height, leaf inclination angle, leaf width, and leaf length, which serve as basic data for the calculation, simulation, and analysis of crop growth models.

[0027] 1. Existing database.

[0028] The query summary is based on multi-source crop databases formed by Web of Science, CNKI, etc., the China Agricultural Science and Technology Literature and Information Service Platform, and the crop databases in the agricultural and rural big data upload and cloud entry platform.

[0029] 2. Actual measurement database.

[0030] Crop leaf temperature and images are collected by thermal infrared imaging device 1 and hyperspectral imaging device 2 deployed diagonally along the field. A laser rangefinder 3 is used to determine the distance between the imaging device and the crop canopy (note: this distance should be 1.5 to 2 times the height of the crop canopy from the ground). Then, an automated electric telescopic pole 4 is used to reach the target height to ensure the reasonableness of the imaging field of view and spatial resolution. The data is then uploaded to the crop database and processed and analyzed by the platform's built-in image recognition and analysis system. The thermal infrared imager accurately captures leaf surface temperature information. Based on this, the identified leaf temperature is input into a pre-constructed maize leaf length model to indirectly invert leaf length, thereby estimating leaf area and growth status. Simultaneously, hyperspectral imaging technology, based on the spectral reflectance characteristics of leaves in different wavelength bands, can not only invert leaf color but also extract morphological parameters such as leaf area, thereby calculating leaf curl rate to assess crop growth status and nutritional requirements. The platform extracts and analyzes indicators from the collected data, including multiple parameters such as normalized thermal difference index, temperature comparison index, and hyperspectral vegetation index, to construct an intelligent diagnostic model for crop water and nutrient status, enabling precise identification and management of crop stress conditions.

[0031] The following points should be noted during setup and shooting: (1) Set the height of the imaging equipment reasonably. The laser rangefinder 3 accurately measures the vertical distance from the equipment to the crop canopy, and the electric telescopic pole 4 is automatically controlled to reach the target height, so as to ensure complete coverage of the imaging field of view and appropriate spatial resolution.

[0032] (2) When the equipment is laid out along the diagonal of the field, the spacing should take into account the field of view of the imaging equipment, the installation height and the diagonal length of the plot. It is recommended that the spacing be 70% to 90% of the width of the ground that each equipment can cover, so as to ensure that there is an overlap area of ​​10% to 30% between images, avoid monitoring blind spots and ensure the scientific and rational nature of sampling.

[0033] (3) The shooting should be carried out under windless or light wind and stable sunlight conditions to avoid leaf shaking and strong light reflection interfering with image and temperature reading. At the same time, the resolution of the shooting equipment should be at least 12 million pixels and the lens should have normal light transmission, without blurring or whitening, so as to ensure the accuracy of subsequent data extraction and the feasibility of establishing a three-dimensional model during crop growth.

[0034] (ii) Meteorological database.

[0035] The system utilizes data from the China Meteorological Network, NASA's meteorological network, and nearby meteorological stations in the farmland area to create a global meteorological database spanning 0.1 km. This database contains essential factors influencing crop growth and development, including temperature, rainfall, wind speed, air pressure, and humidity, serving as foundational data for the water consumption module.

[0036] (iii) Soil Database.

[0037] The soil database includes the China Soil Database, the Soil Science Data Center, and an experimental database established through field data collection, comprehensively containing data on farmland soil conditions across various regions of China. It primarily includes data related to crop growth and development, such as soil depth, soil fertility, soil moisture, soil pH, and soil salinity, serving as foundational data for the soil water and nitrogen transformation module and the crop growth module.

[0038] II. Crop growth model.

[0039] (a) Crop growth module.

[0040] The crop growth module mainly includes a leaf area calculation module, a photosynthesis module, a respiration module, a biomass module, and an assimilate allocation module, which are used to calculate the dynamic changes in leaf area, the total photosynthetic assimilate production, the total respiration consumption, the increase in biomass, and the interconversion relationships of assimilates in different parts of the plant.

[0041] 1. Leaf Area (LAI) Dynamics.

[0042] Based on field trial data, a modified Logistic growth function was used to describe the growth. Under suitable water and nitrogen nutrition conditions:

[0043] .

[0044] .

[0045] .

