Crop water and fertilizer intelligent decision-making method based on multi-modal perception and growth model

Abstract

The application discloses a crop water and fertilizer intelligent decision-making method based on multi-modal perception and growth model and belongs to the technical field of agricultural information intelligent decision-making. Multi-modal perception information of crop plants and gene site expression thresholds related to environmental response are collected, and real-time environmental parameters are obtained. By comparison with the gene site expression thresholds, the expression activity grade of each gene site is determined to generate an environmental influence configuration factor. A preset growth model is called, the perception information is input into the model as an initial state, and the model parameters are corrected by using the environmental influence configuration factor. Future growth state simulation results under different water and fertilizer combination schemes are predicted, deviation accumulations are calculated, the feasibility scores of each scheme are determined, and the highest scheme is selected as a water and fertilizer application instruction output of a current decision-making period. The application introduces an evaluation mechanism of genetic genes responding to the environment, realizes quantitative configuration of the influence of a planting environment on crop growth, and effectively improves the adaptability and reliability of water and fertilizer decision-making.

Description

Technical Field

[0001] This invention discloses a crop water and fertilizer intelligent decision-making method based on multimodal perception and growth model, belonging to the field of agricultural information intelligent decision-making technology. Background Technology

[0002] In precision agriculture, water and fertilizer management is a crucial factor affecting crop yield and quality. Traditional water and fertilizer decision-making methods often rely on growers' experience or trigger irrigation and fertilization based on thresholds from single environmental sensors, lacking a comprehensive understanding of the crop's real-time growth status. While recent attempts have emerged to diagnose crop water requirements using remote sensing images or stem micro-change sensors, existing solutions typically consider only a single sensing modality, failing to comprehensively reflect the crop's water, nutrient, and stress states, resulting in insufficient decision-making basis. Furthermore, most existing water and fertilizer decision-making systems employ fixed crop growth models, generally based on growth curves established under ideal environmental conditions. These models cannot effectively respond to dynamic changes in factors such as temperature, water, and light in real-world planting environments, and do not consider the differentiated response mechanisms of different crop varieties to environmental factors, leading to significant discrepancies between water and fertilizer application instructions and actual crop needs.

[0003] To address the aforementioned issues, some studies have attempted to assimilate crop growth models with environmental sensor data, improving prediction accuracy through continuous model parameter calibration. However, these methods typically rely on extensive historical data for repeated model training or parameter calibration, resulting in high computational complexity and difficulty in scaling to different varieties and planting regions. More importantly, current technology lacks a feasible approach to incorporate the expression thresholds of environmental response gene loci in crop genetic information into water and fertilizer decision-making. Crop growth and development are fundamentally regulated by genetic genes; gene expression activity of the same variety can change under different environmental conditions, thereby affecting water and fertilizer absorption and utilization efficiency.

[0004] Existing decision-making methods ignore this genetic regulatory information, which leads to a disconnect between the decision-making results and the actual tolerance and response characteristics of crops when the planting environment deviates from the suitable range for the variety. This can result in water and fertilizer waste at best, and crop growth and development stagnation or even yield reduction at worst. Summary of the Invention

[0005] To address the aforementioned technical deficiencies, this invention proposes an intelligent water and fertilizer decision-making method that combines multimodal crop perception information, crop genetic environmental response thresholds, dynamic growth model simulation prediction, and environment-gene compatibility evaluation. This method uses multidimensional perception information such as plant canopy thermal infrared images, stem micro-variation diameter, and leaf spectral reflectance as initial states. It then transforms real-time environmental parameters using gene locus expression thresholds of crop varieties to generate environmental impact configuration factors. These factors are used to specifically modify key parameters of the growth model, thereby simulating future growth states under different water and fertilizer combinations. Furthermore, an environment-gene compatibility pre-evaluation mechanism is introduced to dynamically adjust target growth indicators, ensuring that the feasibility evaluation can adapt to a reasonable range of crop growth variation under adverse conditions, ultimately outputting water and fertilizer application instructions. This method does not rely on massive historical data training or data normalization preprocessing operations and can be executed directly on edge computing nodes, significantly improving the adaptability and accuracy of water and fertilizer decisions under different planting environments and crop varieties.

[0006] This application provides a smart crop water and fertilizer decision-making method based on multimodal perception and growth model, including the following steps: Step 1: Collect real-time information on multiple sensory modalities of the crop plant, including thermal infrared images of the plant canopy, micro-variable diameter of the stem, and spectral reflectance of the leaves; Step 2: Extract gene locus expression thresholds related to environmental response based on the genetic information of crop varieties, and obtain real-time parameters of the planting environment, including soil moisture content, air temperature and light intensity. Step 3: Compare the real-time parameters with the gene locus expression thresholds to determine the expression activity level of each gene locus, and generate environmental influence configuration factors based on the expression activity levels; Step 4: Call the preset growth model, input the real-time information of the current crop plant's perception mode as the initial state into the growth model, and use the environmental influence configuration factor to correct the parameters of the growth model to predict the simulation results of the crop's growth state under various water and fertilizer combination schemes. Step 5: For each water and fertilizer combination scheme, extract the key morphological parameters and physiological parameters from the corresponding growth state simulation results, compare them with the target growth indicators at the corresponding time nodes, calculate the cumulative deviation between each parameter and the target indicator, and determine the feasibility score of the water and fertilizer combination scheme based on the cumulative deviation. Step 6: Select the water and fertilizer combination scheme with the highest feasibility score as the water and fertilizer application instruction output for the current decision-making cycle.

[0007] Furthermore, step two specifically includes: The genome sequence of a crop variety is analyzed to identify at least one gene locus that is regulated by environmental factors, including temperature, water and light. For each identified gene locus, its corresponding expression regulatory region is obtained, which consists of the minimum activation threshold and the maximum inhibition threshold; When the environmental factor parameter is below the minimum activation threshold, the gene locus is in a non-expressed state; when the environmental factor parameter is between the minimum activation threshold and the maximum inhibition threshold, the gene locus is in a normally expressed state; when the environmental factor parameter is above the maximum inhibition threshold, the gene locus is in an overexpression and inhibition state. The expression regulatory regions of each gene locus are categorized and stored according to the type of environmental factor, forming a database of gene expression thresholds for crop responses to specific environmental factors under the current planting environment.

[0008] Furthermore, step three specifically includes: For each gene locus, select the real-time parameter value of the environmental factor type corresponding to that gene locus; compare the real-time parameter value with the lowest activation threshold and the highest inhibition threshold of that gene locus. If the real-time parameter value is lower than the minimum activation threshold, the expression activity level of the gene locus is determined to be dormant. If the real-time parameter value is higher than or equal to the minimum activation threshold and lower than or equal to the maximum inhibition threshold, the expression activity level of the gene locus is determined to be activated. If the real-time parameter value is higher than the maximum inhibition threshold, the expression activity level of the gene locus is determined to be inhibited. A weight coefficient is assigned to each expression activity level, with the first weight coefficient corresponding to the dormant level, the second weight coefficient corresponding to the active level, and the third weight coefficient corresponding to the inhibitory level. The weight coefficients of all gene loci are synthesized according to their corresponding environmental factor types to generate environmental impact configuration factor vectors for three types of environmental factors: temperature, moisture, and light.

[0009] Furthermore, step four specifically includes: Read the preset growth model, which includes a physiological development time function to describe crop dry matter accumulation, a morphogenesis function to describe crop organ elongation, a source-sink relationship function to describe photosynthetic product allocation, and an accumulated temperature function to describe phenological stage transitions. The weighted value of the daily temperature contribution to physiological development time is adjusted according to the value of the temperature factor, so that the physiological development time under high temperature or low temperature environment is compressed or stretched relative to the optimal temperature environment. The baseline values ​​of stem diameter growth rate and leaf area expansion rate are adjusted according to the water factor values ​​to proportionally slow down the morphogenesis rate under water stress compared to that under adequate water supply. The photosynthetic carbon fixation efficiency per unit leaf area is adjusted according to the value of the light factor so that the dry matter accumulation rate under low light conditions is proportionally reduced compared to the sufficient light conditions. Adjusting the cumulative temperature threshold required for the transition of each phenological stage based on the combined effects of three environmental factors can advance or delay the phenological process of crops under adverse conditions. After the correction is completed, the real-time information of the current crop plant's perception modality is used as the initial state value and input into the corrected growth model. The water and fertilizer supply parameters corresponding to various water and fertilizer combination schemes are substituted into the model to calculate the values ​​of crop biomass, organ size and phenological stage at multiple time nodes in the future, and the growth state simulation results under each water and fertilizer combination scheme are obtained.

[0010] Furthermore, the key morphological parameters extracted in the growth state simulation results in step five include plant height, stem diameter, and canopy coverage, while the physiological parameters include leaf chlorophyll content and dry matter accumulation. The target growth indicators for the corresponding time nodes are pre-set based on the ideal plant type standard and growth stage standard in the genetic information of the crop variety. The cumulative deviation between each parameter and the target index is calculated, specifically including: for each time node, the absolute value of the first difference between the simulated plant height and the target plant height, the absolute value of the second difference between the simulated stem diameter and the target diameter, the absolute value of the third difference between the simulated canopy coverage and the target coverage, the absolute value of the fourth difference between the simulated leaf chlorophyll content and the target chlorophyll content, and the absolute value of the fifth difference between the simulated dry matter accumulation and the target accumulation. Multiply the absolute values ​​of the first, second, third, fourth, and fifth differences by their respective parameter weighting coefficients and then sum them to obtain the single-point deviation value at that time node. The cumulative deviation is obtained by summing up the single-point deviation values ​​of all future time points. The cumulative deviation is compared with a preset feasibility threshold range. If the cumulative deviation is less than the first threshold, it is judged as a high feasibility score; if it is between the first and second thresholds, it is judged as a medium feasibility score; and if it is greater than the second threshold, it is judged as a low feasibility score.

