A data intelligent analysis method and system for low-temperature freeze injury multi-factor modeling

Abstract

The application provides a data intelligent analysis method and system for freeze injury multi-factor modeling at low temperature, and relates to the technical field of data processing. The method comprises the following steps: analyzing a growth index set of a first target crop to determine a key index set; clustering the growth index set with the key index set clustering center to output an index clustering result; evaluating the index clustering result to output a plurality of factor strengths based on the plurality of clustering results; building a freeze injury growth index model according to the plurality of factor strengths; inputting predicted low-temperature environment information into the freeze injury growth index model for evaluation to output a low-temperature freeze injury evaluation result. The application solves the problem that a single analysis system in the prior art cannot identify the influence of citrus freeze injury, and the difference in fruit tree varieties reduces the accuracy of the analysis result, so that the freeze injury characteristics of different factors under different low-temperature environments can be analyzed, and an intelligent index system evaluation is realized.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a data intelligent analysis method and system for multi-factor modeling of freezing damage at low temperatures. Background Technology

[0002] Frost damage has become a bottleneck threatening the high-quality development of the fruit industry. Citrus fruits have excellent edible and economic value and are favored by many growers. To improve the quality and yield of citrus fruits, it is necessary to improve the efficiency of fruit tree cultivation and the level of frost damage prevention. Due to the impact of climate change and the sensitivity of citrus fruits to low temperatures, frost damage to citrus fruits will lead to reduced yields and affect the economic value of the growing areas. In order to prevent frost damage to fruit trees, it is necessary to pay attention to frost damage prevention of citrus fruits and minimize the losses to the fruit industry. Therefore, how to establish an accurate frost damage analysis system is an urgent problem that needs to be solved.

[0003] Since there are many factors that affect citrus growth after frost damage, existing single analysis systems cannot specifically identify the impact of frost damage on citrus, and the differences between fruit tree varieties further reduce the accuracy of the analysis results. Summary of the Invention

[0004] This application provides a data intelligent analysis method and system for multi-factor modeling of frost damage at low temperatures, which addresses the problem that the single analysis system in the existing technology cannot specifically identify the impact of frost damage on citrus, and that the differences between fruit tree varieties further reduce the accuracy of the analysis results.

[0005] In view of the above problems, this application provides a data intelligent analysis method and system for multi-factor modeling of freezing damage at low temperatures.

[0006] A first aspect of this application provides a data intelligent analysis method for multi-factor modeling of freezing damage under low temperature conditions. The method includes: acquiring a set of growth indicators for a first target crop, wherein the set of growth indicators consists of growth indicators that change under freezing damage conditions; analyzing the set of growth indicators to determine a set of key indicators, wherein the number of key indicators is at least two; clustering the set of growth indicators using the cluster centers of the key indicator sets, and outputting the clustering results, wherein the clustering results include multiple clustering results, and each clustering result includes at least one growth indicator; evaluating the clustering results and outputting multiple factor intensities based on the multiple clustering results, wherein the factor intensities characterize the low temperature sensitivity intensity of the indicator sets included in each clustering result; constructing a freezing damage growth indicator model based on the multiple factor intensities; inputting predicted low temperature environment information into the freezing damage growth indicator model for evaluation, and outputting a low temperature freezing damage evaluation result.