[0046] In the formula, L max This represents the maximum possible green leaf area of ​​corn under certain weather conditions. FD The percentage of yellow leaves. DR Yellow leaf rate (the ratio of yellow leaf weight to total leaf weight); C This is the conversion factor for the yellowing rate.

[0047] Under actual water and nitrogen nutrient stress conditions: .

[0048] .

[0049] .

[0050] In the formula, RLAI(i) , RLAI(i - 1) For actual water and nitrogen nutrient stress i Time and i-1Leaf area index at that time; LAI(i) , LAI(i - 1) Under suitable water and fertilizer conditions i Time and i-1 Leaf area index at that time; [[ID= This represents the increase in leaf area index; ​ , ​ These are the correction factors for water and amino nutrient stress, respectively.

[0051] 2. Photosynthesis.

[0052] Crop photosynthesis is affected by light, temperature, water, fertilizer, and the crop's own population structure, which can be represented as: .

[0053] In the formula, P The assimilation products of photosynthesis per unit time (kg / hm'.d); P s The CO2 assimilation rate (kg / hm'·d) of the canopy crown after correction for actual radiation conditions was determined using the method proposed by deWit and corrected by Goudriaan and van Laar, based on the spring small surface layer model. ​ , ​ , ​ These are the factors influencing temperature, moisture, and nitrogen nutrient stress, respectively.

[0054] Correction for the effect of temperature on photosynthesis: .

[0055] In the formula, ​ For the first i The average temperature (°C) of the ten-day period; T 0 The optimal temperature for photosynthesis; B The coefficients are obtained from experimental data and computer-assigned values.

[0056] 3. Respiration.

[0057] .

[0058] .

[0059] .

[0060] In the formula, R t This represents total respiratory expenditure; R m and R g These are respectively the respiratory expenditure for maintaining respiration and the respiratory expenditure for growth; Rm0 To maintain the respiratory rate, ​ Total dry weight of the plant; ​ This represents the conversion efficiency of primary photosynthetic products.

[0061] 4. Biomass increment (D) ​ ).

[0062] The increase in plant growth depends on photosynthesis ( P ) and respiration ( R The size of ) is expressed as .

[0063] 5. Assimilate allocation and transfer sub-pattern.

[0064] The growth of crop organs is based on the distribution and transfer of assimilates within each organ. The growth process of maize organs can be represented by the following formula: .

[0065] .

[0066] .

[0067] In the formula, W j Representing the roots ( W 1 ),stem( W 2 ),leaf( W 3 ),ear( W 4 The dry weight of ) ​ For the plant in i to i+1 Biomass increase over a period of time; F j , ​ j These are the allocation and transfer coefficients for each organ.

[0068] 6. Crop yield.

[0069] Crop yield depends on the number of ears ( N ear ), number of grains ( N grain ) and particle weight ( W grain The yield of corn can be expressed by the following formula: .

[0070] In the formula, Y is the yield (kg / ha), and the plant density is the number of maize plants per unit area, plants / ha; ​ ear , ​ ear The water and nitrogen stress factors during the panicle formation period (0~1, 1 indicates no stress); k grain The grain formation coefficient (grains / g dry matter) represents the number of grains that can be formed per unit of dry weight of ear; W 3 The dry weight of the spike (g / plant) is calculated by the assimilate allocation module; ​ grain , ​ grain The water and nitrogen stress factors during the grain formation period (0~1); k fill The grain filling efficiency coefficient (g / grain·day) represents the efficiency of assimilation being converted into grains per unit time. ​ 3 The assimilate transfer coefficient (0~1) represents the proportion of biomass allocated to the ear; ​ 3 ⋅W 3 ​ The cumulative dry matter content of the panicle during the grain-filling period (g / plant) is calculated by integral calculation from flowering to maturity.

[0071] (ii) Three-dimensional module The 3D module consists of a leaf length and width calculation module, a leaf area calculation module, a leaf curvature calculation module, and a leaf azimuth angle calculation module, which are used to simulate and calculate the area, curvature, and distribution of the leaves in the 3D model, respectively.

[0072] 1. Leaf Length Calculation The formula for calculating the leaf length of corn can be based on the accumulated temperature during growth, which refers to the effective temperature accumulated during the crop's growth process.

[0073] .