[0011] Furthermore, step six specifically includes: Compare the feasibility scores of all water and fertilizer combination schemes, and select the scheme with the highest feasibility score as the preferred scheme. The water and fertilizer application parameters extracted from this preferred scheme include single irrigation water volume, single fertilizer application volume, water and fertilizer mixing ratio, application interval days, and application time period; The water and fertilizer application parameters are formatted according to the start time of the current decision cycle to generate a water and fertilizer application instruction sequence, which includes multiple application time points and their corresponding water and fertilizer parameters within the decision cycle. The water and fertilizer application instruction sequence is sent to the field water and fertilizer execution equipment, which then performs automatic irrigation and fertilization operations according to the time points and parameters in the instruction sequence.

[0012] Furthermore, before calculating the cumulative deviation for each water-fertilizer combination scheme in step five, an environmental-gene compatibility pre-evaluation is performed, including: The values ​​of temperature, water and light factors in the environmental influence configuration factors were extracted and compared with the normal expression activity range of each environmental response gene locus in the obtained crop genetic information. Determine whether all components of the current environmental impact configuration factor fall within the normal expression activity range of the corresponding gene locus; If any configuration factor component of an environmental factor exceeds the normal expression activity range of the corresponding gene locus, it is determined that the current planting environment is incompatible with the genetic makeup of the crop variety. For each environmental factor that exceeds the normal expression activity range, calculate the minimum deviation of its configuration factor component from the boundary of that range, and then take the geometric mean of the minimum deviation of each environmental factor to obtain the incompatibility value. Based on the range of the incompatibility value, a response strategy is selected from a variety of preset strategies. If the incompatibility value is below the first threshold, the simulation and prediction process of the original water and fertilizer combination scheme is maintained, but an environmental stress penalty term is added to the subsequent calculation of the cumulative deviation. If the incompatibility value is between the first and second thresholds, the parameters of the gene expression regulation module in the growth model are temporarily adjusted to adjust the prediction results of the growth model to include phenotypic plasticity parameters under environmental stress. If the incompatibility value is above the second threshold, the feasibility evaluation of the current water and fertilizer combination scheme is stopped directly, and its feasibility score is set to infeasible. The water and fertilizer combination schemes that have undergone pre-evaluation are sent to the deviation accumulation calculation step, and it is ensured that only schemes that have not been judged as infeasible participate in the feasibility score competition.

[0013] Furthermore, before outputting the water and fertilizer application instruction in step six, a condition verification step is also included: Obtain forecast information on future changes in the planting environment within the current decision-making cycle, including the temperature forecast sequence, precipitation probability, and light intensity forecast for the next few days; The irrigation and fertilization time points required in the selected water and fertilizer combination scheme are compared with the future change prediction information. If a heavy rainfall event exceeding the crop flood tolerance threshold is predicted to occur within the predetermined time window after irrigation, the irrigation amount in the water and fertilizer combination scheme is adjusted to the reduced corrected irrigation amount. If it is predicted that there will be high temperature and drought weather exceeding the crop drought tolerance threshold within the predetermined time window after fertilization, the fertilization method in the water and fertilizer combination scheme will be changed from one-time application to multiple applications. After the adjustments are completed, the final water and fertilizer application instructions will be output.

[0014] Furthermore, the preset growth model is a crop growth simulator based on a function-structure model, which includes a three-dimensional virtual plant structure module, a photosynthesis module, a respiration consumption module, a carbon and nitrogen allocation module, and an organogenesis module. The three-dimensional virtual plant structure module generates plant nodes and internodes according to crop topology rules. The photosynthesis module calculates the photosynthetic rate of each leaf based on light and CO2 concentration. The respiration consumption module calculates the consumption rate of maintenance respiration and growth respiration based on temperature. The carbon and nitrogen allocation module allocates photosynthetic products to various organs such as roots, stems, leaves, and fruits according to source-sink relationships and phenological stages. The organogenesis module controls new leaf germination and flower bud differentiation based on accumulated heat units and genotype parameters. When calling the growth model, the soil moisture supply curves and nitrogen supply curves corresponding to different water and fertilizer combination schemes are used as boundary conditions to input the carbon and nitrogen allocation module and the organ development module. The soil moisture supply curve is determined by irrigation amount and soil evapotranspiration model, and the nitrogen supply curve is determined by fertilizer amount and soil nitrogen mineralization model.

[0015] Furthermore, the target growth indicators for the corresponding time nodes in step five are dynamically adjusted based on the environmental impact configuration factors generated in step three: For temperature factors, if the temperature factor indicates that the current environment deviates from the optimal growth temperature of the crop, the plant height growth rate and leaf area expansion rate in the target growth indicators will be reduced. The reduction magnitude is non-linearly positively correlated with the degree of deviation of the temperature factor, and the more severe the deviation, the faster the reduction magnitude decays. For the water factor, if the water factor indicates that there is water stress in the current environment, the stem diameter and dry matter accumulation in the target growth indicators are lowered, and the root-shoot ratio is raised, as a mapping of the crop's adaptive strategy of prioritizing root growth in a water-deficient environment. For the light factor, if the light factor indicates that the current ambient light intensity is insufficient, the canopy coverage and leaf chlorophyll content in the target growth indicators will be adjusted upwards, and the proportion of photosynthetic products per unit dry matter will be adjusted downwards. The dynamically adjusted target growth indicators are used for subsequent calculations of cumulative deviations, enabling the feasibility assessment to adapt to a reasonable range of crop growth variation under different environmental conditions.

[0016] This invention discloses an intelligent decision-making method for crop water and fertilizer based on multimodal perception and growth models, belonging to the field of agricultural information intelligent decision-making technology. It collects multimodal perception information of crop plants and gene locus expression thresholds related to environmental responses, and obtains real-time environmental parameters. By comparing these parameters with the gene locus expression thresholds, the expression activity level of each gene locus is determined, generating environmental impact configuration factors. A preset growth model is invoked, with the perceived information used as the initial state input to the model. The model parameters are then corrected using the environmental impact configuration factors. The simulation results of future growth states under different water and fertilizer combinations are predicted, the cumulative deviation is calculated, the feasibility score of each scheme is determined, and the scheme with the highest score is selected as the water and fertilizer application instruction for the current decision cycle. This invention, by introducing an evaluation mechanism for genetic responses to the environment, achieves quantitative configuration of the impact of the planting environment on crop growth, effectively improving the adaptability and reliability of water and fertilizer decisions. Attached Figure Description

[0017] Figure 1 This is a flowchart of a crop water and fertilizer intelligent decision-making method based on multimodal perception and growth model disclosed in an embodiment of this application; Figure 2 This is a second flowchart of a crop water and fertilizer intelligent decision-making method based on multimodal perception and growth model disclosed in an embodiment of this application; Figure 3 This is the third workflow diagram of a crop water and fertilizer intelligent decision-making method based on multimodal perception and growth model disclosed in the embodiments of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] The technical solutions disclosed in the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0020] According to the first embodiment of the present invention, referring to Figure 1 This invention claims protection for a crop water and fertilizer intelligent decision-making method based on multimodal perception and growth model, comprising the following steps: Step 1: Collect real-time information on multiple sensory modalities of the crop plant, including thermal infrared images of the plant canopy, micro-variable diameter of the stem, and spectral reflectance of the leaves; Step 2: Extract gene locus expression thresholds related to environmental response based on the genetic information of crop varieties, and obtain real-time parameters of the planting environment, including soil moisture content, air temperature and light intensity. Step 3: Compare the real-time parameters with the gene locus expression thresholds to determine the expression activity level of each gene locus, and generate environmental influence configuration factors based on the expression activity levels; Step 4: Call the preset growth model, input the real-time information of the current crop plant's perception mode as the initial state into the growth model, and use the environmental influence configuration factor to correct the parameters of the growth model to predict the simulation results of the crop's growth state under various water and fertilizer combination schemes. Step 5: For each water and fertilizer combination scheme, extract the key morphological parameters and physiological parameters from the corresponding growth state simulation results, compare them with the target growth indicators at the corresponding time nodes, calculate the cumulative deviation between each parameter and the target indicator, and determine the feasibility score of the water and fertilizer combination scheme based on the cumulative deviation. Step 6: Select the water and fertilizer combination scheme with the highest feasibility score as the water and fertilizer application instruction output for the current decision-making cycle.

[0021] In this embodiment, a mobile monitoring platform equipped with a thermal infrared camera is installed in the crop planting area. This platform collects thermal infrared images of the crop canopy daily at noon according to a preset path, obtaining a canopy temperature distribution map. Simultaneously, a laser micro-displacement sensor is installed at the base of each crop stem to continuously record minute changes in stem diameter over 24 hours, outputting a stem diameter variation signal. Furthermore, using a handheld or fixed spectrometer, on a clear morning at 10:00 AM, the reflectance of crop leaves in the visible to near-infrared bands is measured, obtaining leaf spectral reflectance curves. The sensing information from these three modalities is timestamped and stored in a local database.

