Analysis method for corn quality distribution rule

Abstract

The invention discloses a corn quality distribution rule analysis method, relates to the technical field of corn quality analysis, and aims to solve the problems of inaccurate corn quality data acquisition and unclear corn quality distribution. According to the method, by combining a univariate analysis method and a multivariate analysis method, the internal rule of the abnormal index data can be comprehensively and deeply mined, the abnormal distribution rule decision is made according to the distribution mode of the abnormal index data, and the potential reason of the corn quality index data abnormity can be scientifically deduced; the standard sample collection data, the abnormal index data and the abnormal distribution rule decision result are integrated and displayed on the display terminal, information comprehensiveness is achieved, the standard sample collection data are compared with the standard index data divided in detail, the data deviation of each specific index can be accurately recognized, and the accuracy of the data deviation is improved. Through comprehensive and meticulous preprocessing of the target index data, powerful support is provided for follow-up study of corn quality distribution rules by using a statistical analysis method.

Description

Technical Field

[0001] The invention relates to the technical field of corn quality analysis, and in particular to an analysis method for corn quality distribution law. Background Art

[0002] Corn quality analysis refers to the testing of various quality indicators of corn through a series of tests and evaluation methods to determine its nutritional value, processing characteristics, food safety and commercial value.

[0003] The patent application with publication number CN112541659A discloses a system for collecting and sharing agricultural product quality information based on big data. The system mainly counts the number of corn planting bases in the region, collects environmental parameters of corn at each growth stage for each corn planting base, and extracts the growth characteristics of corn at each growth stage, thereby constructing a set of environmental parameters of the planting bases at the growth stage and a set of growth characteristics of the planting bases at the growth stage. The environmental quality coefficients and growth quality coefficients corresponding to each corn planting base at each growth stage are counted, so that the statistical agricultural product quality information is shared among the various planting farmers, realizing the sharing of agricultural product quality information among the various planting farmers, and making up for the lack of information communication among the various planting farmers. Although the above patent solves the problem of agricultural product quality data collection, the following problems still exist in actual operation: 1. When monitoring and collecting corn quality data, no targeted testing equipment was used for data collection, resulting in inaccurate monitoring of corn quality data.

[0004] 2. After the corn quality data was collected, the data was not further processed, resulting in inaccurate subsequent data analysis.

[0005] 3. The acquired corn quality data was not subjected to targeted abnormal analysis and visualization, which resulted in the staff being unable to obtain data in a timely manner. Summary of the invention

[0006] The purpose of the present invention is to provide an analysis method for the distribution law of corn quality. By combining univariate analysis and multivariate analysis methods, the inherent law of abnormal indicator data can be fully and deeply excavated, and abnormal distribution law decisions can be made according to the distribution pattern of abnormal indicator data. The potential cause of abnormal corn quality can be scientifically inferred, and standard sample collection data, abnormal indicator data and abnormal distribution law decision results are integrated and displayed on a display terminal, thereby achieving comprehensive information. By comparing the standard sample collection data with the detailed standard indicator data, the data deviation on each specific indicator can be accurately identified. By performing comprehensive and detailed preprocessing on the target indicator data, strong support can be provided for the subsequent study of the distribution law of corn quality using statistical analysis methods, and the problems in the prior art can be solved.

[0007] To achieve the above object, the present invention provides the following technical solutions: A method for analyzing corn quality distribution law comprises the following steps: S1: Confirmation of analysis target: According to corn varieties and planting environment, confirm the quality indicators of planted corn, and mark the quality indicators as standard indicator data; S2: Target data collection: According to the indicator attributes in the standard indicator data, corn sample data is collected for planting corn, and the collected data is marked as target indicator data; S3: Data collection and processing: preprocess the target indicator data to obtain standard sample collection data; S4: Data collection analysis: Compare and analyze the standard sample collection data with the standard indicator data, and mark the abnormal indicator distribution data as abnormal indicator data according to the comparison and analysis results; S5: Analysis data presentation: construct a distribution scheme for the standard sample collection data and abnormal indicator data, and transmit the constructed distribution scheme to the display terminal for presentation.

[0008] Preferably, in S1, determining the quality index of the planted corn according to the corn variety and the planting environment includes: Confirm the variety characteristics of the corn to be planted based on the variety of corn to be planted; Variety traits for growing corn include genetic traits and seed quality traits; Confirm the characteristics of the planting environment for corn, including climate conditions, soil conditions and topographic conditions; According to the varieties of corn planted and the characteristics of the planting environment, determine the quality distribution law indicators of the planted corn; Among them, the quality distribution law indicators include yield indicators, morphological indicators, nutritional indicators, processing quality indicators and safety indicators; The confirmed quality distribution indicators of planted corn are marked as standard indicator data.

[0009] Preferably, in S2, corn sample data collection is performed on the planted corn according to the indicator attributes in the standard indicator data, including: Data collection of corn samples from planted corn was conducted based on yield index, morphological index, nutritional index, processing quality index and safety index; The data collection of corn yield index is to collect samples and measure the yield of corn at the fully mature stage at the selected location, and obtain the actual yield per mu at the sampling point; The data collection of corn morphological indicators is to collect the morphological data of corn plants at different growth stages, among which the morphological data include plant height, ear height, leaf area and stem diameter; After the mature corn is collected, corn kernels are obtained, among which the corn kernel yield indicators are: ear length, ear diameter, number of ear rows, number of kernels per row and 100-kernel weight; The data collection of corn kernel nutrition index is to sample corn samples at the mature stage and analyze the nutrition components after sampling. The nutrition components mainly include protein, starch (amylose, amylopectin) and fat. Data collection of food processing quality indicators of corn kernels: Data collection of gelatinization temperature, endosperm texture, waxiness and elasticity of corn samples; Commercial quality of corn kernels: kernel moisture content and bulk density.

[0010] The data collection of corn sample safety indicators is to collect data on pesticide residues, heavy metal content and fungal toxin content in corn samples.

