Method and system for identifying high temperature and strong light disasters causing sunburn of citrus fruits

By combining the historical meteorological record data of the production site of citrus fruits, selecting experimental control parameters, configuring environmental control experiments, generating high-temperature and strong light disaster identification branches, and building an identification network, it solves the low reliability of high-temperature and strong light disaster identification of citrus fruits in the existing technology. and the problem of large error in the identification result, efficient and accurate disaster identification is achieved.

CN118194050BActive Publication Date: 2024-12-06INST OF HORTICULTURE JIANGXI ACAD OF AGRI SCI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410030036.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-09
Publication Date
2024-12-06
Estimated Expiration
2044-01-09

AI Technical Summary

Technical Problem

In the prior art, the recognition of high-temperature and strong light disasters of citrus fruits is low, and the identification network does not fit the actual situation, resulting in large errors in the identification results.

Method used

By obtaining the experimental control parameter interval of the target citrus fruit, combining the historical meteorological record data of the production site to select the control parameter set, configuring environmental control experiments, collecting sample characterization data, generating sensitive feature sets, constructing high-temperature strong light disaster identification branches, and building high-temperature strong light Disaster identification network and match analysis and identification are performed based on the meteorological parameter set.

Benefits of technology

The accuracy of high-temperature and strong light disaster recognition of citrus fruits is improved, and a refined configuration experiment is realized based on the actual situation of the production site, and an efficient and accurate identification network is built, which reduces identification errors.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118194050B_ABST
    Figure CN118194050B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for identifying high temperature and strong light disasters that cause sunburn in citrus fruits, and relates to the technical field of citrus disaster identification. The method includes: obtaining the experimental control parameter interval of the target citrus fruit, performing experimental control parameter optimization, and obtaining N control parameter sets; configuring N groups of environmental control experiments; performing data sensitivity granularity screening, and generating a target sensitive feature set; generating N high temperature and strong light disaster identification branches; constructing a high temperature and strong light disaster identification network; collecting the meteorological parameter set of the target citrus fruit, and performing matching analysis; obtaining a target characterization data set; activating the high temperature and strong light disaster identification branch according to the matching analysis result, and obtaining the target disaster identification result. The present invention solves the technical problems in the prior art that the recognition reliability of high temperature and strong light disasters in citrus fruits is low, the degree of fit between the constructed recognition network and the actual situation of citrus fruits is low, and the recognition result error is large, and achieves the technical effect of improving the accuracy of high temperature and strong light disaster identification.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of citrus disaster identification, and in particular to a method and system for identifying high temperature and strong light disasters that cause sunburn of citrus fruits. Background Art

[0002] The quality of citrus fruit is greatly affected by high temperature and strong light weather. The current methods for identifying high temperature and strong light disasters caused by sunburn on fruits mainly rely on empirical parameters, which are not very refined and cannot provide reliable and accurate disaster identification for fruit farmers. The existing technology has the technical problems of low reliability in identifying high temperature and strong light disasters on citrus fruit, low fit between the constructed identification network and the actual situation of citrus fruit, and large errors in the identification results. Summary of the invention

[0003] The present application provides a method and system for identifying high temperature and strong light disasters that cause sunburn in citrus fruits, which is used to solve the technical problems in the prior art of low reliability in identifying high temperature and strong light disasters in citrus fruits, low fit between the constructed identification network and the actual situation of citrus fruits, and large errors in the identification results.

[0004] In view of the above problems, the present application provides a method and system for identifying high temperature and strong light disasters that cause sunburn on citrus fruits.

[0005] In a first aspect of the present application, a method for identifying high temperature and strong light disasters causing sunburn of citrus fruits is provided, the method comprising:

[0006] Obtaining the experimental control parameter interval of the target citrus fruit, optimizing the experimental control parameters in combination with the historical meteorological record data of the production site of the target citrus fruit, and obtaining N control parameter sets, wherein the N control parameter sets have N parameter adaptation factors;

[0007] Based on N parameter adaptation factors, match the number of N experimental samples and configure N sets of environmental control experiments in combination with N control parameter sets;

[0008] Collect N sample characterization data sets of N groups of environmental control experiments, perform data sensitivity granularity screening based on the N sample characterization data sets, and generate a target sensitive feature set;

[0009] Traversing and collecting N sunburn indexes of the N groups of environmental control experiments, combining N control parameter sets of the N groups of environmental control experiments and N target sensitive feature parameter sets to perform decision analysis, and generate N high temperature and strong light disaster identification branches;

[0010] A high temperature and strong light disaster identification network is constructed based on N high temperature and strong light disaster identification branches;

[0011] Collecting a meteorological parameter set of the target citrus fruit, performing matching analysis with the N control parameter sets, and obtaining a matching analysis result;

[0012] Collecting characteristic data of the target citrus fruit according to the target sensitive characteristic set to obtain a target representation data set;

[0013] The high temperature and strong light disaster identification branch of the high temperature and strong light disaster identification network is activated according to the matching analysis results, and the target characterization data set is sent to the activated high temperature and strong light disaster identification branch for disaster identification, and the identification results are comprehensively analyzed to obtain the target disaster identification results.

[0014] In a second aspect of the present application, a system for identifying high temperature and strong light disasters that cause sunburn in citrus fruits is provided, the system comprising:

[0015] A first obtaining unit is used to obtain an experimental control parameter interval of a target citrus fruit, and to optimize the experimental control parameters in combination with historical meteorological record data of a production site of the target citrus fruit, to obtain N control parameter sets, wherein the N control parameter sets have N parameter adaptation factors;

[0016] A first configuration unit is used to match the number of N experimental samples based on N parameter adaptation factors and configure N groups of environmental control experiments in combination with N control parameter sets;

[0017] A first generating unit is used for generating a target sensitive feature set by collecting N sample representation data sets of N groups of environmental control experiments, performing data sensitivity granularity screening based on the N sample representation data sets, and generating a target sensitive feature set;

[0018] The second generating unit is used to traverse and collect the N sunburn indexes of the N groups of environmental control experiments, combine the N control parameter sets of the N groups of environmental control experiments and the N target sensitive feature parameter sets to perform decision analysis, and generate N high temperature and strong light disaster identification branches;

[0019] A first construction unit is used for first identifying and constructing a high temperature and strong light disaster identification network according to N high temperature and strong light disaster identification branches;

[0020] A second acquisition unit is used to collect the meteorological parameter set of the target citrus fruit, perform matching analysis with the N control parameter sets, and obtain a matching analysis result;

[0021] A third obtaining unit is used to collect the characteristic data of the target citrus fruit according to the target sensitive characteristic set to obtain a target representation data set;

[0022] The fourth acquisition unit is used to activate the high temperature and strong light disaster identification branch of the high temperature and strong light disaster identification network according to the matching analysis results, and send the target characterization data set to the activated high temperature and strong light disaster identification branch for disaster identification, and comprehensively analyze the identification results to obtain the target disaster identification results.

