Root maggot pest classification method and system based on clustering processing

Abstract

The present invention discloses a method and system for classifying root maggot pests based on clustering processing, and the method includes: data collection, feature extraction, cluster analysis, generating cluster reports and formulating pest control strategies. The present invention belongs to the technical field of agricultural pest control, and specifically refers to a method and system for classifying root maggot pests based on clustering processing. The scheme adopts multi-index fusion of features, uses an infinite feature selection method to determine feature weights and adjusts to reduce information redundancy, uses an inflection point detection algorithm to perform feature stratification, and performs probability extraction based on feature weights to avoid excessive neglect of low-weight features and maintain the extraction priority of high-weight features; initializes membership and penalty factors, calculates cluster centers and objective functions, proposes a weighted multi-core fuzzy clustering algorithm based on two variables, optimizes membership, optimizes cluster structure, reduces the dependence of clustering results on initial values, and measures the consistency of clustering results and actual conditions.

Description

Technical Field

[0001] The invention belongs to the technical field of agricultural pest control, and in particular to a root maggot pest classification method and system based on clustering processing. Background Art

[0002] The root maggot pest classification method based on clustering processing is a technical means that uses data clustering algorithms to analyze and classify the characteristic data of root maggot pests. It aims to select appropriate pest control methods. Different types of root maggot pests have different responses to different control measures. Accurate control plans are formulated for specific types of root maggot pests, and appropriate pesticide types, concentrations and application times are selected. This can not only effectively control pests, but also reduce the use of pesticides and reduce pollution to the environment and agricultural products. The life histories and occurrence patterns of various root maggot pests are different. Accurate classification helps to understand the growth and development cycle of pests, such as the hatching period of eggs, the peak feeding period of larvae, and the emergence period of adults. Control can be carried out during the peak feeding period of larvae to minimize the damage of pests to crop roots. However, the existing root maggot pest classification method based on clustering processing has the problem that the feature weights of root maggot pest data are not determined reasonably, resulting in deviations in clustering results, and root maggot pests have multiple related characteristics, which are repeated and highly correlated, and easily cause excessive interference to the clustering results; there is a technical problem that the clustering algorithm is sensitive to noise, which causes the shape and structure of the clusters to be distorted, resulting in a large deviation between the clustering results and the actual situation, and the root maggot pest data distribution is complex, making it difficult to find the inherent distribution law of the data. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a root maggot pest classification method and system based on clustering processing, which aims at the technical problem that the root maggot pest data feature weights are not determined reasonably, resulting in deviations in clustering results, and root maggot pests have multiple related features, these features have problems of duplication and high correlation, and are prone to excessive interference with clustering results. The present invention adopts feature multi-index fusion, comprehensively considers the three aspects of calculating sample feature standard deviation, discrimination criterion and common information, uses infinite feature selection method to determine feature weights and adjusts them to reduce information redundancy, and uses inflection point detection algorithm to perform feature Layered, probability extraction is performed based on feature weights to avoid excessive neglect of low-weight features and maintain the extraction priority of high-weight features; in order to address the technical problems that the clustering algorithm is sensitive to noise, which causes the shape and structure of the clusters to be distorted, resulting in a large deviation between the clustering results and the actual situation, and the root maggot pest data is complex and it is difficult to find the inherent distribution law of the data, the membership degree and penalty factor are initialized, the cluster center and the objective function are calculated, and a weighted multi-core fuzzy clustering algorithm based on two variables is proposed to optimize the membership degree, optimize the clustering structure, reduce the dependence of the clustering results on the initial value, and measure the consistency between the clustering results and the actual situation.

[0004] The technical solution adopted by the present invention is as follows: The root maggot pest classification method based on clustering processing provided by the present invention comprises the following steps:

[0005] Step S1: data collection, specifically soil sampling and crop root sampling to obtain root maggot pest data;

[0006] Step S2: Feature extraction, specifically, to perform feature multi-index fusion, comprehensively consider the three aspects of calculating sample feature standard deviation, discrimination criteria and common information, use infinite feature selection method to determine feature weights and adjust to reduce information redundancy, use inflection point detection algorithm to perform feature stratification, and perform probability extraction based on feature weights to avoid excessive neglect of low-weight features and maintain the extraction priority of high-weight features;

[0007] Step S3: Cluster analysis, specifically using the initial membership and penalty factors, calculating the cluster center and the objective function, proposing a weighted multi-core fuzzy clustering algorithm based on two variables, optimizing the membership, forming a reasonable and compact clustering structure, reducing the dependence of the clustering results on the initial value, and measuring the consistency of the clustering results with the actual situation;

[0008] Step S4: Generate a clustering report, specifically generate clustering protection, visualize the clustering results, and evaluate whether the results meet the expected goals;

[0009] Step S5: Formulate pest control strategies, specifically, comprehensively control root maggot pests based on the clustering results.

[0010] Furthermore, in step S1, the data collection includes the following steps:

[0011] Step S11: soil sampling. The larvae of root maggot pests mainly live in the soil and feed on plant roots. Soil samples are collected at different locations in the field. The larvae of root maggot pests are separated from the soil by screening, rinsing, etc., and the sampling location and soil properties are recorded during sampling.