[0074] in, ​ Right now L For leaf length; a is an empirical constant, representing the initial leaf length (the leaf length when GDD is zero). b is the growth rate constant, representing the leaf length increase per unit GDD; GDD is the accumulated temperature for growth, calculated as follows: .

[0075] T imax For the first i The highest temperature of the day; T i min For the first i The lowest temperature of the day, T base Base temperature for growth, typically 10°C; n It refers to the number of days.

[0076] 2. Calculation of leaf width.

[0077] The shape of a leaf that develops in a plane is described by the relationship between the leaf width and the position of the midrib.

[0078] In the formula, u The distance to the ligule. L The total length of the blade; w yes u Width at the point W It is the maximum width of the blade layer.

[0079] 3. Calculation of the area of ​​the model blades.

[0080] The blades in the model are composed of outlines. The leaf area of ​​the simulated blade can be calculated based on the leaf outline equation, and the blade area in the model can be obtained by integration.

[0081] .

[0082] In the formula, ​ For the blade length, ​ , ​ These are the curve equations for the upper and lower profiles of the blade, respectively. The curve equations can be described by the following quadratic curve equations, with the upper profile curve as an example: .

[0083] Where 'a' is the coefficient of the quadratic term, which controls the curvature of the leaf veins; 'b' is the coefficient of the linear term in the quadratic curve equation, which can be solved using the leaf vein equation when the leaf is straightened; and 'c' is the coefficient of the constant term, representing the height of the leaf's growth point.

[0084] 4. Blade curvature calculation: Blade curling rate ( ​ The value should be determined by the potential leaf area and the sensitivity coefficient, i.e.: .

[0085] in, ​ 0 Potential leaf area (maximum leaf area without curling). ​ The actual leaf area (considering the effect of leaf curling), and the leaf curling rate. ​The leaf curling index ranges from 0 (no curling) to 1 (complete curling).

[0086] 5. Leaf azimuth calculation Based on the analysis of maize variety data in the crop database, the random azimuth angle of the plane containing the first five leaves was considered, and the random deviation of the plane from the leaf level of the upper leaves was plotted.

[0087] With uniform probability density, the deviation angles of blades of orders 6 to 8 are randomly plotted in the interval [-20°, 20°], and the deviation angles of blades of higher orders are plotted in the interval [-60°, 60°].

[0088] Let its distribution function be F(x). By calculating the probability density function f(x) of the leaf, the probability P of the leaf x being distributed in the interval [-20°, 20°] and [-60°, 60°] can be obtained: .

[0089] .

[0090] The simulation results are as follows ​ As shown.

[0091] 6. Integration of 3D Gaussian splash technology.

[0092] Leaf simulation data is provided through a 3D leaf model of the crop, and morphological construction parameters for crop growth and dry weight of crop organs are provided through a 3D physiological model of the crop. The temporal binding of crop leaf simulation data and morphological construction parameters is achieved using 3D Gaussian splashing technology.

[0093] Among them, the three-dimensional leaf model of crops provides leaf simulation data including: using NURBS surface fitting technology to process leaf point cloud data of crops such as corn, generating leaf geometric morphology parameters (aspect ratio, curvature, etc.) through control point algorithm; and combining the PlantGaussian model to capture the changes in leaf topology across the growth period and establish organ-level three-dimensional digital twins.

[0094] The morphological parameters for crop growth and dry weight expression provided by the three-dimensional physiological model of crops include: inputting the output data of the CropGrow model, including diurnal growth indicators such as leaf area index (LAI) and leaf dry weight (LDW); and calculating the distribution ratio of photosynthetic products through the mechanistic model and converting it into morphological parameters such as leaf thickness and density.

[0095] III. Automatic Irrigation and Fertilization Module.

[0096] 1. Irrigation decision model (based on curl rate and photosynthetic rate) Dynamically adjust irrigation volume (​ (Unit: mm / day) to optimize water use efficiency and avoid drought stress. ​ (decline) or over-irrigation. Irrigation decisions follow the principle of "prioritizing leaf curl rate, supplementing with photosynthetic rate, and avoiding over-irrigation": when leaf curl rate significantly increases (decline) ​ And soil moisture is below the drought threshold ( ​ m ​ When the curling is not obvious but the photosynthetic rate is below the threshold, irrigate according to the water shortage ratio; if the curling is not obvious but the photosynthetic rate is below the threshold, irrigate according to the water shortage ratio. ​ ), and the soil did not reach the target moisture content ( ​ ta ​ ge If the soil moisture is sufficient, then supplementary irrigation should be implemented; if the soil moisture is sufficient ( ​ ta ​ ge ) or the crop does not show signs of stress ( ​ and ​ If the water level is 0, irrigation will not be carried out to avoid overwatering.

