Data-visualized tomato yield and growth prediction model and method

Abstract

The invention relates to the field of agricultural informatization and data analysis, and discloses a data-visualized tomato yield and growth prediction model and method.The method comprises the following steps that processing parameters in the tomato growth process are obtained, and the parameters comprise the irrigation amount and the nitrogen application amount; tomato growth data are collected, and tomato yield and dry matter data are generated; analyzing the irrigation amount, the nitrogen application amount and the tomato yield data, and establishing a mathematical model of tomato growth; and graphically displaying the results of the tomato yield and growth prediction model by using a data visualization tool. According to the method, the influence of different kinds of water and fertilizer management on tomato growth can be accurately analyzed, a reference basis is provided for scientific management of water and fertilizer, and the model is evaluated through indexes such as R2, RMSE and MAE and has high prediction precision and reliability.

Description

Technical Field

[0001] The present invention relates to the field of agricultural informatization and data analysis technology, and in particular to a data visualized tomato yield and growth prediction model and method. Background Art

[0002] In modern agriculture, tomato is an important economic crop, and its yield and quality are directly affected by water and fertilizer management strategies. However, the growth process of tomato is complex, and its yield is closely related to many factors, including irrigation amount, nitrogen application amount, soil conditions and climate environment.

[0003] Current tomato yield prediction methods mostly rely on empirical rules or simple regression models, which are difficult to fully and accurately describe the complex effects of water and fertilizer management on yield. Traditional models often ignore the interaction between water and fertilizer, resulting in poor prediction of yield changes under different irrigation and fertilization combinations.

[0004] In addition, existing analysis methods lack effective data visualization methods, making it difficult to intuitively display and understand the prediction results, and it is difficult for agricultural producers to make optimized management decisions based on the prediction results.

[0005] Therefore, how to predict tomato yield more comprehensively and accurately through scientific data analysis and visualization technology has become a technical problem that needs to be urgently solved in modern agricultural production. Summary of the invention

[0006] In view of the shortcomings of the existing technology, the present invention provides a data visualization tomato yield and growth prediction model and method, which uses water and fertilizer management data to accurately predict tomato yield and intuitively display the yield change trend under different management conditions through visualization, so as to support scientific water and fertilizer management decisions.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for predicting tomato yield and growth with data visualization, comprising the following steps:

[0008] Obtaining treatment parameters during tomato growth, the parameters including irrigation amount and nitrogen application amount;

[0009] Collect tomato growth data and generate tomato yield and dry matter data;

[0010] Analyzing the irrigation amount, nitrogen application amount and tomato yield data to establish a mathematical model of tomato growth;

[0011] The results of the tomato yield and growth prediction model are graphically displayed using a data visualization tool.

[0012] Preferably, the step of obtaining processing parameters during tomato growth includes:

[0013] The evapotranspiration of reference crops was calculated by the FAOPenman-Monteith formula, and the daily evapotranspiration of tomatoes was calculated by combining the crop coefficients of tomatoes at different growth stages;

[0014] The total irrigation amount during the experiment was the accumulation of daily evapotranspiration of tomatoes, and if there was rainfall during the experiment, the effective rainfall was deducted;

[0015] Nitrogen application rates were calculated based on nutrient requirements, and multiple levels of irrigation and nitrogen application were set to test the effects of different conditions on yield.

[0016] Preferably, the step of collecting tomato growth data comprises:

[0017] The tomato growth process is divided into several stages, including flowering stage, first fruit expansion stage, second fruit expansion stage, third fruit expansion stage and fourth fruit expansion stage;

[0018] At the fourth fruit expansion stage, several sample plants were collected from each experimental treatment group and divided into roots, stems, leaves and fruits, and the dry matter weight was measured after drying;

[0019] Tomato yield was determined by fruit fresh weight data.

[0020] Preferably, the method further comprises performing a two-way ANOVA on the relationship between the irrigation amount and nitrogen application amount and yield data to quantify their significant effects on tomato yield.

[0021] Preferably, the two-way ANOVA adopts the R-style linear model form:

[0022] Z~X+Y+X:Y

[0023] Among them, Z is the yield, X is the irrigation amount, Y is the nitrogen application amount, and X:Y is the interaction between irrigation and nitrogen application.