[0012] Step S12: sampling the crop roots, collecting pest samples directly from the crop roots, digging the crop roots, and checking whether there are root maggot pests on the roots. The collected pests should be marked with the corresponding crop types and crop growth conditions;

[0013] Step S13: carefully comparing the collected data, removing noise and outliers, and performing data conversion and normalization operations to ensure the accuracy and standardization of the data, thereby obtaining the pre-processed root maggot pest data;

[0014] The pre-processed root maggot pest data includes morphological data, developmental stage data, soil environment data, host plant data and pest distribution data of the root maggot pests.

[0015] Furthermore, in step S2, the feature selection includes the following steps:

[0016] Step S21: feature multi-index fusion, feature multi-index fusion is performed on the pre-processed root maggot pest data, and the indicators from three different angles, standard deviation, discrimination criterion and common information, are comprehensively considered; standard deviation is based on the perspective of data discreteness, discrimination criterion is based on the perspective of category distinguishability, and common information is based on the perspective of the degree of correlation between features and categories; to avoid the one-sidedness of single indicator evaluation, the following steps are included:

[0017] Step S211: Calculate the standard deviation of the sample features. The standard deviation reflects the degree of dispersion of the features in the sample. The discrete characteristics of each feature are preliminarily measured through the standard deviation data. Features with a high degree of dispersion indicate that the value of the feature varies greatly in the sample, which is more helpful for the classification of root maggot pests.

[0018] The characteristics include morphological characteristic data, life history characteristic data and ecological habit characteristic data;

[0019] The morphological characteristic data include body shape, color and body structure;

[0020] The life history characteristic data include the number of generations, development process and overwintering habits of different root maggot pests;

[0021] The ecological habit characteristic data include host plant range, pest tropism and plant damaged parts data;

[0022] Step S212: Calculate the discrimination criterion to measure the distinguishability of features between different categories. The formula used is as follows:

[0023] ;

[0024] In the formula, represents the mean of the i-th feature in the first category of samples, represents the mean of the i-th feature in the second category of samples, represents the standard deviation of the i-th feature in the first category of samples, represents the standard deviation of the i-th feature in the second category of samples, It indicates the distinguishability of the i-th feature between two types of samples. The larger the value, the more obvious the difference between the i-th feature between the two types of samples, and the more helpful it is to distinguish pest samples.

[0025] The first category of samples and the second category of samples are two different categories pre-set according to the classification attributes of the pest samples; for example, the pest samples are divided into two categories according to the development stage of the pests, the first category is the pest samples in the larval stage, and the second category is the pest samples in the adult stage. The pests can also be distinguished according to their gender and the degree of damage to the crops.

[0026] Step S213: Calculate common information, which reflects the degree of association between features and categories. Features with high association are more helpful for classification. The formula used is as follows:

[0027] ;

[0028] In the formula, represents the common information between the i-th feature and the category label, Y represents the category label set, y represents the category label, represents the specific value of the i-th feature, f represents the value set of the feature, represents the joint probability distribution, which indicates the probability of the feature and the category label appearing at the same time, represents the probability of feature occurrence, Indicates the probability of the category label appearing;

[0029] Step S22: Determine feature weights, use an infinite feature selection method to linearly weight the standard deviation, discrimination criterion value, and common information of each feature, analyze the correlation between features, adjust feature weights according to the correlation and redundancy between features, and minimize information redundancy;

[0030] Step S23: Feature stratification is used to reasonably allocate features of different importance during the classification process, making full use of high-weight features while retaining the potential contribution of secondary features. In order to reasonably divide the importance of features, an inflection point detection algorithm is used to determine the significant change points in the feature weight curve. The core idea of ​​the inflection point detection algorithm is to locate the inflection point coordinates by maximizing the vertical distance from each point in the feature weight curve to the straight line, that is, construct a straight line connecting the starting point and the end point of the weight curve, and then calculate the vertical distance from each point on the weight curve to the straight line to calculate the inflection point coordinates. The formula used is as follows:

[0031] ;

[0032] In the formula, represents the coordinates of the inflection point, It means to find the characteristic index i that makes the expression value the largest, h1 represents the slope of the straight line connecting the starting point and the end point of the curve, and h2 represents the intercept of the straight line connecting the starting point and the end point of the curve. represents the characteristic points on the curve, The value representing the feature weight;

[0033] Refine the feature hierarchy. After determining the first inflection point, use the same steps to find the second inflection point and determine the minor change points in the weight curve. The core features are before the first inflection point, the important features are between the two inflection points, and the remaining features are after the second inflection point.

[0034] Step S24: Probabilistic extraction is performed based on feature weights. In order to ensure that the importance of features is reflected while maintaining diversity during feature extraction, a probabilistic extraction method based on feature weights is used. The formula used is as follows:

[0035] ;

[0036] In the formula, represents the probability of the i-th feature being extracted, represents the weight of the i-th feature, and N represents the total number of features;

[0037] This formula ensures that the maximum extraction probability of a feature is 1, and the extraction probabilities of all other features are adjusted proportionally. This rescaling process avoids low-weight features from being excessively ignored while maintaining the priority of high-weight features in extraction.

[0038] Furthermore, in step S3, the cluster analysis includes the following steps:

[0039] Step S31: Data initialization, determining the number of clusters and the total number of samples in the data set, initializing the membership degree and the penalty factor;

[0040] Determine the number of clusters. For example, it is known that root maggot pests have different development stages, different living habits, and different plant root preferences. The number of clusters is preliminarily determined based on different categories.

