A Method for Evaluating Soil Fertility for Tobacco Planting

Abstract

This invention discloses a method for evaluating soil fertility in tobacco-growing areas, belonging to the technical field of soil fertility evaluation methods. The steps are as follows: confirming the full dataset of soil evaluation indicators, determining the minimum dataset, and calculating the weights of each evaluation indicator in the minimum dataset using the entropy weight method; constructing a membership function, and building a soil fertility evaluation model using the soil fertility suitability index method, and calculating the soil fertility suitability index (SFI). The advantages of this invention are: it comprehensively considers the influence of multiple evaluation indicators on soil fertility and the dynamic adjustment of the critical and optimal values ​​of the membership function, allowing the soil fertility evaluation model to be dynamically adjusted according to real-time monitoring data, resulting in more accurate evaluation results. This soil fertility evaluation model can be used to evaluate the soil fertility of the area to be tested and can also serve as a basis for soil improvement.

Description

Technical Field

[0001] This invention relates to the technical field of soil fertility evaluation methods, and in particular to a method for evaluating the soil fertility of tobacco-growing soils. Background Technology

[0002] The long-term sustainable development of modern tobacco agriculture is closely related to soil environmental conditions. Evaluating soil fertility in tobacco-growing areas is of great significance for creating a favorable soil environment to promote tobacco growth and development. However, most current soil fertility evaluation systems for tobacco cultivation employ single evaluation methods, limited to the static changes of a single soil fertility indicator. Due to the different crops grown, their nutrient distribution and fertility requirements in the soil vary. Therefore, using existing single soil fertility evaluation methods makes it difficult to achieve an optimal evaluation system adapted to the target region and crop, leading to inaccurate soil fertility assessments. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a method for evaluating the soil fertility of tobacco-growing areas, which takes into account the influence of multiple evaluation indicators on soil fertility and the dynamic adjustment of the critical and optimal values ​​of the membership function, resulting in more accurate evaluation results.

[0004] To solve the above-mentioned technical problems, the technical solution of the present invention is a method for evaluating soil fertility in tobacco planting areas. Soil samples are collected from the area to be tested, and the detection data of each evaluation index in the minimum dataset MDS of each sample are detected. The detection data are then input into the soil fertility evaluation model for fertility evaluation.

[0005] The method for constructing the soil fertility evaluation model includes the following steps:

[0006] S1. Confirm the full dataset TDS of soil evaluation indicators;

[0007] S2. Determine the minimum dataset MDS:

[0008] S3. Calculate the weights of each evaluation index in the minimum dataset using the entropy weight method;

[0009] S4. Constructing Membership Functions: Based on the positive or negative response of each evaluation index to soil fertility, determine the type of membership function that fits each evaluation index, and then determine the initial critical value and optimal value of each membership function, thereby obtaining the membership function corresponding to each evaluation index.

[0010] S5. Calculate the membership value: Calculate the membership value of each evaluation indicator;

[0011] S6. A soil fertility evaluation model was constructed using the soil fertility suitability index method, and the soil fertility suitability index SFI was calculated.

[0012] Furthermore, the method for determining the minimum dataset MDS in step S2 includes the following steps:

[0013] S21. Evaluation index grouping: Perform principal component analysis on all evaluation indices in the full dataset, screen the principal components that meet the conditions, and group the evaluation indices according to their loadings on different principal components.

[0014] S22. Screening and Evaluation Indicators:

[0015] S221. Based on the Norm value of each evaluation indicator, eliminate evaluation indicators whose Norm value is less than 10% of Normmax; where Normmax refers to the maximum value among all the calculated Norm values ​​of the evaluation indicators.

[0016] S222. For the selected indicators, calculate the correlation coefficient between each evaluation indicator in each group. When the correlation between different evaluation indicators in the same group is less than or equal to 0.3, all evaluation indicators in the group are included in the minimum dataset MDS. When the correlation between two evaluation indicators in the same group is greater than 0.3, the evaluation indicator with the larger Norm value among the two evaluation indicators is selected to be included in the minimum dataset MDS.