[0022] Read the pre-entered whole-genome sequencing data of the current crop variety and screen for known HSP gene loci associated with temperature stress response, DREB gene loci associated with water stress response, and PHY gene loci associated with light response. For each gene locus, extract the expression threshold conditions from its annotation file. For example, for temperature-responsive gene loci, extract the minimum activation threshold as 15 degrees Celsius of daily average temperature and the maximum inhibition threshold as 35 degrees Celsius of daily average temperature; for water-responsive gene loci, extract the minimum activation threshold as 25% of soil volumetric water content and the maximum inhibition threshold as 70% of soil volumetric water content; for light-responsive gene loci, extract the minimum activation threshold as 100 moles of photosynthetically active radiation per square meter per second and the maximum inhibition threshold as 800 moles of photosynthetically active radiation per square meter per second. These extracted thresholds are categorized and stored according to three environmental factors: temperature, moisture, and light, forming a gene expression threshold database specific to this crop variety. Simultaneously, real-time readings of soil moisture content are obtained through soil moisture sensors deployed in the field, air temperature values ​​above the canopy are obtained through air temperature sensors, and instantaneous values ​​of photosynthetically active radiation are obtained through light sensors.

[0023] For temperature-responsive gene loci, the current real-time air temperature is retrieved, assuming it is 28 degrees Celsius. This value is compared with the pre-extracted minimum activation threshold of 15 degrees Celsius and maximum inhibition threshold of 35 degrees Celsius. Since 28 degrees Celsius falls between 15 and 35, the expression activity level of this gene locus is determined to be activated. For water-responsive gene loci, the current soil volumetric water content is retrieved, assuming it is 20%, which is lower than the minimum activation threshold of 25%, thus the expression activity level of this gene locus is determined to be dormant. For light-responsive gene loci, the current photosynthetically active radiation value is retrieved, assuming it is 900 moles per square meter. If the expression level of the gene locus exceeds the maximum inhibition threshold of 800 moles per square meter per second, the expression activity level is determined to be the inhibition level. Next, a preset weight coefficient is assigned to each activity level: the weight coefficient for the dormant level is 0.3, the weight coefficient for the activated level is 1.0, and the weight coefficient for the inhibition level is 0.5. Then, the weight coefficients are synthesized according to the type of environmental factor: the synthesized value corresponding to the temperature factor is 1.0, the synthesized value corresponding to the moisture factor is 0.3, and the synthesized value corresponding to the light factor is 0.5. These synthesized values ​​are combined into a three-dimensional vector, namely the environmental influence configuration factor, denoted as temperature factor = 1.0, moisture factor = 0.3, and light factor = 0.5.

[0024] Step 4 Specific Implementation: An internally pre-stored rice growth model based on physiological development time is used. This model includes an exponential growth function describing dry matter accumulation, a logistic function describing stem elongation, and an accumulated temperature accumulation function describing phenological stages. First, real-time information on the current crop's sensory modality is read: the canopy thermal infrared image shows an average canopy temperature of 26 degrees Celsius, the current stem micro-variable diameter is 5.2 mm, and the normalized difference vegetation index calculated from the leaf spectral reflectance is 0.75. This information is used as the initial state variables of the growth model. Then, the environmental impact configuration factors generated in Step 3 (1.0, 0.3, ...) are applied. The parameter correction process for the growth model is as follows: For the temperature factor of 1.0, the current temperature is considered suitable, and no adjustment is made to the heat effect parameter in the physiological development time function; for the water factor of 0.3, the baseline value of the cell expansion rate in the morphogenesis function is multiplied by 0.3, reducing the daily stem diameter growth rate from the preset 0.2 mm to 0.06 mm; for the light factor of 0.5, the photosynthetic product synthesis rate parameter is multiplied by 0.5, resulting in a halving of the daily dry matter accumulation. This corrected growth model is then substituted into five water and fertilizer combination schemes, each specifying the daily irrigation and fertilization amounts for the next ten days. Starting from the current date, the crop biomass, stem diameter, and phenological stage values ​​for each of the next ten days are calculated sequentially, and the growth state simulation results for each time point are output.

[0025] Extract plant height, stem diameter, canopy coverage, leaf chlorophyll content, and dry matter accumulation from the growth state simulation results for three future time points: day 3, day 7, and day 10. Simultaneously, retrieve the ideal values ​​for the crop variety at the corresponding time points from the target growth index database; for example, the ideal plant height on day 7 is 30 cm, and the ideal stem diameter is 6 mm. For each water and fertilizer combination scheme, calculate the absolute value of the difference between the simulated plant height and the ideal value on day 7, the absolute value of the difference between the simulated stem diameter and the ideal value, and so on. Multiply the absolute value of the difference for each parameter by a pre-set parameter weight, for example, plant height weight 0.2, stem diameter weight 0.3, canopy coverage weight 0.1, chlorophyll content weight 0.2, and dry matter accumulation weight 0.2, and sum them to obtain the single-point deviation value for that time point. Accumulate the single-point deviation values ​​for days 3, 7, and 10 to obtain the cumulative deviation of the water and fertilizer combination scheme. The cumulative deviation is compared with a preset feasibility threshold range: a cumulative deviation of less than 50 dB is a high feasibility score, between 50 and 100 is a medium feasibility score, and greater than 100 is a low feasibility score.

[0026] The feasibility scores of the five water and fertilizer combination schemes were compared. Scheme 3 was assumed to have a cumulative deviation of 45, corresponding to a high feasibility score, higher than the other schemes. The water and fertilizer application parameters recorded in Scheme 3 were read: the single irrigation water volume was 2 liters per plant, the single fertilizer application amount was 5 grams of compound fertilizer per plant, the water and fertilizer mixing ratio was 2.5 grams of fertilizer dissolved in 1 liter of water, the application interval was 3 days, and the application time was from 7:00 to 8:00 in the morning. The start date of the current decision cycle was taken as the first day, and a water and fertilizer application instruction sequence containing three time points was generated: water and fertilizer mixing application was performed once each at 7:30 in the morning on the 1st, 4th, and 7th days. This instruction sequence was sent to the controller of the field water and fertilizer execution equipment in the form of a control message. After receiving the message, the controller parsed each time point and parameter, and drove the water pump and fertilizer valve to perform irrigation and fertilization operations according to the instructions.

[0027] Furthermore, step two specifically includes: The genome sequence of a crop variety is analyzed to identify at least one gene locus that is regulated by environmental factors, including temperature, water and light. For each identified gene locus, its corresponding expression regulatory region is obtained, which consists of the minimum activation threshold and the maximum inhibition threshold; When the environmental factor parameter is below the minimum activation threshold, the gene locus is in a non-expressed state; when the environmental factor parameter is between the minimum activation threshold and the maximum inhibition threshold, the gene locus is in a normally expressed state; when the environmental factor parameter is above the maximum inhibition threshold, the gene locus is in an overexpression and inhibition state. The expression regulatory regions of each gene locus are categorized and stored according to the type of environmental factor, forming a database of gene expression thresholds for crop responses to specific environmental factors under the current planting environment.

[0028] In this embodiment, the genome sequence file of the crop variety is obtained from the variety database provided by the seed company. The file format is the standard FASTA format. A gene annotation module is invoked. This module maintains a gene locus marker library related to environmental factors. The marker library contains multiple known response gene names and their corresponding chromosomal location intervals. The genome sequence is searched for sequence regions that match the marker library. Three temperature-responsive gene loci, such as OsHSP17.3, OsHSP70, and OsHSP90, two water-responsive gene loci, such as OsDREB1A and OsDREB2A, and one light-responsive gene locus, such as OsPHYA, are identified in sequence.

[0029] For each identified gene locus, the cis-acting element sequence in its upstream promoter region is read. For temperature-responsive gene loci, the heat-shock element (HSE) core sequence is located in its promoter region, and the minimum activation threshold and maximum repression threshold of the gene locus are determined based on the conserved base patterns flanking this core sequence. A pre-defined threshold mapping rule table is used, which records the correspondence between combinations of different element types and temperature thresholds. For example, when the promoter region contains two tandem HSEs separated by five bases, the minimum activation threshold is 30 degrees Celsius; when it contains three tandem HSEs separated by ten bases, the minimum activation threshold is 28 degrees Celsius. The specific minimum activation threshold is calculated for each temperature-responsive gene locus according to this rule. To determine the maximum repression threshold, the presence of a negative regulatory element in the terminator region of the gene locus is analyzed. If an ABRE-related motif is present, the maximum repression threshold is set to 40 degrees Celsius.

[0030] For water-responsive gene loci, the dehydration response element DRE and the ABA response element ABRE are searched in their promoter regions. The minimum activation threshold is set according to the copy number of the DRE element: when the copy number is 1, the minimum activation threshold is 30% of the soil moisture content, and when the copy number is 2, the minimum activation threshold is 25% of the soil moisture content. The maximum inhibition threshold is set according to the predicted methylation level of the promoter region: when the methylation level is higher than 60%, the maximum inhibition threshold is 65%.

[0031] For light-responsive gene sites, G-box and I-box elements are located in their promoter regions. The minimum activation threshold is set based on the base distance between the G-box and I-box: when the distance is less than 50 bases, the minimum activation threshold is 50 moles per square meter per second; when the distance is greater than 100 bases, the minimum activation threshold is 200 moles per square meter per second. The maximum inhibition threshold is set based on the presence of a bidirectional transcription termination signal; if present, the maximum inhibition threshold is 800 moles per square meter per second.

[0032] After extracting the aforementioned thresholds, the expression regulation ranges of all gene loci are categorized according to the type of environmental factor, and two index tables are established: one table uses the environmental factor type as the primary key to store all gene loci belonging to that type and their thresholds; the other table uses the gene locus name as the primary key to store its corresponding environmental factor type and threshold. These two tables are persistently stored on the local disk, forming a gene expression threshold database for this crop variety under specific planting conditions. In subsequent steps, when real-time parameters arrive, the thresholds of all relevant gene loci can be quickly retrieved directly based on the environmental factor type.