[0023] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0024] The present application obtains the experimental control parameter interval of the target citrus fruit, and optimizes the experimental control parameters in combination with the historical meteorological record data of the production site of the target citrus fruit to obtain N control parameter sets, wherein the N control parameter sets have N parameter adaptation factors, and then matches the number of N experimental samples based on the N parameter adaptation factors, configures N groups of environmental control experiments in combination with the N control parameter sets, and then collects N sample representation data sets of the N groups of environmental control experiments, performs data sensitivity granularity screening based on the N sample representation data sets, generates a target sensitive feature set, and then traverses and collects N sunburn indexes of the N groups of environmental control experiments, and combines the N control parameter sets of the N groups of environmental control experiments. The decision analysis is performed with N target sensitive characteristic parameter sets to generate N high temperature and strong light disaster identification branches, and a high temperature and strong light disaster identification network is constructed based on the N high temperature and strong light disaster identification branches. Then, the meteorological parameter set of the target citrus fruit is collected, and matching analysis is performed with the N control parameter sets to obtain matching analysis results. Then, the characteristic data of the target citrus fruit is collected based on the target sensitive characteristic set to obtain the target characterization data set. Then, the high temperature and strong light disaster identification branch of the high temperature and strong light disaster identification network is activated based on the matching analysis results, and the target characterization data set is sent to the activated high temperature and strong light disaster identification branch for disaster identification, and the identification results are comprehensively analyzed to obtain the target disaster identification results. The configuration experiment is refined according to the actual situation of the production area of ​​citrus fruit, and an efficient and accurate high temperature and strong light disaster identification network is constructed based on the experimental results to improve the accuracy of disaster identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 A schematic flow chart of a method for identifying high temperature and strong light disasters that cause sunburn on citrus fruits provided in an embodiment of the present application;

[0027] Figure 2 A schematic diagram of a flow chart of obtaining a set of adjustment parameter adaptation factors in a method for identifying high temperature and strong light disasters causing sunburn of citrus fruits provided in an embodiment of the present application;

[0028] Figure 3 A schematic diagram of a process for obtaining N control parameter sets in a method for identifying high temperature and strong light disasters causing sunburn of citrus fruits provided in an embodiment of the present application;

[0029] Figure 4 A schematic diagram of the structure of a system for identifying high temperature and strong light disasters that cause sunburn on citrus fruits provided in an embodiment of the present application.

[0030] Explanation of reference numerals: first obtaining unit 11 , first configuring unit 12 , first generating unit 13 , second generating unit 14 , first constructing unit 15 , second obtaining unit 16 , third obtaining unit 17 , fourth obtaining unit 18 . DETAILED DESCRIPTION

[0031] The present application provides a method and system for identifying high temperature and strong light disasters that cause sunburn in citrus fruits, aiming to solve the technical problems in the prior art of low reliability in identifying high temperature and strong light disasters in citrus fruits, low fit between the constructed identification network and the actual situation of citrus fruits, and large errors in the identification results.

[0032] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0033] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0034] Embodiment 1

[0035] like Figure 1 As shown, the present application provides a method for identifying high temperature and strong light disasters that cause sunburn in citrus fruits, wherein the method comprises:

[0036] Obtaining the experimental control parameter interval of the target citrus fruit, optimizing the experimental control parameters in combination with the historical meteorological record data of the production site of the target citrus fruit, and obtaining N control parameter sets, wherein the N control parameter sets have N parameter adaptation factors;

[0037] Further, such as Figure 2 As shown, the experimental control parameter interval of the target citrus fruit is obtained, and the experimental control parameter is optimized in combination with the historical meteorological record data of the production site of the target citrus fruit to obtain N control parameter sets. The method also includes:

[0038] Based on the historical meteorological record data of the production area of ​​the target citrus fruit within the preset historical window, edge constraint analysis is performed to generate P parameter edge constraints;

[0039] Based on the experimental control parameter interval, a basic control parameter optimization space is constructed, and the basic control parameter optimization space is constrained by using the P parameter edge constraints to generate a target control parameter optimization space, wherein the target control parameter optimization space has P control parameter edge degrees;

[0040] Randomly generating M initial experimental control parameter sets in the target control parameter optimization space;

[0041] According to the P control parameter edge degrees, the M initial experimental control parameter sets are adjusted according to a preset adjustment method to obtain an adjusted experimental control parameter set, wherein the adjusted experimental control parameter set includes multiple adjusted experimental control parameter sets, and the preset adjustment method is to increase or decrease the P experimental control parameters according to the P control parameter edge degrees;

[0042] The adjustment experiment control parameter set is traversed to perform parameter adaptation analysis to obtain an adjustment parameter adaptation factor set.

[0043] Further, the adjustment experiment control parameter set is traversed to perform parameter adaptation analysis to obtain an adjustment parameter adaptation factor set, and the method further includes:

[0044] Construct a parameter adaptation formula, wherein the parameter adaptation formula is:

[0045]

[0046] Where L is the adjustment parameter adaptation factor of the adjustment experiment control parameter set, n is the number of parameters in the adjustment experiment control parameter set, n is an integer greater than or equal to 1, P(x i ) is the number of times the ith adjustment experimental control parameter in the adjustment experimental control parameter set appears in the historical meteorological record data of the production site, The total parameter quantity of the control parameter type corresponding to the i-th adjustment experimental control parameter in the adjustment experimental control parameter set in the historical meteorological record data of the production site;

[0047] The adjustment experiment control parameter set is input into the parameter adaptation formula for calculation to obtain an adjustment parameter adaptation factor set.