[0097] .

[0098] In the formula, ​ This refers to irrigation volume, expressed in mm / day, representing the daily irrigation water volume. ​ max Maximum permissible irrigation volume, in mm / day; threshold to prevent over-irrigation. ​ The leaf curl rate is a dimensionless value of 0-1, where 0 represents no curl and 1 represents complete curl. ​ th The curl rate threshold is dimensionless, ranging from 0.3 to 0.6. Exceeding this value indicates moisture stress. P The current photosynthetic rate is expressed in g / m²·day, reflecting crop productivity. P opt The optimal photosynthetic rate, in g / m²·day, is the maximum photosynthetic rate under no stress. P th This is the photosynthetic rate threshold, typically taken as 0.7. P opt A value below this indicates limited photosynthesis; ​ Soil moisture content, expressed in % or mm, indicates the current soil moisture status; ​ min This is the lower limit of soil moisture, expressed as a percentage or mm; below this value, crops wilt. ​ target Target soil moisture, in % or mm, ideal moisture content; ​The water stress factor is dimensionless, ranging from 0 to 1, where 1 represents no stress and 0 represents complete stress. ​ opt The optimal water stress factor is fixed at 1, indicating no water stress.

[0099] 2. Fertilization decision model (based on yield forecast and leaf yellowing rate) Dynamically adjust nitrogen fertilizer application rate ( ​ (Unit: kg N / ha) to balance yield targets and leaf senescence ( ​ (Signal) to avoid over-fertilization. The fertilization strategy is based on the principles of "yield-oriented, leaf yellowing ratio response, and avoidance of over-fertilization": when the predicted yield is significantly lower than the target value ( Y p ​ ed <0.8Y ta ​ ge And the soil nitrogen is insufficient. N so ​ c When the proportion of yellow leaves is high, apply fertilizer according to the nitrogen deficiency ratio; if the proportion of yellow leaves is too high ( ​ ), then based on the original foundation, according to ( ​ The weighting increases the amount of fertilizer applied in response to potential nutrient decline signals; when soil nitrogen is sufficient ( N so ​ c ) or the predicted output is close to the target ( Y p ​ ed ≥0.9Y ta ​ ge When the fertilizer application is ineffective, stop the fertilization process to avoid wasting resources and over-application.

[0100] .

[0101] In the formula: ​ This refers to the amount of fertilizer applied, expressed in kg N / ha, indicating the amount of nitrogen fertilizer applied. ​ max The maximum amount of fertilizer applied at one time, expressed in kg N / ha, is the threshold for preventing seedling burn. Y pred For predicted yield, the unit is kg / ha, and the estimated yield is based on the current biomass; Y target Target yield, in kg / ha, is the expected yield set according to variety and climate. ​ The percentage of yellow leaves is a dimensionless value of 0-1, reflecting the degree of nitrogen stress.​ th The threshold for the proportion of yellow leaves is 0.2-0.4, which is dimensionless. Exceeding this value will trigger additional fertilization. N soil Soil nitrogen content, unit: kg N / ha; N c This is the critical soil nitrogen concentration, expressed in kg N / ha. Fertilizer application is required if the concentration is below this value.

[0102] IV. Intelligent Decision-Making System.

[0103] The intelligent decision-making system mainly consists of the following components: data acquisition module, data transmission and storage module, data processing and analysis module, decision support system module, control execution system module, user interface, and system maintenance and optimization. ​ As shown.