[0011] Preferably, in S2, the corn sample data collection for planting corn is performed according to the indicator attributes in the standard indicator data, and further includes: Data on yield indicators of corn samples were collected using harvesters or manual collection; The data of morphological indexes of corn samples were measured using tape measure and vernier caliper; The protein data in the nutritional index data of corn samples was collected using the Kjeldahl nitrogen determination method; the starch data and dietary fiber data were collected using the enzymatic method; the fat data were collected using the Soxhlet extraction method; the vitamin data were collected using high performance liquid chromatography; the mineral data were collected using atomic absorption spectrometry; and the moisture data were collected using a moisture meter. The gelatinization temperature of the corn sample processing quality index data was collected using a differential scanning calorimeter; the endosperm soil was collected using a shear force tester; the glutinousness was collected using a viscosity meter; and the elasticity was collected using a texture analyzer. The pesticide residue data in the safety index data of corn samples were tested by liquid chromatography-mass spectrometry; the heavy metal content was tested by atomic fluorescence spectroscopy; the fungal toxin content was tested by enzyme-linked immunosorbent assay; After the yield index, morphological index, nutritional index, processing quality index and safety index data are collected, the quality data of the planted corn seed corn samples are obtained and marked as target index data.

[0012] Preferably, data preprocessing is performed on the target indicator data in S3, including: Perform data preprocessing on the acquired target indicator data; The data preprocessing process is to first clean the target indicator data, use pivot tables, sorting, and filtering functions to check and clean duplicate data, fill or mark missing values ​​in the target indicator data, and identify outliers; After data cleaning is completed, data conversion is performed to standardize and normalize the target indicator data; After data conversion is completed, use data processing software to integrate the data; Conduct logic check and range check on the target indicator data after data integration; After the logic check and range check are completed, the target indicator data after data preprocessing is obtained and marked as standard sample collection data.

[0013] Preferably, the data comparison and analysis of the standard sample collection data and the standard indicator data in S4 includes: Compare the standard sample collection data with the standard indicator data; Among them, the yield index, morphological index, nutritional index, processing quality index and safety index in the standard sample collection data and the standard index data are retrieved respectively; Compare the retrieved standard sample collection data with the yield index, morphological index, nutritional index, processing quality index and safety index in the standard index data; Determine abnormal data in the standard sample collection data based on the comparison results; Among them, when the yield index, morphological index, nutritional index, processing quality index and safety index in the standard sample collection data are not within the threshold range of the yield index, morphological index, nutritional index, processing quality index and safety index in the standard index data, the corresponding attribute data in the standard sample collection data is abnormal quality data; The abnormal range of the abnormal quality data is confirmed, the abnormal range and the corresponding abnormal quality data are extracted separately, and marked as abnormal indicator data.

[0014] Preferably, for S5, the standard sample collection data and abnormal index data are distributed to construct a program, and the constructed distribution program is transmitted to the display terminal for presentation, including; Performing visualization conversion on the abnormal indicator data, wherein the visualization conversion is converting the abnormal indicator data into charts, and the charts include box plots, histograms, and scatter plots; Analyze the abnormal value distribution law of the abnormal indicator data after visualization conversion; The analysis of the distribution pattern of outliers includes univariate analysis and multivariate analysis; Among them, the histogram and box plot were subjected to univariate analysis, and the scatter plot was subjected to multivariate analysis; The univariate analysis of the histogram is to determine the peak and skewness in the diagram based on the distribution shape of the data in the histogram; the univariate analysis of the box plot is to define the indicators based on the median, quartiles and outliers of the data indicators in the box plot; the multivariate analysis of the scatter plot is to draw a scatter plot matrix and confirm the relationship between different indicators.

[0015] Preferably, the distribution scheme for the standard sample collection data and the abnormal index data is constructed in S5, and the constructed distribution scheme is transmitted to the display terminal for presentation, and further includes: A distribution module for determining abnormal indicator data according to the analysis results of univariate analysis and multivariate analysis, wherein the distribution modes include random distribution, concentrated distribution, trend distribution and periodic distribution; Making abnormal distribution law decisions based on the distribution pattern, wherein the abnormal distribution law decisions are to judge the abnormal causes of corn quality based on the distribution pattern of abnormal indicator data; Then the standard sample collection data is converted visually; The standard sample collection data and abnormal index data after visual conversion, as well as the abnormal distribution law decision data of the abnormal index data are transmitted to the display terminal for display.

[0016] Preferably, the abnormal indicator data after the visualization conversion is analyzed for abnormal value distribution law, including: The abnormal indicator data after the visualization conversion is input into the univariate analysis model, and the first analysis result is output; wherein the training method of the univariate analysis model is: obtaining sample data; preprocessing the sample data based on the data layer of the univariate analysis model to determine the abnormal factors; performing data analysis on the abnormal factors based on the indicator layer of the univariate analysis model to determine several indicators, and screening to obtain abnormal indicators; combining various abnormal indicators based on the model parameter layer of the univariate analysis model to obtain multiple combination results, and screening the combination result with the highest abnormal probability prediction accuracy as the target combination result; outputting the prediction result of the sample data based on the output layer of the univariate analysis model, when the prediction result is consistent with the actual result corresponding to the sample data, it indicates that the training is completed; Constructing an indicator data list based on the abnormal indicator data and other indicator data completed by the visualization conversion, mapping the list structure information of the indicator data list to a preset coordinate plane in the form of graph data, so as to draw a graph data set of the indicator data list in the coordinate plane; Perform feature extraction on each graph data node in the graph data set to obtain the plane feature corresponding to each graph data node, establish the association relationship of the graph data nodes according to the feature dimension of the plane feature of each graph data node, and generate a data system; input the data system into the multivariate analysis model, and output the second analysis result; the training method of the multivariate analysis model includes: determining the big data of the corn quality distribution law, and constructing the map data of the corn quality distribution law according to the big data; extracting each graph data node in the map data, and determining the association relationship between each graph data node, and constructing the knowledge map of the data analysis chain corresponding to the corn quality distribution law; determining the abnormal indicator information of each graph data node; extracting the associated feature vector in the knowledge map, and performing feature fusion according to the associated feature vector and the abnormal indicator information to obtain the fusion vector; training the multivariate analysis model based on the fusion vector; Based on the first analysis result and the second analysis result, a final result of performing an outlier value distribution law analysis on the abnormal indicator data is determined.