[0041] Initialize membership, randomly assign membership values ​​to each sample for each cluster, ranging from [0, 1];

[0042] Initialize the penalty factor. Based on prior experience, initialize the penalty factor to 0.5, run the clustering algorithm, observe the rationality of the clustering results, observe the compactness of the clusters and the separation between classes, and gradually adjust the value of the penalty factor.

[0043] The inter-class separation degree indicates the interval degree between different clusters, and a high inter-class separation degree indicates a high degree of distinction between different clusters;

[0044] Step S32: Calculate the cluster center and determine the location of the cluster center based on the sample and the degree of membership of the sample to the cluster center. The formula used is as follows:

[0045] ;

[0046] In the formula, V represents the cluster center, represents the ath cluster center, d represents the number of clusters in the data set, a represents the cluster index, b represents the sample index, represents the membership of the bth sample to the ath cluster center, and m represents the weighted index for determining the clustering fuzziness level;

[0047] Update the cluster center by considering the samples in the cluster at different times and the membership of the samples to improve the stability of the clustering algorithm in a dynamic data environment and avoid the problem of being unable to calculate the cluster center in special cases, such as when the denominator is zero. The formula used is as follows:

[0048] ;

[0049] ;

[0050] In the formula, represents the updated cluster center, Indicates the identifier of the a-th cluster center, which is used to distinguish whether the sample is in the cluster. represents the cluster at time t, represents the cluster at time t-1;

[0051] As the data is updated and the clustering process proceeds, the clusters will become outdated due to changes in the data. The distribution of data points no longer conforms to the characteristics of the original clusters. Delete the outdated clusters to improve clustering accuracy. After the clusters are deleted, there may be less than two clusters left, making the clustering algorithm unstable. Therefore, at least two clusters must be left after the deletion. Find the period T when no clusters are created. During this period, delete the clusters that meet the following conditions. The formula used is as follows:

[0052] ;

[0053] In the formula, represents the jth cluster at time t, represents the jth cluster at time tT, represents the total number of clusters at time t, and T represents the period without any cluster creation;

[0054] Step S33: Calculate the objective function to measure the clustering quality. Considering the sensitivity of the fuzzy clustering algorithm to noise points, a weighted multi-core fuzzy clustering algorithm based on two variables is proposed. The formula used is as follows:

[0055] ;

[0056] In the formula, U represents the objective function, represents the total number of samples, represents the distance metric from the a-th cluster center to the b-th sample, k represents the adaptive coefficient of the distance metric, n represents the noise impact factor, L2 represents the membership constraint, represents the penalty factor of the a-th cluster center, Represents the weight of the kernel function of the bth sample;

[0057] Step S34: Optimize the membership. During the iteration process, the membership of the sample to the cluster center is continuously adjusted so that the objective function reaches the minimum value. The formula used is as follows:

[0058] ;

[0059] In the formula, represents the optimized membership, o represents the index of the number of iterations, Q represents the total number of iterations, Indicates the degree of difference between the ath cluster center and the updated cluster center;

[0060] Step S35: consistency evaluation, measuring the consistency between the clustering results and the actual situation, the formula used is as follows:

[0061] ;

[0062] In the formula, represents the consistency assessment value, represents the clustering results to be compared, represents the actual clustering result, represents the t1th clustering result to be compared, represents the t2th true clustering result, Indicates the total number of clustering results.

[0063] Furthermore, in step S4, generating a clustering report specifically generates a report title and overview, characterizes the root maggot pest data, records the clustering analysis process in detail, presents the clustering results to relevant technical personnel using visualization methods, evaluates the clustering effect, and determines whether the clustering results meet expectations.

[0064] Furthermore, in step S5, the formulation of the pest control strategy includes the following steps:

[0065] Step S51: prevention and treatment based on development stage, including the following contents:

[0066] For clusters with a high proportion of larval stage pests, since the larvae mainly move around the roots of crops and harm the growth and development of the root system, soil treatment methods are used and pesticides are used during crop growth;

[0067] For clusters with a high proportion of adult pests, since the adults have the ability to fly and will spread in the field, sex pheromones are used to trap the adults. According to the sex pheromone composition of the pests in the cluster, sex pheromone traps are made to interfere with the mating behavior of the adults during the adult breeding season and reduce the number of pests that reproduce.

[0068] Step S52: Chemical control, determining the differences in sensitivity of pests to chemical agents based on cluster analysis results, and reasonably arranging the spraying range and frequency of chemical agents;

[0069] Step S53: biological control, considering biological control measures from the perspective of the entire farmland ecosystem, introducing natural enemy insects and beneficial insects, and building a stable biological community;

[0070] Step S54: Agricultural prevention and control: the pests in the clusters are related to soil moisture and fertility. Reasonable irrigation and fertilization are used to improve the soil environment and inhibit the growth and reproduction of pests.