[0024] Preferably, the method further comprises using the linear interpolation method of the scipy library to interpolate the irrigation amount, nitrogen application amount and yield data to generate continuous data for modeling.

[0025] Preferably, the mathematical model of tomato yield adopts a quadratic polynomial model, which is expressed as:

[0026]

[0027] in, represents the predicted yield, x represents the irrigation amount, y represents the nitrogen application amount, and a, b, c, d, e, and f are model parameters.

[0028] Preferably, the model parameters are solved by the least square method so that the residual sum of squares between the model prediction value and the actual value is minimized, and the model parameters are solved in the following matrix form:

[0029]

[0030] Among them, X1 = x, X2 = y, X3 = xy, X4 = x 2 、X5=y 2 .

[0031] Preferably, the method further comprises evaluating the fitting accuracy of the model, and the evaluation index comprises the determination coefficient R 2 , root mean square error, and mean absolute error, where:

[0032] Coefficient of determination R 2 Used to evaluate the model's ability to explain yield variation;

[0033] The root mean square error is used to measure the deviation between the model prediction value and the actual value;

[0034] The mean absolute error is a measure of the absolute deviation between the model's predicted values ​​and the actual values.

[0035] The present invention also provides a data-visualized tomato yield and growth prediction model, which is based on irrigation amount and nitrogen application amount data, and solves model parameters through the least squares method to achieve prediction of tomato yield.

[0036] The present invention provides a data visualization tomato yield and growth prediction model and method. It has the following beneficial effects:

[0037] 1. The present invention provides a method for accurately predicting tomato yield. The method comprehensively considers the effects of irrigation amount and nitrogen application amount on tomato yield, and establishes a tomato yield prediction model based on water-fertilizer coupling effect. The method can more accurately estimate the effects of different water-fertilizer management strategies on tomato yield, provide agricultural producers with reliable prediction data, and help make more effective decisions.

[0038] 2. The present invention establishes a prediction model for tomato yield by using data interpolation and mathematical modeling methods. It innovatively uses the data interpolation method to perform linear interpolation on the existing data. The data interpolation method further optimizes the simulation model accuracy. The R 2 Both are greater than 0.80.

[0039] 3. The present invention creates a visualization of the yield model, which enables agricultural producers to more clearly observe and understand the relationship between tomato yield and water and fertilizer management strategies. Such visualization helps to quickly identify changing trends in tomato production and make the decision-making process more information-based. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0041] Figure 2 A schematic diagram of a method flow of an embodiment of the present invention;

[0042] Figure 3 This is a distribution diagram of the output data of an embodiment of the present invention;

[0043] Figure 4 This is a variance analysis diagram of an embodiment of the present invention;

[0044] Figure 5 This is a fitting model diagram of an embodiment of the present invention;

[0045] Figure 6 It is a contour map of the fitting model of the embodiment of the present invention;

[0046] Figure 7 It is a cumulative distribution function diagram of the test value and the model prediction value of the embodiment of the present invention;

[0047] Figure 8 It is a box plot of the relationship between the test value and the model prediction value of the embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Please see attached Figure 1 The present invention provides a method for predicting tomato yield and growth with data visualization. By introducing key factors in the growth process such as irrigation amount and nitrogen application amount, combined with data visualization technology, it helps agricultural producers to intuitively understand the growth status and yield prediction of tomatoes. This method covers the entire process of parameter acquisition, data collection, mathematical modeling and data visualization, which helps to accurately control water and fertilizer, increase yield, and optimize the efficiency of agricultural resource utilization. The following is a detailed description of each step of the present invention.

[0050] S1. Obtaining processing parameters during tomato growth

[0051] Step S1 involves obtaining treatment parameters such as irrigation amount and nitrogen application amount, which directly affect the water absorption and nutrient supply of tomatoes, and have an important impact on the growth rate and final yield. In order to accurately evaluate the specific impact of these factors on tomato growth, the present invention obtains relevant data through a standardized formula calculation method at different growth stages to ensure the accuracy and consistency of the model input.