[0017] Preferably, the association relationship of the graph data nodes is established according to the feature dimension of the plane feature of each graph data node to generate a data system, including: Determine the phase space reconstruction results of the graph data nodes ;

[0018] in, It is the first graph data node; is the nth graph data node; n is the number of graph data nodes; is the preset time delay, the value is (0,1); Based on the phase space reconstruction results of the graph data nodes, the characteristic dimension of the plane feature of each graph data node is used as the basis for phase space reconstruction, and the one-dimensional time series is expanded in the multi-dimensional phase space. ;

[0019] in, for The i-th characteristic dimension of is also the phase point; Determine the phase point in the reconstructed m-dimensional phase space To other points The number of points whose distance does not exceed r ;

[0020] in, () is the Heaviside function; r is the preset distance information, which is used to indicate the threshold of the distance of the association relationship; Determine the chaotic time series correlation dimension feature extraction function of graph data nodes based on point information ;

[0021] Based on the chaotic time series correlation dimension feature extraction function, the correlation relationship of graph data nodes is determined to generate a data system.

[0022] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention provides an analysis method for the distribution law of corn quality. The multi-dimensional evaluation method is helpful to discover the inherent laws and potential problems of quality changes, and provides an important reference for quality improvement and variety improvement. The professional collection and detection methods ensure the accuracy and reliability of the data. Data collection is carried out according to standard detection and analysis methods, which ensures the scientificity and standardization of the data collection process.

[0023] 2. The present invention provides a method for analyzing the distribution law of corn quality. By comparing the standard sample collection data with the detailed standard indicator data, the data deviation on each specific indicator can be accurately identified, thereby ensuring the accuracy of the analysis results. By performing comprehensive and detailed preprocessing of the target indicator data, strong support can be provided for the subsequent study of the distribution law of corn quality using statistical analysis, machine learning and other methods.

[0024] 3. The present invention provides a method for analyzing the distribution law of corn quality, which combines univariate analysis and multivariate analysis methods, can comprehensively and deeply explore the inherent laws of abnormal indicator data, make abnormal distribution law decisions based on the distribution patterns of abnormal indicator data, and can scientifically infer the potential causes of abnormal corn quality. The standard sample collection data, abnormal indicator data and abnormal distribution law decision results are integrated and displayed on the display terminal, thereby realizing comprehensive and rapid transmission of information. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 Schematic diagram of the analysis steps of the corn quality distribution law of the present invention; Figure 2 The figure shows the analysis process of the corn quality distribution law of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] In order to solve the problem that in the prior art, when monitoring and collecting corn quality data, no targeted detection equipment is used for data collection, resulting in inaccurate corn quality data monitoring, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions: A method for analyzing corn quality distribution law comprises the following steps: S1: Confirmation of analysis target: According to corn varieties and planting environment, confirm the quality indicators of planted corn, and mark the quality indicators as standard indicator data; Among them, multi-dimensional evaluation methods help to discover the inherent laws and potential problems of quality changes; S2: Target data collection: According to the indicator attributes in the standard indicator data, corn sample data is collected for planting corn, and the collected data is marked as target indicator data; Among them, professional instruments are used to analyze nutritional content and processing characteristics, and high-precision testing technology is used to evaluate safety indicators; S3: Data collection and processing: preprocess the target indicator data to obtain standard sample collection data; Among them, comprehensive and detailed preprocessing of target indicator data can provide strong support for the subsequent use of statistical analysis, machine learning and other methods to study the distribution law of corn quality; S4: Data collection analysis: Compare and analyze the standard sample collection data with the standard indicator data, and mark the abnormal indicator distribution data as abnormal indicator data according to the comparison and analysis results; Among them, by comparing and analyzing the subdivided indicators, quality management is made more refined; S5: Analytical data presentation: construct a distribution scheme for the standard sample collection data and abnormal indicator data, and transmit the constructed distribution scheme to the display terminal for presentation; Among them, making abnormal distribution law decisions based on the distribution pattern of abnormal indicator data can scientifically infer the potential causes of abnormal corn quality.

[0028] According to the corn variety and planting environment in S1, the quality indicators of corn to be planted are determined, including: Confirm the variety characteristics of the corn to be planted based on the variety of corn to be planted; Variety traits for growing corn include genetic traits and seed quality traits; Confirm the characteristics of the planting environment for corn, including climate conditions, soil conditions and topographic conditions; According to the varieties of corn planted and the characteristics of the planting environment, determine the quality distribution law indicators of the planted corn; Among them, the quality distribution law indicators include yield indicators, morphological indicators, nutritional indicators, processing quality indicators and safety indicators; The confirmed quality distribution indicators of planted corn are marked as standard indicator data.

[0029] Specifically, the comprehensiveness and accuracy of the quality distribution law analysis are ensured based on the two key factors of corn variety characteristics and planting environment characteristics. By carefully analyzing the genetic potential of the variety and the initial quality of the seeds, combined with the specific climate, soil and terrain conditions, the quality changes of corn during growth can be more accurately grasped. Based on scientific methods and detailed data collection, the subjectivity of judgment based on experience or intuition is avoided. The quality distribution law indicators cover multiple aspects such as yield, morphology, nutrition, processing and safety, and achieve a comprehensive evaluation of corn quality. This multi-dimensional evaluation method helps to discover the inherent laws and potential problems of quality changes, and provides an important reference for quality improvement and variety improvement. By analyzing the quality distribution law, the future quality of corn can be predicted and evaluated, and forward-looking planting guidance and market strategy suggestions can be provided to farmers. This helps farmers to adjust the planting structure and optimize resource allocation in a timely manner to adapt to market changes and changes in consumer demand.