[0104] The data acquisition module uses crop data calculated from three major databases and a crop growth model, pre-setting the farmland environment and crop growth status. The data transmission and storage module transmits data via a wireless communication module and performs storage analysis and management. The data processing and analysis module cleans and structures the collected data, further introducing a random forest model, trained based on historical crop growth, water and fertilizer input, and output data, to generate prediction results for water and fertilizer application combinations under similar environmental and crop conditions. Simultaneously, the system's crop growth model can simulate the dynamic changes in water and fertilizer requirements during crop growth based on physiological parameters such as crop leaves, photosynthesis, and respiration, possessing strong physical driving and mechanistic explanation capabilities. To achieve complementary integration of prediction and simulation, the system introduces a data assimilation algorithm, dynamically integrating and parameterizing the data-driven prediction results with the simulation results of the mechanistic model over time: on the one hand, the random forest model provides estimates of crop status and water and fertilizer demand trends based on big data learning; on the other hand, the crop growth model provides continuous simulation of the crop growth process and deduction of change mechanisms. The data assimilation method weights and fuses the results from both methods at each time step, and then back-optimizes the model parameters to improve the model's prediction accuracy and stability, thereby achieving intelligent decision-making driven by both mechanisms and data. The decision support module uses the fused prediction and simulation results to formulate precise and efficient water and fertilizer management plans. The control execution module connects hardware devices such as water pumps, pipelines, and solenoid valves, and automatically adjusts irrigation and fertilization behaviors based on the decision results. The user interface provides data monitoring and decision display. System maintenance and optimization ensure the system operates normally and is continuously improved.

[0105] Finally, the system continuously adjusts and optimizes irrigation and fertilization plans based on crop growth and environmental changes to ensure healthy crop growth and yield. Simultaneously, the system can also continuously improve and optimize decision-making based on user feedback and market changes.

[0106] This invention addresses the problem that traditional agricultural planting and monitoring consume a lot of time, manpower, and resources, and lack precise fertilization and irrigation control decisions. It proposes to use consumer-grade non-measuring cameras to capture and identify various indicators of crops, calculate relevant parameters through crop growth models, and use an intelligent decision-making system for scientific fertilization and irrigation. This establishes a camera-based three-dimensional digital farmland irrigation and fertilization decision-making platform (CRD).

[0107] The first step involves calculating the height between the equipment and the canopy using a laser rangefinder 3, and then automatically controlling an electric telescopic pole 4 to reach the target height. The second step involves deploying an infrared thermal imager 1 and a hyperspectral imaging device 2 diagonally across the farmland. The system automatically collects image data at each growth stage of the crop based on preset shooting frequencies and lighting conditions, ensuring the stability and coverage of information acquisition. The third step involves extracting key parameters such as crop leaf temperature and spectral characteristics from the images, importing them into a built-in crop growth model, calculating the geometric features and physiological state of the leaves, and performing 3D modeling. The fourth step involves the system's decision-making module calculating crop moisture content, nitrogen content, and photosynthetic energy based on the crop growth model. The system uses indicators such as force and leaf curling rate, combined with historical data prediction results, and employs data assimilation algorithms to dynamically determine whether fertilization or irrigation is needed. In the fifth step, the intelligent decision-making system controls remotely connected irrigation and fertilization equipment, including pumps, pipes, and solenoid valves, based on the generated management plan. Simultaneously, it obtains real-time feedback through water and nitrogen sensors buried in the field to determine if thresholds are met and automatically stops operation. In the sixth step, throughout the entire crop growth cycle, infrared thermal imager 1 and hyperspectral imaging device 2 continuously acquire new images to update crop status information and calibrate and dynamically optimize the crop growth model, improving the overall simulation and decision-making accuracy of the system.

[0108] This invention provides a crop growth simulation system, comprising: The data acquisition module is used to acquire relevant data from crop databases, meteorological databases, and soil databases.

[0109] The model simulation module is used to simulate a three-dimensional leaf model of a crop based on its leaf length, leaf width, leaf area, leaf curling rate, and leaf distribution; and to simulate a three-dimensional physiological model of a crop based on the degree of change in leaf area, assimilation products of the crop under photosynthesis, respiratory consumption of the crop under respiration, growth weight increment of the crop, and organ dry weight of the crop.

[0110] The crop growth simulation module is used to simulate crop growth using 3D Gaussian splashing technology, based on the crop's three-dimensional leaf model and three-dimensional physiological model.