[0048] In the embodiments of the present application, the sunburn phenomenon of citrus fruits is closely related to the meteorological environment in which they grow. In order to reliably identify the high temperature and strong light disasters that cause sunburn of citrus fruits, it is necessary to determine the experimental control parameters that are more in line with the actual situation. Therefore, the information of the light source used in the experiment is analyzed to determine the power and model of the light source, and the temperature range is determined according to the model of the circulating water bath used for heating in the experiment, so as to obtain the experimental control parameter interval. Among them, the experimental control parameter interval is used to describe the experimental control parameter selection range used in the high temperature and strong light disaster identification and analysis experiment that causes sunburn of citrus fruits. The experimental control parameters include parameters such as ambient light intensity, illumination time, temperature, and ambient light intensity gradient change amplitude. By retrieving the historical meteorological record data of the production site of the target citrus fruit within the preset historical window, the experimental control parameters in the experimental control parameter interval are optimized and screened, so as to obtain N control parameter sets that are more in line with the meteorological changes in the production site of the target citrus fruit. Among them, the N control parameter sets have N parameter adaptation factors. The N parameter adaptation factors are the degree of fit between the experimental environment under the configuration of the N control parameter sets and the meteorological conditions of the production site of the target citrus fruit. The larger the parameter adaptation factor, the higher the degree of fit between the experimental environment configured by the corresponding control parameter set and the meteorological conditions of the production site of the target citrus fruit. This achieves the goal of providing a reliable experimental control parameter set for subsequent experiments and improving the reliability of the experiment from the perspective of experimental configuration, and achieves the technical effect of paving the way for the subsequent construction of an intelligent high temperature and strong light disaster identification network based on experiments and improving the accuracy of identification.

[0049] In one embodiment, the historical meteorological record data of the production site is used to describe the conditions of environmental parameters such as light intensity, light time, and ambient temperature of the production site of the target citrus fruit within a preset historical window, and the parameter values ​​of each environmental parameter are counted respectively, and the parameter values ​​of each environmental parameter exceeding the environmental parameter threshold and the parameter values ​​whose number is less than the preset parameter value are eliminated, thereby achieving the goal of eliminating abnormal values ​​and low-frequency values, and avoiding the interference of abnormal values ​​on the analysis results. According to the end values ​​of the parameter values ​​of each environmental parameter after elimination, the two edge values ​​of the parameter constraint interval in the corresponding parameter edge constraint are constructed. Among them, the P parameter edge constraints are used to describe the climate change fluctuation range of the production site of the target citrus fruit within the preset historical window. Each parameter edge constraint corresponds to an environmental parameter, and each environmental parameter corresponds to an experimental control parameter in a control parameter set. The preset historical window is a historical time period set by technicians in this field, which can be half a year, one year, two years, etc. According to the experimental control parameter interval, the range that can be selected when setting the experimental parameters for each experimental control parameter is determined, and the basic control parameter optimization space is constructed. Furthermore, the P parameter edge constraints are used to constrain the basic control parameter optimization space, that is, according to the range of each environmental parameter in the P parameter edge constraints, the range of the corresponding experimental control parameter in the basic control parameter optimization space is limited, and the target control parameter optimization space is obtained according to the range limitation result. In this way, the technical effect of constraining the experimental control parameter selection range according to the actual production site conditions of the target citrus fruit is achieved, so that the parameter selection is more in line with the actual conditions of the production site.

[0050] In one embodiment, the parameter selection range values ​​corresponding to each experimental control parameter in the target control parameter preferred space are divided by the total value of the parameter selection range values ​​corresponding to each experimental control parameter, and the calculation results are used as the P control parameter edge degrees of the P experimental control parameters in the target control parameter preferred space. Among them, the P control parameter edge degrees reflect the range of the P experimental control parameters that can be selected. The larger the control parameter edge degree, the larger the range of the corresponding experimental control parameter that can be selected. The P control parameter edge degrees are multiplied by the preset adjustment step size respectively to obtain the amplitude of a single adjustment of each experimental control parameter according to the preset adjustment method, and P adjustment amplitudes are obtained.

[0051] Furthermore, M initial experimental control parameter sets are randomly generated in the target control parameter optimization space, and the M initial experimental control parameter sets are adjusted according to the preset adjustment method, so as to obtain the adjusted experimental control parameter set. Each initial experimental control parameter set includes P experimental control parameters. The adjusted experimental control parameter set includes multiple adjusted experimental control parameter sets, and the preset adjustment method is to increase or decrease the P experimental control parameters according to the edge degrees of the P control parameters, and the increase or decrease amplitude is determined according to the P adjustment amplitudes. By adjusting the M initial experimental control parameter sets, the technical effect of enriching the parameter set selection space is achieved.

[0052] In one possible embodiment, the parameter adaptation formula is used to quantitatively analyze the uniformity of the frequency of occurrence of the control parameters in the adjustment experiment control parameter set in the historical production situation of the production site of the target citrus fruit. The higher the uniformity, the larger the adjustment parameter adaptation factor. Since the adjustment experiment control parameter set is obtained after random adjustment based on the randomly generated M initial experiment control parameter sets, the frequency of occurrence of the experimental control parameters in each adjustment experiment control parameter set in the historical time is different and may not be consistent. In the experiment, it is necessary to balance the experimental environment configured in the experiment. For example, when the frequency of occurrence of the control parameter set of the configured experimental environment in the historical time is low, the experimental environment configured at this time is the low-frequency growth environment of the target citrus fruit; when the frequency of occurrence of the control parameter set of the configured experimental environment in the historical time is high, the experimental environment configured at this time is the general growth environment of the target citrus fruit. By using the parameter adaptation formula for quantitative calculation, a basis is provided for whether the experimental environment configured by each adjustment experiment control parameter set is in line with the actual situation. When the adjustment parameter adaptation factor of the control parameter set for configuring the experimental environment is low, it indicates that some experimental control parameters in the control parameter set have a high frequency of occurrence in the historical time, while other experimental control parameters have a low frequency of occurrence in the historical time. At this time, the configured experimental environment is not representative, and the error of the experimental results is high.

[0053] In one embodiment, the adjustment experiment control parameter set is input into the parameter adaptation formula for quantitative calculation, thereby obtaining an adjustment parameter adaptation factor set, wherein the adjustment parameter adaptation factor set provides support for the rationality and reliability of subsequent experimental environment configuration.