[0030] According to the indicator attributes in the standard indicator data in S2, corn sample data will be collected for planting corn, including: Data collection of corn samples from planted corn was conducted based on yield index, morphological index, nutritional index, processing quality index and safety index; The data collection of corn yield index is to collect samples and measure the yield of corn at the fully mature stage at the selected location, and obtain the actual yield per mu at the sampling point; The data collection of corn morphological indicators is to collect the morphological data of corn plants at different growth stages, among which the morphological data include plant height, ear height, leaf area and stem diameter; After the mature corn is collected, corn kernels are obtained, among which the corn kernel yield indicators are: ear length, ear diameter, number of ear rows, number of kernels per row and 100-kernel weight; The data collection of corn kernel nutrition index is to sample corn samples at the mature stage and analyze the nutrition components after sampling. The nutrition components mainly include protein, starch (amylose, amylopectin) and fat. Data collection of food processing quality indicators of corn kernels: Data collection of gelatinization temperature, endosperm texture, waxiness and elasticity of corn samples; Commercial quality of corn kernels: kernel moisture content and bulk density.

[0031] The data collection of corn sample safety indicators is to collect data on pesticide residues, heavy metal content and fungal toxin content in corn samples.

[0032] Data on yield indicators of corn samples were collected using harvesters or manual collection; The data of morphological indexes of corn samples were measured using tape measure and vernier caliper; The protein data in the nutritional index data of corn samples was collected using the Kjeldahl nitrogen determination method; the starch data and dietary fiber data were collected using the enzymatic method; the fat data were collected using the Soxhlet extraction method; the vitamin data were collected using high performance liquid chromatography; the mineral data were collected using atomic absorption spectrometry; and the moisture data were collected using a moisture meter. The gelatinization temperature of the corn sample processing quality index data was collected using a differential scanning calorimeter; the endosperm soil was collected using a shear force tester; the glutinousness was collected using a viscosity meter; and the elasticity was collected using a texture analyzer. The pesticide residue data in the safety index data of corn samples were tested by liquid chromatography-mass spectrometry; the heavy metal content was tested by atomic fluorescence spectroscopy; the fungal toxin content was tested by enzyme-linked immunosorbent assay; After the yield index, morphological index, nutritional index, processing quality index and safety index data are collected, the quality data of the planted corn seed corn samples are obtained and marked as target index data.

[0033] Specifically, according to multiple key dimensions of corn quality assessment, including yield, morphology, nutrition, processing and safety indicators. This comprehensive data collection method can ensure a comprehensive and in-depth understanding of corn quality, providing a solid foundation for subsequent distribution law analysis. Special collection methods and tools are used for different indicators, such as using tape measures and vernier calipers to measure morphological data, using professional instruments to analyze nutritional components and processing characteristics, and evaluating safety indicators through high-precision testing technology. These professional collection and testing methods ensure the accuracy and reliability of the data. Data collection is carried out according to standard testing and analysis methods to ensure the scientificity and standardization of the data collection process. This not only improves the credibility of the data, but also facilitates data comparison and sharing between different studies. Through continuous data collection and analysis, the changing trend of corn quality can be dynamically monitored, and problems can be discovered and solved in a timely manner.

[0034] In order to solve the problem in the prior art that after the corn quality data is collected, the data is not further processed, resulting in inaccurate data analysis in the later stage, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions: Preprocess the target indicator data in S3, including: Perform data preprocessing on the acquired target indicator data; The data preprocessing process is to first clean the target indicator data, use pivot tables, sorting, and filtering functions to check and clean duplicate data, fill or mark missing values ​​in the target indicator data, and identify outliers; After data cleaning is completed, data conversion is performed to standardize and normalize the target indicator data; After data conversion is completed, use data processing software to integrate the data; Conduct logic check and range check on the target indicator data after data integration; After the logic check and range check are completed, the target indicator data after data preprocessing is obtained and marked as standard sample collection data.

[0035] Specifically, through data cleaning steps, such as using pivot tables, sorting, and filtering to check and clean duplicate data, redundancy and errors in the data can be significantly reduced, thereby improving the quality of the data. Filling or labeling missing values ​​and identifying and processing outliers ensure the integrity and consistency of the data, laying a solid foundation for subsequent analysis. Standardization and normalization in the data conversion process enable data of different dimensions and distribution ranges to be compared and analyzed on the same scale. This is crucial for understanding the relative relationship between corn quality indicators and studying distribution patterns. The data preprocessing process is systematized, from cleaning to conversion to integration, and each step is supported by clear methods and tools, which not only simplifies the complexity of data analysis, but also greatly improves the efficiency of analysis. Logical checks and range checks are the last step of the data preprocessing process to ensure the rationality and accuracy of the data. These checks can detect and correct potential errors or unreasonable aspects in the data, thereby avoiding the impact of erroneous data on the analysis results and ensuring the accuracy and reliability of the analysis results. By conducting comprehensive and detailed preprocessing of the target indicator data, it can provide strong support for the subsequent use of statistical analysis, machine learning and other methods to study the distribution laws of corn quality, thereby promoting progress in the fields of corn quality evaluation and breeding improvement.

[0036] Compare and analyze the standard sample collection data with the standard indicator data in S4, including: Compare the standard sample collection data with the standard indicator data; Among them, the yield index, morphological index, nutritional index, processing quality index and safety index in the standard sample collection data and the standard index data are retrieved respectively; Compare the retrieved standard sample collection data with the yield index, morphological index, nutritional index, processing quality index and safety index in the standard index data; Determine abnormal data in the standard sample collection data based on the comparison results; Among them, when the yield index, morphological index, nutritional index, processing quality index and safety index in the standard sample collection data are not within the threshold range of the yield index, morphological index, nutritional index, processing quality index and safety index in the standard index data, the corresponding attribute data in the standard sample collection data is abnormal quality data; The abnormal range of the abnormal quality data is confirmed, the abnormal range and the corresponding abnormal quality data are extracted separately, and marked as abnormal indicator data.