[0111] A specific example is as follows: The experiment was conducted from 2022 to 2023 at the Jiuzhuang Water-Saving Comprehensive Experimental Station in Shuanghe Town, Hetao Irrigation District, Inner Mongolia (107°18′E, 40°41′N). This area has a typical mid-latitude semi-arid continental climate, with an average annual precipitation of 138.8 mm, an average temperature of 6.8 ℃, large diurnal temperature range, and up to 3230 hours of sunshine, making it one of the regions with the longest sunshine hours in China. The experimental area is mainly composed of silty loam soil, with a soil bulk density of 1.42 g / cm³ in the 0-100 cm depth. -3 The average field water holding capacity was 0.285 cm³. 3 cm -3 The total nitrogen, total phosphorus, and total potassium content (by mass) of the soil were 0.093%, 0.07%, and 1.60%, respectively, the organic matter content was 1.2%, and the pH value was 7.6.

[0112] The corn was sown on May 1, 2022, and May 4, 2023, and harvested on September 14, 2022, and September 20, 2023, respectively. The planting method was "one film, one pipe, two rows," with a crop spacing of 30cm and a row spacing of 50cm, resulting in a planting density of 80,040 plants / ha. -1 The tested mulch film was a fully biodegradable film with a thickness of 0.008 mm and a width of 80 cm. The experiment used 60% field moisture content as the lower limit for irrigation. The basal fertilizer mainly consisted of urea, diammonium phosphate, and potassium sulfate, with 120 kg / ha of diammonium phosphate and potassium sulfate each applied. -1 The amount of urea applied is 20% of the total nitrogen. Topdressing uses liquid urea fertilizer (N: 32%), applied at the jointing stage, tasseling stage, and grain-filling stage, respectively, at 20%, 30%, and 20% of the total nitrogen.

[0113] I. The CropGrow (CG) model was used to determine the leaf area, dry matter, and crop yield of maize.

[0114] Table 1 compares the accuracy of different maize growth estimation methods, showing that the simulation accuracy of this invention (CRD) is relatively high. Taking 2023 data as an example, the coefficient of determination (R²) for the CRD-estimated LAI simulation value is... 2 The relative error (MRE), root mean square error (RMSE), and standard root mean square error (NRMSE) can reach 0.88, 11.3%, 0.41 cm³ / cm⁻³, and 0.19, respectively; the accuracy of AB is 0.86, 6.7%, 19.95 g, and 0.12, respectively; and the accuracy of Y is 7.2%, 595.29 kg / ha, respectively. -1 And 5.79. The CRD simulation accuracy is significantly higher than that of CG, LAI and AB. 2Compared to CG, the MRE of LAI, AB, and Y were improved by 5.0% and 6.1%, respectively; the RMSE of LAI, AB, and Y were reduced by 81.3%, 74.3%, and 67.4%, respectively; the RMSE of LAI, AB, and Y were reduced by 78%, 69.1%, and 67.4%, respectively; and the NRMSE of LAI, AB, and Y were reduced by 78.4%, 70%, and 67%, respectively. Overall, this invention can accurately simulate changes in maize growth.

[0115] Table 1. Comparison of accuracy of different maize growth estimation methods The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method of simulating growth of a crop, characterized by, The method comprises the following steps: obtaining relevant data of a crop database, a meteorological database and a soil database; simulating a three-dimensional leaf model of the crop according to a leaf length of the crop, a leaf width of the crop, a leaf area of the crop, a leaf curling rate of the crop and a leaf distribution of the crop; and simulating a three-dimensional physiological model of the crop according to a leaf area change degree of the crop, an assimilation product of the crop under photosynthesis, a respiration consumption of the crop under respiration, a growth weight increment of the crop and an organ dry weight of the crop; simulating a growth condition of the crop by using a 3D Gaussian splatting technology according to the three-dimensional leaf model of the crop and the three-dimensional physiological model of the crop.

2. A method of simulating the growth of a crop as claimed in claim 1, wherein, The simulating of the three-dimensional leaf model of the crop according to the leaf length of the crop, the leaf width of the crop, the leaf area of the crop, the leaf curling rate of the crop and the leaf distribution of the crop comprises the following steps: inputting an environmental temperature into a leaf length model for describing a leaf growth affected by the environmental temperature to obtain the leaf length; inputting the leaf length into a leaf width model for describing a correlation between the leaf length and the leaf width to obtain the leaf width; simulating the leaf area according to an upper and lower profile curve equation of the leaf; inputting the leaf area of the leaf in the stretched and curled states into a leaf curling rate model for describing an influence of the leaf area on the leaf curling to obtain the leaf curling rate; simulating the leaf distribution according to a probability distribution.