[0052] In this embodiment, step S1 is described in detail as follows:

[0053] 1. Obtaining irrigation amount

[0054] In this embodiment, the irrigation amount is obtained based on the FAO Penman-Monteith formula to calculate the reference crop evapotranspiration (ET0) during tomato growth. The specific calculation process is as follows:

[0055] First, ET0 is calculated using the FAO Penman-Monteith formula, as follows:

[0056]

[0057] in:

[0058] Δ is the slope of the saturated water vapor pressure curve (kPa / ℃);

[0059] R n is the net radiation (MJ / m 2 / d);

[0060] G is the soil heat flux (MJ / m 2 / d);

[0061] γ is the humidity constant (kPa / ℃);

[0062] T is the average temperature (℃);

[0063] u2 is the wind speed at a height of 2 meters (m / s);

[0064] e s -e a is the water vapor pressure difference (kPa).

[0065] Next, combined with the crop coefficient (kc) of tomatoes at different growth stages, the actual daily evapotranspiration (ET) of tomatoes can be obtained. C ):

[0066] ET C =ET0×kc

[0067] The crop coefficient kc value is adjusted according to the growth stage of tomatoes to reflect the changes in water demand at different times. For example, the water demand is usually higher during the flowering and fruit expansion period, and the corresponding kc value is also higher.

[0068] Finally, the daily ET during the entire experimental period C The total irrigation amount of tomatoes is accumulated. If natural rainfall occurs during the irrigation cycle, the effective rainfall is deducted from the total irrigation amount to avoid water waste due to excessive irrigation. The calculation of effective rainfall will be determined based on local rainfall data, thereby further optimizing the irrigation amount setting.

[0069] 2. Calculation of nitrogen application amount

[0070] In this embodiment, the calculation of nitrogen application amount is based on the nutrient requirement formula of tomatoes at different growth stages. The nitrogen application amount directly affects the growth and yield of the plants, and appropriate amounts of nitrogen fertilizers need to be applied at different growth stages to ensure stable growth.

[0071] In the experimental design, the nitrogen application rate is set at different levels to study its effect on yield. Generally, the nitrogen application rate is set at 3 to 4 levels, increasing gradually from low to high, so that a comprehensive analysis of the nitrogen requirements of tomato growth can be carried out.

[0072] The calculation formula for nitrogen application is:

[0073] N app =N req ×A

[0074] in:

[0075] N app is the amount of nitrogen fertilizer applied,

[0076] N req is the nitrogen requirement of tomatoes in a specific growth period,

[0077] A is the area of ​​the test area.

[0078] In addition, in the present invention, by rationally allocating different combination levels of irrigation and nitrogen application, it is ensured that data are available for analysis under various water and fertilizer conditions.

[0079] In a preferred embodiment, the irrigation amount is set to 3 levels and the nitrogen application amount is set to 4 levels.

[0080] Specifically:

[0081] 1. Setting the irrigation level

[0082] In this embodiment, the irrigation amount is divided into three levels: low, medium and high, so as to study the effect of different water supply conditions on tomato growth. The irrigation amount setting of each level is determined according to the reference crop evapotranspiration calculated by the FAO Penman-Monteith formula and the crop coefficient of each growth period.

[0083] Under low-level irrigation conditions, the irrigation amount is ET C The lower level is ET to simulate drier growing conditions; the medium level is ET C The moderate ratio of ET provides the water required for normal growth of tomatoes; the high-level irrigation amount is slightly higher than ET C , provide adequate water supply.

[0084] 2. Setting the nitrogen application level

[0085] The nitrogen application rate was divided into four levels, also set as low, medium, higher and high, so as to observe the growth response of tomatoes under different nitrogen fertilizer application conditions.

[0086] The low nitrogen application level provides basic nutrient supply to simulate nitrogen-poor conditions; the medium nitrogen application level is the appropriate nutrient requirement of tomatoes; the higher and high levels increase nitrogen application to provide additional nutrient supply conditions, making it possible to analyze the effects of excessive nitrogen application on growth and yield.

[0087] This distribution of water and fertilizer levels helps to fully understand the combined effects of different combinations of irrigation and nitrogen application on tomato growth and yield, providing rich data support for subsequent modeling and variance analysis.