[0071] The root maggot pest classification system based on clustering processing provided by the present invention comprises a data acquisition module, a feature extraction module, a cluster analysis module, a cluster report generation module and a pest control strategy formulation module;

[0072] The data collection module specifically performs soil sampling and crop root sampling to obtain root maggot pest data;

[0073] The feature extraction module specifically performs feature multi-index fusion, comprehensively considers the three aspects of calculating sample feature standard deviation, discrimination criteria and common information, adopts infinite feature selection method to determine feature weights and adjusts them to reduce information redundancy, uses inflection point detection algorithm to perform feature stratification, and performs probability extraction based on feature weights to avoid excessive neglect of low-weight features and maintain the extraction priority of high-weight features;

[0074] The cluster analysis module specifically adopts the initialization membership and penalty factor, calculates the cluster center and the objective function, proposes a weighted multi-core fuzzy clustering algorithm based on two variables, optimizes the membership, forms a reasonable and compact clustering structure, reduces the dependence of the clustering results on the initial value, and measures the consistency between the clustering results and the actual situation;

[0075] The cluster report generation module is specifically used to generate cluster protection, visualize cluster results, and evaluate whether the results meet the expected goals;

[0076] The module for formulating pest control strategies is specifically for comprehensively controlling root maggot pests based on clustering results.

[0077] The beneficial results achieved by the present invention using the above scheme are as follows:

[0078] (1) In order to solve the technical problem that the clustering results are biased due to the unreasonable determination of the feature weights of the root maggot pest data, and that the root maggot pests have multiple related features, which are repeated and highly correlated, and easily cause excessive interference to the clustering results, we adopted the feature multi-index fusion method, comprehensively considered the three aspects of calculating the sample feature standard deviation, discrimination criteria and common information, used the infinite feature selection method to determine the feature weights and adjusted them to reduce information redundancy, used the inflection point detection algorithm to perform feature stratification, and performed probability extraction based on feature weights to avoid excessive neglect of low-weight features and maintain the extraction priority of high-weight features;

[0079] (2) In order to address the technical problems that the clustering algorithm is sensitive to noise, which causes the shape and structure of the clusters to be distorted, resulting in a large deviation between the clustering results and the actual situation, and the root maggot pest data is complex and it is difficult to find the intrinsic distribution law of the data, this paper uses the initial membership and penalty factors to calculate the cluster center and the objective function, and proposes a weighted multi-core fuzzy clustering algorithm based on two variables to optimize the membership, form a reasonable and compact clustering structure, reduce the dependence of the clustering results on the initial value, and measure the consistency between the clustering results and the actual situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 A schematic diagram of the process of the root maggot pest classification method based on clustering provided by the present invention;

[0081] Figure 2A schematic diagram of a root maggot pest classification system based on clustering provided by the present invention;

[0082] Figure 3 is a schematic flow chart of step S1;

[0083] Figure 4 is a schematic flow chart of step S2;

[0084] Figure 5 It is a schematic diagram of the process of step S3.

[0085] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0086] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0087] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0088] Example 1, see Figure 1 The present invention provides a method for classifying root maggot pests based on clustering processing, the method comprising the following steps:

[0089] Step S1: data collection, specifically soil sampling and crop root sampling to obtain root maggot pest data;

[0090] Step S2: Feature extraction, specifically, to perform feature multi-index fusion, comprehensively consider the three aspects of calculating sample feature standard deviation, discrimination criteria and common information, use infinite feature selection method to determine feature weights and adjust to reduce information redundancy, use inflection point detection algorithm to perform feature stratification, and perform probability extraction based on feature weights to avoid excessive neglect of low-weight features and maintain the extraction priority of high-weight features;

[0091] Step S3: Cluster analysis, specifically using the initial membership and penalty factors, calculating the cluster center and the objective function, proposing a weighted multi-core fuzzy clustering algorithm based on two variables, optimizing the membership, forming a reasonable and compact clustering structure, reducing the dependence of the clustering results on the initial value, and measuring the consistency of the clustering results with the actual situation;

[0092] Step S4: Generate a clustering report, specifically generate clustering protection, visualize the clustering results, and evaluate whether the results meet the expected goals;

[0093] Step S5: Formulate pest control strategies, specifically, comprehensively control root maggot pests based on the clustering results.

[0094] Example 2, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S1, the data collection includes the following steps:

[0095] Step S11: soil sampling. The larvae of root maggot pests mainly live in the soil and feed on plant roots. Soil samples are collected at different locations in the field. The larvae of root maggot pests are separated from the soil by screening, rinsing, etc., and the sampling location and soil properties are recorded during sampling.

[0096] Step S12: sampling the crop roots, collecting pest samples directly from the crop roots, digging the crop roots, and checking whether there are root maggot pests on the roots. The collected pests should be marked with the corresponding crop types and crop growth conditions;

[0097] Step S13: carefully comparing the collected data, removing noise and outliers, and performing data conversion and normalization operations to ensure the accuracy and standardization of the data, thereby obtaining the pre-processed root maggot pest data;

[0098] The pre-processed root maggot pest data includes morphological data, developmental stage data, soil environment data, host plant data and pest distribution data of the root maggot pests.

[0099] Example 3, see Figure 1 and Figure 4 This embodiment is based on the above embodiment. In step S2, the feature selection includes the following steps:

[0100] Step S21: feature multi-index fusion, feature multi-index fusion is performed on the pre-processed root maggot pest data, and the indicators from three different angles, standard deviation, discrimination criterion and common information, are comprehensively considered; standard deviation is based on the perspective of data discreteness, discrimination criterion is based on the perspective of category distinguishability, and common information is based on the perspective of the degree of correlation between features and categories; to avoid the one-sidedness of single indicator evaluation, the following steps are included:

[0101] Step S211: Calculate the standard deviation of the sample features. The standard deviation reflects the degree of dispersion of the features in the sample. The discrete characteristics of each feature are preliminarily measured through the standard deviation data. Features with a high degree of dispersion indicate that the value of the feature varies greatly in the sample, which is more helpful for the classification of root maggot pests.