[0054] Further, such as Figure 3 As shown, the step of obtaining the adjustment parameter adaptation factor set further includes:

[0055] Based on the adjustment parameter adaptation factor set, Q leading adjustment control parameter sets and Q following adjustment control parameter sets are obtained, wherein each leading control parameter set and each following adjustment control parameter set has a corresponding adjustment parameter adaptation factor;

[0056] Taking the guided adjustment control parameter set corresponding to the maximum value of the adjustment parameter adaptation factor in the Q guided adjustment control parameter sets as the guided optimization direction, adjusting the Q guided adjustment control parameter sets according to Q first free paces, taking the Q guided adjustment control parameter sets as the following optimization direction, adjusting the Q following adjustment control parameter sets according to Q second free pace sets, and updating the adjustment experiment control parameter set according to the adjustment result;

[0057] After a preset number of iterative adjustments, K adjustment experiment control parameter sets whose adjustment parameter adaptation factors are in the top K positions are obtained;

[0058] The K adjustment experiment control parameter sets are similarly screened to obtain N control parameter sets.

[0059] In one embodiment, after obtaining the adjustment parameter adaptation factor set, the adjustment parameters are sorted in descending order according to the factors, and the Q adjustment experimental control parameter sets corresponding to the adjustment parameter adaptation factor sets in the first Q positions are used as Q guiding adjustment control parameter sets. Then, the remaining adjustment control parameter sets are respectively calculated with the Q guiding adjustment control parameter sets for cosine similarity, and the remaining adjustment control parameter sets are respectively divided into the guiding adjustment control parameter sets corresponding to the maximum cosine similarity in the calculation results as the following adjustment control parameter sets, thereby obtaining Q following adjustment control parameter sets. Among them, each guiding control parameter set and each following adjustment control parameter set have corresponding adjustment parameter adaptation factors. The guiding adjustment control parameter set corresponding to the maximum value of the adjustment parameter adaptation factor in the Q guiding adjustment control parameter sets is used as the guiding optimization direction, and the Q guiding adjustment control parameter sets are adjusted according to Q first free steps. Among them, the first free step is the adjustment amplitude when the guiding adjustment control parameter set performs a single parameter adjustment. Preferably, the inverse of the ratio of the Q adjustment parameter adaptation factors corresponding to the Q guide adjustment control parameter sets to the adjustment parameter adaptation factor of the guide optimization direction is calculated respectively, and the calculation result is multiplied by the preset free step, and the Q first free steps are obtained according to the multiplication result. The preset free step is the adjustment range set by those skilled in the art when performing a single parameter adjustment.

[0060] Preferably, the reciprocals of the multiple adjustment parameter adaptation factors of the multiple follow-up adjustment control parameter sets in the Q follow-up adjustment control parameter sets over the corresponding guide adjustment control parameter sets are calculated respectively, and the calculation results are multiplied by the preset free pace to obtain Q second free pace sets. Then, the Q guide adjustment control parameter sets are respectively used as the follow-up optimization directions, and the Q follow-up adjustment control parameter sets are adjusted according to the Q second free pace sets, and the adjustment experimental control parameter set is updated according to the adjustment results of the guide optimization direction and the follow-up optimization direction. The technical effect of optimizing and updating the multiple adjustment experimental control parameters in accordance with the actual adaptation conditions of each adjustment control parameter set is achieved, and the experimental control parameters that are more in line with the production conditions of the target citrus fruit are obtained, and the reliability of the experiment is improved.

[0061] In one embodiment, when the number of update iterative adjustments meets the preset number, K adjustment experiment control parameter sets whose adjustment parameter adaptation factors are in the top K positions in the updated adjustment experiment control parameter set are obtained. Then, the K adjustment experiment control parameter sets are enumerated for cosine similarity calculations in pairs, and the two adjustment experiment control parameter sets whose calculation results meet the preset cosine similarity threshold are retained as the parameter set with the higher adjustment parameter adaptation factor, and N control parameter sets are obtained. Thus, the technical effect of removing duplicate data from the adjustment experiment control parameter set and avoiding experiments in the same experimental environment is achieved.

[0062] Based on N parameter adaptation factors, match the number of N experimental samples and configure N sets of environmental control experiments in combination with N control parameter sets;

[0063] In one possible embodiment, the ratio of the N parameter adaptation factors to the sum of the N parameter adaptation factors is calculated respectively, and the calculation result is multiplied by the total number of experimental samples to obtain the N number of experimental samples. In this way, more experimental samples are matched to the experiments that are more in line with the production site conditions of the target citrus fruits, thereby optimizing the allocation of experimental resources. Furthermore, N groups of environmental control experiments are configured according to the N number of experimental samples and the N control parameter sets. After the configuration is completed, the citrus fruit sunburn experiment is carried out.

[0064] Collect N sample characterization data sets of N groups of environmental control experiments, perform data sensitivity granularity screening based on the N sample characterization data sets, and generate a target sensitive feature set;

[0065] Furthermore, the method further comprises collecting N sample characterization data sets of N groups of environmental control experiments, performing data sensitivity granularity screening based on the N sample characterization data sets, and generating a target sensitive feature set. The method further comprises:

[0066] Using the N control parameter sets as indexes, retrieve and obtain N benchmark characterization data sets;

[0067] Traversing the N sample characterization data sets and the N benchmark characterization data sets to perform sensitivity granularity calculation to obtain N sample sensitivity granularity sets;

[0068] N sensitive granularity analysis spaces are constructed using N sample sensitive granularity sets, wherein each sensitive granularity analysis space includes multiple particles, and each particle corresponds to a sample sensitive granularity;

[0069] Data sensitivity granularity screening is performed based on the N sensitivity granularity analysis spaces to obtain a target sensitive feature set.

[0070] Furthermore, the method further comprises:

[0071] Traversing the N sensitive granularity analysis spaces to search for central particles, and obtaining N central particles, wherein each central particle has a corresponding central sample sensitive granularity;

[0072] Taking the N central particles as starting points respectively, and configuring N screening steps according to the N central sample sensitive granularities to search in the N sensitive granularity analysis space, to obtain N stage particles;

[0073] Determine whether the central sample sensitivity granularity of the N central particles is greater than the stage sample sensitivity granularity of the N stage particles respectively. If so, update the N central particles to N stage particles according to a certain probability;

[0074] According to the N stage sample sensitive granularities corresponding to the N stage particles, N stage screening steps are configured, and the N stage particles are used as the starting point to search in the N sensitive granularity analysis space until the preset search times are met, and the stage particle corresponding to the maximum value of the stage sample sensitive granularity during the search process is used as the N target particles;

[0075] Collect a preset number of particles in the order of the distance from the N target particles from near to far, and obtain a corresponding N target sample representation data set;

[0076] Based on the N target sample representation data sets, weighted calculation is performed to obtain the target sensitive feature set.