[0007] In a second aspect, this application also provides a data intelligent analysis system for multi-factor modeling of freezing damage under low temperature conditions. The system includes: a growth index acquisition module, used to acquire a set of growth indicators for a first target crop, wherein the set of growth indicators comprises growth indicators that change under freezing damage conditions; a key indicator analysis module, used to analyze the set of growth indicators to determine a set of key indicators, wherein the set of key indicators contains at least two indicators; and a growth index clustering module, used to cluster the set of growth indicators using the cluster centers of the key indicator sets. The system comprises: a clustering module, which outputs clustering results for the indicators, wherein the clustering results include multiple clustering results, each of which includes at least one growth indicator; a sensitivity identification module, which evaluates the clustering results and outputs multiple factor intensities based on the multiple clustering results, wherein the factor intensities characterize the low-temperature sensitivity intensity of the set of indicators included in each clustering result; a model building module, which builds a frost damage growth indicator model based on the multiple factor intensities; and a frost damage assessment module, which inputs predicted low-temperature environmental information into the frost damage growth indicator model for assessment and outputs a low-temperature frost damage assessment result.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] This application embodiment collects data on the growth process of the target crop planted in the orchard to determine all indicators affected by frost damage. Key indicators are extracted from these affected indicators, and then clustered around these key indicators as cluster centers. Low-temperature sensitivity analysis is performed on each cluster, and the low-temperature sensitivity intensity of each cluster is determined. A frost damage growth index model is then generated based on this intensity. When a preset frost damage temperature is input, the low-temperature frost damage assessment result is output based on the low-temperature sensitivity intensity of each cluster in the frost damage growth index model. This addresses the limitations of existing single-analysis systems that cannot specifically identify the impact of frost damage on citrus trees, and the fact that differences between fruit tree varieties further reduce the accuracy of the analysis results. By clustering all indicators and training the assessment model through low-temperature sensitivity analysis, a targeted analysis of fruit tree growth indicators under different low-temperature conditions can be performed, resulting in an intelligent system assessment of the impact of frost damage under different low-temperature environments.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0012] Figure 1 A flowchart illustrating a data intelligent analysis method for multi-factor modeling of freezing damage at low temperatures, provided in an embodiment of this application;

[0013] Figure 2 A flowchart illustrating the identification of key indicators in a data intelligent analysis method for multi-factor modeling of freezing damage at low temperatures, provided in an embodiment of this application;

[0014] Figure 3 A flowchart illustrating the growth index clustering process in a data intelligent analysis method for multi-factor modeling of freezing damage at low temperatures, provided in an embodiment of this application.

[0015] Figure 4 A schematic diagram of the structure of a data intelligent analysis system for multi-factor modeling of freezing damage at low temperatures, provided in an embodiment of this application;

[0016] Figure labeling: Growth index acquisition module 11, key index analysis module 12, growth index clustering module 13, sensitivity identification module 14, model building module 15, frost damage assessment module 16. Detailed Implementation

[0017] This application provides a data intelligent analysis method and system for multi-factor modeling of frost damage at low temperatures, which solves the problem that the single analysis system in the existing technology cannot identify the impact of frost damage on citrus in a targeted manner, and the differences between fruit tree varieties further reduce the accuracy of the analysis results, making it impossible to effectively control the impact of frost damage on fruit trees.

[0018] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0019] Example 1

[0020] like Figure 1As shown in the embodiments of this application, a data intelligent analysis method for multi-factor modeling of freezing damage at low temperatures is provided, the method comprising:

[0021] Step S100: Obtain a set of growth indicators for the first target crop, wherein the set of growth indicators consists of growth indicators of the first target crop that change under frost damage conditions.

[0022] Currently, citrus is widely cultivated in many regions. In recent years, severe climate anomalies have led to varying degrees of frost damage in orchards across the country. Generally, frost damage to citrus trees can cause yellowing of leaves, stunted growth, and even complete death of some orchards, resulting in reduced fruit yield. Frost damage generally refers to the damage caused to crops by temperatures below 0°C. Different types of fruit trees have different cold resistance, thus the indicators of frost damage to fruit trees vary significantly. It is necessary to develop corresponding evaluation indicators for different types of disasters and different fruit tree varieties. This application provides a data intelligent analysis method for multi-factor modeling of frost damage under low temperature conditions. This method solves the problem that the single analysis system in the existing technology cannot specifically identify the impact of frost damage on citrus, and the differences between fruit tree varieties further reduce the accuracy of the analysis results. It achieves targeted analysis of the growth indicators of fruit trees under different low temperature conditions, thereby outputting intelligent system evaluation of the impact of frost damage under different low temperature environments.

[0023] In this embodiment, the first target crop is a fruit tree planted in a regional environment, which can be any citrus crop. The set of growth indicators is collected based on artificial environmental control experiments and experiments on crop growth status and physiological changes. This reflects the growth data of the first target crop under frost damage conditions, such as the cell membrane permeability, photosynthetic characteristics, enzyme metabolism system, antioxidant capacity, endogenous hormones, and low-temperature induced proteins corresponding to different growth stages (flowering period, fruit setting period, fruit enlargement period, and ripening period). These growth indicators are then combined to further analyze the effects of frost damage on growth delay, stagnation, etc.