[0102] The characteristics include morphological characteristic data, life history characteristic data and ecological habit characteristic data;

[0103] The morphological characteristic data include body shape, color and body structure;

[0104] The life history characteristic data include the number of generations, development process and overwintering habits of different root maggot pests;

[0105] The ecological habit characteristic data include host plant range, pest tropism and plant damaged parts data;

[0106] Step S212: Calculate the discrimination criterion to measure the distinguishability of features between different categories. The formula used is as follows:

[0107] ;

[0108] In the formula, represents the mean of the i-th feature in the first category of samples, represents the mean of the i-th feature in the second category of samples, represents the standard deviation of the i-th feature in the first category of samples, represents the standard deviation of the i-th feature in the second category of samples, It indicates the distinguishability of the i-th feature between two types of samples. The larger the value, the more obvious the difference between the i-th feature between the two types of samples, and the more helpful it is to distinguish pest samples.

[0109] The first category of samples and the second category of samples are two different categories pre-set according to the classification attributes of the pest samples; for example, the pest samples are divided into two categories according to the development stage of the pests, the first category is the pest samples in the larval stage, and the second category is the pest samples in the adult stage. The pests can also be distinguished according to their gender and the degree of damage to the crops.

[0110] Step S213: Calculate common information, which reflects the degree of association between features and categories. Features with high association are more helpful for classification. The formula used is as follows:

[0111] ;

[0112] In the formula, represents the common information between the i-th feature and the category label, Y represents the category label set, y represents the category label, represents the specific value of the i-th feature, f represents the value set of the feature, represents the joint probability distribution, which indicates the probability of the feature and the category label appearing at the same time. represents the probability of feature occurrence, Indicates the probability of the category label appearing;

[0113] Step S22: Determine feature weights, use an infinite feature selection method to linearly weight the standard deviation, discrimination criterion value, and common information of each feature, analyze the correlation between features, adjust the initial weights according to the correlation and redundancy between features, and minimize information redundancy;

[0114] Step S23: Feature stratification is used to reasonably allocate features of different importance during the classification process, making full use of high-weight features while retaining the potential contribution of secondary features. In order to reasonably divide the importance of features, an inflection point detection algorithm is used to determine the significant change points in the feature weight curve. The core idea of ​​the inflection point detection algorithm is to locate the inflection point coordinates by maximizing the vertical distance from each point in the feature weight curve to the straight line, that is, construct a straight line connecting the starting point and the end point of the weight curve, and then calculate the vertical distance from each point on the weight curve to the straight line to calculate the inflection point coordinates. The formula used is as follows:

[0115] ;

[0116] In the formula, represents the coordinates of the inflection point, It means to find the characteristic index i that makes the expression value the largest, h1 represents the slope of the straight line connecting the starting point and the end point of the curve, and h2 represents the intercept of the straight line connecting the starting point and the end point of the curve. represents the characteristic points on the curve, The value representing the feature weight;

[0117] Refine the feature hierarchy. After determining the first inflection point, use the same steps to find the second inflection point and determine the minor change points in the weight curve. The core features are before the first inflection point, the important features are between the two inflection points, and the remaining features are after the second inflection point.

[0118] Step S24: Probabilistic extraction is performed based on feature weights. In order to ensure that the importance of features is reflected while maintaining diversity during feature extraction, a probabilistic extraction method based on feature weights is used. The formula used is as follows:

[0119] ;

[0120] In the formula, represents the probability of the i-th feature being extracted, represents the weight of the i-th feature, and N represents the total number of features;

[0121] This formula ensures that the maximum extraction probability of a feature is 1, and the extraction probabilities of all other features are adjusted proportionally. This rescaling process avoids low-weight features from being excessively ignored while maintaining the priority of high-weight features in extraction.

[0122] By executing the above operations, multi-index fusion of features is adopted, and comprehensive consideration is given to three aspects: calculating the standard deviation of sample features, discrimination criteria and common information. The infinite feature selection method is used to determine the feature weights and adjust them to reduce information redundancy. The inflection point detection algorithm is used for feature stratification. Probabilistic extraction is performed based on feature weights to avoid excessive neglect of low-weight features and maintain the extraction priority of high-weight features. The technical problem that the unreasonable determination of feature weights of root maggot pest data leads to deviations in clustering results, and that root maggot pests have multiple related features, which are repeated and highly correlated, and easily cause excessive interference to the clustering results is solved.

[0123] Example 4, see Figure 1 and Figure 5 This embodiment is based on the above embodiment. In step S3, the cluster analysis includes the following steps:

[0124] Step S31: Data initialization, determining the number of clusters and the total number of samples in the data set, initializing the membership degree and the penalty factor;

[0125] Determine the number of clusters. For example, it is known that root maggot pests have different development stages, different living habits, and different plant root preferences. The number of clusters is preliminarily determined based on different categories.

[0126] Initialize membership, randomly assign membership values ​​to each sample for each cluster, ranging from [0, 1];

[0127] Initialize the penalty factor. Based on prior experience, initialize the penalty factor to 0.5, run the clustering algorithm, observe the rationality of the clustering results, observe the compactness of the clusters and the separation between classes, and gradually adjust the value of the penalty factor.

