Cut-flower chrysanthemum growth decision-making method and system based on multi-model fusion

Abstract

The invention discloses a cut-flower chrysanthemum growth decision-making method and system based on multi-model fusion. The method comprises the steps that greenhouse environment parameters and soil parameters are collected in real time through an environment sensor; calling a corresponding calibration model parameter set in a preset model parameter library according to the planted cut-flower chrysanthemum variety; synchronously executing the following sub-models through a multi-model fusion decision engine; and decision output: outputting a comprehensive decision result containing the model calculation result. According to the method, a growth and development index (DVI) is adopted as a key driving variable, and a growth prediction and management decision model of the cut-flower chrysanthemum is constructed. According to the method, the calculation complexity is reduced, and the prediction precision, timeliness and scientificity are remarkably improved; by coupling the leaf area index of the cut-flower chrysanthemum and referring to key parameters such as crop evapotranspiration, moisture evaporation and soil moisture content, accurate irrigation decision suggestions are output in real time; the blindness of irrigation with traditional experience is effectively avoided, the utilization efficiency of water resources is remarkably improved, and resource waste is reduced.

Description

Technical Field

[0001] This invention relates to a method and system for decision-making regarding the growth of cut chrysanthemums, and more particularly to a method and system for decision-making regarding the growth of cut chrysanthemums based on multi-model fusion. Background Technology

[0002] As a high-value ornamental crop, cut chrysanthemums are subject to dynamic coupling of environmental, soil, and physiological factors in a closed greenhouse environment. Therefore, their growth management is crucial for improving yield and quality.

[0003] Traditional cultivation management mainly relies on growers' experience, which has three major bottlenecks: (1) Strong reliance on experience and lack of quantitative basis: Decisions on environmental control (such as ventilation / supplementary lighting) and water and fertilizer application lack real-time data support, which can easily lead to resource waste or growth stress. For example, inaccurate control of flowering period may cause the peak sales season to be missed, and blind irrigation may cause root diseases. (2) Variety specificity is ignored: There are significant differences in light and temperature response, nutrient requirements and growth period division between single-headed chrysanthemums and multi-headed chrysanthemums, and even between different varieties of the same type. Existing technologies have not established a mapping relationship between variety and model parameters, and the prediction error of general models is as high as 20%-30%. (3) Fragmented multi-dimensional decision-making: Growth prediction, photosynthetic efficiency analysis and water and fertilizer management are usually handled by independent systems, which lack coordination. For example, the supplementary lighting scheme does not consider the real-time photosynthetic rate, and the amount of nitrogen applied is not related to the needs of flower bud differentiation period, resulting in the imbalance of environment-physiology-nutrition coupling. In the greenhouse environment, environmental parameters (such as temperature, light intensity, humidity, CO2 concentration) and soil parameters (such as soil temperature, water content) have a significant impact on the growth of cut chrysanthemums. Therefore, how to accurately obtain these parameters and make scientific decisions based on the characteristics of different cut chrysanthemum varieties has become an urgent problem to be solved in facility agriculture.

[0004] With the development of facility agriculture, intelligent and precise management has become a trend. However, traditional cut chrysanthemum cultivation and management lacks intelligent decision support systems, making it difficult to achieve real-time monitoring and scientific analysis of environmental parameters and soil conditions, and unable to provide personalized management suggestions based on plant growth status and varietal characteristics. In conclusion, traditional cut chrysanthemum cultivation and management methods can no longer meet the modern agricultural pursuit of yield, quality, and economic benefits, and a scientific, precise, and intelligent management method is urgently needed to address these challenges. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to propose a multi-model fusion-based decision-making method and system for cut chrysanthemum growth, in order to improve prediction accuracy and timeliness.

[0006] Technical solution: This invention includes the following steps:

[0007] S1. Real-time collection of greenhouse environmental and soil parameters using environmental sensors;

[0008] S2. Based on the planted cut chrysanthemum variety, call the corresponding calibration model parameter set from the preset model parameter library;

[0009] S3. Simultaneously execute the following sub-models through a multi-model fusion decision engine:

[0010] (a) Fertility prediction model: using the growth and development index (DVI) as the driving variable, the fertility period is predicted through a piecewise function;

[0011] (b) Growth index prediction model: Based on the growth and development index (DVI), predict the daily plant height, stem diameter, number of leaves, leaf area index, and biomass.

[0012] (c) Photosynthetic effect model: The relative temperature effect, relative light intensity effect, relative humidity effect, relative carbon dioxide effect and instantaneous photosynthetic rate are calculated using the leaf-level photosynthetic response function;

[0013] (d) Nutrient decision model: The critical nitrogen accumulation, nitrogen requirement, nitrogen application rate, phosphorus application rate and potassium application rate are predicted by calculating the critical nitrogen accumulation, nitrogen application rate and potassium application rate through the critical nitrogen accumulation model of DVI and the nitrogen, phosphorus and potassium application ratios at each growth stage.