[0033] Furthermore, step three specifically includes: For each gene locus, select the real-time parameter value of the environmental factor type corresponding to that gene locus; compare the real-time parameter value with the lowest activation threshold and the highest inhibition threshold of that gene locus. If the real-time parameter value is lower than the minimum activation threshold, the expression activity level of the gene locus is determined to be dormant. If the real-time parameter value is higher than or equal to the minimum activation threshold and lower than or equal to the maximum inhibition threshold, the expression activity level of the gene locus is determined to be activated. If the real-time parameter value is higher than the maximum inhibition threshold, the expression activity level of the gene locus is determined to be inhibited. A weight coefficient is assigned to each expression activity level, with the first weight coefficient corresponding to the dormant level, the second weight coefficient corresponding to the active level, and the third weight coefficient corresponding to the inhibitory level. The weight coefficients of all gene loci are synthesized according to their corresponding environmental factor types to generate environmental impact configuration factor vectors for three types of environmental factors: temperature, moisture, and light.

[0034] In this embodiment, a two-dimensional array is established, with rows corresponding to each identified gene locus and columns corresponding to three environmental factor types. For the first locus, OsHSP17.3, in the temperature response gene locus, the current air temperature value is retrieved from the real-time environmental parameters, assuming it to be 22 degrees Celsius. The gene expression threshold database shows that the minimum activation threshold of OsHSP17.3 is 15 degrees Celsius, and the maximum inhibition threshold is 35 degrees Celsius. A three-branch comparison operation is performed: the first branch checks if 22 degrees Celsius is less than 15 degrees Celsius, and the result is no; the second branch checks if 22 degrees Celsius is greater than or equal to 15 degrees Celsius and less than or equal to 35 degrees Celsius, and the result is yes; therefore, the expression activity level of this gene locus is set to the activated level. For the second temperature-responsive gene locus, OsHSP70, assuming its minimum activation threshold is 18 degrees Celsius and its maximum inhibition threshold is 30 degrees Celsius, the current 22 degrees Celsius is also between the two, so it is also determined to be in the activated level. For the third temperature-responsive gene locus, OsHSP90, assuming its minimum activation threshold is 23 degrees Celsius and its maximum inhibition threshold is 32 degrees Celsius, the current 22 degrees Celsius is below the minimum activation threshold, so it is determined to be in the dormant level.

[0035] For the water-responsive gene locus OsDREB1A, with a current soil volumetric water content of 18%, its minimum activation threshold is 25%, and its maximum inhibition threshold is 70%. A comparison was performed: 18% is less than 25%, so the result is true, and it is classified as dormant. For OsDREB2A, its minimum activation threshold is 20%, and 18% is less than 20%, so it is also classified as dormant.

[0036] For the light-responsive gene locus OsPHYA, the current photosynthetically active radiation is 900 moles per square meter per second, its minimum activation threshold is 100, and its maximum inhibition threshold is 800; a comparison is performed: 900 is greater than 800, so it is judged to be in the inhibition level.

[0037] After determining the activity level of all gene loci, the weighting coefficient allocation stage begins. A mapping table from level to weight is pre-defined: dormant level corresponds to the first weight coefficient of 0.2, active level corresponds to the second weight coefficient of 1.0, and inhibited level corresponds to the third weight coefficient of 0.4. For the temperature factor, the activity levels of all temperature-responsive gene loci are collected: OsHSP17.3 is the active level of 1.0, OsHSP70 is the active level of 1.0, and OsHSP90 is the dormant level of 0.2. The arithmetic mean of these weight coefficients is taken, i.e., 1.0 + 1.0 + 0.2 / 3 = 0.733, which is used as the environmental influence configuration factor component of the temperature factor. For the moisture factor, OsDREB1A and OsDREB2A are both dormant levels of 0.2 and 0.2, respectively, with an average of 0.2, which is used as the moisture factor component. For the illumination factor, OsPHYA is the suppression level of 0.4, which is used as the illumination factor component. These three components are combined into a three-dimensional vector of 0.733, 0.2, 0.4, and this vector is used as the basis for the correction of the growth model parameters in step four.

[0038] If no gene loci are identified under a certain environmental factor type, the configuration factor component of that environmental factor is set to 0.8 by default, indicating a neutral effect without genetic information constraints; at the same time, the calculation source of each factor is recorded so that it can be used for environmental-gene compatibility judgment in subsequent steps.

[0039] Furthermore, referring to Figure 2 Step four specifically includes: Read the preset growth model, which includes a physiological development time function to describe crop dry matter accumulation, a morphogenesis function to describe crop organ elongation, a source-sink relationship function to describe photosynthetic product allocation, and an accumulated temperature function to describe phenological stage transitions. The weighted value of the daily temperature contribution to physiological development time is adjusted according to the value of the temperature factor, so that the physiological development time under high temperature or low temperature environment is compressed or stretched relative to the optimal temperature environment. The baseline values ​​of stem diameter growth rate and leaf area expansion rate are adjusted according to the water factor values ​​to proportionally slow down the morphogenesis rate under water stress compared to that under adequate water supply. The photosynthetic carbon fixation efficiency per unit leaf area is adjusted according to the value of the light factor so that the dry matter accumulation rate under low light conditions is proportionally reduced compared to the sufficient light conditions. Adjusting the cumulative temperature threshold required for the transition of each phenological stage based on the combined effects of three environmental factors can advance or delay the phenological process of crops under adverse conditions. After the correction is completed, the real-time information of the current crop plant's perception modality is used as the initial state value and input into the corrected growth model. The water and fertilizer supply parameters corresponding to various water and fertilizer combination schemes are substituted into the model to calculate the values ​​of crop biomass, organ size and phenological stage at multiple time nodes in the future, and the growth state simulation results under each water and fertilizer combination scheme are obtained.

[0040] In this embodiment, the preset growth model read is a crop simulator based on the WOFOST model. This simulator contains four core sub-functions: the physiological development time function takes the daily average temperature as input and outputs the cumulative physiological development days; the morphogenesis function takes the physiological development time and water stress coefficient as input and outputs the daily increment of stem diameter and leaf area; the source-sink relationship function takes the leaf area index and light intensity as input and outputs the total amount of photosynthetic products; and the accumulated temperature function compares the daily accumulated temperature value with the phenological transition threshold to determine whether to enter the next phenological stage.

[0041] The first step in the correction process is to extract the temperature factor of 0.733 from the environmental impact configuration factors and transform it: the new temperature response coefficient = 0.733 × 1.2. This coefficient is then applied to the effective accumulated temperature calculation parameter in the physiological development time function. The original calculation method for the physiological development time function was: Daily physiological development time = Daily average temperature / Baseline temperature (set to 25 degrees Celsius). This method is modified to: Daily physiological development time = Temperature response coefficient × Daily average temperature / Baseline temperature. When the daily average temperature is 20 degrees Celsius, the daily physiological development time before the modification was 0.8, and after the modification it is 0.733 × 0.8 = 0.586. This means that under the same temperature conditions, the physiological development time of the crop has been compressed, corresponding to the growth delay when the temperature factor is below the optimum value.

[0042] The second step of the correction operation is to extract the water factor of 0.2 from the environmental influence configuration factors and apply it to the cell expansion rate parameter in the morphogenesis function. The morphogenesis function originally has a stem diameter daily growth rate parameter with a default value of 0.15 mm per day. Multiplying the default value by the water factor of 0.2, the corrected stem diameter daily growth rate is 0.03 mm per day. Similarly, the baseline value of the leaf area expansion rate is 10 square centimeters per plant per day, which is corrected to 10 multiplied by 0.2 equals 2 square centimeters per day. Replacing the original parameters in the function with these corrected rate values ​​significantly reduces the organ expansion rate under water stress.

[0043] The third step of the correction operation is to extract the light factor of 0.4 from the environmental impact configuration factors and apply it to the photosynthetic product synthesis rate parameter in the source-sink relationship function. The source-sink relationship function has a light response curve subfunction, whose default value for the maximum photosynthetic rate parameter is five grams of dry matter synthesized per square meter of leaf area per hour. Multiplying this parameter by the light factor of 0.4, the corrected maximum photosynthetic rate is two grams of dry matter per square meter per hour. At the same time, the light saturation point parameter is also adjusted, reducing the original light saturation point value to 2 / 5 of the original value. This means that in an environment with insufficient light, the photosynthetic capacity is proportionally inhibited.

[0044] The fourth step of the correction operation involves applying the temperature factor (0.733), water factor (0.2), and light factor (0.4) to the phenological threshold parameter in the accumulated temperature function. The original accumulated temperature function required 280 degree-days from emergence to jointing and 350 degree-days from jointing to heading. The combined product of the three environmental factors is calculated as: 0.733 × 0.2 × 0.4 = 0.05864. This product is used as the threshold correction coefficient, and the original accumulated phenological threshold is multiplied by this coefficient. For example, the accumulated threshold from emergence to jointing becomes approximately 16.4 degree-days (280 multiplied by 0.05864). This extremely low product means that under multiple adverse conditions, the phenological process will be extremely accelerated, and the crop will enter the next stage earlier. To prevent logical errors caused by an excessively small threshold, a lower limit is set before multiplication: if the product of the three factors is less than 0.2, then 0.2 is forcibly used as the correction coefficient. Therefore, 0.2 is actually used, and the adjusted threshold is 56 degree-days.

[0045] After completing the above four corrections, the stem diameter of 5.2 mm in the current sensing modality is read as the initial stem diameter, and the leaf area index of 2.3 calculated from spectral reflectance is read as the initial leaf area. The corrected growth model is then combined with all water and fertilizer combination schemes one by one. For each scheme, starting from the current date, photosynthetic products are calculated daily based on the temperature and light intensity of the day, and morphological increments are calculated based on the water factor of the day. The stem diameter and leaf area are updated, and the cumulative degree-days are updated simultaneously. It is then checked whether the phenological threshold has been crossed. All state variables at multiple future time points are calculated sequentially, and the output results at each time point are stored.