[0088] S2. Collect tomato growth data and generate tomato yield and dry matter data

[0089] The present invention is directed to a method for collecting tomato growth data, including the acquisition of yield and dry matter data. These data provide basic information for the establishment of subsequent mathematical models and yield prediction. In step S2, the collection of tomato growth data covers key biological characteristics at different growth stages, specifically including dry matter measurement and yield data recording of roots, stems, leaves and fruits. To ensure the integrity and accuracy of the data, this step records in detail various types of data of tomatoes during the growth process through a staged collection method.

[0090] In this embodiment, the specific implementation of step S2 is as follows:

[0091] 1. Stages of tomato growth

[0092] In this embodiment, in order to accurately reflect the physiological characteristics of each stage in the growth process of tomatoes, the entire growth process of tomatoes is divided into multiple key stages, including:

[0093] Flowering stage: Tomatoes enter the initial stage of reproductive growth, and preliminary growth data is recorded at this time to serve as the basis for subsequent growth trends.

[0094] The first fruit expansion period: The tomato fruit begins to expand and gradually accumulate dry matter.

[0095] The second fruit expansion period: the fruit continues to expand, dry matter accumulation further increases, and yield characteristics gradually emerge.

[0096] The third fruit expansion stage: the fruit is close to maturity and the dry matter accumulation reaches a relatively high level.

[0097] Fourth fruit expansion stage: The fruits are fully mature and sampling is carried out to ensure that the data reflects the final yield and dry matter accumulation.

[0098] This phased collection method ensures the integrity of the growth data, enables the characteristics of each growth stage to be accurately captured in subsequent analysis, and provides detailed data support for model building.

[0099] 2. Selection of sample plants and sampling by location

[0100] In the fourth fruit expansion stage, this example selected 5 sample plants from each experimental treatment group for sampling to ensure the representativeness of the data. Each sample plant was divided into the following four parts for sampling:

[0101] Roots: Mainly records the development of the plant's root system, representing the plant's ability to absorb nutrients.

[0102] Stem: records the growth of the stem, reflecting the overall structural support capacity of the plant.

[0103] Leaf: records the dry matter mass of the leaves, reflecting the efficiency of plant photosynthesis.

[0104] Fruit: Record the biomass of the fruit, which directly reflects the final data of yield.

[0105] Samples from each part are dried after sampling to remove moisture and measure their dry matter content. Dry matter data can provide accurate quantitative information on nutrient accumulation at different growth stages, ensuring the validity and consistency of data in subsequent analysis.

[0106] 3. Measurement of production data

[0107] In this embodiment, the fresh weight of the fruit at the fourth fruit expansion stage is used as the tomato yield indicator. The reason for choosing the fresh weight of the fruit at this stage is that the fruit is close to being fully mature at this time, and its biomass can accurately reflect the yield of each tomato plant.

[0108] The fresh weight of the fruit of each sample plant is recorded by standard measurement methods and input into the model as yield data. The acquisition of yield data provides key output variables for subsequent modeling, ensuring the accuracy and precision of the model's yield prediction.

[0109] Furthermore, in order to ensure the accuracy and data quality of tomato yield and dry matter data, the present invention also performs an outlier test and distribution analysis on the collected data in step S2. Specifically, the outliers in the data are detected and screened by the Grubbs'test method, and the distribution of tomato yield and other data is analyzed to provide more stable data input for subsequent modeling and prediction.

[0110] In this embodiment, the further specific implementation of step S2 is as follows:

[0111] 1. Inspection and screening of data outliers

[0112] In this example, to ensure data quality, the collected tomato yield and dry matter data were subjected to a Grubbs'test outlier test, with the significance level set at α = 0.05. The Grubbs'test method is used to detect a single outlier in a sample, is suitable for the analysis of small samples, and helps to identify possible extreme values.

[0113] For a data set (X1, X2, …, X n ), assuming the mean is The standard deviation is s, and the calculation formula for the outlier G is:

[0114]

[0115] Among them, X i represents a single data point in the sample, is the mean of the data, and s is the standard deviation of the sample.