[0014] (e) Water decision model: predicting water demand and irrigation volume for a given period;

[0015] S4. Decision Output: The output includes the comprehensive decision results calculated by models (a) to (e) in S3.

[0016] The environmental parameters include temperature, light intensity, humidity, and CO2 concentration, while the soil parameters include soil temperature and moisture content.

[0017] The preset model parameter library contains multiple sets of calibration model parameters, corresponding to several representative varieties of single-headed cut chrysanthemums and several varieties of multi-headed cut chrysanthemums.

[0018] The reproductive period is predicted using the following function:

[0019]

[0020] In the formula, ba-bd are the model parameters fitted for the corresponding varieties of cut chrysanthemum based on a large number of experiments.

[0021] The growth and development index (DVI) is calculated based on the cumulative photothermal effect. The specific calculation formula is as follows:

[0022]

[0023]

[0024] In the formula, T i,jThe temperature at the i-th half hour of day j, in °C; PAR i,j Photosynthetically active radiation (PAR) during the i-th half-hour of day j, μmol·m -2 ·s -1 ;RTE i,j For relative thermal effects; Pn i,j For absolute light effect; PTP j The cumulative photothermal effect is denoted by α; aa-ag represents the model parameters fitted to the corresponding varieties of cut chrysanthemum based on numerous experiments.

[0025] The growth index prediction model includes sub-models for predicting leaf area, plant height, stem diameter, number of leaves, biomass, and flower quality. Each sub-model uses DVI as the driving variable for prediction, and the specific calculation formula for each sub-model is as follows:

[0026] Leaf area prediction: LAI = be / (1 + bf × exp(-bg × DVI))

[0027] Plant height prediction: PH=bh / (1+bi×exp(-bj×DVI))

[0028] Stem diameter prediction: SD = bk / (1 + bl × exp(-bm × DVI))

[0029] Leaf number prediction: LN = bn / (1 + bo × exp(-bp × DVI))

[0030] Biomass prediction: TFH = bq / (1 + br × exp(-bs × DVI))

[0031] Flower quality prediction:

[0032]

[0033] In the formula, be~bv are model parameters fitted for the corresponding varieties of cut chrysanthemum based on a large number of experiments.

[0034] The formula for calculating the relative temperature effect is:

[0035]

[0036] The formula for calculating the relative light intensity effect is:

[0037]

[0038] The formula for calculating the relative humidity effect is:

[0039]

[0040] The formula for calculating the relative carbon dioxide effect is: RC i,j =al×RC i,jam

[0041] The formula for calculating the instantaneous photosynthetic rate is: IRn i,j =RTE i,j ×RPn i,j ×RRH i,j ×RC i,j ×an

[0042] In the formula, RH i,j The humidity (%) for the i-th half hour of day j; C i,j The carbon dioxide concentration in ppm at the i-th half hour of day j; aa-an are model parameters fitted for the corresponding varieties of cut chrysanthemum based on a large number of experiments.

[0043] The specific calculation formula for the nutrient decision model is as follows:

[0044] Critical nitrogen accumulation prediction: N uptc =ca / (1+cb×e) -cc×DVI )

[0045] Nitrogen requirement forecast for day j: N uptc,j =N uptc,j -N uptc,j-1

[0046] Nitrogen application rate prediction for day j: NUE = -cd × DVI 2 +ce×DVI-cf

[0047]

[0048] Phosphorus application rate prediction for day j:

[0049] Potassium application rate prediction for day j:

[0050] In the formula, NUE is nitrogen fertilizer utilization rate, and ca~cl are model parameters fitted for the corresponding varieties of cut chrysanthemum based on a large number of experiments.

[0051] The water decision model generates an irrigation plan based on reference crop evaporation and water evaporation, combined with the real-time growth status of the cut chrysanthemum and soil moisture content. Specifically, it includes:

[0052] Water demand ET based on ET0 c1 Prediction: ET c1 =da×LAI db ×ET0

[0053] Based on Epan's water demand ET c2 Prediction: ET c2 =dc×LAI dd ×Epan

[0054] Prediction of irrigation volume I1 using a coupled moisture sensor:

[0055] I1=de×DVI -df ×ET c1 +Δ w

[0056]

[0057] Prediction of irrigation volume I2 using coupled moisture sensor: I2 = de × DVI -df ×ET c2 +Δ w

[0058] In the formula, VWC is the soil volumetric water content, %; Δ w da-df represents the soil moisture deficit; da-df represents the model parameters fitted for the corresponding varieties of cut chrysanthemum based on numerous experiments.

[0059] A multi-model fusion-based decision-making system for cut chrysanthemum growth includes:

[0060] Environmental data acquisition module: Real-time acquisition of greenhouse environmental parameters and soil parameters;

[0061] Variety identification and parameter matching module: calls the corresponding calibration model parameter set from the preset model parameter library according to the planted variety;

[0062] Multi-model fusion decision engine module: integrates and runs growth index prediction model, photosynthetic effect model, nutrient decision model and water decision model, receives data from environmental data acquisition module and parameter set output from variety identification and parameter matching module, and performs calculations;

[0063] Decision output module: Generates and outputs comprehensive decision results including growth prediction, environmental effect assessment and cultivation management recommendations.