[0046] Furthermore, the key morphological parameters extracted in the growth state simulation results in step five include plant height, stem diameter, and canopy coverage, while the physiological parameters include leaf chlorophyll content and dry matter accumulation. The target growth indicators for the corresponding time nodes are pre-set based on the ideal plant type standard and growth stage standard in the genetic information of the crop variety. The cumulative deviation between each parameter and the target index is calculated, specifically including: for each time node, the absolute value of the first difference between the simulated plant height and the target plant height, the absolute value of the second difference between the simulated stem diameter and the target diameter, the absolute value of the third difference between the simulated canopy coverage and the target coverage, the absolute value of the fourth difference between the simulated leaf chlorophyll content and the target chlorophyll content, and the absolute value of the fifth difference between the simulated dry matter accumulation and the target accumulation. Multiply the absolute values ​​of the first, second, third, fourth, and fifth differences by their respective parameter weighting coefficients and then sum them to obtain the single-point deviation value at that time node. The cumulative deviation is obtained by summing up the single-point deviation values ​​of all future time points. The cumulative deviation is compared with a preset feasibility threshold range. If the cumulative deviation is less than the first threshold, it is judged as a high feasibility score; if it is between the first and second thresholds, it is judged as a medium feasibility score; and if it is greater than the second threshold, it is judged as a low feasibility score.

[0047] In this embodiment, the record corresponding to each future time node is read from the generated growth state simulation result file. Assuming that the fifth, tenth, fifteenth, and thirtieth days are selected as sampling points for deviation calculation, the following five parameters are parsed for each sampling point: plant height (in centimeters, directly read from the plant height field of the simulation results); stem diameter (in millimeters, read from the stem thickness field); canopy coverage (in fractions between 0 and 1, read from the projected coverage ratio field); leaf chlorophyll content (in chlorophyll SPAD value field); and accumulated dry matter (in grams per plant, read from the aboveground biomass field).

[0048] Next, target growth indicators are acquired; these indicators are pre-stored in a target database categorized by crop variety and growth stage. For the current crop variety, the database records target values ​​for the fifth, tenth, fifteenth, and thirtieth days after sowing. For example, on the fifth day, the target plant height is 15 cm, the target stem diameter is 3 mm, the target canopy coverage is 0.1, the target chlorophyll SPAD value is 25, and the target dry matter accumulation is 3 g; on the tenth day, the target plant height is 28 cm, the target stem diameter is 4.5 mm, the target canopy coverage is 0.25, the target chlorophyll SPAD value is 30, and the target dry matter accumulation is 8 g; on the fifteenth day, the target plant height is 40 cm, the target stem diameter is 6 mm, the target canopy coverage is 0.45, the target chlorophyll SPAD value is 35, and the target dry matter accumulation is 15 g; on the thirtieth day, the target plant height is 65 cm, the target stem diameter is 8 mm, the target canopy coverage is 0.7, the target chlorophyll SPAD value is 40, and the target dry matter accumulation is 30 g.

[0049] For each water and fertilizer combination scheme, the following deviation calculation operations are performed at each sampling point: Calculate the absolute value of the difference between the simulated plant height and the target value, denoted as difference A; calculate the absolute value of the difference between the simulated stem diameter and the target value, denoted as difference B; calculate the absolute value of the difference between the simulated canopy coverage and the target value, denoted as difference C; calculate the absolute value of the difference between the simulated chlorophyll content and the target value, denoted as difference D; calculate the absolute value of the difference between the simulated dry matter accumulation and the target value, denoted as difference E. A set of preset weighting coefficients is used: plant height weight 0.15, stem diameter weight 0.25, canopy coverage weight 0.1, chlorophyll content weight 0.2, and dry matter accumulation weight 0.3; the single-point deviation value calculation formula for this time node is: single-point deviation value = 0.15 × difference A + 0.25 × difference B + 0.1 × difference C + 0.2 × difference D + 0.3 × difference E. The single-point deviation values ​​of the above four sampling points are calculated and then added in chronological order to obtain the cumulative deviation of the water and fertilizer combination scheme.

[0050] After the cumulative deviation is calculated, it is compared with three preset thresholds: the acceptable threshold T1 is set to 25, the warning threshold T2 is set to 65, and the failure threshold T3 is set to 120. Feasibility scoring rules are set: if the cumulative deviation is less than 25, the feasibility score is high; if the cumulative deviation is between 25 and 65, the score is medium; if the cumulative deviation is between 65 and 120, the score is low; if the cumulative deviation is greater than or equal to 120, the score is infeasible, and this score is associated with and stored as the corresponding water and fertilizer combination scheme.

[0051] Furthermore, step six specifically includes: Compare the feasibility scores of all water and fertilizer combination schemes, and select the scheme with the highest feasibility score as the preferred scheme. The water and fertilizer application parameters extracted from this preferred scheme include single irrigation water volume, single fertilizer application volume, water and fertilizer mixing ratio, application interval days, and application time period; The water and fertilizer application parameters are formatted according to the start time of the current decision cycle to generate a water and fertilizer application instruction sequence, which includes multiple application time points and their corresponding water and fertilizer parameters within the decision cycle. The water and fertilizer application instruction sequence is sent to the field water and fertilizer execution equipment, which then performs automatic irrigation and fertilization operations according to the time points and parameters in the instruction sequence.

[0052] In this embodiment, all water-fertilizer combination schemes with calculated feasibility scores are iterated, with scheme numbers ranging from Scheme 1 to Scheme 10. A current optimal scheme variable is maintained, initially assigned the values ​​of Scheme 1 and its score. For Schemes 2 to 10, their feasibility scores are compared sequentially with the score of the current optimal scheme. The comparison rules are: higher score is better than medium score, which is better than low score, which is better than infeasibility score. If two schemes have the same score, the cumulative deviation is further compared, and the scheme with the smaller cumulative deviation is better. Assuming that after the comparison, Scheme 7 is determined to be the scheme with the highest feasibility score.

[0053] Read the configuration description file of Scheme 7, which is stored in XML format and contains the following fields: irrigation water amount, with a value of 1.8 liters per plant; fertilizer amount, with a value of 0.6 grams of compound fertilizer per plant; water-fertilizer mixing ratio, with a value of 0.33 grams of fertilizer dissolved per liter of water; application interval, with a value of two days; and application time, with a value of between 6:00 and 7:00 AM every day.

[0054] Obtain the start time of the current decision-making cycle, assuming it's 00:00 on May 1, 2025. Based on the application interval field (two days) and the application time (6:00-7:00 AM), generate a ten-day instruction sequence. The instructions for the first day, May 1st, are: application time 6:30 AM, irrigation water volume 1.8 liters, fertilizer amount 0.6 grams, and mixing ratio of 1:0.33. The instructions for the third day, May 3rd, May 5th, May 7th, and May 9th are the same. Write these instructions in chronological order into a JSON format control file. Each instruction record includes a timestamp, valve number, water pump start-up time, and fertilizer pump start-up time.

[0055] The control file is sent to the controller of the field fertigation machine via a wired Ethernet or LoRa wireless communication module. Upon receiving the complete control file, the controller parses the timestamp of the first instruction and compares it with its local real-time clock. When the local time reaches 6:30 AM on May 1, 2025, the controller opens the main water valve and the corresponding area valve, simultaneously starting the water pump. After the timer reaches the duration calculated based on the irrigation water volume, the water pump is shut off. Then, the fertilizer pump is started, and after the timer reaches the duration calculated based on the fertilizer application amount, the fertilizer pump and all valves are shut off. The controller waits for the time point of the next instruction and repeats the above operation.

[0056] Furthermore, referring to Figure 3 Before calculating the cumulative deviation for each water-fertilizer combination scheme in step five, an environmental-gene compatibility pre-evaluation is performed, including: The values ​​of temperature, water and light factors in the environmental influence configuration factors were extracted and compared with the normal expression activity range of each environmental response gene locus in the obtained crop genetic information. Determine whether all components of the current environmental impact configuration factor fall within the normal expression activity range of the corresponding gene locus; If any configuration factor component of an environmental factor exceeds the normal expression activity range of the corresponding gene locus, it is determined that the current planting environment is incompatible with the genetic makeup of the crop variety. For each environmental factor that exceeds the normal expression activity range, calculate the minimum deviation of its configuration factor component from the boundary of that range, and then take the geometric mean of the minimum deviation of each environmental factor to obtain the incompatibility value. Based on the range of the incompatibility value, a response strategy is selected from a variety of preset strategies. If the incompatibility value is below the first threshold, the simulation and prediction process of the original water and fertilizer combination scheme is maintained, but an environmental stress penalty term is added to the subsequent calculation of the cumulative deviation. If the incompatibility value is between the first and second thresholds, the parameters of the gene expression regulation module in the growth model are temporarily adjusted to adjust the prediction results of the growth model to include phenotypic plasticity parameters under environmental stress. If the incompatibility value is above the second threshold, the feasibility evaluation of the current water and fertilizer combination scheme is stopped directly, and its feasibility score is set to infeasible. The water and fertilizer combination schemes that have undergone pre-evaluation are sent to the deviation accumulation calculation step, and it is ensured that only schemes that have not been judged as infeasible participate in the feasibility score competition.