[0024] Step S200: Analyze the set of growth indicators to determine the set of key indicators, wherein the set of key indicators has at least two indicators;

[0025] Specifically, since there is a large amount of data measured in the set of growth indicators, a set of key indicators is extracted from the set of growth indicators based on the degree of influence of each indicator on growth and development. The number of key indicators can be 2 or any integer number greater than or equal to 2, and the number of key indicators is much smaller than the number of growth indicators. For example, leaf area index, light energy utilization efficiency, fruit shape index, and quality indicators.

[0026] Furthermore, such as Figure 2 As shown, the set of growth indicators is analyzed to determine the set of key indicators. In this embodiment, step S200 further includes:

[0027] Step S210: By analyzing each growth index in the set of growth indicators, output the set of index influence degrees, wherein the set of index influence degrees represents the degree of influence of each growth index on the growth quality of the first target crop.

[0028] Step S220: Based on the set of indicator influence degrees, determine multiple indicator influence degrees that are greater than or equal to the preset indicator influence degree;

[0029] Step S230: Identify the multiple growth indicators corresponding to the influence of the multiple indicators, and output the identified growth indicator set;

[0030] Step S240: Output the key indicator set based on the identified growth indicator set.

[0031] Furthermore, in the embodiment of this application, step S200 further includes outputting the set of key indicators:

[0032] Step S250: Perform coupling analysis on the set of identifier growth indicators and output multiple coupling indicators;

[0033] Step S260: Based on the multiple coupling indices, merge adjacent indices with a coupling greater than or equal to a preset coupling to obtain an index merging result, wherein the index merging result includes a merged result and a non-merged result;

[0034] Step S270: Output the set of key indicators according to the merged results of the indicators.

[0035] Specifically, in the above steps, a series of growth indicators are measured to reflect the relationship between various indicators of the first target crop and the response mechanism of frost damage. The influence degree of each growth indicator on the growth quality of the first target crop is then output as an indicator influence degree set. Each growth indicator has a corresponding influence degree parameter. The indicator influence degree set is judged by a pre-set indicator influence degree, and multiple indicator influence degrees that are greater than or equal to the pre-set indicator influence degree are extracted. The growth indicators corresponding to the extracted multiple indicator influence degrees are then output as key indicators.

[0036] The preset index influence can be determined by analyzing the relationship between the growth status of the first target crop and various key growth indicators, thereby determining the threshold that will reduce the yield of the first target crop, and then setting the preset index influence based on the threshold.

[0037] Since the extracted set of key indicators is used as cluster centers, in order to ensure the accuracy of clustering, the key indicator set is required to have low coupling. When the coupling between key growth indicators is high, the targeted analysis utility of the indicators is reduced. In order to ensure the independence of cluster centers, coupling analysis is performed on the key growth indicator set. In the coupling analysis process, the key growth indicator sets are paired up for coupling analysis, thereby obtaining multiple coupling indicators C, where each coupling indicator C represents the promotion / inhibition relationship between two growth indicators.

[0038] For example, the coupling index C can be measured by changing one of the growth indices to determine the changes in the remaining growth indices. The coupling can be determined based on the changes. For instance, the closer C is to 1, the greater the coupling between the indices; when C is closer to 0, the minimum coupling is indicated.

[0039] After obtaining the coupling indicators corresponding to each key growth indicator, if the coupling of an indicator is greater than the preset coupling, the two corresponding key growth indicators need to be merged. That is, the two key growth indicators are treated as the same key growth indicator class. The merged result is the merged indicator, and the unmerged result is the initial key indicator with coupling less than the preset coupling. The optimized set of key indicators is output. It should be understood that the number of key indicators in the optimized set can be the same as the number of key indicators in the unoptimized set, or the number of key indicators in the optimized set can be less than the number of key indicators in the unoptimized set. This achieves the goal of increasing the relevance of the indicator model by performing coupling analysis on the set of key growth indicators and optimizing the indicator set based on the coupling analysis results.