[0064] Beneficial effects: The present invention has the following advantages:

[0065] 1. This invention uses the Developmental Growth Index (DVI) as a key driving variable to construct a growth prediction and management decision-making model for cut chrysanthemums. This method reduces computational complexity and significantly improves prediction accuracy, timeliness, and scientific rigor.

[0066] 2. This invention, by coupling key parameters such as leaf area index of cut chrysanthemum, reference crop evapotranspiration, water evaporation, and soil moisture content, outputs accurate irrigation decision suggestions in real time; effectively avoiding the blindness of traditional experience-based irrigation, significantly improving water resource utilization efficiency, and reducing resource waste;

[0067] 3. This invention establishes a critical nitrogen accumulation model for cut chrysanthemums, calculates and outputs the dynamic nitrogen requirement of plants in real time, and generates scientific fertilization recommendations in real time by combining dynamically changing nitrogen fertilizer utilization rate and optimized nitrogen, phosphorus and potassium application ratio; it significantly improves nitrogen fertilizer utilization efficiency and reduces fertilizer loss and environmental pollution risks.

[0068] 4. By constructing a photosynthetic effect model for cut chrysanthemums, this invention can accurately quantify and assess the impact of key environmental factors on instantaneous photosynthesis, and generate scientific command thresholds for greenhouse environmental regulation in real time, which significantly improves the utilization efficiency of greenhouse environmental resources and lays a key environmental foundation for the high-yield and high-quality production of cut chrysanthemums.

[0069] 5. This invention, through its innovative parameter library design (10 sets covering 85 varieties) and multi-model fusion engine (simultaneous calculation of growth / photosynthesis / nutrients / water), solves the pain points of large variety differences, multiple management dimensions, and fragmented decision-making in greenhouse cut chrysanthemum cultivation, providing a powerful scientific decision-making tool for the precise, intelligent, and standardized production of cut chrysanthemums. Attached Figure Description

[0070] Figure 1 This is a flowchart of the present invention;

[0071] Figure 2 The light response curves A-Li and carbon dioxide response curves A-Ci are shown at different temperatures: (A) is the light response curve at 10-25℃; (B) is the light response curve at 30-45℃; (C) is the carbon dioxide response curve at 10-25℃; and (D) is the carbon dioxide response curve at 30-45℃.

[0072] Figure 3 Temperature response curves for the maximum net photosynthetic rate of cut chrysanthemum leaves;

[0073] Figure 4 The light response curves A-Li and carbon dioxide response curves A-Ci are shown under different humidity conditions: (A) is the light response curve A-Li; (B) is the carbon dioxide response curve A-Ci.

[0074] Figure 5 (A) is a graph showing the relationship between dry matter and cumulative nitrogen accumulation; (B) is a graph showing the relationship between dry matter and cumulative photothermal effect.

[0075] Figure 6 The curve for predicting critical nitrogen accumulation based on the cumulative photothermal effect;

[0076] Figure 7 The relationship between the water requirements of cut chrysanthemums, ETC, ET0, and Epan. Detailed Implementation

[0077] The invention will now be further described with reference to the accompanying drawings.

[0078] Example 1

[0079] like Figure 1 As shown, the multi-model fusion-based cut flower chrysanthemum growth decision method in this embodiment includes the following steps:

[0080] S1. Real-time collection of greenhouse environmental and soil parameters through environmental sensors, including temperature, light intensity, humidity and CO2 concentration, and soil parameters including soil temperature and moisture content;

[0081] S2. Based on the planted cut chrysanthemum variety, call the corresponding calibration model parameter set from the preset model parameter library.

[0082] The pre-defined model parameter library contains 10 sets of calibrated model parameters. Five sets correspond to five representative varieties of single-headed cut chrysanthemums, while the other five sets are categorized to correspond to 80 varieties of multi-headed cut chrysanthemums. The 10 calibrated parameter sets in the model parameter library cover the main categories of both single-headed and multi-headed cut chrysanthemums. Other varieties were matched to the calibration parameter set that best matches their growth characteristics by conducting basic growth parameter measurements.

[0083] S3. Simultaneously execute the following sub-models through a multi-model fusion decision engine:

[0084] (a) Fertility period prediction model: Using the growth and development index (DVI) as the driving variable, the fertility period is predicted through a piecewise function:

[0085]

[0086] In the formula, ba-bd are the model parameters fitted for the corresponding varieties of cut chrysanthemum based on a large number of experiments.

[0087] The growth and development index (DVI) is calculated based on the cumulative photothermal effect, and the specific calculation formula is as follows:

[0088]

[0089] In the formula, T i,j The temperature at the i-th half hour of day j, in °C; PAR i,j Photosynthetically active radiation (PAR) during the i-th half-hour of day j, μmol·m -2 ·s -1 ;RTE i,j For relative thermal effects; Pn i,j For absolute light effect; PTP j The cumulative photothermal effect is denoted by α; aa-ag represents the model parameters fitted to the corresponding varieties of cut chrysanthemum based on numerous experiments.