[0037] Specifically, by comparing the standard sample collection data with the detailed standard indicator data (yield, morphology, nutrition, processing, safety), the data deviation on each specific indicator can be accurately identified, thereby ensuring the accuracy of the analysis results. A clear threshold range is set to judge abnormal data. This rule-based detection method can quickly identify data points that exceed the normal range, that is, abnormal quality data, which improves the efficiency and accuracy of abnormal detection. It not only identifies abnormal data, but also further confirms the abnormal range, and extracts and annotates these abnormal indicator data separately. This helps to conduct in-depth analysis of the causes of abnormalities, such as planting conditions, processing or environmental factors, and provides specific improvement directions for quality control. By comparing and analyzing the subdivided indicators, quality management can be more refined. More targeted improvement measures can be formulated for different indicators to improve the overall quality control level. Through continuous data comparison and analysis, quality change trends can be monitored, potential problems can be discovered and corrected in a timely manner, and the continuous improvement and optimization of the quality management system can be promoted.

[0038] In order to solve the problem that the existing technology does not conduct targeted abnormal analysis and visualization of the acquired corn quality data, which results in the staff being unable to obtain the data in a timely manner, please refer to Figure 1 and Figure 2 , this embodiment provides the following technical solutions: Constructing a distribution scheme for the standard sample collection data and abnormal indicator data in S5, and transmitting the constructed distribution scheme to the display terminal for presentation, including; Performing visualization conversion on the abnormal indicator data, wherein the visualization conversion is converting the abnormal indicator data into charts, and the charts include box plots, histograms, and scatter plots; Analyze the abnormal value distribution law of the abnormal indicator data after visualization conversion; The analysis of the distribution pattern of outliers includes univariate analysis and multivariate analysis; Among them, the histogram and box plot were subjected to univariate analysis, and the scatter plot was subjected to multivariate analysis; The univariate analysis of the histogram is to determine the peak and skewness in the diagram based on the distribution shape of the data in the histogram; the univariate analysis of the box plot is to define the indicators based on the median, quartiles and outliers of the data indicators in the box plot; the multivariate analysis of the scatter plot is to draw a scatter plot matrix and confirm the relationship between different indicators.

[0039] A distribution module for determining abnormal indicator data according to the analysis results of univariate analysis and multivariate analysis, wherein the distribution modes include random distribution, concentrated distribution, trend distribution and periodic distribution; Making abnormal distribution law decisions based on the distribution pattern, wherein the abnormal distribution law decisions are to judge the abnormal causes of corn quality based on the distribution pattern of abnormal indicator data; Then the standard sample collection data is converted visually; The standard sample collection data and abnormal index data after visual conversion, as well as the abnormal distribution law decision data of the abnormal index data are transmitted to the display terminal for display.

[0040] Specifically, by visually converting abnormal indicator data into charts (such as box plots, histograms, and scatter plots), complex data can be made more intuitive and easy to understand. This display method helps non-professional data analysts to quickly understand data characteristics, and facilitates decision makers and researchers to quickly capture key information in the data. Combining univariate analysis and multivariate analysis methods, it can comprehensively and deeply explore the inherent laws of abnormal indicator data. Univariate analysis (such as histograms and box plots) focuses on the distribution characteristics of a single indicator, while multivariate analysis (such as scatter plots) can reveal the correlation between different indicators, thereby providing more comprehensive data insights. Through the accurate identification and definition of outliers by box plots and histograms, the abnormal points in the data can be quickly locked, providing a clear direction for subsequent abnormal cause analysis and processing. This precise positioning capability is crucial to improving the efficiency and accuracy of corn quality monitoring. According to the distribution pattern of abnormal indicator data (such as random distribution, concentrated distribution, trend distribution, and periodic distribution), abnormal distribution law decisions can be made, which can scientifically infer the potential causes of abnormal corn quality. This data-based decision support method helps to improve the scientificity and accuracy of decision-making, optimize corn quality management strategies, integrate and display standard sample collection data, abnormal index data and abnormal distribution law decision results on the display terminal, and realize comprehensive and rapid information transmission. This integrated display method not only improves work efficiency, but also facilitates relevant personnel to conduct comprehensive analysis and comparison of data, so as to make more reasonable judgments and decisions.

[0041] Preferably, the abnormal indicator data after the visualization conversion is analyzed for abnormal value distribution law, including: The abnormal indicator data after the visualization conversion is input into the univariate analysis model, and the first analysis result is output; wherein the training method of the univariate analysis model is: obtaining sample data; preprocessing the sample data based on the data layer of the univariate analysis model to determine the abnormal factors; performing data analysis on the abnormal factors based on the indicator layer of the univariate analysis model to determine several indicators, and screening to obtain abnormal indicators; combining various abnormal indicators based on the model parameter layer of the univariate analysis model to obtain multiple combination results, and screening the combination result with the highest abnormal probability prediction accuracy as the target combination result; outputting the prediction result of the sample data based on the output layer of the univariate analysis model, when the prediction result is consistent with the actual result corresponding to the sample data, it indicates that the training is completed; Constructing an indicator data list based on the abnormal indicator data and other indicator data completed by the visualization conversion, mapping the list structure information of the indicator data list to a preset coordinate plane in the form of graph data, so as to draw a graph data set of the indicator data list in the coordinate plane; Perform feature extraction on each graph data node in the graph data set to obtain the plane feature corresponding to each graph data node, establish the association relationship of the graph data nodes according to the feature dimension of the plane feature of each graph data node, and generate a data system; input the data system into the multivariate analysis model, and output the second analysis result; the training method of the multivariate analysis model includes: determining the big data of the corn quality distribution law, and constructing the map data of the corn quality distribution law according to the big data; extracting each graph data node in the map data, and determining the association relationship between each graph data node, and constructing the knowledge map of the data analysis chain corresponding to the corn quality distribution law; determining the abnormal indicator information of each graph data node; extracting the associated feature vector in the knowledge map, and performing feature fusion according to the associated feature vector and the abnormal indicator information to obtain the fusion vector; training the multivariate analysis model based on the fusion vector; Based on the first analysis result and the second analysis result, a final result of performing an outlier value distribution law analysis on the abnormal indicator data is determined.