[0128] The inter-class separation degree indicates the interval degree between different clusters, and a high inter-class separation degree indicates a high degree of distinction between different clusters;

[0129] Step S32: Calculate the cluster center and determine the location of the cluster center based on the sample and the degree of membership of the sample to the cluster center. The formula used is as follows:

[0130] ;

[0131] In the formula, V represents the cluster center, represents the ath cluster center, d represents the number of clusters in the data set, a represents the cluster index, b represents the sample index, represents the membership of the bth sample to the ath cluster center, and m represents the weighted index for determining the clustering fuzziness level;

[0132] Update the cluster center by considering the samples in the cluster at different times and the membership of the samples to improve the stability of the clustering algorithm in a dynamic data environment and avoid the problem of being unable to calculate the cluster center in special cases, such as when the denominator is zero. The formula used is as follows:

[0133] ;

[0134] ;

[0135] In the formula, represents the updated cluster center, Indicates the identifier of the a-th cluster center, which is used to distinguish whether the sample is in the cluster. represents the cluster at time t, represents the cluster at time t-1;

[0136] As the data is updated and the clustering process proceeds, the clusters will become outdated due to changes in the data. The distribution of data points no longer conforms to the characteristics of the original clusters. Delete the outdated clusters to improve clustering accuracy. After the clusters are deleted, there may be less than two clusters left, making the clustering algorithm unstable. Therefore, at least two clusters must be left after the deletion. Find the period T when no clusters are created. During this period, delete the clusters that meet the following conditions. The formula used is as follows:

[0137] ;

[0138] In the formula, represents the jth cluster at time t, represents the jth cluster at time tT, represents the total number of clusters at time t, and T represents the period without any cluster creation;

[0139] Step S33: Calculate the objective function to measure the clustering quality. Considering the sensitivity of the fuzzy clustering algorithm to noise points, a weighted multi-core fuzzy clustering algorithm based on two variables is proposed. The formula used is as follows:

[0140] ;

[0141] In the formula, U represents the objective function, represents the total number of samples, represents the distance metric from the a-th cluster center to the b-th sample, k represents the adaptive coefficient of the distance metric, n represents the noise impact factor, L2 represents the membership constraint, represents the penalty factor of the a-th cluster center, Represents the weight of the kernel function of the bth sample;

[0142] Step S34: Optimize the membership. During the iteration process, the membership of the sample to the cluster center is continuously adjusted so that the objective function reaches the minimum value. The formula used is as follows:

[0143] ;

[0144] In the formula, represents the optimized membership, o represents the index of the number of iterations, Q represents the total number of iterations, Indicates the degree of difference between the ath cluster center and the updated cluster center;

[0145] Step S35: consistency evaluation, measuring the consistency between the clustering results and the actual situation, the formula used is as follows:

[0146] ;

[0147] In the formula, represents the consistency assessment value, represents the clustering results to be compared, represents the actual clustering result, represents the t1th clustering result to be compared, represents the t2th true clustering result, Indicates the total number of clustering results.

[0148] By performing the above operations, the membership degree and penalty factor are initialized, the cluster center and the objective function are calculated, and a weighted multi-core fuzzy clustering algorithm based on two variables is proposed to optimize the membership degree, form a reasonable and compact clustering structure, reduce the dependence of the clustering results on the initial value, and measure the consistency of the clustering results and the actual situation. The technical problems that the clustering algorithm is sensitive to noise, which causes the shape and structure of the cluster to be distorted, resulting in a large deviation between the clustering results and the actual situation, and the root maggot pest data distribution is complex, making it difficult to find the inherent distribution law of the data.

[0149] Example 5, see Figure 1 This embodiment is based on the above embodiment. In step S4, the generation of the clustering report specifically generates a report title and overview, describes the characteristics of the root maggot pest data, records the clustering analysis process in detail, presents the clustering results to relevant technical personnel using visualization methods, evaluates the clustering effect, and determines whether the clustering results meet expectations.

[0150] Example 6, see Figure 1 and Figure 5 This embodiment is based on the above embodiment. In step S5, the formulation of the pest control strategy includes the following steps:

[0151] Step S51: prevention and control based on cluster characteristics, including the following contents:

[0152] For clusters with a high proportion of larval stage pests, since the larvae mainly move around the roots of crops and harm the growth and development of the root system, soil treatment methods are used and pesticides are used during crop growth;

[0153] For clusters with a high proportion of adult pests, since the adults have the ability to fly and will spread in the field, sex pheromones are used to trap the adults. According to the sex pheromone composition of the pests in the cluster, sex pheromone traps are made to interfere with the mating behavior of the adults during the adult breeding season and reduce the number of pests that reproduce.

[0154] Step S52: Chemical control, determining the differences in sensitivity of pests to chemical agents based on cluster analysis results, and reasonably arranging the spraying range and frequency of chemical agents;

[0155] Step S53: biological control, considering biological control measures from the perspective of the entire farmland ecosystem, introducing natural enemy insects and beneficial insects, and building a stable biological community;

[0156] Step S54: Agricultural prevention and control: the pests in the clusters are related to soil moisture and fertility. Reasonable irrigation and fertilization are used to improve the soil environment and inhibit the growth and reproduction of pests.