[0116] Determination of significance level: Under the condition of significance level α = 0.05, calculate the corresponding critical value G critical If the calculated G value exceeds G critical , then the data is considered to be an outlier and should be removed.

[0117] Outlier screening process: When processing data, calculate the G value of each data point in turn and compare it with the critical value G critical By comparison, the data exceeding the critical value are marked as outliers and removed. This process ensures the uniformity of the data input into the model and reduces the impact of extreme values ​​on the prediction accuracy of the model.

[0118] 2. Checking the data distribution

[0119] In this embodiment, after outlier processing, the distribution of tomato yield and dry matter data was further tested. The distribution characteristics of the data have a direct impact on the model fitting method and the reliability of the prediction results.

[0120] Normality test: The Shapiro-Wilk normality test is used for data such as output to determine whether the data conforms to the normal distribution. The result of the normality test determines whether the assumptions of the model are valid, ensuring that the appropriate data fitting method is used in the subsequent modeling process.

[0121] Distribution analysis: If the data does not conform to the normal distribution, this embodiment will select other appropriate distribution models (such as lognormal distribution or gamma distribution) for modeling based on the distribution characteristics to improve the adaptability of the model to non-normal distribution data.

[0122] Through the above-mentioned outlier test and distribution analysis steps, the high quality and consistency of the collected data are ensured, which provides a reliable data basis for the establishment of the subsequent mathematical model and improves the stability of the prediction results.

[0123] S3. Analyze the irrigation amount, nitrogen application amount and tomato yield data to establish a mathematical model for tomato growth

[0124] By analyzing key data such as irrigation amount, nitrogen application amount and yield, a mathematical model of tomato growth was established by combining two-factor variance analysis with mathematical modeling, providing a scientific basis for data prediction. In step S3, the model was established based on the comprehensive effect of irrigation and nitrogen application on yield, and the parameters were optimized by the least squares method so that the model could reflect the complex effects of water and fertilizer on tomato yield.

[0125] In this embodiment, the specific implementation of step S3 is as follows:

[0126] 1. Data Analysis and Two-Way ANOVA

[0127] In this example, in order to quantify the independent and interactive effects of irrigation amount and nitrogen application amount on tomato yield, a two-way ANOVA was used to analyze the significance and synergistic effect of these two factors. The analysis process is as follows:

[0128] Use the R-style formula in the statsmodels library to build a linear model. The formula is:

[0129] Z~X+Y+X:Y

[0130] Among them, Z is yield (dependent variable), X is irrigation amount (independent variable), Y is nitrogen application amount (independent variable), and X:Y is the interaction between irrigation and nitrogen application.

[0131] Through two-way ANOVA, the main effects of each factor (independent effects of irrigation and nitrogen application) and the interaction effects can be quantified, revealing the significant effects of the two factors on yield under different combinations. The results of the variance analysis will provide parameter references for subsequent modeling to ensure that the model's response to changes in irrigation and nitrogen application levels is consistent with actual conditions.

[0132] 2. Data interpolation processing

[0133] In data analysis, there may be some missing values ​​or discontinuities in the irrigation amount and nitrogen application amount levels. Therefore, this embodiment uses a linear interpolation method to process the data to ensure the integrity of the model input.

[0134] Specifically, the griddata function of the scipy library was used to generate continuous irrigation and nitrogen application data through linear interpolation. This method can effectively fill the gaps in the test data and make the data points evenly distributed, thereby improving the stability of the model and the prediction accuracy.

[0135] 3. Establishment of mathematical model and parameter solution

[0136] In this embodiment, the mathematical model of tomato growth is constructed based on the combined effects of irrigation amount and nitrogen application amount, and the least squares method is used to solve the model parameters to minimize the residual sum of squares between the model prediction value and the measured value. The specific steps are as follows:

[0137] The model established is a multi-factor regression model, and its mathematical expression is:

[0138]

[0139] in, represents the predicted yield, x represents the irrigation amount, y represents the nitrogen application amount, and a, b, c, d, e, and f are model parameters.

[0140] The model parameters are calculated by minimizing the residual sum of squares using the least squares method. The calculation process is as follows:

[0141] Convert z=f(x,y) into the following linear model:

[0142]

[0143] Among them, X1 = x, X2 = y, X3 = xy, X4 = x 2 , X5=y 2 .