[0017] Furthermore, the full dataset TDS consists of soil pH, SOM, AN, AP, AK, Zn, B, Cl, and Mg; the minimum dataset MDS consists of soil pH, SOM, AP, AK, B, and Cl.

[0018] Furthermore, in step S21, the conditions for selecting principal components are to select principal components with a cumulative contribution rate greater than or equal to 80% or to select principal components with an eigenvalue greater than 1.

[0019] Furthermore, the specific method for grouping the evaluation indicators in step S21 is as follows: if the evaluation indicator has a loading of 0.5 or more on only one principal component or a loading of 0.5 or less on all principal components, then the evaluation indicator is assigned to the principal component corresponding to the maximum loading value, and all evaluation indicators assigned to the same principal component are determined to be a group; if the evaluation indicator has a loading of 0.5 or more on at least two principal components, then the maximum correlation coefficient between the evaluation indicator and other evaluation indicators in the group is calculated, and the indicator is assigned to the group with the lowest maximum correlation coefficient.

[0020] Furthermore, the calculation method for determining the weights using the entropy weight method in step S3 is as follows:

[0021] ;

[0022] In the formula, k = 1 / ln(m); ;i=1,2,...,m;j=1,2,...,n;

[0023] The original data represents i samples and j evaluation indicators; represent Data after standardization;

[0024] In the formula W j Let be the weight of the j-th indicator; j = 1, 2, ..., n.

[0025] The above will Standardization process is required to obtain The process is as follows:

[0026] If the evaluation indicator is a positive indicator, meaning the higher the indicator value, the better, then the standardized formula is:

[0027] ;

[0028] If the evaluation indicator is a negative indicator, meaning the smaller the indicator value, the better, the standardized formula is:

[0029] ;

[0030] in This represents all sample data for the j-th indicator.

[0031] Furthermore, the contents of AP, AK, and B are matched with S-type membership functions; pH, SOM, and Cl are matched with parabolic membership functions.

[0032] Furthermore, the calculation formula for S6 when constructing the soil fertility evaluation model using the soil fertility suitability index method is as follows:

[0033] j = 1, 2, ..., n;

[0034] In the formula: SFI represents the soil fertility suitability index; N j W represents the membership value of the j-th evaluation index; j This represents the weight of the j-th evaluation indicator.

[0035] Furthermore, in step S4, the critical and optimal values ​​of the membership function corresponding to each evaluation index are dynamically adjusted. The dynamic adjustment method includes the following steps:

[0036] 1. Analyze the real-time monitoring data of each evaluation indicator and determine whether it exceeds the range of historical monitoring data;

[0037] 2. If the result of step one is "exceeding", then calculate the proportion of real-time monitoring data that exceeds the limit of historical monitoring data range;

[0038] 3. Set the adjustment gradients for the critical and optimal values ​​of each membership function in advance, and set the correspondence between the adjustment gradients and the excess ratios;

[0039] Fourth, based on the correspondence between the excess ratio and the adjustment gradient, redetermine the critical and optimal values ​​of each membership function to obtain new membership functions;

[0040] 5. Compare the accuracy of the new membership function with the original membership function. If the accuracy of the new membership function is higher, then use the new membership function to replace the original membership function; if the accuracy of the original membership function is higher, then do not update the membership function.

[0041] Furthermore, the methods for determining accuracy in step five include comparative experiments, cross-validation, expert evaluation, practical application effect evaluation, or statistical testing.

[0042] Beneficial effects of this invention:

[0043] This invention constructs a soil fertility evaluation model that comprehensively considers the influence of multiple evaluation indicators on soil fertility and proposes a method for selecting the minimum dataset, thereby avoiding redundancy among indicators. This invention also considers the dynamic adjustment of the critical and optimal values ​​of the membership function in the soil fertility evaluation model, allowing the model to dynamically adjust with real-time monitoring data, resulting in more accurate evaluation results. This soil fertility evaluation model can be used to evaluate the soil fertility of the tested area and can also serve as a basis for soil improvement. Attached Figure Description

[0044] Figure 1 The principal component analysis and Norm value calculation results are shown in the example.