[0057] In this embodiment, the generated environmental impact configuration factor vector is extracted, assuming its three components are temperature factor = 0.733, moisture factor = 0.2, and light factor = 0.4. Simultaneously, the normal expression activity range corresponding to each environmental factor is extracted from the gene expression threshold database constructed in step two. The normal expression activity range is defined as follows: under this environmental factor type, the maximum value among the lowest activation thresholds of all gene loci is used as the lower limit of the range, and the minimum value among the highest inhibition thresholds of all gene loci is used as the upper limit of the range. For example, for the temperature factor, the lowest activation thresholds of the three temperature-responsive gene loci are 15, 18, and 23, respectively, with the maximum value being 23; the highest inhibition thresholds are 35, 30, and 32, respectively, with the minimum value being 30; therefore, the normal expression activity range of the temperature factor is [23, 30]. For the moisture factor, the lowest activation thresholds of the two loci are 25 and 20, respectively, with the maximum value being 25; the highest inhibition thresholds are 70 and 65, respectively, with the minimum value being 65; the normal range is [25, 65]. For the light factor, the normal range for one locus is [100, 800].

[0058] The actual environmental impact configuration factor components are compared with their respective interval boundaries: the temperature factor is 0.733, while the lower limit of the interval is 23 degrees Celsius and the upper limit is 30 degrees Celsius. Note that the temperature factor 0.733 is a dimensionless coefficient, while the interval is in degrees Celsius, resulting in a dimensional inconsistency. Therefore, in actual implementation, the temperature component of the environmental impact configuration factor is converted back to a temperature value. The conversion method is: temperature component = current average temperature / optimal temperature. The optimal temperature is 25 degrees Celsius, so the current average temperature = 0.733 × 25 ≈ 18.3 degrees Celsius. Comparing 18.3 degrees Celsius with the interval [23, 30], 18.3 is lower than the lower limit of 23, indicating that the temperature environment exceeds the normal expression range. For the moisture factor, the actual configuration factor 0.2 corresponds to soil moisture content. The conversion rule is: moisture factor = current soil moisture content / optimal soil moisture content. The optimal value is set to 30%, so the current moisture content = 0.2 × 30% = 6%, which is lower than the lower limit of 25%, indicating that it exceeds the range. With a light factor of 0.4, the optimal light intensity is assumed to be 500 moles per square meter per second. The current light intensity is 0.4 × 500 = 200 moles, which is within the normal range [100, 800] and does not exceed the limit.

[0059] Temperature and moisture factors were detected to be outside their normal expression ranges, indicating incompatibility between the planting environment and the crop variety's genetic makeup. Further calculations of the incompatibility degree were performed. For the temperature factor, the deviation of the converted temperature value of 18.3 from the lower limit of the range of 23 was calculated to be 23 - 18.3 = 4.7 degrees Celsius. A pre-defined conversion table for deviation ranges to dimensionless deviation coefficients was used: 0-2 degrees corresponds to 0.1, 2-5 degrees to 0.3, 5-10 degrees to 0.6, and 4.7 degrees falls within the 2-5 degree range, corresponding to a deviation coefficient of 0.3. For the moisture factor, the deviation of the current moisture content of 6% from the lower limit of the range of 25% was calculated to be 19 percentage points. The corresponding deviation coefficients were: 0-5 points to 0.1, 5-15 points to 0.4, and 15-25 points to 0.7, satisfying the 15-25 range, with a coefficient of 0.7. The geometric mean of these two deviation coefficients was calculated as: 0.3 × 0.7 = approximately 0.458. Thresholds were set for the degree of incompatibility: 0.2 for the first degree and 0.6 for the second degree. The current incompatibility value of 0.458 falls between 0.2 and 0.6, therefore the second response strategy is chosen: temporarily adjusting the parameters of the gene expression regulation module in the growth model. Specifically, a phenotypic plasticity coefficient is added to the growth model, which dynamically changes the organogenesis rate based on the degree of incompatibility. The originally fixed organogenesis rate in the growth model is multiplied by an adaptation factor (adaptation factor = 1 + 0.458 × 2 = 1.916), making the crop model exhibit stronger plasticity, i.e., adapting to the environment by accelerating or delaying development under adversity. The adjusted growth model is used for subsequent calculations of cumulative bias, while the other two response strategies either add an environmental stress penalty or directly stop evaluation in this scenario. All schemes not deemed infeasible continue to participate in the subsequent feasibility scoring competition.

[0060] Furthermore, before outputting the water and fertilizer application instruction in step six, a condition verification step is also included: Obtain forecast information on future changes in the planting environment within the current decision-making cycle, including the temperature forecast sequence, precipitation probability, and light intensity forecast for the next few days; The irrigation and fertilization time points required in the selected water and fertilizer combination scheme are compared with the future change prediction information. If a heavy rainfall event exceeding the crop flood tolerance threshold is predicted to occur within the predetermined time window after irrigation, the irrigation amount in the water and fertilizer combination scheme is adjusted to the reduced corrected irrigation amount. If it is predicted that there will be high temperature and drought weather exceeding the crop drought tolerance threshold within the predetermined time window after fertilization, the fertilization method in the water and fertilizer combination scheme will be changed from one-time application to multiple applications. After the adjustments are completed, the final water and fertilizer application instructions will be output.

[0061] In this embodiment, weather forecast data for the next five days of the current decision-making period is obtained from the API interface of a local meteorological station or weather forecast service provider. When calling the interface, latitude and longitude and the current date are passed in, and a JSON object is returned, containing the daily predicted temperature sequence, the highest and lowest temperatures at three-hour intervals, the percentage of precipitation probability, and the predicted daily average photosynthetically active radiation. Parsing the returned data yields the following information: 80% probability of precipitation on the second day, with an expected precipitation of 15 mm; 30% probability of precipitation on the third day; 90% probability of precipitation on the fourth day, with an expected precipitation of 25 mm; 10% probability of precipitation on the fifth day; and the temperature forecast shows that the afternoon high temperature on the third day will reach 38 degrees Celsius, and the high temperature on the fourth day will be 35 degrees Celsius.

[0062] Irrigation and fertilization plans are extracted from the selected optimal water and fertilizer combination scheme. Assume the original plan was to irrigate once on the first and fifth days of the decision-making cycle, with each irrigation being two liters per plant. The fertilization plan was to apply ten grams of slow-release fertilizer per plant once on the first day. Irrigation condition verification was performed: For the first day's irrigation, the rainfall forecast for the next 24 hours was checked; the next 24 hours corresponded to the second day, with an 80% probability of rainfall and a rainfall of 15 mm; the pre-set crop flood tolerance threshold was that rainfall exceeding 10 mm within 24 hours was considered potentially flooding; the current forecast of 15 mm exceeded 10 mm, thus triggering an irrigation adjustment. The adjustment method was: multiplying the original irrigation amount of two liters by a rainfall substitution coefficient, calculated based on the ratio of rainfall to the threshold, set to 0.5, resulting in a corrected irrigation amount of one liter. The first day's irrigation amount was updated to one liter. For the fifth day's irrigation, the probability of rainfall on the sixth day within the following 24 hours was 10%, which did not exceed the flood tolerance threshold, so the original irrigation amount of two liters was maintained.

[0063] Fertilization condition verification: The original plan was to apply 10 grams of slow-release fertilizer in a single application on the first day. The system checked for high-temperature drought exceeding the drought tolerance threshold within seven days after fertilization; the forecast for the third day was a high of 38 degrees Celsius with zero precipitation; the forecast for the fourth day was a high of 35 degrees Celsius with zero precipitation; the preset drought tolerance threshold was two consecutive days with temperatures above 35 degrees Celsius and no rainfall. The third and fourth days met these conditions, thus triggering a fertilization method adjustment. The adjustment was to change the single application to a multi-application application, dividing it into three applications of 3.3 grams each, applied on the first, third, and fifth days respectively. The fertilization instruction sequence was updated: 3.3 grams on the first day, 3.3 grams on the third day, and 3.3 grams on the fifth day.

[0064] After all adjustments are completed, the water and fertilizer application instructions are reorganized; the new instruction sequence includes: irrigate one liter and apply three .3 grams of fertilizer on the first day, irrigate no more but apply three .3 grams of fertilizer on the third day, and irrigate two liters and apply three .3 grams of fertilizer on the fifth day. The final water and fertilizer application instruction sequence is then output to the execution device.

[0065] Furthermore, the preset growth model is a crop growth simulator based on a function-structure model, which includes a three-dimensional virtual plant structure module, a photosynthesis module, a respiration consumption module, a carbon and nitrogen allocation module, and an organogenesis module. The three-dimensional virtual plant structure module generates plant nodes and internodes according to crop topology rules. The photosynthesis module calculates the photosynthetic rate of each leaf based on light and CO2 concentration. The respiration consumption module calculates the consumption rate of maintenance respiration and growth respiration based on temperature. The carbon and nitrogen allocation module allocates photosynthetic products to various organs such as roots, stems, leaves, and fruits according to source-sink relationships and phenological stages. The organogenesis module controls new leaf germination and flower bud differentiation based on accumulated heat units and genotype parameters. When calling the growth model, the soil moisture supply curves and nitrogen supply curves corresponding to different water and fertilizer combination schemes are used as boundary conditions to input the carbon and nitrogen allocation module and the organ development module. The soil moisture supply curve is determined by irrigation amount and soil evapotranspiration model, and the nitrogen supply curve is determined by fertilizer amount and soil nitrogen mineralization model.

[0066] In this embodiment, an open-source functional-structural modeling software library is integrated, which constructs a three-dimensional virtual plant based on L-system syntax. The simulator contains five modules: the three-dimensional virtual plant structure module uses a recursive algorithm to generate the crop's topology, where each node corresponds to a stem node, and each stem node carries a leaf unit. The module's inputs are the internode elongation rate and leaf germination interval from the crop's genetic parameters, and its output is a three-dimensional mesh data structure containing node coordinates, leaf orientation, and leaf area.