[0157] Embodiment 7, see Figure 2 , this embodiment is based on the above embodiment, and the root maggot pest classification system based on clustering processing provided by the present invention includes a data acquisition module, a feature extraction module, a cluster analysis module, a cluster report generation module and a pest control strategy formulation module;

[0158] The data collection module specifically performs soil sampling and crop root sampling to obtain root maggot pest data;

[0159] The feature extraction module specifically performs feature multi-index fusion, comprehensively considers the three aspects of calculating sample feature standard deviation, discrimination criteria and common information, adopts infinite feature selection method to determine feature weights and adjusts them to reduce information redundancy, uses inflection point detection algorithm to perform feature stratification, and performs probability extraction based on feature weights to avoid excessive neglect of low-weight features and maintain the extraction priority of high-weight features;

[0160] The cluster analysis module specifically adopts the initialization membership and penalty factor, calculates the cluster center and the objective function, proposes a weighted multi-core fuzzy clustering algorithm based on two variables, optimizes the membership, forms a reasonable and compact clustering structure, reduces the dependence of the clustering results on the initial value, and measures the consistency between the clustering results and the actual situation;

[0161] The cluster report generation module is specifically used to generate cluster protection, visualize cluster results, and evaluate whether the results meet the expected goals;

[0162] The module for formulating pest control strategies is specifically for comprehensively controlling root maggot pests based on clustering results.

[0163] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0164] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0165] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A method for classifying root maggot pests based on clustering, characterized in that: The method comprises the following steps: Step S1: data collection, obtaining root maggot pest data after pre-processing; Step S2: feature extraction, specifically, to perform feature multi-index fusion, comprehensively consider the three aspects of calculating sample feature standard deviation, discrimination criteria and common information, use infinite feature selection method to determine feature weights, use inflection point detection algorithm to perform feature stratification, and perform probability extraction based on feature weights; Step S3: Cluster analysis, specifically using the initial membership and penalty factors, calculating the cluster center and the objective function, proposing a weighted multi-core fuzzy clustering algorithm based on two variables, optimizing the membership, and measuring the consistency between the clustering results and the actual situation; Step S4: generating a clustering report, specifically generating a pest clustering report; Step S5: Develop pest control strategies; In step S2, the feature selection includes the following steps: Step S21: feature multi-index fusion, which is to perform feature multi-index fusion on the pre-processed root maggot pest data, and comprehensively consider the indicators from three different angles: standard deviation, discrimination criterion and common information, including the following steps: Step S211: Calculate the standard deviation of the sample feature, which reflects the degree of dispersion of the feature in the sample. The standard deviation data is used to preliminarily measure the discrete characteristics of each feature. A feature with a high degree of dispersion indicates that the value of the feature varies greatly in the sample, which is more helpful for the classification of root maggot pests. Step S212: Calculate the discrimination criterion to measure the distinguishability of features between different categories. The formula used is as follows: ； In the formula, represents the mean of the i-th feature in the first category of samples, represents the mean of the i-th feature in the second category of samples, represents the standard deviation of the i-th feature in the first category of samples, represents the standard deviation of the i-th feature in the second category of samples, Indicates the distinguishability of the i-th feature between two types of samples; The first type of samples and the second type of samples are two different categories pre-set according to the classification attributes of the pest samples; Step S213: Calculate common information, which reflects the degree of association between features and categories. The formula used is as follows: ； In the formula, represents the common information between the i-th feature and the category label, Y represents the category label set, y represents the category label, represents the specific value of the i-th feature, f represents the value set of the feature, represents the joint probability distribution, which indicates the probability of the feature and the category label appearing at the same time, represents the probability of feature occurrence, Indicates the probability of the category label appearing; Step S22: determining feature weights, using an infinite feature selection method to linearly weight the standard deviation, discrimination criterion value, and common information of each feature, analyzing the correlation between features, and adjusting feature weights according to the correlation and redundancy between features; Step S23: Feature stratification is used to reasonably allocate features of different importance during the classification process, fully utilize high-weight features while retaining the potential contribution of secondary features, and locate the inflection point coordinates by maximizing the vertical distance from each point in the feature weight curve to the straight line. The formula used is as follows: ； In the formula, represents the coordinates of the inflection point, It means to find the characteristic index i that makes the expression value the largest, h1 represents the slope of the straight line connecting the starting point and the end point of the curve, and h2 represents the intercept of the straight line connecting the starting point and the end point of the curve. represents the characteristic points on the curve, represents the value of feature weight; Step S24: Probabilistic extraction is performed based on feature weights to maintain diversity while reflecting the importance of features during feature extraction. The formula used is as follows: ； In the formula, represents the probability of the i-th feature being extracted, represents the weight of the i-th feature, and N represents the total number of features.