[0090] (b) Growth index prediction model: Based on the growth and development index (DVI), predict the daily plant height, stem diameter, number of leaves, leaf area index, and biomass.

[0091] The growth index prediction model includes sub-models for predicting leaf area, plant height, stem diameter, number of leaves, biomass, and flower quality. Each sub-model uses DVI as the driving variable for prediction, and the specific calculation formulas for each sub-model are as follows:

[0092] Leaf area prediction: LAI = be / (1 + bf × exp(-bg × DVI))

[0093] Plant height prediction: PH=bh / (1+bi×exp(-bj×DVI))

[0094] Stem diameter prediction: SD = bk / (1 + bl × exp(-bm × DVI))

[0095] Leaf number prediction: LN=bn / (1+bo×exp(-bp×DVI))

[0096] Biomass prediction: TFH = bq / (1 + br × exp(-bs × DVI))

[0097] Flower quality prediction:

[0098]

[0099] In the formula, be-bv represents the model parameters fitted to the corresponding varieties of cut chrysanthemum based on a large number of experiments.

[0100] (c) Photosynthetic effect model: The photosynthetic response function at the leaf level is used for calculation, including the calculation of relative temperature effect, relative light intensity effect, relative humidity effect, relative carbon dioxide effect, and instantaneous photosynthetic rate. The specific calculation formula is as follows:

[0101] Relative temperature effect:

[0102]

[0103] Relative light intensity effect:

[0104]

[0105] Relative humidity effect:

[0106]

[0107] Relative carbon dioxide effect: RC i,j =al×RC i,j am

[0108] Instantaneous photosynthetic rate: IRn i,j=RTE i,j ×RPn i,j ×RRH i,j ×RC i,j ×an

[0109] In the formula, RH i,j The humidity (%) for the i-th half hour of day j; C i,j The carbon dioxide concentration in ppm at the i-th half hour of day j; aa-an are model parameters fitted for the corresponding varieties of cut chrysanthemum based on a large number of experiments.

[0110] (d) Nutrient decision model: predicting the daily application rates of nitrogen, phosphorus, and potassium;

[0111] The nutrient decision model calculates the critical nitrogen accumulation (CVI) using the critical nitrogen accumulation model and the nitrogen, phosphorus, and potassium application ratios for each growth stage. This includes predictions of the CVI, nitrogen requirement, nitrogen application rate, phosphorus application rate, and potassium application rate. The specific calculation formula is as follows:

[0112] Critical nitrogen accumulation prediction: N uptc =ca / (1+cb×e) -cc×DVI )

[0113] Nitrogen requirement forecast for day j: N uptc,j =N uptc,j -N uptc,j-1

[0114] Nitrogen application rate prediction for day j: NUE = -cd × DVI 2 +ce×DVI-cf

[0115]

[0116] Phosphorus application rate prediction for day j:

[0117] Potassium application rate prediction for day j:

[0118] In the formula, NUE is nitrogen fertilizer utilization rate, and ca-cl are model parameters fitted for the corresponding varieties of cut chrysanthemum based on a large number of experiments.

[0119] (e) Water decision model: predict daily or hourly water demand and irrigation volume.

[0120] The water decision model, based on reference crop evaporation and water evaporation, combined with the real-time growth status of cut chrysanthemums and soil moisture content, generates a precise irrigation plan, specifically including:

[0121] Water demand ET based on ET0 c1 Prediction: ET c1 =da×LAI db ×ET0

[0122] Based on Epan's water demand ET c2 Prediction: ET c2 =dc×LAI dd ×Epan

[0123] Prediction of irrigation volume I1 using a coupled moisture sensor:

[0124] I1=de×DVI -df ×ET c1 +Δ w

[0125]

[0126] Prediction of irrigation volume I2 using coupled moisture sensor: I2 = de × DVI -df ×ET c2 +Δ w .

[0127] In the formula, VWC is the soil volumetric water content, %; Δ w da-df represents the soil moisture deficit; da-df represents the model parameters fitted for the corresponding varieties of cut chrysanthemum based on numerous experiments.

[0128] When there is misting activity in the greenhouse, it is recommended to use water demand ET based on ET0. c1 And irrigation amount I1; when there is no misting activity in the greenhouse, it is recommended to use the Epan-based water requirement ET. c2 And irrigation amount I2.

[0129] S4. Decision Output: The output includes the comprehensive decision results calculated by models (a) to (e) in S3, including growth prediction values, environmental effect assessment values, and cultivation management recommendations.