[0042] The working principle of the above technical solution: when the abnormal indicator data after the visualization conversion is analyzed for the distribution law of abnormal values, the analysis of the distribution law of abnormal values ​​includes univariate analysis and multivariate analysis. First, univariate analysis is performed: the abnormal indicator data after the visualization conversion is input into the univariate analysis model, and the first analysis result is output; the univariate analysis model is used to analyze the abnormal information when there is only one variable. When training the univariate analysis model, the sample data is first preprocessed based on the data layer of the univariate analysis model to determine the abnormal factors, and the abnormal factors are analyzed based on the indicator layer of the univariate analysis model to determine several indicators, and then screened to obtain abnormal indicators; various abnormal indicators are combined based on the model parameter layer of the univariate analysis model to obtain multiple combination results, and the combination result with the highest abnormal probability prediction accuracy is screened out as the target combination result; the output layer based on the univariate analysis model outputs the prediction result of the sample data, and when the prediction result is consistent with the actual result corresponding to the sample data, it indicates that the training is completed.

[0043] Secondly, multivariate analysis is performed. First, an indicator data list is constructed based on the abnormal indicator data and other indicator data completed by the visualization conversion. The other indicator data is the indicator data other than the abnormal indicator data. The list structure information of the indicator data list is mapped to a preset coordinate plane in the form of graph data, so as to draw a graph data set of the indicator data list in the coordinate plane; the data is integrated into a graph data set; each graph data node in the graph data set is feature extracted to obtain the plane feature corresponding to each graph data node, and the association relationship of the graph data nodes is established according to the feature dimension of the plane feature of each graph data node to generate a data system; the integration of multiple variables is realized as a data system. Input into the multivariate analysis model, and output the second analysis result; when training the variable analysis model, a large amount of data on corn quality is collected, which may come from different farmlands, different planting seasons, different varieties, etc. The data should include various quality indicators of corn, such as protein content, fat content, starch content, moisture content, etc., as well as factors that may affect these qualities, such as soil type, fertilizer application amount, irrigation amount, climatic conditions, etc. Based on the collected big data, the atlas data of the distribution law of corn quality can be constructed. Graph data graphically represents the relationship between data. Each node represents a data point (such as corn quality data of a farm), and the lines between nodes represent the association between data points (such as close geographical location, similar planting conditions, etc.). In the graph data, it is necessary to extract the specific information of each graph data node, that is, each data point. At the same time, it is also necessary to determine the association relationship between these nodes, which may be based on geographical location, planting conditions, variety similarity, etc. By extracting and determining these relationships, a knowledge graph of the data analysis chain corresponding to the distribution law of corn quality can be constructed. In the knowledge graph, each graph data node corresponds to a set of corn quality data. It is necessary to analyze these data to determine which indicators are abnormal, that is, values ​​that are significantly different from the overall distribution or expected values. These abnormal indicator information is crucial for subsequent feature fusion and model training. In the knowledge graph, in addition to the abnormal indicator information, it is also necessary to extract the associated feature vectors between nodes. These feature vectors reflect the similarity or difference between nodes, which is of great significance for understanding the distribution law of corn quality. By fusing the associated feature vector with the abnormal indicator information, a more comprehensive and information-rich fusion vector can be obtained. The obtained fusion vector is used to train the multivariate analysis model. The multivariate analysis model can handle the relationship between multiple variables and extract the combination of variables that has a significant impact on the results. Through training, the model can learn the inherent characteristics of the distribution law of corn quality, so that it can accurately predict and analyze new data. Construct a multivariate analysis model based on big data and graph data technology to analyze and predict the distribution law of corn quality.

[0044] Beneficial effects of the above technical solution: The univariate analysis model and the multivariate analysis model conduct in-depth analysis of abnormal indicator data from different perspectives. The univariate analysis model focuses on the abnormal detection of a single indicator, while the multivariate analysis model considers the correlation and mutual influence between multiple indicators. Combining the analysis results of the two, the distribution law and characteristics of abnormal indicator data can be more comprehensively understood, thereby drawing more accurate conclusions.

[0045] Preferably, the association relationship of the graph data nodes is established according to the feature dimension of the plane feature of each graph data node to generate a data system, including: Determine the phase space reconstruction results of the graph data nodes ;

[0046] in, It is the first graph data node; is the nth graph data node; n is the number of graph data nodes; is the preset time delay, the value is (0,1); Based on the phase space reconstruction results of the graph data nodes, the characteristic dimension of the plane feature of each graph data node is used as the basis for phase space reconstruction, and the one-dimensional time series is expanded in the multi-dimensional phase space. ;

[0047] in, for The i-th characteristic dimension of is also the phase point; Determine the phase point in the reconstructed m-dimensional phase space To other points The number of points whose distance does not exceed r ;

[0048] in, () is the Heaviside function; r is the preset distance information, which is used to indicate the threshold of the distance of the association relationship; Determine the chaotic time series correlation dimension feature extraction function of graph data nodes based on point information ;

[0049] Based on the chaotic time series correlation dimension feature extraction function, the correlation relationship of graph data nodes is determined to generate a data system.