2. The root maggot pest classification method based on clustering processing according to claim 1, characterized in that: In step S3, the cluster analysis includes the following steps: Step S31: Data initialization, determining the number of clusters and the total number of samples in the data set, initializing the membership degree and penalty factor; Step S32: Calculate the cluster center and determine the location of the cluster center based on the sample and the degree of membership of the sample to the cluster center. The formula used is as follows: ； In the formula, V represents the cluster center, represents the ath cluster center, d represents the number of clusters in the data set, a represents the cluster index, b represents the sample index, represents the membership of the bth sample to the ath cluster center, and m represents the weighted index for determining the clustering fuzziness level; Update the cluster center by considering the samples in the cluster at different times and the membership of the samples. The formula used is as follows: ； ； In the formula, represents the updated cluster center, Indicates the identifier of the a-th cluster center, which is used to distinguish whether the sample is in the cluster. represents the cluster at time t, represents the cluster at time t-1; As data is updated and the clustering process proceeds, clusters will become outdated due to changes in data. The distribution of data points no longer conforms to the characteristics of the original clusters. Outdated clusters are deleted. Clusters that meet the following conditions are deleted. The formula used is as follows: ； In the formula, represents the jth cluster at time t, represents the jth cluster at time tT, represents the total number of clusters at time t, and T represents the period without any cluster creation; Step S33: Calculate the objective function to measure the clustering quality. Considering the sensitivity of the fuzzy clustering algorithm to noise points, a weighted multi-core fuzzy clustering algorithm based on two variables is proposed. The formula used is as follows: ； In the formula, U represents the objective function, represents the total number of samples, represents the distance metric from the a-th cluster center to the b-th sample, k represents the adaptive coefficient of the distance metric, n represents the noise impact factor, L2 represents the membership constraint, represents the penalty factor of the a-th cluster center, Represents the weight of the kernel function of the bth sample; Step S34: Optimize the membership. During the iteration process, the membership of the sample to the cluster center is continuously adjusted so that the objective function reaches the minimum value. The formula used is as follows: ； In the formula, represents the optimized membership, o represents the index of the number of iterations, and Q represents the total number of iterations; Step S35: consistency evaluation, measuring the consistency between the clustering results and the actual situation, the formula used is as follows: ； In the formula, represents the consistency assessment value, represents the clustering results to be compared, represents the actual clustering result, represents the t1th clustering result to be compared, represents the t2th true clustering result, Indicates the total number of clustering results.

3. The root maggot pest classification method based on clustering processing according to claim 1, characterized in that: In step S1, the data collection includes the following steps: Step S11: soil sampling. The larvae of root maggot pests mainly live in the soil and feed on plant roots. Soil samples are collected at different locations in the field. The larvae of root maggot pests are separated from the soil by screening and rinsing. The sampling location and soil properties are recorded during sampling. Step S12: sampling the crop roots, collecting pest samples directly from the crop roots, digging the crop roots, checking whether there are root maggot pests in the roots, and marking the corresponding crop types and crop growth conditions for the collected pests; Step S13: carefully comparing the collected data, removing noise and outliers, and performing data conversion and normalization operations to obtain pre-processed root maggot pest data; The pre-processed root maggot pest data includes morphological data, developmental stage data, soil environment data, host plant data and pest distribution data of the root maggot pests.

4. The root maggot pest classification method based on clustering processing according to claim 1, characterized in that: In step S4, generating a cluster report specifically includes generating a report title and overview, characterizing the root maggot pest data, recording the cluster analysis process in detail, presenting the clustering results to relevant technical personnel using visualization methods, evaluating the clustering effect, and determining whether the clustering results meet expectations.

5. The root maggot pest classification method based on clustering processing according to claim 1, characterized in that: In step S5, the formulation of a pest control strategy includes the following steps: Step S51: prevention and treatment based on development stage, including the following contents: For clusters with a high proportion of larval stage pests, since the larvae mainly move around the roots of crops and harm the growth and development of the root system, soil treatment methods are used to spray pesticides during crop growth; For clusters with a high proportion of adult pests, since the adults have the ability to fly and will spread in the field, sex pheromones are used to trap the adults. According to the sex pheromone composition of the pests in the cluster, sex pheromone traps are made to interfere with the mating behavior of the adults during the adult breeding season and reduce the number of pests that reproduce. Step S52: Chemical control, determining the differences in sensitivity of pests to chemical agents based on the cluster analysis results, and arranging the spraying range and frequency of the chemical agents; Step S53: biological control, considering biological control measures from the perspective of the entire farmland ecosystem, introducing natural enemy insects and beneficial insects, and building a stable biological community; Step S54: Agricultural prevention and control: the pests in the clusters are related to soil moisture and fertility. Irrigation and fertilization are used to improve the soil environment and inhibit the growth and reproduction of pests.

6. A root maggot pest classification system based on clustering processing, used to implement the root maggot pest classification method based on clustering processing as described in any one of claims 1 to 5, characterized in that: It includes a data collection module, a feature extraction module, a cluster analysis module, a cluster report generation module and a pest control strategy formulation module.

7. The root maggot pest classification system based on clustering processing according to claim 6, characterized in that: The data collection module specifically performs soil sampling and crop root sampling to obtain root maggot pest data; The feature extraction module specifically performs feature multi-index fusion, comprehensively considers the three aspects of calculating sample feature standard deviation, discrimination criteria and common information, adopts infinite feature selection method to determine feature weight, uses inflection point detection algorithm to perform feature stratification, and performs probability extraction based on feature weight; The cluster analysis module specifically uses the initial membership and penalty factor to calculate the cluster center and the objective function, proposes a weighted multi-core fuzzy clustering algorithm based on two variables, optimizes the membership, and measures the consistency between the clustering results and the actual situation; The cluster report generation module specifically generates a pest cluster report, visualizes the cluster results, and evaluates whether the results meet the expected goals; The module for formulating pest control strategies is specifically for comprehensively controlling root maggot pests based on clustering results.

Images