[0040] Step S300: Cluster the growth index set using the cluster center of the key index set, and output the index clustering result, wherein the index clustering result includes multiple clustering results, and each clustering result includes at least one growth index;

[0041] To reduce the dimensionality of all growth indicator sets, indicator clustering can be used to preserve the characteristics of all growth indicators. This requires clustering the growth indicator sets using the key indicator set as cluster centers and outputting the indicator clustering results. In other words, when clustering the key indicator set, the key indicator set is used as a pre-set cluster center to cluster the remaining indicator sets in all growth indicator sets. For example, the clustering method can be Gaussian Mixture Model (GMM) Expectation-Maximum (EM) clustering, which can accurately cluster complex cases and ensure that the output indicator clustering results have high accuracy.

[0042] Furthermore, such as Figure 3As shown, step S300 in this embodiment further includes:

[0043] Step S310: Initialize the Gaussian distribution parameters of the key indicator set, wherein the Gaussian distribution parameters include variance and mean;

[0044] Step S320: Based on the Gaussian distribution parameters, calculate the probability that each indicator in the remaining growth indicator set belongs to the key indicator set, classify the indicators according to their probability, and so on, and output the indicator clustering result.

[0045] Specifically, clustering using a Gaussian mixture model first assumes that the data points are Gaussian distributed, using two parameters to describe the shape of the clusters: mean and standard deviation. Clustering should begin by finding the mean and standard deviation of the dataset, and then employing an optimization algorithm called Expectation-Maximization (EM) to perform the Expectation-Maximization clustering process.

[0046] Step S321: Initialize the Gaussian distribution parameters of the key indicator set, i.e., set the initial variance and mean; Step S322: Obtain the Gaussian distribution parameters with the maximum probability based on the probability of each indicator in the remaining growth indicator set, wherein the remaining growth indicators are the remaining indicator set excluding the key indicator set in the growth indicator set; Step S323: Repeatedly iterate the probability based on the Gaussian distribution parameters with the maximum probability until the change of the Gaussian distribution parameters is in a convergent state, stop the iteration, and output the indicator clustering result.

[0047] In detail, besides selecting the number of clusters (similar to K-Means) and randomly initializing the Gaussian distribution parameters (mean and variance) for each cluster, we can also first observe the indicators to obtain a relatively accurate mean and variance, and then, given the Gaussian distribution for each cluster, calculate the probability that each indicator belongs to each cluster. The closer an indicator is to the center of the Gaussian distribution, the more likely it is to belong to that cluster. Based on these probabilities, we calculate the Gaussian distribution parameters to maximize the probability of the indicator. For example, we can use a weighted average of the probabilities of data points to calculate these parameters. We then iterate through steps S322 and S323 until the changes during iteration are minimal, thereby improving the accuracy of the clustering results and refining the indicator system.

[0048] Step S400: Evaluate the index clustering results and output multiple factor intensities based on the multiple clustering results, wherein the factor intensities characterize the low-temperature sensitivity intensity of the index set contained in each clustering result;

[0049] Furthermore, when evaluating the clustering results of the indicators, sensitivity analysis is performed on each growth indicator in each clustering result to obtain the frost damage sensitivity index corresponding to each clustering result. Based on the frost damage sensitivity index, the factor intensity of each clustering result is output to obtain multiple factor intensities based on the multiple clustering results.

[0050] When identifying factor strengths, the first step is to determine the indicators included in each clustering result, and then perform a comprehensive calculation on each clustering result. For example, the indicator set P = {x1, x2, ... x} in a clustering result is... n}, where n is the total number of clusters, x i Characterize specific growth indicators to obtain the low-temperature sensitivity of each growth indicator, and output the set P = {x1, x2, ... x...} n The corresponding set of low-temperature susceptibility is Q = {y1, y2, ... y} n This process calculates the factor strength of each data point in set Q, outputting a clustering result. This process is repeated to obtain multiple factor strengths corresponding to the clustering results of the indicators. Low-temperature sensitivity is established by collecting data on the changes in growth parameters of each growth indicator under low-temperature conditions, including the time of change under low-temperature conditions and the magnitude of data changes in the parameters themselves due to low temperatures. This establishes the sensitivity of each clustering result to low-temperature conditions, providing a better identification basis for the next stage of evaluation.