[0144] The residual sum of squares of the model is:

[0145]

[0146] According to the least squares principle, to make SSe To reach the minimum, the following conditions are met:

[0147]

[0148] This gives the following matrix:

[0149]

[0150] Among them, a, b, c, d, e, and f are the coefficients of the fitting function, x is the irrigation amount, y is the nitrogen application amount, and z is the tomato yield.

[0151] Through the data analysis, interpolation processing and mathematical modeling in step S3, a quantitative analysis of the effects of irrigation and nitrogen application on tomato yield is achieved, providing a reliable mathematical basis for subsequent processing of data visualization and prediction models.

[0152] S4. Using data visualization tools, graphically display the results of the tomato yield and growth prediction model

[0153] By graphically displaying the data output by the model, the yield change trend under different combinations of irrigation and nitrogen application can be more intuitively displayed, which is convenient for observation and analysis. Step S4 combines 3D scatter plots and surface fitting to present the prediction results of the model with clear visual effects, thereby supporting precise management in agricultural production.

[0154] In this embodiment, the specific implementation of step S4 is as follows:

[0155] 1. 3D scatter plot display of data

[0156] In this embodiment, in order to preliminarily display the yield data distribution under different irrigation and nitrogen application levels, a 3D scatter plot is used for visual display. The specific implementation process is as follows:

[0157] In the matplotlib library, import the mplot3d module and use the scatter function to plot the irrigation amount, nitrogen application amount, and yield data into a 3D scatter plot.

[0158] The X-axis represents the amount of irrigation, the Y-axis represents the amount of nitrogen applied, and the Z-axis represents the yield. The positions of the respective data points reflect the yield under different water-fertilizer combinations.

[0159] This 3D scatter plot display helps to observe the overall distribution of data points and provide a reference for subsequent model fitting.

[0160] 2.3D surface fitting and display

[0161] In this embodiment, in order to more intuitively demonstrate the effect of water-fertilizer combination on yield, a 3D surface fitting method is further used to smooth and visualize the data. The specific steps are as follows:

[0162] First, based on the griddata function of the scipy library, linear interpolation is performed on the irrigation amount, nitrogen application amount, and yield data to generate continuous surface data from discrete data points. This interpolation process can make up for the uneven distribution of the test data and make the prediction surface smoother.

[0163] Then, the interpolated data was plotted into a 3D surface using the plot_surface function in the matplotlib library. The surface was generated by fitting a quadratic polynomial model, which expressed the nonlinear effects of different combinations of irrigation and nitrogen application on yield.

[0164] The mathematical expression of the surface model is:

[0165]

[0166] in, To predict the yield, x is the irrigation amount, y is the nitrogen application amount, and the parameters a, b, c, d, e, and f have been solved by the least squares method in the above step S3.

[0167] 3. Evaluation indicators of model fitting effect

[0168] In this embodiment, in order to evaluate the surface fitting effect and the prediction accuracy of the model, the following indicators are introduced:

[0169] Coefficient of determination R 2 : It is used to measure the model's ability to explain changes in production. The calculation formula is:

[0170]

[0171] Among them, z i is the experimental measurement value of tomato yield, is the model prediction for tomato yield, is the average weight of dried tomatoes.

[0172] Root mean square error (RMSE): It is used to measure the deviation between the model prediction value and the actual value. The formula is:

[0173]

[0174] Among them, z i is the experimental measurement value of tomato yield, is the model prediction value of tomato yield, and n is the sample size.

[0175] Mean absolute error (MAE): It is used to measure the average absolute deviation between the model prediction value and the actual value. The formula is:

[0176]

[0177] Among them, z i is the experimental measurement value of tomato yield, is the model prediction value of tomato yield, and n is the sample size.

[0178] Through the above steps, the data visualization process presents the model prediction results in the form of intuitive 3D scatter plots and surface plots, and measures the accuracy of the model through a variety of error evaluation indicators, providing agricultural producers with a reliable reference for water and fertilizer management.