[0045] Figure 2 The results show the correlation coefficients among the nine evaluation indicators in the full dataset in this example.

[0046] Figure 3 These are the initial critical values ​​and optimal values ​​of the S-shaped and parabolic membership functions for matching each evaluation index in the embodiments. Detailed Implementation

[0047] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that these descriptions are for the purpose of aiding understanding the present invention, but do not constitute a limitation thereof. Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0048] A method for evaluating soil fertility in tobacco-growing areas involves collecting soil samples from the area to be tested, detecting the data of each evaluation index in the minimum dataset (MDS) of each sample, and inputting the detection data into a soil fertility evaluation model for fertility evaluation.

[0049] The method for constructing the soil fertility evaluation model includes the following steps:

[0050] S1. Confirm the full dataset TDS of soil evaluation indicators; In this embodiment, the full dataset TDS is the soil's: pH, SOM, AN, AP, AK, Zn, B, Cl, Mg.

[0051] The availability and form transformation of soil nutrients are closely related to soil pH. The optimal soil pH range for tobacco growth is 5.5-6.5. Sodium phosphate (SOM), as an important source of soil nutrients for tobacco cultivation, directly affects the growth, development, yield, and quality of tobacco, playing a crucial role in improving soil physicochemical properties and microbial energy supply. Phosphorus peroxide (AP) is one of the essential macronutrients for tobacco. Due to the low phosphorus fertilizer utilization rate in flue-cured tobacco, tobacco-growing soils have a more severe phosphorus surplus than ordinary farmland. Tobacco is a potassium-loving crop, and potassium phosphate (AK), as an important quality element, restricts the overall improvement of tobacco leaf quality. Anthocyanins (AN) are key factors limiting the productivity of agricultural ecosystems and have a dual effect of promoting productivity and affecting environmental quality. Zinc (Zn) is a component and activator of various enzymes in flue-cured tobacco, participating in the synthesis of chlorophyll, protein, and auxin. Zinc deficiency leads to stunted growth and slow development of tobacco plants. Botanicals (B) can increase leaf transpiration rate and photosynthesis. Botanical deficiency in tobacco plants leads to terminal bud death and leaf wrinkling and curling. Tobacco is also a chloride-sensitive crop. Excessive Mg content will reduce the quality of flue-cured tobacco; Mg is an important component of chlorophyll and is crucial for the photosynthesis of tobacco. Mg deficiency will lead to the inhibition of chlorophyll synthesis and accelerated decomposition, thereby reducing the photosynthetic function.

[0052] In this embodiment, the sampling site was located in Honghe Hani and Yi Autonomous Prefecture, Yunnan Province, China, which belongs to the plateau subtropical monsoon climate zone. Honghe Prefecture is one of the high-quality tobacco producing areas in China, with tobacco leaves characterized by a prominent "sweet and fragrant" style and excellent "yellow, bright, and soft" quality. To ensure the reliability of data for each year, soil sample collection and testing were organized by Honghe Prefecture Tobacco Company. Topsoil samples (0-20cm) were collected from six counties (cities) in Honghe Prefecture: Mile City, Luxi County, Jianshui County, Shiping County, Gejiu City, and Mengzi City. Sampling was completed in March and April each year. The number of sampling points in each county (city) from 2011 to 2021 is shown in the table below:

[0053] Statistical table of the number of sampling points in each county (city)

[0054] years Maitreya Luxi Shiping Jianshui Mengzi Gejiu Jeonju 2011 64 50 50 50 31 10 255 2012 79 50 54 60 40 15 298 2013 65 60 45 50 12 10 242 2014 87 76 42 39 23 10 277 2015 71 70 50 45 30 13 279 2016 72 70 50 45 30 13 280 2017 240 240 165 70 50 15 780 2018 251 246 166 76 68 21 828 2019 192 226 176 76 54 21 755 2020 263 231 181 83 66 22 846 2021 361 354 218 124 80 18 1155

[0055] S2. Determining the Minimum Data Set (MDS): When evaluating differences in soil fertility, the more evaluation indicators selected, the more comprehensively the overall soil fertility can be reflected. However, there are generally correlations and even overlaps between various soil attributes, resulting in redundancy among indicators and increasing data acquisition costs. Therefore, soil indicators need to be screened when conducting soil fertility evaluation. Generally, appropriate evaluation indicator screening methods are selected according to different evaluation purposes and regions to obtain the minimum data set (MDS).