[0067] The photosynthesis module receives leaf area and orientation angle data for each leaf unit from the 3D virtual plant structure module, combined with light intensity and carbon dioxide concentration data obtained from meteorological sensors. Internally, the module uses a simplified version of the Farquhar photosynthetic biochemical model to calculate the instantaneous photosynthetic rate for each leaf unit, then sums these rates across all leaves to obtain the total canopy photosynthetic rate. The module's output is the total daily photosynthetic product amount, expressed as glucose equivalents.

[0068] The respiration consumption module uses a standard calculation formula that doubles the maintenance respiration coefficient for every ten-degree increase in temperature, based on the daily average temperature. The module reads the total biomass of the current crop, calculates the dry matter consumed for maintenance respiration, and simultaneously calculates the growth respiration consumption based on the growth rate. The module's output is the daily net primary productivity, which is the remaining dry matter after subtracting respiration consumption from photosynthetic products.

[0069] The carbon and nitrogen allocation module distributes net primary productivity to the four sink organs (roots, stems, leaves, and fruits) based on the crop's phenological stage and source-sink relationship. The allocation ratio is determined by a set of source-sink strength parameters, which are genetic characteristics of the crop variety. When calling this module, the soil moisture supply curves and nitrogen supply curves corresponding to different water and fertilizer combinations in step one are used as boundary conditions. The soil moisture supply curve is determined jointly by irrigation amount and a soil evapotranspiration model. This model uses the Penman-Monteith formula to calculate evapotranspiration and updates soil moisture content daily in conjunction with irrigation events. The nitrogen supply curve is determined jointly by fertilizer application amount and a soil nitrogen mineralization model. This model calculates the mineralization rate of organic nitrogen based on soil temperature and humidity, and adds the inorganic nitrogen input from fertilizer application to obtain the daily available nitrogen content in the soil. The carbon and nitrogen allocation module adjusts the allocation ratio according to changes in soil moisture and nitrogen content: increasing the root allocation ratio under water stress and increasing the root allocation ratio while decreasing the leaf allocation ratio under nitrogen stress.

[0070] The organogenesis module controls new leaf germination and flower bud differentiation based on accumulated heat units and genotype parameters. The module stores a developmental rate table that records the amount of accumulated heat units required for each phenological stage. The module accumulates effective accumulated temperature daily; when the accumulated value crosses a threshold, it triggers a new leaf primordium formation or flower organ differentiation event. The module's output is the 3D virtual plant structure update instruction for the next time step.

[0071] When using this growth model for prediction, the irrigation and fertilization plans corresponding to each water and fertilizer combination scheme are converted into the two boundary condition curves mentioned above: the soil moisture supply curve provides the average water content in the root zone daily, and the nitrogen supply curve provides the average nitrate nitrogen concentration in the root zone daily. These curves are input into the carbon and nitrogen allocation module and the organogenesis module, and then the entire simulator is driven to iterate and calculate at a daily step size, outputting the final three-dimensional structural data and the biomass of each organ.

[0072] Furthermore, the target growth indicators for the corresponding time nodes in step five are dynamically adjusted based on the environmental impact configuration factors generated in step three: For temperature factors, if the temperature factor indicates that the current environment deviates from the optimal growth temperature of the crop, the plant height growth rate and leaf area expansion rate in the target growth indicators will be reduced. The reduction magnitude is non-linearly positively correlated with the degree of deviation of the temperature factor, and the more severe the deviation, the faster the reduction magnitude decays. For the water factor, if the water factor indicates that there is water stress in the current environment, the stem diameter and dry matter accumulation in the target growth indicators are lowered, and the root-shoot ratio is raised, as a mapping of the crop's adaptive strategy of prioritizing root growth in a water-deficient environment. For the light factor, if the light factor indicates that the current ambient light intensity is insufficient, the canopy coverage and leaf chlorophyll content in the target growth indicators will be adjusted upwards, and the proportion of photosynthetic products per unit dry matter will be adjusted downwards. The dynamically adjusted target growth indicators are used for subsequent calculations of cumulative deviations, enabling the feasibility assessment to adapt to a reasonable range of crop growth variation under different environmental conditions.

[0073] In this embodiment, the environmental impact configuration factors generated in step three are read as follows: temperature factor = 0.733, moisture factor = 0.2, and light factor = 0.4. Simultaneously, the static ideal values ​​from the target growth index database are read; these ideal values ​​are set under optimal environmental conditions.

[0074] For the temperature factor of 0.733, it is converted into a deviation degree. The conversion method is: Temperature Deviation Degree = |0.733 - 1.0| / 0.5, where 0.5 is the maximum allowable deviation range, and the calculated deviation degree is 0.534. A non-linear decay function is used to adjust the target values ​​of plant height growth rate and leaf area expansion rate. The decay function is defined as: Adjusted target value = Original target value × 0.5 + 0.5 / 1 + 5 × Deviation Degree. Assuming the original target plant height growth rate is 1.5 cm per day, then the adjusted value = 1.5 × 0.5 + 0.5 / 1 + 5 × 0.534 = 1.5 × 0.5 + 0.5 / 3.67 = 1.5 × 0.5 + 0.136 = 0.954 cm per day. This adjusted plant height growth rate is applied to the calculation of the target plant height value for all subsequent time nodes, that is, the target plant height value for each node is regenerated at this rate based on the original cumulative value. A similar non-linear downward adjustment is performed for the leaf area expansion rate.

[0075] With a moisture factor of 0.2, water stress is identified. Two adjustments are made: First, the target stem diameter is lowered. The static ideal stem diameter target is 6 mm at a certain point. The adjustment factor is the square root of the moisture factor multiplied by 0.8, i.e., sqrt(0.2)×0.8=0.357. The adjusted target diameter is 6×0.357=2.14 mm. Second, the root-to-shoot ratio is raised. The static ideal root-to-shoot ratio is 0.2. The adjustment factor is 1 / moisture factor + 0.1=1 / 0.3≈3.33. The adjusted root-to-shoot ratio is 0.2×3.33=0.666. This adjusted root-to-shoot ratio is used as an additional evaluation indicator. A root-to-shoot ratio deviation term is added to the deviation accumulation calculation, with a weight of 0.15. At the same time, the sum of the weights of other parameters is reduced accordingly.

[0076] With a light factor of 0.4, a weak light environment is identified. Two adjustments are made: First, the target value for canopy coverage is increased. The static ideal coverage is 0.5 at a certain point, so the adjustment factor is 1 / light factor = 2.5. After adjustment, the target coverage = 0.5 × 2.5 = 1.25, but coverage cannot exceed 1.0, so it is truncated to 1.0. Second, the target value for leaf chlorophyll content is increased. The static SPAD target is 35, so the adjustment factor is 1.2, resulting in SPAD = 42. Simultaneously, the proportion of photosynthetic products per unit of dry matter is adjusted downward: originally, 0.5 grams of photosynthetic products per gram of dry matter were reserved for structural growth; after adjustment, this proportion, multiplied by the light factor of 0.4, means that only 0.2 grams per gram of dry matter are used for structural growth, with the remainder stored as non-structural carbohydrates.

[0077] All target growth indicators, after the aforementioned dynamic adjustments, replace the original static indicators. Then, the old and new indicators are compared using the cumulative deviation calculation method. In this way, under adverse conditions, the cumulative deviation will not be artificially inflated simply because the crop cannot reach its optimal value. Instead, it is evaluated based on reasonable, environmentally adaptable targets, improving the rationality of the feasibility score. The adjusted target indicators are then made permanent for use throughout the entire decision-making cycle.

[0078] This embodiment provides a smart decision-making method for crop water and fertilizer based on multimodal sensing and growth models, applicable to greenhouse or field crop cultivation environments. The system implementing this method includes a multimodal sensing node deployed in the field, an edge computing unit, and an actuator connected to control valves or irrigation and fertilization equipment.

[0079] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A smart decision-making method for crop water and fertilizer based on multimodal perception and growth models, characterized in that, Includes the following steps: Step 1: Collect real-time information on multiple sensory modalities of the crop plant, including thermal infrared images of the plant canopy, micro-variable diameter of the stem, and spectral reflectance of the leaves; Step 2: Extract gene locus expression thresholds related to environmental response based on the genetic information of crop varieties, and obtain real-time parameters of the planting environment, including soil moisture content, air temperature and light intensity. Step 3: Compare the real-time parameters with the gene locus expression thresholds to determine the expression activity level of each gene locus, and generate environmental influence configuration factors based on the expression activity levels; Step 4: Call the preset growth model, input the real-time information of the current crop plant's perception mode as the initial state into the growth model, and use the environmental influence configuration factor to correct the parameters of the growth model to predict the simulation results of the crop's growth state under various water and fertilizer combination schemes. Step 5: For each water and fertilizer combination scheme, extract the key morphological parameters and physiological parameters from the corresponding growth state simulation results, compare them with the target growth indicators at the corresponding time nodes, calculate the cumulative deviation between each parameter and the target indicator, and determine the feasibility score of the water and fertilizer combination scheme based on the cumulative deviation. Step 6: Select the water and fertilizer combination scheme with the highest feasibility score as the water and fertilizer application instruction output for the current decision-making cycle.

2. The method according to claim 1, characterized in that, Step two specifically includes: The genome sequence of a crop variety is analyzed to identify at least one gene locus that is regulated by environmental factors, including temperature, water and light. For each identified gene locus, its corresponding expression regulatory region is obtained, which consists of the minimum activation threshold and the maximum inhibition threshold; When the environmental factor parameter is below the minimum activation threshold, the gene locus is in a non-expressed state; when the environmental factor parameter is between the minimum activation threshold and the maximum inhibition threshold, the gene locus is in a normally expressed state; when the environmental factor parameter is above the maximum inhibition threshold, the gene locus is in an overexpression and inhibition state. The expression regulatory regions of each gene locus are categorized and stored according to the type of environmental factor, forming a database of gene expression thresholds for crop responses to specific environmental factors under the current planting environment.