[0077] In one possible embodiment, the characterization data of sample citrus fruits in N groups of environmental control experiments are collected respectively to obtain N sample characterization data sets. The N sample characterization data sets are used to describe the status of multiple sample citrus fruits in the N groups of environmental control experiments, including peel sugar and acid, peel cell membrane permeability, pigment content, antioxidant content, endogenous hormones, chlorophyll fluorescence and other data. By analyzing the degree of deviation between multiple characterization data in the N sample characterization data sets and the benchmark characterization data, the characterization data is cleaned to obtain the features that best reflect the status of the sample citrus fruits, and generate the target sensitive feature set. The technical effect of reducing the dimension of the features and improving the efficiency of subsequent analysis is achieved.

[0078] In one possible embodiment, by using the N control parameter sets as indexes, N benchmark characterization data sets are retrieved in the database to obtain N benchmark characterization data sets. The N benchmark characterization data sets are characterization data sets of normal growth of citrus fruits under N groups of environmental control experiments. The N sample characterization data sets and the N benchmark characterization data sets are traversed to calculate the data deviation and obtain N sample sensitive granularity sets. The N sample sensitive granularity sets reflect the degree of deviation between the experimental samples and the benchmark requirements. N sensitive granularity analysis spaces are constructed using the N sample sensitive granularity sets, wherein each sensitive granularity analysis space includes multiple particles, each particle corresponds to a sample sensitive granularity, and data sensitive granularity screening is performed based on the N sensitive granularity analysis spaces to obtain a target sensitive feature set. By retaining data features with larger data sensitive granularity, the technical effect of providing an identification basis for subsequent citrus fruit disaster identification is achieved.

[0079] In one possible embodiment, the mean sensitive granularity in N sensitive granularity analysis spaces is calculated respectively, and the calculated result is used for one particle as the central particle, thereby obtaining N central particles. Each central particle has a corresponding central sample sensitive granularity. Further, the N central particles are taken as the starting point, and N screening steps are configured according to the N central sample sensitive granularities to perform a search in the N sensitive granularity analysis space to obtain N stage particles. The inverse of the ratio of the N central sample sensitive granularity to the sum of the N central sample sensitive granularities is calculated respectively, and multiplied by the preset screening step to obtain N screening steps. The preset screening step is the distance of a single movement when performing data sensitive granularity screening set by a technician in this field. The N screening steps reflect the distance of a single movement of the N central particles when performing data screening. The higher the central sample sensitive granularity, the smaller the distance the corresponding central particle moves. Then, it is determined whether the central sample sensitivity granularity of the N central particles is greater than the stage sample sensitivity granularity of the N stage particles. If so, it indicates that the deviation of the central particle is large. In order to avoid falling into local optimization, the N central particles are updated to N stage particles according to a certain probability. Then, N stage screening steps are configured according to the N stage sample sensitivity granularity corresponding to the N stage particles, and the N stage particles are used as the starting point to search in the N sensitive granularity analysis space until the preset number of searches is met. The stage particles corresponding to the maximum value of the stage sample sensitivity granularity during the search process are used as N target particles. According to the order from near to far from the N target particles, a preset number of particles are collected to obtain the corresponding N target sample characterization data sets. Because the target particle is the characterization data with the largest deviation after optimization in each sensitive granularity analysis space, by collecting a preset number of particles around it, multiple reliable auxiliary judgment characterization data can be obtained. According to the N parameter adaptation factors of the N control parameter sets corresponding to the N groups of environmental control experiments, the N target sample characterization data sets are weighted calculated to obtain the target sensitive feature set. The weight value of the weighted calculation is the ratio of the N parameter adaptation factors to the sum of the N parameter adaptation factors.

[0080] Traversing and collecting N sunburn indexes of the N groups of environmental control experiments, combining N control parameter sets of the N groups of environmental control experiments and N target sensitive feature parameter sets to perform decision analysis, and generate N high temperature and strong light disaster identification branches;

[0081] A high temperature and strong light disaster identification network is constructed based on N high temperature and strong light disaster identification branches;

[0082] In one possible embodiment, N sunburn indexes are obtained according to the ratio of the sample size of sunburned samples of multiple experimental samples in N groups of environmental control experiments to the total number of multiple experimental samples. Among them, the N sunburn indexes reflect the degree of sunburn caused by high temperature and strong light to citrus fruits under N groups of environmental control experiments. Then, the outputs of N high temperature and strong light disaster identification branches constructed based on convolutional neural networks are constrained by N sunburn indexes, and the inputs of N high temperature and strong light disaster identification branches constructed based on convolutional neural networks are constrained by N control parameter sets and N target sensitive feature parameter sets of N groups of environmental control experiments. After supervised training, when the output reaches convergence, N high temperature and strong light disaster identification branches are obtained. Among them, the N high temperature and strong light disaster identification branches are used to intelligently identify the degree of sunburn disaster in the citrus growth environment under N groups of environmental control experiments. The N target sensitive feature parameter sets are parameter sets obtained after the samples in the N groups of environmental control experiments are collected according to the target sensitive feature sets. The high temperature and strong light disaster identification network is constructed by parallelly fully connecting the obtained N high temperature and strong light disaster identification branches. The technical effect of intelligently identifying the degree of sunburn caused by high temperature and strong light based on the citrus growth environment and the characteristics of citrus fruit has been achieved.

[0083] Collecting a meteorological parameter set of the target citrus fruit, performing matching analysis with the N control parameter sets, and obtaining a matching analysis result;

[0084] Furthermore, the method further comprises:

[0085] Calculate the cosine similarity between the meteorological parameter set of the target citrus fruit and the N control parameter sets respectively to obtain N similarities;

[0086] Filter L control parameter sets that meet a preset similarity threshold;

[0087] The L high temperature and strong light disaster identification branches corresponding to the L control parameter sets are taken as the matching analysis results.

[0088] Collecting characteristic data of the target citrus fruit according to the target sensitive characteristic set to obtain a target representation data set;

[0089] In one embodiment, the growth environment of the target citrus fruit is collected according to a plurality of environmental parameters to obtain a meteorological parameter set. The meteorological parameter set is respectively calculated for similarity with N control parameter sets to obtain N similarities. The N similarities reflect the degree of similarity between the meteorological parameter set and the N control parameter sets. In other words, it reflects the degree of similarity between the actual growth environment of the target citrus fruit and the experimental control environment. Optionally, the similarity calculation method includes calculating cosine similarity, Euclidean distance, etc., and adding the control parameter set whose similarity calculation result satisfies the preset matching similarity into the matching analysis result. Among them, the matching analysis result reflects the experimental environment that is more in line with the target citrus fruit, including L control parameter sets. Furthermore, according to the L control parameter sets, a basis can be provided for the subsequent activation of the corresponding high temperature and strong light disaster identification branch in the high temperature and strong light disaster identification network.