[0179] In general, the present invention obtains key growth parameters such as irrigation amount and nitrogen application amount, combines growth data collection, two-way variance analysis and mathematical modeling, and establishes a prediction model that reflects the impact of water and fertilizer on tomato yield. The method uses the FAOPenman-Monteith formula to calculate water demand, and collects dry matter and yield data in layers according to the growth stage to ensure the scientificity and integrity of the data. The model uses the least squares method to optimize parameters, and combines 3D scatter plots and surface plots for data visualization, so that producers can intuitively understand the yield changes under different water and fertilizer conditions, provide accurate water and fertilizer management guidance, and improve resource utilization efficiency and yield prediction accuracy.

[0180] The present invention also provides a data-visualized tomato yield and growth prediction model, which is based on the relationship between key parameters in the growth process such as irrigation amount and nitrogen application amount, determines the significance of each factor on yield through two-way variance analysis, and combines mathematical modeling with data visualization technology to achieve prediction of tomato yield. The model is established using growth data collected in the early stage, and the model parameters are optimized by the least squares method to improve the prediction accuracy. Through the generated 3D scatter plot and surface plot, the model can intuitively display the influence trend of different water-fertilizer combinations on yield, so that users can better apply data guidance in production management, optimize water-fertilizer configuration, and achieve scientific prediction of yield and efficient use of resources.

[0181] Example:

[0182] Please see attached Figure 2 -Attached Figure 8 , an embodiment of the present invention provides a method for predicting tomato yield and growth based on data visualization, comprising the following steps:

[0183] S1. Obtaining experimental treatment parameters of tomatoes (irrigation amount, nitrogen application amount);

[0184] S2. Obtaining parameters such as tomato yield at maturity (root, stem, leaf, fruit dry matter, single plant yield);

[0185] S3. Test the outliers of output data and the distribution of output data;

[0186] S4. Perform variance analysis on the relationship between irrigation amount, nitrogen application amount and yield data;

[0187] S5. Model fitting is performed after data interpolation, and visualization is achieved through the matplotlib library.

[0188] The experimental processing parameters of this embodiment are shown in Table 1.

[0189] Table 1 Experimental treatment parameters

[0190]

[0191] The parameters such as the yield of the fourth fruit expansion period test of this embodiment are shown in Table 2.

[0192] Table 2 Parameters of yield and other parameters during the fourth fruit expansion period

[0193]

[0194] It is to be noted that the values ​​of each index in the table are the average of 5 samples; the same below. The variance analysis of tomato yield per plant and irrigation amount and nitrogen application amount is shown in Table 3.

[0195] Table 3 Variance analysis of single plant yield, irrigation amount and nitrogen application amount

[0196] Source of Difference SS df F P-value Significance Watering amount 1.39 2.0 37.12 1.806×10-10 ** Nitrogen application rate 6.09 3.0 108.24 2.277×10-21 ** Irrigation amount × nitrogen application amount 1.17 6.0 10.42 2.224×10-7 ** error 0.90 48.0 - - - sum 9.55 59.0 - - -

[0197] The embodiment uses the form z=f(x,y)=a+bx+cy+dxy+ex 2 +fy 2 The mathematical model was used to fit the tomato yield and dry matter. The 3D surface diagram of the tomato yield fitting model is shown in Figure 5 As shown, the contour map of the fitted model is as follows Figure 6 The fitting model is shown in Table 4. The coefficient of determination R 2 The value close to 1 indicates a good fitting effect; the smaller the root mean square error RMSE and mean absolute error MAE values ​​are, the better the consistency between the model prediction value and the test value is, and the smaller the deviation between the prediction value and the test value is, that is, the more accurate and reliable the model fitting result is. The calculation results are shown in Table 4. The R 2 They are all greater than 0.80, and the RMSE and MAE values ​​are small, which shows that the model fitting effect is good.

[0198] Table 4 Fitting model parameters and effects

[0199]

[0200] It should be noted that the unit of dry weight of roots, stems, leaves and fruits is gram (g), and the unit of single plant yield is kilogram (kg); the same below.