[0056] S21. Evaluation Index Grouping: Principal component analysis is performed on all evaluation indicators in the full dataset. Principal components that meet the criteria are selected, and the evaluation indicators are grouped according to their loadings on different principal components. The criteria for selecting principal components are that the cumulative contribution rate is greater than or equal to 80%, and the criteria can be adjusted to principal components with eigenvalues ​​greater than 1 as needed. The specific method for grouping each evaluation indicator is as follows: If an evaluation indicator has a loading greater than or equal to 0.5 on only one principal component or a loading less than 0.5 on all principal components, then the evaluation indicator is assigned to the principal component corresponding to the maximum loading value, and all evaluation indicators assigned to the same principal component are grouped together; if an evaluation indicator has loadings greater than or equal to 0.5 on at least two principal components, then the maximum correlation coefficient between the evaluation indicator and other evaluation indicators in the group is calculated, and the indicator is assigned to the group with the lowest maximum correlation coefficient.

[0057] In this embodiment, principal component analysis was performed on the nine initially selected evaluation indicators using RStudio 4.2.0 software, and the nine evaluation indicators were divided into five groups according to their loadings on different PCs. The results of principal component analysis and Norm value calculation are presented below. Figure 1 As shown. Based on the above grouping method, the nine evaluation indicators were divided into five groups: the first group is SOM and AN, the second group is pH, AK, and Mg, the third group is Cl, the fourth group is AP and Zn, and the fifth group is B.

[0058] S22. Screening and Evaluation Indicators:

[0059] S221. Based on the Norm value of each evaluation indicator, eliminate evaluation indicators whose Norm value is less than 10% Normmax; where Normmax refers to the maximum value among all the calculated Norm values ​​of the evaluation indicators; in this embodiment, the Normmax value is 1.07, and the Norm values ​​of all indicators are greater than 10% Normmax.

[0060] S222. For the selected indicators, calculate the correlation coefficient between each evaluation indicator in each group. When the correlation between different evaluation indicators in the same group is less than or equal to 0.3, all evaluation indicators in the group are included in the minimum dataset MDS. When the correlation between two evaluation indicators in the same group is greater than 0.3, the evaluation indicator with the larger Norm value among the two evaluation indicators is selected to be included in the minimum dataset MDS.

[0061] In this embodiment, the correlation coefficient results among the nine evaluation indicators are as follows: Figure 2 As shown in the diagram. In the first group, the correlation between SOM and AN is greater than 0.3, and the norm of SOM is greater than that of AN. Therefore, AN is removed, and SOM is included in the minimum dataset. In the second group, the correlation between pH and Mg is greater than 0.3, and the norm of pH is greater than that of Mg. Therefore, Mg is removed, and pH is included in the minimum dataset. In the fourth group, the correlation between Zn and AP is greater than 0.3, and the norm of AP is greater than that of Zn. Therefore, Zn is removed, and AP is included in the minimum dataset. The final determined minimum dataset MDS includes: pH, SOM, AP, AK, B, and Cl, a total of 6 evaluation indicators.

[0062] S3. Calculate the weights of each evaluation index in the minimum dataset using the entropy weight method; the calculation method is as follows:

[0063] ;

[0064] In the formula, k = 1 / ln(m); ; i=1, 2,..., m; j=1, 2,..., n;

[0065] The original data represents i samples and j evaluation indicators; represent Data after standardization;

[0066] In the formula W j Let be the weight of the j-th indicator; j = 1, 2, ..., n.

[0067] The above will Standardization process is required to obtain The process is as follows:

[0068] If the evaluation indicator is a positive indicator, meaning the higher the indicator value, the better, then the standardized formula is:

[0069] ;

[0070] If the evaluation indicator is a negative indicator, meaning the smaller the indicator value, the better, the standardized formula is:

[0071] ;

[0072] in This represents all sample data for the j-th indicator.