3. A method of simulating the growth of a crop as claimed in claim 1, wherein, The simulating of the three-dimensional physiological model of the crop according to the leaf area change degree of the crop, the assimilation product of the crop under photosynthesis, the respiration consumption of the crop under respiration, the growth weight increment of the crop and the organ dry weight of the crop comprises the following steps: inputting water and nitrogen into a leaf area dynamic model for describing an influence of the water and the nitrogen on the leaf area change to obtain the leaf area change degree; inputting solar radiation, the environmental temperature, the water and the nitrogen into a photosynthesis model for describing an influence of multiple factors on the photosynthesis to obtain the assimilation product under the photosynthesis; obtaining the total respiration consumption according to a respiration consumption required for maintaining life and a respiration consumption required for growth; inputting the assimilation product and the respiration consumption required for maintaining life into an increment model for describing an influence of the photosynthesis and the respiration on the growth weight increment to obtain the growth weight increment; inputting the assimilation product into a dry weight model for describing an influence of a distribution of the assimilation product in various organs on the organ dry weight to obtain the organ dry weight.

4. A method of simulating the growth of a crop as claimed in claim 2, wherein, The leaf curling rate of the crop comprises the following steps: The blade curl RR, The formula is: ； wherein, LA 0 is the potential leaf area, i.e. the maximum leaf area without curling; LA is the actual leaf area, the range of the leaf curling ratio RR is 0-1.

5. A method of simulating the growth of a crop as claimed in claim 2, wherein, The simulating of the leaf distribution according to the probability distribution comprises the following steps: By the probability distribution function f ( x ), the probability of the blade in the interval [-20°, 20°] and [-60°, 60°] is obtained P ; The leaf is divided into multiple stages from low to high, with a probability P Draw a 6-stage or 8-stage leaf in the interval [-20°, 20°]; draw a 6-stage or 8-stage leaf above in the interval [-60°, 60°].

6. A method of simulating the growth of a crop as claimed in claim 3, wherein, The leaf area change degree of the crop comprises the following steps: the leaf area change degree is obtained according to a leaf area index; Leaf area index under conditions of adequate moisture and nitrogen nutrition LAI The acquisition formula is: wherein, L max is the maximum green leaf area of the corn under the desired weather condition, FD is the proportion of the yellow leaf area, DR is the yellow leaf rate, C is the conversion factor of the yellow leaf rate, DVS is the crop biological growth time, a and b are both coefficients; Leaf area index under water and nitrogen nutrient stress conditions RLAI The acquisition formula is: wherein, RLAI(i) , RLAI(i-1) LAI is the leaf area index at time t; i LAI is the leaf area index at time t; i-1 LAI is the leaf area index at time t; LAI(i) , LAI(i-1) LAI is the leaf area index at time t; i LAI is the leaf area index at time t; i-1 LAI is the leaf area index at time t; DLAI LAI is the leaf area index increment; SWF LAI is the leaf area index increment; SNF1 LAI is the leaf area index increment; 7. A method of simulating the growth of a crop as claimed in claim 1, wherein, The simulating of the growth condition of the crop by using the 3D Gaussian splatting technology comprises the following steps: providing leaf simulation data by the three-dimensional leaf model of the crop and providing crop growth and organ dry weight expression morphological construction parameters by the three-dimensional physiological model of the crop; realizing time sequence binding of the leaf simulation data and the morphological construction parameters of the crop by using the 3D Gaussian splatting technology.

8. A simulation system for crop growth, characterized by, The method comprises the following steps: The data acquisition module is configured to acquire relevant data of a crop database, a meteorological database and a soil database. The model simulation module is configured to simulate a three-dimensional leaf model of the crop according to a leaf length of the crop, a leaf width of the crop, a leaf area of the crop, a leaf curling rate of the crop and a leaf distribution of the crop, and simulate a three-dimensional physiological model of the crop according to a leaf area variation degree of the crop, an assimilation product of the crop under photosynthesis, a respiration consumption of the crop under respiration, a growth weight increment of the crop and an organ dry weight of the crop. The crop growth simulation module is configured to simulate a growth condition of the crop by using a 3D Gaussian splashing technology according to the three-dimensional leaf model of the crop and the three-dimensional physiological model of the crop.

Images