[0050] The working principle and beneficial effects of the above technical solution are as follows: First, the phase space reconstruction results of the graph data nodes are determined. Phase space reconstruction is a method of expanding one-dimensional time series data to multi-dimensional phase space in order to better analyze the dynamic characteristics of the data. Y represents the phase space reconstruction results of all graph data nodes, where Y 1 to Yn represent the reconstruction results of the 1st to nth graph data nodes respectively. τ is the preset time delay, which determines the time interval between adjacent data points when reconstructing the phase space. Here, the value range of τ is (0,1). In practical applications, τ will be more finely selected according to the characteristics of the data and the analysis requirements. Based on the phase space reconstruction results of the graph data nodes, the characteristic dimension of the plane feature of each graph data node is used as the foundation for phase space reconstruction, and the one-dimensional time series is expanded in the multidimensional phase space, which is a data vector containing multiple time points. This vector represents a phase point in the multidimensional phase space, which reflects the state of the data at different time points. In the reconstructed m-dimensional phase space, the distance from each phase point yj to other phase points yi is calculated, and the number of points Q whose distance does not exceed the preset threshold r is determined. The Heaviside function H() is used to determine whether the distance is less than r. If the distance is less than r, the function value is 1, otherwise it is 0. By performing such calculations on all possible phase point pairs, the number of points Q of each phase point can be obtained. The chaotic time series correlation dimension feature extraction function of the graph data nodes is determined based on the point number information. This function reflects the degree of correlation between phase points in the phase space under a given distance threshold r, which is convenient for displaying the dynamic characteristics of the data. The chaotic time series correlation dimension feature extraction function is used to determine the correlation relationship between graph data nodes. These correlation relationships reflect the similarity and dynamic connection between data points. Based on these correlation relationships, a data system can be constructed, which can more comprehensively reflect the intrinsic characteristics and laws of the data, facilitate accurate generation of data systems, and facilitate multivariate analysis and processing.

[0051] 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.

[0052] 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.

Claims

1. A method for analyzing the distribution law of corn quality, characterized in that: The steps include: S1: Confirmation of analysis target: According to corn varieties and planting environment, confirm the quality indicators of planted corn, and mark the quality indicators as standard indicator data; S2: Target data collection: According to the indicator attributes in the standard indicator data, corn sample data is collected for planting corn, and the collected data is marked as target indicator data; S3: Data collection and processing: preprocess the target indicator data to obtain standard sample collection data; S4: Data collection analysis: Compare and analyze the standard sample collection data with the standard indicator data, and mark the abnormal indicator distribution data as abnormal indicator data according to the comparison and analysis results; S5: Analysis data presentation: construct a distribution scheme for the standard sample collection data and abnormal indicator data, and transmit the constructed distribution scheme to the display terminal for presentation.

2. The method for analyzing corn quality distribution law according to claim 1, characterized in that: According to the corn variety and planting environment in S1, the quality indicators of corn to be planted are determined, including: Confirm the variety characteristics of the corn to be planted based on the variety of corn to be planted; Variety traits for growing corn include genetic traits and seed quality traits; Confirm the characteristics of the planting environment for corn, including climate conditions, soil conditions and topographic conditions; According to the varieties of corn planted and the characteristics of the planting environment, determine the quality distribution law indicators of the planted corn; Among them, the quality distribution law indicators include yield indicators, morphological indicators, nutritional indicators, processing quality indicators and safety indicators; The confirmed quality distribution indicators of planted corn are marked as standard indicator data.

3. The method for analyzing the distribution law of corn quality according to claim 2, characterized in that: According to the indicator attributes in the standard indicator data in S2, corn sample data will be collected for planting corn, including: Data collection of corn samples from planted corn was conducted based on yield index, morphological index, nutritional index, processing quality index and safety index; The data collection of corn yield index is to collect samples and measure the yield of corn at the fully mature stage at the selected location, and obtain the actual yield per mu at the sampling point; The data collection of corn morphological indicators is to collect the morphological data of corn plants at different growth stages, among which the morphological data include plant height, ear height, leaf area and stem diameter; After the mature corn is collected, corn kernels are obtained, among which the corn kernel yield indicators are: ear length, ear diameter, number of ear rows, number of kernels per row and 100-kernel weight; The data collection of corn kernel nutrition index is to sample corn samples at the mature stage and analyze the nutrition components after sampling, among which the nutrition components mainly include protein, starch and fat; Data collection of food processing quality indicators of corn kernels: Data collection of gelatinization temperature, endosperm texture, waxiness and elasticity of corn samples; Corn kernel commercial quality: kernel moisture content and bulk density; The data collection of corn sample safety indicators is to collect data on pesticide residues, heavy metal content and fungal toxin content in corn samples.

4. The method for analyzing corn quality distribution law according to claim 3, characterized in that: According to the indicator attributes in the standard indicator data in S2, corn sample data collection is carried out for planting corn, which also includes: Data on yield indicators of corn samples were collected using harvesters or manual collection; The data of morphological indexes of corn samples were measured using tape measure and vernier caliper; The protein data in the nutritional index data of corn samples was collected using the Kjeldahl nitrogen determination method; the starch data and dietary fiber data were collected using the enzymatic method; the fat data were collected using the Soxhlet extraction method; the vitamin data were collected using high performance liquid chromatography; the mineral data were collected using atomic absorption spectrometry; and the moisture data were collected using a moisture meter. The gelatinization temperature of the corn sample processing quality index data was collected using a differential scanning calorimeter; the endosperm soil was collected using a shear force tester; the glutinousness was collected using a viscosity meter; and the elasticity was collected using a texture analyzer. The pesticide residue data in the safety index data of corn samples were tested by liquid chromatography-mass spectrometry; the heavy metal content was tested by atomic fluorescence spectroscopy; the fungal toxin content was tested by enzyme-linked immunosorbent assay; After the yield index, morphological index, nutritional index, processing quality index and safety index data are collected, the quality data of the planted corn seed corn samples are obtained and marked as target index data.

5. The method for analyzing the distribution law of corn quality according to claim 4, characterized in that: Preprocess the target indicator data in S3, including: Perform data preprocessing on the acquired target indicator data; The data preprocessing process is to first clean the target indicator data, use pivot tables, sorting, and filtering functions to check and clean duplicate data, fill or mark missing values ​​in the target indicator data, and identify outliers; After data cleaning is completed, data conversion is performed to standardize and normalize the target indicator data; After data conversion is completed, use data processing software to integrate the data; Conduct logic check and range check on the target indicator data after data integration; After the logic check and range check are completed, the target indicator data after data preprocessing is obtained and marked as standard sample collection data.