[0090] In a possible embodiment, the characteristic data of the target citrus fruit is collected according to the target sensitive characteristic set, the physiological characteristics of the target citrus fruit that can be used to analyze the high temperature and strong light disaster identification are obtained, and the target characterization data set is obtained. The technical effect of obtaining a high-fitness identification branch for reliable disaster identification and obtaining reliable identification analysis data is achieved.

[0091] The high temperature and strong light disaster identification branch of the high temperature and strong light disaster identification network is activated according to the matching analysis results, and the target characterization data set is sent to the activated high temperature and strong light disaster identification branch for disaster identification, and the identification results are comprehensively analyzed to obtain the target disaster identification results.

[0092] In an embodiment of the present application, according to the L control parameter sets in the matching analysis results, the L high temperature and strong light disaster identification branches of the high temperature and strong light disaster identification network are activated, and the target characterization data set is sent to the activated L high temperature and strong light disaster identification branches for disaster identification, and then the output L sunburn indexes are weighted according to the corresponding similarities, and the target disaster identification result is obtained according to the weighted calculation result. The target disaster identification result reflects the degree of sunburn caused by high temperature and strong light stress on the target citrus fruit. The technical effect of intelligently identifying high temperature and strong light disasters and improving the accuracy of identification is achieved.

[0093] In summary, the embodiments of the present application have at least the following technical effects:

[0094] The present application obtains the experimental control parameter interval of the target citrus fruit, optimizes the experimental control parameters in combination with the historical meteorological record data of the production site of the target citrus fruit, obtains N control parameter sets, and achieves the goal of paving the way for subsequent experiments that fit the actual meteorological conditions of the production site. Then, the number of N experimental samples is matched based on N parameter adaptation factors, and N groups of environmental control experiments are configured in combination with N control parameter sets. Then, N sample characterization data sets of the N groups of environmental control experiments are collected, and data sensitivity granularity screening is performed based on the N sample characterization data sets to generate target sensitive feature sets, thereby achieving the goal of dimensionality reduction analysis of features for disaster identification, reducing the amount of analysis data, and improving analysis efficiency. Then, N sunburn indices of the N groups of environmental control experiments are traversed and collected, and decision analysis is performed in combination with the N control parameter sets of the N groups of environmental control experiments and the N target sensitive feature parameter sets. Generate N high temperature and strong light disaster identification branches, and build a high temperature and strong light disaster identification network based on the N high temperature and strong light disaster identification branches, so as to achieve the goal of building an intelligent analysis identification network and improving the intelligent level of disaster identification. Then, collect the meteorological parameter set of the target citrus fruit, match and analyze it with the N control parameter sets, obtain the matching analysis results, and then collect the characteristic data of the target citrus fruit according to the target sensitive feature set to obtain the target representation data set, and then activate the high temperature and strong light disaster identification branch of the high temperature and strong light disaster identification network according to the matching analysis results, so as to achieve the goal of further improving the refinement of intelligent analysis, and use the high temperature and strong light disaster identification branch constructed by the experimental environment data with high similarity for identification, and send the target representation data set to the activated high temperature and strong light disaster identification branch for disaster identification, and conduct a comprehensive analysis of the identification results to obtain the target disaster identification results. The configuration experiment is refined according to the actual situation of the production area of ​​citrus fruit, and an efficient and accurate high temperature and strong light disaster identification network is built according to the experimental results to improve the accuracy of disaster identification.

[0095] Embodiment 2

[0096] Based on the same inventive concept as the method for identifying high temperature and strong light disasters causing sunburn of citrus fruits in the aforementioned embodiment, Figure 4 As shown, the present application provides a high temperature and strong light disaster identification system that causes sunburn in citrus fruits. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0097] The first obtaining unit 11 is used to obtain the experimental control parameter interval of the target citrus fruit, and to optimize the experimental control parameters in combination with the historical meteorological record data of the production area of ​​the target citrus fruit to obtain N control parameter sets, wherein the N control parameter sets have N parameter adaptation factors;

[0098] A first configuration unit 12 is used to match the number of N experimental samples based on N parameter adaptation factors and configure N groups of environmental control experiments in combination with N control parameter sets;

[0099] The first generating unit 13 is used for generating a target sensitive feature set by collecting N sample representation data sets of N groups of environmental control experiments, performing data sensitivity granularity screening based on the N sample representation data sets, and generating a target sensitive feature set;

[0100] The second generating unit 14 is used to traverse and collect N sunburn indexes of the N groups of environmental control experiments, combine the N control parameter sets of the N groups of environmental control experiments and the N target sensitive feature parameter sets to perform decision analysis, and generate N high temperature and strong light disaster identification branches;

[0101] A first construction unit 15 is used for first identifying and constructing a high temperature and strong light disaster identification network according to N high temperature and strong light disaster identification branches;

[0102] The second obtaining unit 16 is used to collect the meteorological parameter set of the target citrus fruit, perform matching analysis with the N control parameter sets, and obtain matching analysis results;

[0103] A third obtaining unit 17 is used to collect the characteristic data of the target citrus fruit according to the target sensitive characteristic set to obtain a target representation data set;

[0104] The fourth acquisition unit 18 is used to activate the high temperature and strong light disaster identification branch of the high temperature and strong light disaster identification network according to the matching analysis results, and send the target characterization data set to the activated high temperature and strong light disaster identification branch for disaster identification, and comprehensively analyze the identification results to obtain the target disaster identification results.

[0105] Furthermore, the system also includes:

[0106] Based on the historical meteorological record data of the production area of ​​the target citrus fruit within the preset historical window, edge constraint analysis is performed to generate P parameter edge constraints;

[0107] Based on the experimental control parameter interval, a basic control parameter optimization space is constructed, and the basic control parameter optimization space is constrained by using the P parameter edge constraints to generate a target control parameter optimization space, wherein the target control parameter optimization space has P control parameter edge degrees;

[0108] Randomly generating M initial experimental control parameter sets in the target control parameter optimization space;

[0109] According to the P control parameter edge degrees, the M initial experimental control parameter sets are adjusted according to a preset adjustment method to obtain an adjusted experimental control parameter set, wherein the adjusted experimental control parameter set includes multiple adjusted experimental control parameter sets, and the preset adjustment method is to increase or decrease the P experimental control parameters according to the P control parameter edge degrees;

[0110] The adjustment experiment control parameter set is traversed to perform parameter adaptation analysis to obtain an adjustment parameter adaptation factor set.