[0051] Step S500: Based on the intensity of multiple factors, construct a frost damage growth index model;

[0052] Step S600: Input the predicted low temperature environment information into the freezing damage growth index model for evaluation, and output the low temperature freezing damage evaluation result.

[0053] Furthermore, step S600 in this embodiment of the application also includes:

[0054] Step S610: Collect historical low-temperature environment datasets for the first target crop;

[0055] Step S620: Input the historical low temperature environment dataset into the environmental simulation module for simulation, and output the low temperature environment simulation model;

[0056] Step S630: Connect the low-temperature environment simulation model with the frost damage growth index model, and output the predicted low-temperature environment information based on the low-temperature environment simulation model;

[0057] Step S640: Input the predicted low temperature environment information into the frost damage growth index model for analysis.

[0058] Specifically, by studying the growth and development indicators of the first target crop, the relationship between each cluster result and the low-temperature climate can be established. Based on the factor strength corresponding to each cluster result, the frost damage sensitivity relationship of each cluster result can be determined. Based on this, a frost damage growth index model can be built. The frost damage growth index model can assess the growth status of the first target crop based on the low-temperature environment information to be predicted, and output the low-temperature frost damage assessment result. The low-temperature frost damage assessment result can output the impact of each cluster result on the growth of the first target crop. Since the factor strength of each cluster result is different and the sensitivity to different low-temperature environments is different, the degree of impact based on each cluster result can be output, thereby obtaining the low-temperature frost damage assessment result.

[0059] The frost damage growth index model is a neural network trained on a fully connected three-layer network, including a sensitivity factor mapping layer, a weight training layer, and an evaluation output layer. The neural network includes multiple sets of training data, which include clustering results, factor strengths corresponding to each clustering result, and indicators of the degree of growth impact. In the low-temperature frost damage evaluation results output in this embodiment, the model is mainly used to output the degree of decline in growth performance, so that anti-frost measures can be taken in a timely manner to intervene in the growth of the orchard and ensure yield.

[0060] The low-temperature freezing damage assessment results can be used to train a weighted network layer based on the factor strength of the clustering results. The weighted network layer has a low-temperature-sensitivity mapping relationship. Since the factor strength of the sensitivity is different in each clustering result under different low-temperature environments, the training of the weighted network layer is automatically identified and converged when different predicted low-temperature environment information is input. This is used to reduce the dimensionality of all growth indicators based on different clustering results, which not only preserves the feature strength but also reduces the feature quantity, reduces the training load of the model, and improves the usability of the low-temperature freezing damage assessment results.

[0061] Furthermore, historical low-temperature environment datasets can be obtained by collecting climate data from the growing area of ​​the first target crop, including daily maximum temperature, daily minimum temperature, and hourly air temperature within the orchard. Then, an environmental simulation module can be used to simulate these daily maximum temperature, daily minimum temperature, and hourly air temperature within the orchard, constructing a low-temperature environment simulation model. This model can be used to forecast meteorological conditions within the orchard. Based on the low-temperature environment simulation model, forecasted low-temperature environment information can be output. Additionally, if the orchard currently has a climate forecasting system, the corresponding forecasted low-temperature environment information can be obtained by connecting to the existing system. The forecasted low-temperature environment information is then input into the frost damage growth index model for analysis, outputting a low-temperature frost damage assessment result.

[0062] In summary, the data intelligent analysis method for multi-factor modeling of freezing damage at low temperatures provided in this application has the following technical effects:

[0063] 1. This application embodiment retains the characteristics of all growth indicators by adopting an indicator clustering method. It is necessary to cluster the growth indicator set with the cluster center of the key indicator set, output the indicator clustering results, and further analyze the indicator clustering results before outputting the final frost damage assessment results. By clustering all indicators and training the assessment model by analyzing low temperature sensitivity, it is possible to perform targeted analysis on fruit tree growth indicators under different low temperature conditions, thereby outputting frost damage impact indicators for different low temperature environments and realizing intelligent system assessment.

[0064] 2. In this embodiment of the application, key indicators are extracted from all growth indicators. When the coupling between key growth indicators is large, the effectiveness of targeted analysis of the indicators is reduced. In order to ensure the accuracy after clustering, the key indicator set is required to have small coupling. By performing coupling analysis on the key indicators, the effect and accuracy of indicator clustering are improved.