3. The method according to claim 1 or 2, characterized in that, Step three specifically includes: For each gene locus, select the real-time parameter value of the environmental factor type corresponding to that gene locus; compare the real-time parameter value with the lowest activation threshold and the highest inhibition threshold of that gene locus. If the real-time parameter value is lower than the minimum activation threshold, the expression activity level of the gene locus is determined to be dormant. If the real-time parameter value is higher than or equal to the minimum activation threshold and lower than or equal to the maximum inhibition threshold, the expression activity level of the gene locus is determined to be activated. If the real-time parameter value is higher than the maximum inhibition threshold, the expression activity level of the gene locus is determined to be inhibited. A weight coefficient is assigned to each expression activity level, with the first weight coefficient corresponding to the dormant level, the second weight coefficient corresponding to the active level, and the third weight coefficient corresponding to the inhibitory level. The weight coefficients of all gene loci are synthesized according to their corresponding environmental factor types to generate environmental impact configuration factor vectors for three types of environmental factors: temperature, moisture, and light.

4. The method according to claim 1, characterized in that, Step four specifically includes: Read the preset growth model, which includes a physiological development time function to describe crop dry matter accumulation, a morphogenesis function to describe crop organ elongation, a source-sink relationship function to describe photosynthetic product allocation, and an accumulated temperature function to describe phenological stage transitions. The weighted value of the daily temperature contribution to physiological development time is adjusted according to the value of the temperature factor, so that the physiological development time under high temperature or low temperature environment is compressed or stretched relative to the optimal temperature environment. The baseline values ​​of stem diameter growth rate and leaf area expansion rate are adjusted according to the water factor values ​​to proportionally slow down the morphogenesis rate under water stress compared to that under adequate water supply. The photosynthetic carbon fixation efficiency per unit leaf area is adjusted according to the value of the light factor so that the dry matter accumulation rate under low light conditions is proportionally reduced compared to the sufficient light conditions. Adjusting the cumulative temperature threshold required for the transition of each phenological stage based on the combined effects of three environmental factors can advance or delay the phenological process of crops under adverse conditions. After the correction is completed, the real-time information of the current crop plant's perception modality is used as the initial state value and input into the corrected growth model. The water and fertilizer supply parameters corresponding to various water and fertilizer combination schemes are substituted into the model to calculate the values ​​of crop biomass, organ size and phenological stage at multiple time nodes in the future, and the growth state simulation results under each water and fertilizer combination scheme are obtained.

5. The method according to claim 1, characterized in that, The key morphological parameters extracted in the growth state simulation results in step five include plant height, stem diameter and canopy coverage, while the physiological parameters include leaf chlorophyll content and dry matter accumulation. The target growth indicators for the corresponding time nodes are pre-set based on the ideal plant type standard and growth stage standard in the genetic information of the crop variety. The cumulative deviation between each parameter and the target index is calculated, specifically including: for each time node, the absolute value of the first difference between the simulated plant height and the target plant height, the absolute value of the second difference between the simulated stem diameter and the target diameter, the absolute value of the third difference between the simulated canopy coverage and the target coverage, the absolute value of the fourth difference between the simulated leaf chlorophyll content and the target chlorophyll content, and the absolute value of the fifth difference between the simulated dry matter accumulation and the target accumulation. Multiply the absolute values ​​of the first, second, third, fourth, and fifth differences by their respective parameter weighting coefficients and then sum them to obtain the single-point deviation value at that time node. The cumulative deviation is obtained by summing up the single-point deviation values ​​of all future time points. The cumulative deviation is compared with a preset feasibility threshold range. If the cumulative deviation is less than the first threshold, it is judged as a high feasibility score; if it is between the first and second thresholds, it is judged as a medium feasibility score; and if it is greater than the second threshold, it is judged as a low feasibility score.

6. The method according to claim 1, characterized in that, Step six specifically includes: Compare the feasibility scores of all water and fertilizer combination schemes, and select the scheme with the highest feasibility score as the preferred scheme. The water and fertilizer application parameters extracted from this preferred scheme include single irrigation water volume, single fertilizer application volume, water and fertilizer mixing ratio, application interval days, and application time period; The water and fertilizer application parameters are formatted according to the start time of the current decision cycle to generate a water and fertilizer application instruction sequence, which includes multiple application time points and their corresponding water and fertilizer parameters within the decision cycle. The water and fertilizer application instruction sequence is sent to the field water and fertilizer execution equipment, which then performs automatic irrigation and fertilization operations according to the time points and parameters in the instruction sequence.

7. The method according to claim 1, characterized in that, Before calculating the cumulative deviation for each water-fertilizer combination scheme in step five, an environmental-gene compatibility pre-evaluation is performed, including: The values ​​of temperature, water and light factors in the environmental influence configuration factors were extracted and compared with the normal expression activity range of each environmental response gene locus in the obtained crop genetic information. Determine whether all components of the current environmental impact configuration factor fall within the normal expression activity range of the corresponding gene locus; If any configuration factor component of an environmental factor exceeds the normal expression activity range of the corresponding gene locus, it is determined that the current planting environment is incompatible with the genetic makeup of the crop variety. For each environmental factor that exceeds the normal expression activity range, calculate the minimum deviation of its configuration factor component from the boundary of that range, and then take the geometric mean of the minimum deviation of each environmental factor to obtain the incompatibility value. Based on the range of the incompatibility value, a response strategy is selected from a variety of preset strategies. If the incompatibility value is below the first threshold, the simulation and prediction process of the original water and fertilizer combination scheme is maintained, but an environmental stress penalty term is added to the subsequent calculation of the cumulative deviation. If the incompatibility value is between the first and second thresholds, the parameters of the gene expression regulation module in the growth model are temporarily adjusted to adjust the prediction results of the growth model to include phenotypic plasticity parameters under environmental stress. If the incompatibility value is above the second threshold, the feasibility evaluation of the current water and fertilizer combination scheme is stopped directly, and its feasibility score is set to infeasible. The water and fertilizer combination schemes that have undergone pre-evaluation are sent to the deviation accumulation calculation step, and it is ensured that only schemes that have not been judged as infeasible participate in the feasibility score competition.

8. The method according to claim 1, characterized in that, Before outputting the water and fertilizer application instruction in step six, a condition verification step is also included: Obtain forecast information on future changes in the planting environment within the current decision-making cycle, including the temperature forecast sequence, precipitation probability, and light intensity forecast for the next few days; The irrigation and fertilization time points required in the selected water and fertilizer combination scheme are compared with the future change prediction information. If a heavy rainfall event exceeding the crop flood tolerance threshold is predicted to occur within the predetermined time window after irrigation, the irrigation amount in the water and fertilizer combination scheme is adjusted to the reduced corrected irrigation amount. If it is predicted that there will be high temperature and drought weather exceeding the crop drought tolerance threshold within the predetermined time window after fertilization, the fertilization method in the water and fertilizer combination scheme will be changed from one-time application to multiple applications. After the adjustments are completed, the final water and fertilizer application instructions will be output.

9. The method according to claim 1, characterized in that, The preset growth model is a crop growth simulator based on a function-structure model, which includes a three-dimensional virtual plant structure module, a photosynthesis module, a respiration consumption module, a carbon and nitrogen allocation module, and an organogenesis module. The three-dimensional virtual plant structure module generates plant nodes and internodes according to crop topology rules. The photosynthesis module calculates the photosynthetic rate of each leaf based on light and CO2 concentration. The respiration consumption module calculates the consumption rate of maintenance respiration and growth respiration based on temperature. The carbon and nitrogen allocation module allocates photosynthetic products to various organs such as roots, stems, leaves, and fruits according to source-sink relationships and phenological stages. The organogenesis module controls new leaf germination and flower bud differentiation based on accumulated heat units and genotype parameters. When calling the growth model, the soil moisture supply curves and nitrogen supply curves corresponding to different water and fertilizer combination schemes are used as boundary conditions to input the carbon and nitrogen allocation module and the organ development module. The soil moisture supply curve is determined by irrigation amount and soil evapotranspiration model, and the nitrogen supply curve is determined by fertilizer amount and soil nitrogen mineralization model.

10. The method according to claim 1, characterized in that, The target growth indicators for the corresponding time nodes in step five are dynamically adjusted based on the environmental impact configuration factors generated in step three. For temperature factors, if the temperature factor indicates that the current environment deviates from the optimal growth temperature of the crop, the plant height growth rate and leaf area expansion rate in the target growth indicators will be reduced. The reduction magnitude is non-linearly positively correlated with the degree of deviation of the temperature factor, and the more severe the deviation, the faster the reduction magnitude decays. For the water factor, if the water factor indicates that there is water stress in the current environment, the stem diameter and dry matter accumulation in the target growth indicators are lowered, and the root-shoot ratio is raised, as a mapping of the crop's adaptive strategy of prioritizing root growth in a water-deficient environment. For the light factor, if the light factor indicates that the current ambient light intensity is insufficient, the canopy coverage and leaf chlorophyll content in the target growth indicators will be adjusted upwards, and the proportion of photosynthetic products per unit dry matter will be adjusted downwards. The dynamically adjusted target growth indicators are used for subsequent calculations of cumulative deviations, enabling the feasibility assessment to adapt to a reasonable range of crop growth variation under different environmental conditions.

Images