[0130] Example 2

[0131] The multi-model fusion-based cut flower chrysanthemum growth decision system of this embodiment includes:

[0132] Environmental data acquisition module: Real-time acquisition of greenhouse environmental parameters and soil parameters;

[0133] Variety identification and parameter matching module: calls the corresponding calibration model parameter set from the preset model parameter library according to the planted variety;

[0134] Multi-model fusion decision engine module: integrates and runs growth index prediction model, photosynthetic effect model, nutrient decision model and water decision model, receives data from environmental data acquisition module and parameter set output from variety identification and parameter matching module, and performs calculations;

[0135] Decision output module: Generates and outputs comprehensive decision results including growth prediction, environmental effect assessment and cultivation management recommendations.

[0136] The system collects environmental parameters (temperature, light intensity, humidity, CO2 concentration) and soil parameters (soil temperature, moisture content) in real time through meteorological stations deployed in the target greenhouse. After the user sets the planting variety and time, the system automatically matches the corresponding calibration parameter set in the preset model parameter library. The core multi-model fusion decision engine simultaneously executes the growth index prediction model, the growth period prediction model, the photosynthetic effect model, the nutrient decision model, and the water decision model (the latter dynamically generates a spray / non-spray dual-mode irrigation scheme).

[0137] The system displays comprehensive decision-making results in real time on a large screen, covering greenhouse environmental data monitoring, crop growth status environmental effect assessment, and daily water and nutrient management recommendations. Users can view the greenhouse status in real time, understand predicted growth data and growth progress; manage water according to the recommendations of two irrigation modes; perform precise fertilization based on the recommended daily application rates of nitrogen, phosphorus, and potassium; and refer to photosynthetic effect data to guide greenhouse shading, ventilation, and other environmental control operations.

[0138] This system not only possesses the basic functions of a greenhouse Internet of Things, but its core advantage lies in integrating the growth decision system disclosed in this invention as its core, providing users with intelligent suggestion output and greenhouse management support.

[0139] Example 3

[0140] I. Construction of Growth Indicators and Growth Period Prediction Model for Cut Chrysanthemums

[0141] This embodiment involves a total of twelve batches of cut chrysanthemum cultivation experiments conducted in Kaiyuan, Yunnan; Hushu, Nanjing; and Baguazhou, Nanjing. During the experiment, the growth period was observed every 2 days; plant height, stem diameter, and number of leaves were measured every 7 days; and the fresh and dry weight of each organ (stem, leaf, and flower) and leaf area per plant were measured every 14 days.

[0142] Greenhouse environmental data is automatically collected by a data logger (CR1000X, Campbell Scientific Inc., USA) equipped with sensors. The collected data includes meteorological parameters such as temperature, light intensity, humidity, and CO2 concentration, as well as soil parameters such as soil temperature and moisture content.

[0143] DVI (Digital Vibration Index) Calculation for Cut Chrysanthemums

[0144] The Growth and Development Index (DVI) is calculated based on the cumulative photothermal effect. Under suitable cultivation and management conditions, the growth and development of cut chrysanthemums are mainly determined by temperature, light, and developmental factors. Based on the light response and temperature response curves of cut chrysanthemum photosynthesis, the relative heat effect of cut chrysanthemum growth is defined as the ratio of the maximum net photosynthetic rate of leaves at the actual temperature to the maximum net photosynthetic rate of leaves at the optimum temperature; the relative light effect of cut chrysanthemum growth is defined as the ratio of the maximum net photosynthetic rate of leaves under actual photosynthetically active radiation to the maximum net photosynthetic rate of leaves under optimum photosynthetically active radiation. The cumulative photothermal effect is the product of the relative heat effect and the relative light effect. Based on the light response and temperature response curves of the multi-headed cut chrysanthemum 'Nanjing Agricultural University Little Golden Star', the calculation formula for the DVI of cut chrysanthemums is constructed as follows:

[0145] PTP i,j =Pn i,j ×RTE i,j

[0146]

[0147] DVI = APTP j / 610

[0148] In the formula, T i,j The temperature at the i-th half hour of day j, in °C; PAR i,j Photosynthetically active radiation (PAR) during the i-th half-hour of day j, μmol·m -2 ·s -1 .

[0149] Construction of Chrysanthemum Fertility Prediction Model

[0150] Using DVI as the driving variable, a model for predicting the reproductive period of cut chrysanthemums is established through piecewise functions:

[0151] Prediction of fertility period:

[0152]

[0153] Construction of a Prediction Model for Growth Indicators of Cut Chrysanthemums

[0154] The study found that, with the cumulative light and temperature index as the X-axis, the growth trends of various growth indicators of cut chrysanthemums all showed an S-shaped pattern of "slow-fast-slow".