[0073] The weights calculated for each evaluation index in this embodiment are as follows:

[0074] index SOM (g / kg) AP (mg / kg) AK (mg / kg) Cl (mg / kg) pH B (mg / kg) Weight 0.099 0.277 0.166 0.296 0.139 0.023

[0075] S4. Constructing Membership Functions: Based on the positive or negative response of each evaluation index to soil fertility, determine the type of membership function that fits each evaluation index, and then determine the initial critical value and optimal value of each membership function, thereby obtaining the membership function corresponding to each evaluation index.

[0076] In this embodiment, the contents of AP, AK, and B are matched with S-type membership functions; pH value, SOM, and Cl are matched with parabolic membership functions.

[0077] The expression for the S-type membership function is as follows:

[0078]

[0079] The expression for the parabolic membership function is as follows:

[0080]

[0081] In the formula: x is the measured value of each indicator, x4 and x1 are the upper and lower critical values ​​of each indicator, respectively, and x3 and x2 are the optimal values ​​of the upper and lower limits of each indicator;

[0082] In this embodiment, the initial critical values ​​and optimal values ​​of the membership functions of each evaluation index are initially determined as follows: Figure 3 As shown;

[0083] S5. Calculate the membership value: Calculate the membership value of each evaluation indicator;

[0084] S6. A soil fertility evaluation model is constructed using the soil fertility suitability index method, and the soil fertility suitability index SFI is calculated. The calculation formula is as follows:

[0085] j = 1, 2, ..., n;

[0086] In the formula: SFI represents the soil fertility suitability index; N j W represents the membership value of the j-th evaluation index; j This represents the weight of the j-th evaluation indicator.

[0087] Furthermore, in step S4, the critical and optimal values ​​of the membership function corresponding to each evaluation index are dynamically adjusted. The dynamic adjustment method includes the following steps:

[0088] 1. Analyze the real-time monitoring data of each evaluation indicator and determine whether it exceeds the range of historical monitoring data;

[0089] 2. If the result of step one is "exceeding", then calculate the proportion of real-time monitoring data that exceeds the limit of historical monitoring data range;

[0090] 3. Set the adjustment gradients for the critical and optimal values ​​of each membership function in advance, and set the correspondence between the adjustment gradients and the excess ratios;

[0091] Fourth, based on the correspondence between the excess ratio and the adjustment gradient, redetermine the critical and optimal values ​​of each membership function to obtain new membership functions;

[0092] 5. Compare the accuracy of the new membership function with the original membership function. If the accuracy of the new membership function is higher, then use the new membership function to replace the original membership function; if the accuracy of the original membership function is higher, then do not update the membership function.

[0093] Furthermore, the methods for determining accuracy in step five include comparative experiments, cross-validation, expert evaluation, practical application effect evaluation, or statistical testing.

[0094] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and these variations still fall within the protection scope of the present invention.

Claims

1. A method for evaluating soil fertility for tobacco cultivation, characterized by: Soil samples were collected from the area to be tested, and the detection data of each evaluation index in the minimum dataset MDS of each sample were detected. The detection data were then input into the soil fertility evaluation model for fertility evaluation. The method for constructing the soil fertility evaluation model includes the following steps: S1. Confirm the full dataset TDS of soil evaluation indicators; the full dataset TDS is the soil's: pH, SOM, AN, AP, AK, Zn, B, Cl, Mg; S2. Determine the minimum dataset MDS: The minimum dataset MDS is the pH, SOM, AP, AK, B, and Cl of the soil. S3. Calculate the weights of each evaluation index in the minimum dataset using the entropy weight method; S4. Constructing Membership Functions: Based on the positive or negative response of each evaluation index to soil fertility, determine the type of membership function suitable for each evaluation index, then determine the initial critical value and optimal value of each membership function, and dynamically adjust the critical value and optimal value of the membership function corresponding to each evaluation index to obtain the membership function corresponding to each evaluation index; AP, AK, and B are matched with S-type membership functions; pH, SOM, and Cl are matched with parabolic membership functions. The method for dynamically adjusting the critical and optimal values ​​of the membership function includes the following steps:

1. Analyze the real-time monitoring data of each evaluation indicator and determine whether it exceeds the range of historical monitoring data; 2. If the result of step one is "exceeding", then calculate the proportion of real-time monitoring data that exceeds the limit of historical monitoring data range; 3. Set the adjustment gradients for the critical and optimal values ​​of each membership function in advance, and set the correspondence between the adjustment gradients and the excess ratios; Fourth, based on the correspondence between the excess ratio and the adjustment gradient, redetermine the critical and optimal values ​​of each membership function to obtain new membership functions; 5. Compare the accuracy of the new membership function with the original membership function. If the accuracy of the new membership function is higher, then use the new membership function to replace the original membership function; if the accuracy of the original membership function is higher, then do not update the membership function. S5. Calculate the membership value: Calculate the membership value of each evaluation indicator; S6. A soil fertility evaluation model was constructed using the soil fertility suitability index method, and the soil fertility suitability index SFI was calculated.

2. The method for evaluating soil fertility in tobacco-growing areas according to claim 1, characterized in that: Step S2, the method for determining the minimum dataset MDS, includes the following steps: S21. Evaluation index grouping: Perform principal component analysis on all evaluation indices in the full dataset, screen the principal components that meet the conditions, and group the evaluation indices according to their loadings on different principal components. S22. Screening and Evaluation Indicators: S221. Based on the Norm value of each evaluation indicator, eliminate evaluation indicators whose Norm value is less than 10% of Normmax; where Normmax refers to the maximum value among all the calculated Norm values ​​of the evaluation indicators. S222. For the selected indicators, calculate the correlation coefficient between each evaluation indicator in each group. When the correlation between different evaluation indicators in the same group is less than or equal to 0.3, all evaluation indicators in that group are included in the minimum dataset MDS. When the correlation between two evaluation indicators in the same group is greater than 0.3, the evaluation indicator with the larger Norm value among the two evaluation indicators is selected and included in the minimum dataset MDS.

3. The method for evaluating soil fertility in tobacco-growing areas according to claim 2, characterized in that: The conditions for selecting principal components in step S21 are to select principal components with a cumulative contribution rate greater than or equal to 80% or to select principal components with an eigenvalue greater than 1.

4. The method for evaluating soil fertility in tobacco-growing areas according to claim 2, characterized in that: The specific method for grouping the evaluation indicators in step S21 is as follows: if the evaluation indicator has a loading of 0.5 or more on only one principal component or a loading of 0.5 or less on all principal components, then the evaluation indicator is assigned to the principal component corresponding to the maximum loading value, and all evaluation indicators assigned to the same principal component are determined to be a group; if the evaluation indicator has a loading of 0.5 or more on at least two principal components, then the maximum correlation coefficient between the evaluation indicator and other evaluation indicators in the group is calculated, and the indicator is assigned to the group with the lowest maximum correlation coefficient.

5. The method for evaluating soil fertility in tobacco-growing areas according to claim 1, characterized in that: The calculation method for determining the weights using the entropy weight method in step S3 is as follows: ； In the formula, k=1 / ln(m); ;i=1,2,...,m;j=1,2,...,n; The original data represents i samples and j evaluation indicators; represent Data after standardization; In the formula W j Let be the weight of the j-th indicator; j = 1, 2, ..., n.

6. The method for evaluating soil fertility in tobacco-growing areas according to claim 1, characterized in that: The calculation formula for S7 when constructing the soil fertility evaluation model using the soil fertility suitability index method is as follows: j=1, 2, ..., n; In the formula: SFI represents the soil fertility suitability index; N j W represents the membership value of the j-th evaluation index; j This represents the weight of the j-th evaluation indicator.

7. The method for evaluating soil fertility in tobacco-growing areas according to claim 1, characterized in that: The methods for determining accuracy in step five include comparative experiments, cross-validation, expert evaluation, evaluation of practical application effects, or statistical testing.

Images