[0201] The yield prediction model was used to calculate the yield prediction value corresponding to the irrigation amount and the nitrogen application amount, and the degree of conformity between the model prediction value and the test value was statistically analyzed. The predicted value (SIM) and the test value (OBS) calculated by the model are shown in Table 5, and the statistical analysis results are shown in Table 5. Figure 7 , Figure 8 As shown, the results show that the deviation between the model prediction value and the experimental value is small and the degree of consistency is high.

[0202] Table 5 Predicted values ​​of fitting model and experimental values

[0203]

[0204]

[0205] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for predicting tomato yield and growth based on data visualization, characterized in that: The following steps are involved: Obtaining treatment parameters during tomato growth, the parameters including irrigation amount and nitrogen application amount; Collect tomato growth data and generate tomato yield and dry matter data; Analyzing the irrigation amount, nitrogen application amount and tomato yield data to establish a mathematical model of tomato growth; The results of the tomato yield and growth prediction model are graphically displayed using a data visualization tool.

2. A method for predicting tomato yield and growth based on data visualization according to claim 1, characterized in that: The step of obtaining the processing parameters during the growth of tomatoes includes: The evapotranspiration of reference crops was calculated using the FAO Penman-Monteith formula, and the daily evapotranspiration of tomatoes was calculated using the crop coefficients of tomatoes at different growth stages; The total irrigation amount during the experiment was the accumulation of daily evapotranspiration of tomatoes, and if there was rainfall during the experiment, the effective rainfall was deducted; Nitrogen application rates were calculated based on nutrient requirements, and multiple levels of irrigation and nitrogen application were set to test the effects of different conditions on yield.

3. The method for predicting tomato yield and growth based on data visualization according to claim 1, characterized in that: The step of collecting tomato growth data comprises: The tomato growth process is divided into several stages, including flowering stage, first fruit expansion stage, second fruit expansion stage, third fruit expansion stage and fourth fruit expansion stage; At the fourth fruit expansion stage, several sample plants were collected from each experimental treatment group and divided into roots, stems, leaves and fruits, and the dry matter weight was measured after drying; Tomato yield was determined by fruit fresh weight data.

4. The method for predicting tomato yield and growth based on data visualization according to claim 1, characterized in that: The method also includes performing a two-way ANOVA on the relationship between the irrigation amount and nitrogen application amount and yield data to quantify their significant effects on tomato yield.

5. A method for predicting tomato yield and growth based on data visualization according to claim 4, characterized in that: The two-way ANOVA adopts the R-style linear model form: Z~X+Y+X:Y Among them, Z is the yield, X is the irrigation amount, Y is the nitrogen application amount, and X:Y is the interaction between irrigation and nitrogen application.

6. The method for predicting tomato yield and growth based on data visualization according to claim 1, characterized in that: The method also includes using a linear interpolation method of the scipy library to interpolate the irrigation amount, nitrogen application amount and yield data to generate continuous data for modeling.

7. The method for predicting tomato yield and growth based on data visualization according to claim 1, characterized in that: The mathematical model of tomato yield adopts a quadratic polynomial model, which is expressed as: in, represents the predicted yield, x represents the irrigation amount, y represents the nitrogen application amount, and a, b, c, d, e, and f are model parameters.

8. The method for predicting tomato yield and growth based on data visualization according to claim 7, characterized in that: The model parameters are solved by the least square method to minimize the residual sum of squares between the model prediction value and the actual value, and the model parameters are solved in the following matrix form: For example,X1=x、X2=y、X3=xy、X4=x 2 、X5=y 2 。 9. The method for predicting tomato yield and growth based on data visualization according to claim 1, characterized in that: The method also includes evaluating the fitting accuracy of the model, and the evaluation index includes the determination coefficient R 2 , root mean square error, and mean absolute error, where: Coefficient of determination R 2 Used to evaluate the model's ability to explain yield variation; The root mean square error is used to measure the deviation between the model prediction value and the actual value; The mean absolute error is a measure of the absolute deviation between the model's predicted values ​​and the actual values.

10. A data visualization model for predicting tomato yield and growth. The method for predicting tomato yield and growth by data visualization according to any one of claims 1 to 9 is established, characterized in that: Based on the irrigation amount and nitrogen application amount data, the model parameters were solved by the least squares method to predict the tomato yield.

Images