[0111] Furthermore, the system also includes:

[0112] Based on the adjustment parameter adaptation factor set, Q leading adjustment control parameter sets and Q following adjustment control parameter sets are obtained, wherein each leading control parameter set and each following adjustment control parameter set has a corresponding adjustment parameter adaptation factor;

[0113] Taking the guided adjustment control parameter set corresponding to the maximum value of the adjustment parameter adaptation factor in the Q guided adjustment control parameter sets as the guided optimization direction, adjusting the Q guided adjustment control parameter sets according to Q first free paces, taking the Q guided adjustment control parameter sets as the following optimization direction, adjusting the Q following adjustment control parameter sets according to Q second free pace sets, and updating the adjustment experiment control parameter set according to the adjustment result;

[0114] After a preset number of iterative adjustments, K adjustment experiment control parameter sets whose adjustment parameter adaptation factors are in the top K positions are obtained;

[0115] The K adjustment experiment control parameter sets are similarly screened to obtain N control parameter sets.

[0116] Furthermore, the system also includes:

[0117] Construct a parameter adaptation formula, wherein the parameter adaptation formula is:

[0118]

[0119] Where L is the adjustment parameter adaptation factor of the adjustment experiment control parameter set, n is the number of parameters in the adjustment experiment control parameter set, n is an integer greater than or equal to 1, P(x i ) is the number of times the ith adjustment experimental control parameter in the adjustment experimental control parameter set appears in the historical meteorological record data of the production site, The total parameter quantity of the control parameter type corresponding to the i-th adjustment experimental control parameter in the adjustment experimental control parameter set in the historical meteorological record data of the production site;

[0120] The adjustment experiment control parameter set is input into the parameter adaptation formula for calculation to obtain an adjustment parameter adaptation factor set.

[0121] Furthermore, the system also includes:

[0122] Using the N control parameter sets as indexes, retrieve and obtain N benchmark characterization data sets;

[0123] Traversing the N sample characterization data sets and the N benchmark characterization data sets to perform sensitivity granularity calculation to obtain N sample sensitivity granularity sets;

[0124] N sensitive granularity analysis spaces are constructed using N sample sensitive granularity sets, wherein each sensitive granularity analysis space includes multiple particles, and each particle corresponds to a sample sensitive granularity;

[0125] Data sensitivity granularity screening is performed based on the N sensitivity granularity analysis spaces to obtain a target sensitive feature set.

[0126] Furthermore, the system also includes:

[0127] Traversing the N sensitive granularity analysis spaces to search for central particles, and obtaining N central particles, wherein each central particle has a corresponding central sample sensitive granularity;

[0128] Taking the N central particles as starting points respectively, and configuring N screening steps according to the N central sample sensitive granularities to search in the N sensitive granularity analysis space, to obtain N stage particles;

[0129] Determine whether the central sample sensitivity granularity of the N central particles is greater than the stage sample sensitivity granularity of the N stage particles respectively. If so, update the N central particles to N stage particles according to a certain probability;

[0130] According to the N stage sample sensitive granularities corresponding to the N stage particles, N stage screening steps are configured, and the N stage particles are used as the starting point to search in the N sensitive granularity analysis space until the preset search times are met, and the stage particle corresponding to the maximum value of the stage sample sensitive granularity during the search process is used as the N target particles;

[0131] Collect a preset number of particles in the order of the distance from the N target particles from near to far, and obtain a corresponding N target sample representation data set;

[0132] Based on the N target sample representation data sets, weighted calculation is performed to obtain the target sensitive feature set.

[0133] Furthermore, the system also includes:

[0134] Calculate the cosine similarity between the meteorological parameter set of the target citrus fruit and the N control parameter sets respectively to obtain N similarities;

[0135] Filter L control parameter sets that meet a preset similarity threshold;

[0136] The L high temperature and strong light disaster identification branches corresponding to the L control parameter sets are taken as the matching analysis results.

[0137] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. Other embodiments are within the scope of the attached claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0138] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0139] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for identifying high temperature and strong light disasters that cause sunburn on citrus fruits, characterized in that: The method comprises: Obtaining the experimental control parameter interval of the target citrus fruit, optimizing the experimental control parameters in combination with the historical meteorological record data of the production site of the target citrus fruit, and obtaining N control parameter sets, wherein the N control parameter sets have N parameter adaptation factors; Based on N parameter adaptation factors, match the number of N experimental samples and configure N sets of environmental control experiments in combination with N control parameter sets; Collect N sample characterization data sets of N groups of environmental control experiments, perform data sensitivity granularity screening based on the N sample characterization data sets, and generate a target sensitive feature set; Traversing and collecting N sunburn indexes of the N groups of environmental control experiments, combining N control parameter sets of the N groups of environmental control experiments and N target sensitive feature parameter sets to perform decision analysis, and generate N high temperature and strong light disaster identification branches; A high temperature and strong light disaster identification network is constructed based on N high temperature and strong light disaster identification branches; Collecting a meteorological parameter set of the target citrus fruit, performing matching analysis with the N control parameter sets, and obtaining a matching analysis result; Collecting characteristic data of the target citrus fruit according to the target sensitive characteristic set to obtain a target representation data set; The high temperature and strong light disaster identification branch of the high temperature and strong light disaster identification network is activated according to the matching analysis results, and the target characterization data set is sent to the activated high temperature and strong light disaster identification branch for disaster identification, and the identification results are comprehensively analyzed to obtain the target disaster identification results.