[0155] A growth index prediction model was established for cut chrysanthemum 'Nanjing Agricultural University Little Golden Star' using DVI as the driving variable and based on the Logstic model. This model includes sub-models for predicting leaf area, plant height, stem diameter, number of leaves, and biomass. The specific calculation formulas for each sub-model are as follows:

[0156] Leaf area prediction: LAI = 8.8 / (1 + 26.77 × exp(-8.06 × DVI))

[0157] Plant height prediction: pH = 93.5 / (1 + 17.32 × exp(-6.82 × DVI))

[0158] Stem diameter prediction: SD = 5.96 / (1 + 1.73 × exp(-4.39 × DVI))

[0159] Leaf number prediction: LN = 35.2 / (1 + 6.58 × exp(-5.09 × DVI))

[0160] Biomass prediction: TFH = 61.50 / (1 + 25.34 × exp(-6.01 × DVI))

[0161] Cut flower quality indicators include: number of flowers, flower diameter, fresh flower weight, and flower neck length. These indicators are calculated starting when cut flowers show buds. For example, the formula for calculating flower diameter is:

[0162]

[0163] II. Construction of a Prediction Model for Photosynthetic Effect of Cut Chrysanthemums

[0164] Multiple batches of photosynthetic assays of cut chrysanthemums were conducted in Kaiyuan, Yunnan; Hushu, Nanjing; and Baguazhou, Nanjing. During the vegetative and reproductive growth stages of each batch of cut chrysanthemums, the light response curves and carbon dioxide response curves of healthy functional leaves (7 leaves from the bottom) under different temperatures and humidity were measured using a Li-6400 photosynthesis system.

[0165] Based on the light response curves and carbon dioxide response curves of cut chrysanthemums at different temperatures, such as Figure 2 As shown, the optimal temperature for cut chrysanthemums is around 25℃ to 30℃. The maximum photosynthetic rate decreases with both rising and falling temperatures. The temperature response curve for the maximum photosynthetic rate can be fitted using trigonometric functions, as shown in the figure. Figure 3 Based on the light response curves and carbon dioxide response curves of cut chrysanthemums at different temperatures, such as... Figure 4 It was found that photosynthesis in cut chrysanthemums was not significantly inhibited under high humidity (60%–80%), but decreased significantly when humidity was below 40%. However, studies have shown that when humidity exceeds 90%, the stomata of cut chrysanthemum leaves close, and photosynthesis gradually ceases.

[0166] Using the leaf-level photosynthetic response function, single-factor photosynthetic effect response models were established, including response models for relative temperature effect, relative light intensity effect, relative humidity effect, and relative carbon dioxide effect, as well as a comprehensive effect model: instantaneous photosynthetic rate, the specific calculation formula is as follows:

[0167] Relative temperature effect:

[0168]

[0169] Relative light intensity effect:

[0170]

[0171] Relative humidity effect:

[0172]

[0173] Relative carbon dioxide effect: RC i,j =0.1229×C i,j 0.350

[0174] Instantaneous photosynthetic rate: IRn i,j =RTE i,j ×RPn i,j ×RRH i,j ×RC i,j ×16

[0175] III. Construction of Nutrient Decision Model for Cut Chrysanthemums

[0176] Multiple batches of nutrient treatment experiments on cut chrysanthemums were conducted in Kaiyuan, Yunnan and Hushu, Nanjing. Destructive sampling was performed every 7 days. After sampling, the fresh weight of the above-ground stems, leaves, and flowers was measured using a 0.01g precision balance. The samples were then blanched at 105℃ for 15 minutes and dried at 75℃ to constant weight. The dry weight of each organ was then measured using a 0.0001g precision balance. To determine the nitrogen content of the organs, the dry weight of the destructive samples taken every 14 days was used. The organs were then ground using a small grinder and further pulverized in test tubes using a high-throughput grinder. The plant samples were digested using the H₂SO₄-H₂O₂ method, and the total nitrogen content in the digest was determined using an AA3 flow analyzer.

[0177] Critical nitrogen accumulation (also known as critical nitrogen content) is defined as the minimum nitrogen accumulation required to maximize all growth indicators of cut chrysanthemum. By analyzing the relationship between dry matter and critical concentration dilution models, a dry matter-driven critical nitrogen accumulation prediction model can be established, such as... Figure 5 As shown, a critical nitrogen accumulation prediction model was further established through model fusion, using dry matter prediction as the process model and cumulative photothermal effect as the driving variable. Figure 6 .

[0178] Nitrogen fertilizer utilization rates vary at different growth stages of cut chrysanthemums, and the optimal ratios of nitrogen, phosphorus, and potassium application also differ. A critical nitrogen accumulation prediction model for cut chrysanthemums was established using the Developmental Growth Index (DVI) as the driving variable. Then, nutrient decision models for daily nitrogen requirements, applied nitrogen rate, applied phosphorus rate, and applied potassium rate were established. The specific calculation formulas are as follows:

[0179] Critical nitrogen accumulation prediction: N uptc =382 / (1+46.661e) -5.662×DVI )

[0180] Nitrogen requirement forecast for day j: N uptc,j =N uptc,j -N uptc,j-1

[0181] Nitrogen application rate prediction for day j:

[0182] NUE = -0.76 × DVI 2 +1.43×DVI-0.0038

[0183] Prediction of P application rate on day j:

[0184] Prediction of K application rate on day j:

[0185] IV. Construction of Water Decision Model for Cut Chrysanthemums

[0186] Multiple batches of water requirement determination experiments for cut chrysanthemums were conducted in Kaiyuan, Yunnan and Hushu, Nanjing. Meteorological data were monitored using environmental sensors, and the reference crop evaporation rate (ET0) was calculated based on a modified Penman formula. Water requirement (ETc) and evaporation rate (Epan) of the cut chrysanthemums were collected in real time using weight sensors. The water requirement of cut chrysanthemums increased with plant growth, mainly due to the increase in leaf area index (LAI), which increased the transpiration area. The relationship between water requirement (ETc), leaf area index (LAI), reference crop evaporation rate (ET0), and evaporation rate (Epan) of cut chrysanthemums is shown in the figure below. Figure 7 As shown.