[0065] Example 2

[0066] Based on the data intelligent analysis method for multi-factor modeling of freezing damage at low temperatures described in the foregoing embodiments, and using the same inventive concept, this invention also provides a data intelligent analysis system for multi-factor modeling of freezing damage at low temperatures, such as... Figure 4 As shown, the system includes:

[0067] The growth index acquisition module 11 is used to acquire a set of growth indicators of the first target crop, wherein the set of growth indicators is the growth indicators of the first target crop that change under frost damage conditions.

[0068] Key indicator analysis module 12, the key indicator analysis module 12 is used to analyze the growth indicator set and determine the key indicator set, wherein the number of the key indicator set is at least two indicators;

[0069] The growth index clustering module 13 is used to cluster the growth index set with the cluster center of the key index set and output the index clustering result. The index clustering result includes multiple clustering results, and each clustering result includes at least one growth index.

[0070] Sensitivity identification module 14 is used to evaluate the index clustering results and output multiple factor intensities based on the multiple clustering results, wherein the factor intensities characterize the low temperature sensitivity intensity of the index set contained in each clustering result.

[0071] Model building module 15, which is used to build a frost damage growth index model based on the intensity of multiple factors;

[0072] The freezing damage assessment module 16 is used to input predicted low temperature environment information into the freezing damage growth index model for assessment and output low temperature freezing damage assessment results.

[0073] Furthermore, the key indicator analysis module 12 is also used to implement the following functions:

[0074] By analyzing each growth index in the set of growth indicators, an index influence set is output, wherein the index influence set represents the degree of influence of each growth index on the growth quality of the first target crop.

[0075] Based on the set of indicator influence degrees, determine the influence degrees of multiple indicators that are greater than or equal to the preset indicator influence degree;

[0076] Identify the multiple growth indicators corresponding to the influence of the multiple indicators, and output the identified growth indicator set;

[0077] Based on the set of growth indicators, output the set of key indicators.

[0078] Furthermore, the key indicator analysis module 12 is also used to implement the following functions:

[0079] A coupling analysis is performed on the aforementioned set of growth indicators to output multiple coupling indicators;

[0080] Based on the multiple coupling indices, adjacent indices with a coupling greater than or equal to a preset coupling are merged to obtain an index merging result, wherein the index merging result includes merged results and non-merged results;

[0081] Based on the merged results of the aforementioned indicators, the set of key indicators is output.

[0082] Furthermore, the growth index clustering module 13 is also used to implement the following functions:

[0083] Initialize the Gaussian distribution parameters of the key indicator set, wherein the Gaussian distribution parameters include variance and mean;

[0084] Based on the Gaussian distribution parameters, the probability of each indicator in the remaining growth indicator set belonging to the key indicator set is calculated. Based on the probability of each indicator in the remaining growth indicator set, they are classified, and so on, to output the indicator clustering results.

[0085] The remaining growth indicators are the set of remaining indicators in the set of growth indicators excluding the set of key indicators.

[0086] Furthermore, the growth index clustering module 13 is also used to implement the following functions:

[0087] Based on the probabilities of each indicator in the remaining set of growth indicators, obtain the Gaussian distribution parameters that maximize the probability; and

[0088] Repeatedly iterate based on the Gaussian distribution parameters that are at the maximum probability until the changes in the Gaussian distribution parameters converge, then stop iterating and output the index clustering results.

[0089] Furthermore, the frost damage assessment module 16 is also used to perform the following functions:

[0090] Collect historical low-temperature environment datasets for the first target crop;

[0091] The historical low-temperature environment dataset is input into the environmental simulation module for simulation, and a low-temperature environment simulation model is output.

[0092] Connect the low-temperature environment simulation model with the frost damage growth index model, and output the predicted low-temperature environment information based on the low-temperature environment simulation model;

[0093] The predicted low-temperature environment information is input into the frost damage growth index model for analysis.