2. The method according to claim 1, characterized in that The method further comprises: obtaining the experimental control parameter interval of the target citrus fruit, optimizing the experimental control parameters in combination with the historical meteorological record data of the production site of the target citrus fruit, and obtaining N control parameter sets. Based on the historical meteorological record data of the production area of ​​the target citrus fruit within the preset historical window, edge constraint analysis is performed to generate P parameter edge constraints; Based on the experimental control parameter interval, a basic control parameter optimization space is constructed, and the basic control parameter optimization space is constrained by using the P parameter edge constraints to generate a target control parameter optimization space, wherein the target control parameter optimization space has P control parameter edge degrees; Randomly generating M initial experimental control parameter sets in the target control parameter optimization space; According to the P control parameter edge degrees, the M initial experimental control parameter sets are adjusted according to a preset adjustment method to obtain an adjusted experimental control parameter set, wherein the adjusted experimental control parameter set includes multiple adjusted experimental control parameter sets, and the preset adjustment method is to increase or decrease the P experimental control parameters according to the P control parameter edge degrees; The adjustment experiment control parameter set is traversed to perform parameter adaptation analysis to obtain an adjustment parameter adaptation factor set.

3. The method according to claim 2, characterized in that The step of obtaining the adjustment parameter adaptation factor set further includes: Based on the adjustment parameter adaptation factor set, Q leading adjustment control parameter sets and Q following adjustment control parameter sets are obtained, wherein each leading control parameter set and each following adjustment control parameter set has a corresponding adjustment parameter adaptation factor; Taking the guided adjustment control parameter set corresponding to the maximum value of the adjustment parameter adaptation factor in the Q guided adjustment control parameter sets as the guided optimization direction, adjusting the Q guided adjustment control parameter sets according to Q first free paces, taking the Q guided adjustment control parameter sets as the following optimization direction, adjusting the Q following adjustment control parameter sets according to Q second free pace sets, and updating the adjustment experiment control parameter set according to the adjustment result; After a preset number of iterative adjustments, K adjustment experiment control parameter sets whose adjustment parameter adaptation factors are in the top K positions are obtained; The K adjustment experiment control parameter sets are similarly screened to obtain N control parameter sets.

4. The method according to claim 2, characterized in that Traversing the adjustment experiment control parameter set to perform parameter adaptation analysis to obtain an adjustment parameter adaptation factor set, the method further includes: Construct a parameter adaptation formula, wherein the parameter adaptation formula is: Where L is the adjustment parameter adaptation factor of the adjustment experiment control parameter set, n is the number of parameters in the adjustment experiment control parameter set, n is an integer greater than or equal to 1, P(x i ) is the number of times the ith adjustment experimental control parameter in the adjustment experimental control parameter set appears in the historical meteorological record data of the production site, The total parameter quantity of the control parameter type corresponding to the i-th adjustment experimental control parameter in the adjustment experimental control parameter set in the historical meteorological record data of the production site; The adjustment experiment control parameter set is input into the parameter adaptation formula for calculation to obtain an adjustment parameter adaptation factor set.

5. The method according to claim 3, characterized in that The method further comprises collecting N sample characterization data sets of N groups of environmental control experiments, performing data sensitivity granularity screening based on the N sample characterization data sets, and generating a target sensitive feature set. The method further comprises: Using the N control parameter sets as indexes, retrieve and obtain N benchmark characterization data sets; Traversing the N sample characterization data sets and the N benchmark characterization data sets to perform sensitivity granularity calculation to obtain N sample sensitivity granularity sets; N sensitive granularity analysis spaces are constructed using N sample sensitive granularity sets, wherein each sensitive granularity analysis space includes multiple particles, and each particle corresponds to a sample sensitive granularity; Data sensitivity granularity screening is performed based on the N sensitivity granularity analysis spaces to obtain a target sensitive feature set.

6. The method according to claim 5, characterized in that The method further comprises: Traversing the N sensitive granularity analysis spaces to search for central particles, and obtaining N central particles, wherein each central particle has a corresponding central sample sensitive granularity; Taking the N central particles as starting points respectively, and configuring N screening steps according to the N central sample sensitive granularities to search in the N sensitive granularity analysis space, to obtain N stage particles; Determine whether the central sample sensitivity granularity of the N central particles is greater than the stage sample sensitivity granularity of the N stage particles respectively. If so, update the N central particles to N stage particles according to a certain probability; According to the N stage sample sensitive granularities corresponding to the N stage particles, N stage screening steps are configured, and the N stage particles are used as the starting point to search in the N sensitive granularity analysis space until the preset search times are met, and the stage particle corresponding to the maximum value of the stage sample sensitive granularity during the search process is used as the N target particles; Collect a preset number of particles in the order of the distance from the N target particles from near to far, and obtain a corresponding N target sample representation data set; Based on the N target sample representation data sets, weighted calculation is performed to obtain the target sensitive feature set.

7. The method according to claim 1, characterized in that The method further comprises: Calculate the cosine similarity between the meteorological parameter set of the target citrus fruit and the N control parameter sets respectively to obtain N similarities; Filter L control parameter sets that meet a preset similarity threshold; The L high temperature and strong light disaster identification branches corresponding to the L control parameter sets are taken as the matching analysis results.

8. A system for identifying high temperature and strong light disasters that cause sunburn on citrus fruits, characterized in that: The system comprises: A first obtaining unit is used to obtain an experimental control parameter interval of a target citrus fruit, and to optimize the experimental control parameters in combination with historical meteorological record data of a production site of the target citrus fruit, to obtain N control parameter sets, wherein the N control parameter sets have N parameter adaptation factors; A first configuration unit is used to match the number of N experimental samples based on N parameter adaptation factors and configure N groups of environmental control experiments in combination with N control parameter sets; A first generating unit is used for generating a target sensitive feature set by collecting N sample representation data sets of N groups of environmental control experiments, performing data sensitivity granularity screening based on the N sample representation data sets, and generating a target sensitive feature set; The second generating unit is used to traverse and collect the N sunburn indexes of the N groups of environmental control experiments, combine the N control parameter sets of the N groups of environmental control experiments and the N target sensitive feature parameter sets to perform decision analysis, and generate N high temperature and strong light disaster identification branches; A first construction unit is used for first identifying and constructing a high temperature and strong light disaster identification network according to N high temperature and strong light disaster identification branches; A second acquisition unit is used to collect the meteorological parameter set of the target citrus fruit, perform matching analysis with the N control parameter sets, and obtain a matching analysis result; A third obtaining unit is used to collect the characteristic data of the target citrus fruit according to the target sensitive characteristic set to obtain a target representation data set; The fourth acquisition unit is used to activate the high temperature and strong light disaster identification branch of the high temperature and strong light disaster identification network according to the matching analysis results, and send the target characterization data set to the activated high temperature and strong light disaster identification branch for disaster identification, and comprehensively analyze the identification results to obtain the target disaster identification results.