[0187] Water Decision Model Based on ET0

[0188] Since there is no wind in the greenhouse, the reference crop evaporation ET0 needs to be calculated using the modified Penman formula, as follows:

[0189] T mean =mean{T 1,j ···T 48,j}

[0190] T min =min{T 1,j ···T 48,j}

[0191] T max =max{T 1,j ···T 48,j}

[0192] RH mean=mean{RH 1,j RH 48,j}

[0193] RH min =min{RH 1,j RH 48,j}

[0194] RH max =max{RH 1,j RH 48,j}

[0195]

[0196] In the formula: T i,j The temperature at the i-th half hour of day j, in °C; PAR i,j Photosynthetically active radiation (PAR) during the i-th half-hour of day j, μmol·m -2 ·s -1 RH i,j The humidity is % for the i-th half hour of day j.

[0197] Combined with leaf matrix index to predict crop coefficient K c1 Predicting the water requirement ET for cut chrysanthemums c1 The calculation formula is as follows:

[0198] Water demand ET based on ET0 c1 Prediction: ET c1 =K c1 ×ET0

[0199] K c1 =0.634×LAI 0.505

[0200] Prediction of irrigation volume I1 using coupled moisture sensor: I1 = θ × ET c1 +Δ w

[0201] θ = 0.826 DVI -0.12

[0202]

[0203] Water Decision Model Based on Epan

[0204] Both water surface evaporation and plant transpiration are influenced by environmental meteorological factors and show a consistent trend. Therefore, using Epan to predict ETC (evaporation-to-transpiration rate) is highly accurate and cost-effective. However, since current water surface evaporation is frequently affected by greenhouse humidification, pesticide application, or sprinkler irrigation, the accuracy of water management decisions is higher when there is no spraying. The specific calculation formula is as follows:

[0205] Daily water surface evaporation Epan calculation model: Epan = EP 48,j -EP 48,j-1

[0206] Note: EP 48,j The water level is the 48th on day j.

[0207] Based on Epan's water demand ET c2 Prediction: ET c2 =K c2 ×E pan

[0208] K c2 =1.38×LAI 0.55

[0209] Prediction of irrigation volume I2 using coupled moisture sensor: I2 = θ × ET c2 +Δ w

[0210] θ = 0.826 DVI -0.12

[0211]

[0212] V. Chrysanthemum Growth Decision Model Library

[0213] Based on the above growth decision models, a set of growth decision model parameters for cut chrysanthemums was constructed. This set of parameters is applicable to the growth decision management of varieties such as 'Nanjing Agricultural University Little Golden Star', 'Nanjing Agricultural University Pink Oriole', 'Kennedy Milk Yellow', 'Nanjing Agricultural University Hengqun', 'Salmon', 'Pink Beauty', 'Flower Fairy', 'Pure Heart', and 'Colorful Cloud'.

[0214] Table 1. Parameters of the Chrysanthemum Growth Decision Model

[0215]

Claims

1. A cut chrysanthemum growth decision-making method based on multi-model fusion, characterized in that, The method comprises the following steps: S1. Collecting greenhouse environment parameters and soil parameters in real time through environment sensors; S2. According to the planted cut flower chrysanthemum variety, calling the corresponding calibration model parameter set in the preset model parameter library; S3. The following sub-models are executed synchronously through a multi-model fusion decision engine: (a) Growth period prediction model: using growth development index DVI as the driving variable, predicting the growth period through a segmented function; (b) Growth index prediction model: predicting daily plant height, stem diameter, leaf number, leaf area index, and biomass based on growth development index DVI; (c) Photosynthetic effect model: using a leaf-level photosynthetic response function to calculate relative temperature effect, relative light intensity effect, relative humidity effect, relative carbon dioxide effect, and instantaneous photosynthetic rate; (d) Nutrient decision model: calculating the critical nitrogen accumulation amount, nitrogen requirement amount, nitrogen application amount, phosphorus application amount, and potassium application amount through the critical nitrogen accumulation amount model of DVI and the nitrogen, phosphorus, and potassium application ratio at each growth stage; (e) Water decision model: predicting the water requirement and irrigation amount in a set period; S4. Decision output: outputting the comprehensive decision results including the calculation results of models (a)-(e) in S3.