6. The method for analyzing the distribution law of corn quality according to claim 5, characterized in that: Compare and analyze the standard sample collection data with the standard indicator data in S4, including: Compare the standard sample collection data with the standard indicator data; Among them, the yield index, morphological index, nutritional index, processing quality index and safety index in the standard sample collection data and the standard index data are retrieved respectively; Compare the retrieved standard sample collection data with the yield index, morphological index, nutritional index, processing quality index and safety index in the standard index data; Determine abnormal data in the standard sample collection data based on the comparison results; Among them, when the yield index, morphological index, nutritional index, processing quality index and safety index in the standard sample collection data are not within the threshold range of the yield index, morphological index, nutritional index, processing quality index and safety index in the standard index data, the corresponding attribute data in the standard sample collection data is abnormal quality data; The abnormal range of the abnormal quality data is confirmed, the abnormal range and the corresponding abnormal quality data are extracted separately, and marked as abnormal indicator data.

7. The method for analyzing corn quality distribution law according to claim 6, characterized in that: Constructing a distribution scheme for the standard sample collection data and abnormal indicator data in S5, and transmitting the constructed distribution scheme to the display terminal for presentation, including; Performing visualization conversion on the abnormal indicator data, wherein the visualization conversion is converting the abnormal indicator data into charts, and the charts include box plots, histograms, and scatter plots; Analyze the abnormal value distribution law of the abnormal indicator data after visualization conversion; The analysis of the distribution pattern of outliers includes univariate analysis and multivariate analysis; Among them, the histogram and box plot were subjected to univariate analysis, and the scatter plot was subjected to multivariate analysis; The univariate analysis of the histogram is to determine the peak and skewness in the diagram based on the distribution shape of the data in the histogram; the univariate analysis of the box plot is to define the indicators based on the median, quartiles and outliers of the data indicators in the box plot; the multivariate analysis of the scatter plot is to draw a scatter plot matrix and confirm the relationship between different indicators.

8. The method for analyzing corn quality distribution law according to claim 7, characterized in that: Constructing a distribution scheme for the standard sample collection data and the abnormal index data in S5, and transmitting the constructed distribution scheme to the display terminal for presentation, further comprising: A distribution module for determining abnormal indicator data according to the analysis results of univariate analysis and multivariate analysis, wherein the distribution modes include random distribution, concentrated distribution, trend distribution and periodic distribution; Making abnormal distribution law decisions based on the distribution pattern, wherein the abnormal distribution law decisions are to judge the abnormal causes of corn quality based on the distribution pattern of abnormal indicator data; Then the standard sample collection data is converted visually; The standard sample collection data and abnormal index data after visual conversion, as well as the abnormal distribution law decision data of the abnormal index data are transmitted to the display terminal for display.

9. The method for analyzing corn quality distribution law according to claim 7, characterized in that: The abnormal indicator data after visualization conversion is analyzed for abnormal value distribution rules, including: The abnormal indicator data after the visualization conversion is input into the univariate analysis model, and the first analysis result is output; wherein the training method of the univariate analysis model is: obtaining sample data; preprocessing the sample data based on the data layer of the univariate analysis model to determine the abnormal factors; performing data analysis on the abnormal factors based on the indicator layer of the univariate analysis model to determine several indicators, and screening to obtain abnormal indicators; combining various abnormal indicators based on the model parameter layer of the univariate analysis model to obtain multiple combination results, and screening the combination result with the highest abnormal probability prediction accuracy as the target combination result; outputting the prediction result of the sample data based on the output layer of the univariate analysis model, when the prediction result is consistent with the actual result corresponding to the sample data, it indicates that the training is completed; Constructing an indicator data list based on the abnormal indicator data and other indicator data completed by the visualization conversion, mapping the list structure information of the indicator data list to a preset coordinate plane in the form of graph data, so as to draw a graph data set of the indicator data list in the coordinate plane; Perform feature extraction on each graph data node in the graph data set to obtain the plane feature corresponding to each graph data node, establish the association relationship of the graph data nodes according to the feature dimension of the plane feature of each graph data node, and generate a data system; input the data system into the multivariate analysis model, and output the second analysis result; the training method of the multivariate analysis model includes: determining the big data of the corn quality distribution law, and constructing the map data of the corn quality distribution law according to the big data; extracting each graph data node in the map data, and determining the association relationship between each graph data node, and constructing the knowledge map of the data analysis chain corresponding to the corn quality distribution law; determining the abnormal indicator information of each graph data node; extracting the associated feature vector in the knowledge map, and performing feature fusion according to the associated feature vector and the abnormal indicator information to obtain the fusion vector; training the multivariate analysis model based on the fusion vector; Based on the first analysis result and the second analysis result, a final result of performing an outlier value distribution law analysis on the abnormal indicator data is determined.

10. The method for analyzing corn quality distribution law according to claim 9, characterized in that: The association relationship of the graph data nodes is established based on the characteristic dimension of the plane characteristics of each graph data node to generate a data system, including: Determine the phase space reconstruction results of the graph data nodes ; ； in, It is the first graph data node; is the nth graph data node; n is the number of graph data nodes; is the preset time delay, the value is (0,1); Based on the phase space reconstruction results of the graph data nodes, the characteristic dimension of the plane feature of each graph data node is used as the basis for phase space reconstruction, and the one-dimensional time series is expanded in the multi-dimensional phase space. ; ； in, for The i-th characteristic dimension of is also the phase point; Determine the phase point in the reconstructed m-dimensional phase space To other points The number of points whose distance does not exceed r ; ； in, () is the Heaviside function; r is the preset distance information, which is used to indicate the threshold of the distance of the association relationship; Determine the chaotic time series correlation dimension feature extraction function of graph data nodes based on point information ; ； Based on the chaotic time series correlation dimension feature extraction function, the correlation relationship of graph data nodes is determined to generate a data system.

Images