[0094] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0095] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A data intelligent analysis method for multi-factor modeling of freezing damage at low temperatures, characterized in that, The method includes: Obtain a set of growth indicators for a first target crop, wherein the set of growth indicators consists of growth indicators that change under frost damage conditions. The set of growth indicators is analyzed to determine the set of key indicators, wherein the set of key indicators has at least two indicators. The growth index set is clustered using the cluster center of the key index set, and the index clustering result is output. The index clustering result includes multiple clustering results, and each clustering result includes at least one growth index. The clustering results of the indicators are evaluated, and multiple factor intensities based on the multiple clustering results are output, wherein the factor intensities characterize the low-temperature sensitivity intensity of the set of indicators contained in each clustering result; Based on the strength of multiple factors, a growth index model for frost damage was constructed. The predicted low-temperature environment information is input into the freezing damage growth index model for evaluation, and the low-temperature freezing damage evaluation result is output. The method further includes: Initialize the Gaussian distribution parameters of the key indicator set, wherein the Gaussian distribution parameters include variance and mean; Based on the Gaussian distribution parameters, the probability of each indicator in the remaining growth indicator set belonging to the key indicator set is calculated. Based on the probability of each indicator in the remaining growth indicator set, they are classified, and so on, to output the indicator clustering results. Wherein, the remaining growth indicators are the set of remaining indicators in the set of growth indicators excluding the set of key indicators; Based on the probabilities of each indicator in the remaining set of growth indicators, obtain the Gaussian distribution parameters that maximize the probability; and Repeatedly iterate based on the Gaussian distribution parameters that are at the maximum probability until the changes in the Gaussian distribution parameters are in a convergent state, then stop iterating and output the index clustering results. The method further includes: Collect historical low-temperature environment datasets for the first target crop; The historical low-temperature environment dataset is input into the environmental simulation module for simulation, and a low-temperature environment simulation model is output. Connect the low-temperature environment simulation model with the frost damage growth index model, and output the predicted low-temperature environment information based on the low-temperature environment simulation model; The predicted low-temperature environment information is input into the frost damage growth index model for analysis.

2. The method as described in claim 1, characterized in that, The set of growth indicators is analyzed to determine the set of key indicators. The methods include: By analyzing each growth index in the set of growth indicators, an index influence set is output, wherein the index influence set represents the degree of influence of each growth index on the growth quality of the first target crop. Based on the set of indicator influence degrees, determine the influence degrees of multiple indicators that are greater than or equal to the preset indicator influence degree; Identify the multiple growth indicators corresponding to the influence of the multiple indicators, and output the identified growth indicator set; Based on the set of growth indicators, output the set of key indicators.

3. The method as described in claim 2, characterized in that, The method for outputting the set of key indicators also includes: A coupling analysis is performed on the aforementioned set of growth indicators to output multiple coupling indicators; Based on the multiple coupling indices, adjacent indices with a coupling greater than or equal to a preset coupling are merged to obtain an index merging result, wherein the index merging result includes merged results and non-merged results; Based on the merged results of the aforementioned indicators, the set of key indicators is output.

4. The method as described in claim 1, characterized in that, The method includes evaluating the clustering results of the indicators and outputting the strengths of multiple factors based on the multiple clustering results, comprising: Sensitivity analysis was performed on each growth index in each cluster of the index clustering results to obtain the frost damage sensitivity index corresponding to each cluster. The factor strength of each clustering result is output based on the frost damage sensitivity index, and multiple factor strengths based on the multiple clustering results are obtained.

5. A data intelligent analysis system for multi-factor modeling of freezing damage at low temperatures, applied to the method as described in any one of claims 1 to 4, characterized in that, The system includes: A growth index acquisition module is used to acquire a set of growth indicators for a first target crop, wherein the set of growth indicators consists of growth indicators that change under frost damage conditions. A key indicator analysis module is used to analyze the set of growth indicators and determine the set of key indicators, wherein the set of key indicators has at least two indicators. A growth index clustering module is used to cluster the growth index set with the cluster center of the key index set and output the index clustering results. The index clustering results include multiple clustering results, and each clustering result includes at least one growth index. A sensitivity identification module is used to evaluate the index clustering results and output multiple factor intensities based on the multiple clustering results, wherein the factor intensities characterize the low-temperature sensitivity intensity of the index set contained in each clustering result. A model building module is used to build a frost damage growth index model based on the intensity of multiple factors. The freezing damage assessment module is used to input predicted low-temperature environment information into the freezing damage growth index model for assessment and output low-temperature freezing damage assessment results.

Images