2. The cut chrysanthemum growth decision method based on multi-model fusion according to claim 1, characterized in that, The environment parameters include temperature, light intensity, humidity, and CO2 concentration, and the soil parameters include soil temperature and water content. 3.The cut chrysanthemum growth decision method based on multi-model fusion according to claim 1, characterized in that, The preset model parameter library contains multiple sets of calibration model parameters, respectively corresponding to multiple representative varieties of single-headed cut flowers and multiple varieties of multi-headed cut flowers.

4. The multi-model fusion based cut chrysanthemum growth decision method according to claim 1, characterized in that, The growth period is predicted by the following function: In the formula, ba-bd are model parameters fitted for the corresponding variety of cut flowers according to a large number of tests.

5. The multi-model fusion based cut chrysanthemum growth decision method according to claim 1 or 4, characterized in that, The growth development index DVI is calculated by cumulative light and heat effect, and the specific calculation formula is: In the formula, T i,j The temperature at the i-th half hour of day j, in °C; PAR i,j Photosynthetically active radiation (PAR) during the i-th half-hour of day j, μmol·m -2 ·s -1 ;RTE i,j For relative thermal effects; Pn i,j For absolute light effect; PTP j The cumulative photothermal effect is denoted by α; aa-ag represents the model parameters fitted to the corresponding varieties of cut chrysanthemum based on numerous experiments.

6. The multi-model fusion based cut chrysanthemum growth decision method according to claim 5, characterized in that, The growth index prediction model includes leaf area, plant height, stem diameter, leaf number, biomass, and flower quality prediction sub-models, each of which is predicted by taking DVI as the driving variable, and the specific calculation formula of each sub-model is: Leaf area prediction: LAI = be / (1+bf×exp(-bg×DVI) Plant height prediction: PH = bh / (1+bi×exp(-bj×DVI) Stem diameter prediction: SD = bk / (1+bl×exp(-bm×DVI) Leaf number prediction: LN = bn / (1+bo×exp(-bp×DVI) Biomass prediction: TFH = bq / (1+br×exp(-bs×DVI) Flower quality prediction: In the formula, be-bv are model parameters fitted for the corresponding variety of cut flowers according to a large number of tests.

7. The multi-model fusion based cut chrysanthemum growth decision method according to claim 1, characterized in that, The calculation formula of the relative temperature effect is: The calculation formula of the relative light intensity effect is: The calculation formula of the relative humidity effect is: The formula for calculating the relative carbon dioxide effect is: RC i,j = al x RC i,j am The formula for calculating the instantaneous rate of photosynthesis is: IRn i,j = RTE i,j x RPn i,j x RRH i,j x RC i,j x an where RH i,j Ci is the humidity, %; C i,j Ci is the carbon dioxide concentration, ppm; aa-an are model parameters fitted for the respective variety of cut chrysanthemum according to a large number of experiments. 8.The cut chrysanthemum growth decision method based on multi-model fusion according to claim 1, characterized in that, The specific calculation formula of the nutrient decision model is: Critical nitrogen accumulation prediction: N uptc = ca / (1 + cb x e -cc×DVI ) Nitrogen requirement prediction for day j: N uptc,j = N uptc,j - N uptc,j-1 Day j nitrogen application prediction: NUE = -cd x DVI + ce x DVI - cf 2 + ce x DVI - cf Phosphorus application amount prediction on day j: Potassium application amount prediction on day j: In the formula, NUE is the nitrogen fertilizer utilization rate, and ca-cl are model parameters fitted for the corresponding variety of cut flowers according to a large number of tests. 9.The cut chrysanthemum growth decision method based on multi-model fusion according to claim 1, characterized in that, The water decision model generates an irrigation scheme based on the reference crop evaporation and water evaporation, combined with the real-time growth state of cut flowers and soil water content, and specifically includes: ET0-based water requirement ET c1 Forecast: ET c1 = da x LAI db x ET0 Epan-based water requirement ET c2 Prediction: ET c2 = dc x LAI dd x Epan Coupling water sensor irrigation amount I1 prediction: I1 = de x DVI -df x ET c1 + Δ w Coupling moisture sensor irrigation volume I2 prediction: I2 = de x DVI -df x ET c2 + Δ w where VWC is the volumetric water content of the soil, %; Δ w is the soil water deficit; da-dfare model parameters fitted for the respective cultivar of cut chrysanthemum according to a large number of experiments.

10. A cut chrysanthemum growth decision system based on multi-model fusion, characterized in that, The system is suitable for the cut flower chrysanthemum growth decision method based on multi-model fusion according to any one of claims 1-9, comprising: An environmental data collection module: real-time collection of greenhouse environmental parameters and soil parameters; A variety identification and parameter matching module: calling corresponding calibration model parameter sets in a preset model parameter library according to a planting variety; A multi-model fusion decision engine module: integrating and running a growth index prediction model, a photosynthetic effect model, a nutrient decision model and a water decision model, receiving data of the environmental data collection module and parameter sets output by the variety identification and parameter matching module, and performing calculation; A decision output module: generating and outputting comprehensive decision results including growth prediction, environmental effect evaluation and cultivation management suggestions.

Images