Volatile substance dynamic acquisition and analysis system and method for early detection of tomato root rot

Abstract

The invention relates to the technical field of agricultural disease control, in particular to a system and a method for dynamically collecting and analyzing volatile substances for early detection of tomato root rot. The method comprises the following steps: setting acquisition parameters, acquiring gas information of a plurality of targets according to the acquisition parameters, preprocessing to obtain preprocessed information, performing gas chromatography on the preprocessed information, creating a gas characteristic model in combination with a gas chromatography result, and inputting the acquired real-time gas information into the gas characteristic model. The root rot score of the real-time gas information output by the gas characteristic model is obtained, the collection period of the real-time gas information is judged according to the root rot score of the real-time gas information, and early warning is given out according to the collection period. And setting a relatively high weight for the gas type with high concentration, thereby calculating the root rot index of the gas information, and determining the root rot period of the gas information according to the root rot index.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of agricultural disease prevention and treatment, and particularly relates to a volatile substance dynamic collection and analysis system and method for early detection of tomato root rot. BACKGROUND

[0002] Root rot is a disease caused by fungi, which can cause root rot, gradually weaken the function of water and nutrient absorption, and finally die, mainly showing yellowing and wilting of the whole plant. At the early stage of the disease, only individual branch roots and fibrous roots are infected, and gradually spread to the main roots. After the main roots are infected, the function of water and nutrient absorption gradually weakens, and the aboveground part cannot recover at night due to insufficient nutrients. The new leaves turn yellow first, and when the disease is serious, the wilting condition cannot recover at night, and the whole plant turns yellow and withers. Root rot can cause plants to produce volatile gases.

[0003] The Chinese patent with the publication number CN114287424B discloses the application of volatile substances in the prevention and treatment of tomato neck rot and root rot, and discloses that acetoin and 2-heptanol have obvious inhibitory effect on the pathogenic bacteria of tomato neck rot and root rot, and can be used as effective ingredients for the prevention and treatment of tomato neck rot and root rot. However, the prior art does not mention how to detect the root rot of tomatoes, so that the root rot of tomatoes is difficult to be found in time, and the best treatment opportunity is easily missed. SUMMARY

[0004] The application aims to solve the problems in the background art and provides a volatile substance dynamic collection and analysis system and method for early detection of tomato root rot.

[0005] The technical scheme of the application comprises the following steps:

[0006] On the one hand, the application provides a volatile substance dynamic collection and analysis method for early detection of tomato root rot, which comprises the following steps:

[0007] Setting collection parameters, collecting gas information of multiple targets according to the collection parameters, and preprocessing the gas information to remove invalid gas information and obtain preprocessed information;

[0008] Performing gas chromatography analysis on the preprocessed information, and creating a gas feature model combined with the gas chromatography analysis result;

[0009] Collecting real-time gas information, inputting the collected real-time gas information into the gas feature model, and obtaining the root rot score of the real-time gas information output by the gas feature model;

[0010] Determining the collection period of the real-time gas information through the root rot score of the real-time gas information, and issuing a warning according to the collection period.

[0011] Preferably, the collection parameters are set, the gas information of the multiple targets is collected according to the collection parameters respectively, and the gas information is preprocessed to remove invalid gas information and obtain preprocessed information, including:

[0012] creating a gas data table;

[0013] corresponding collection parameters are set for each collection target; the collection parameters include a collection period and a collection frequency;

[0014] The gas data and gas samples of each collection target at different collection periods are collected according to the collection parameters, and all the collected gas data are put into the gas data table.

[0015] Preferably, the collection parameters are set, the gas information of the multiple targets is collected according to the collection parameters respectively, and the gas information is preprocessed to remove invalid gas information and obtain preprocessed information, further including:

[0016] selecting one collection target;

[0017] selecting one collection period of the collection target and all the gas data under the collection period from the gas data table;

[0018] setting a frequency threshold, and sequentially passing all the gas data under the collection period through the frequency threshold to filter the gas data and obtain a filtered signal;

[0019] calculating the standard deviation of the filtered signal, and setting a signal range based on the standard deviation;

[0020] judging whether the filtered signal is in the signal range;

[0021] if the filtered signal is not in the signal range, deleting the filtered signal;

[0022] if the filtered signal is in the signal range, recording the filtered signal as a preprocessed signal and putting it into the gas data table;

[0023] returning to selecting one collection period of the collection target and all the gas data under the collection period from the gas data table until all the collection periods of the collection target are selected and completed;

[0024] returning to selecting one collection target until all the collection targets are selected and completed, and obtaining multiple preprocessed signals of each collection target.

[0025] Preferably, the preprocessed information is subjected to gas chromatography analysis, and a gas characteristic model is created in combination with the gas chromatography analysis result, including:

[0026] selecting one preprocessed signal from the gas data table;

[0027] obtaining a gas sample corresponding to the preprocessed signal;

[0028] performing gas chromatography analysis on the gas sample to divide the gas sample into multiple sub-samples;

[0029] respectively performing ion mobility spectrum analysis on each sub-sample to obtain an analysis result of the gas sample;

[0030] returning to select a preprocessed signal from the gas data table until all preprocessed signals in the gas data table are selected to obtain the analysis result of the gas sample corresponding to each preprocessed signal.

[0031] Preferably, the preprocessed information is subjected to gas chromatography analysis, and a gas feature model is created in combination with the gas chromatography analysis result, further comprising:

[0032] creating a gas feature model;

[0033] For each collection target, multiple preprocessed signals of each collection period of each collection target and the analysis result of the gas sample corresponding to each preprocessed signal are obtained, and a coupling relationship of collection period-preprocessed signal-gas sample-analysis result of gas sample is created for each preprocessed signal, and the coupling relationship of collection period-preprocessed signal-gas sample-analysis result of gas sample is recorded as a training sample, thereby obtaining multiple training samples;

[0034] All training samples are divided into a training set and a test set according to a random ratio;

[0035] The training set is input into the gas feature model to train the gas feature model, and a trained gas feature model is obtained;

[0036] The test set is input into the trained gas feature model to verify whether the trained gas feature model is trained;

[0037] For each collection period, a gas weight corresponding to a reference gas is set respectively;

[0038] The root rot index of each collection period is calculated by a formula in combination with the gas weight corresponding to the reference gas;

[0039]

[0040] wherein, P i is the root rot score of the real-time gas information, Q i is the gas score of the i-th reference gas, A i is the weight parameter of the i-th reference gas, and M is the total number of reference gases.

[0041] Preferably, the training set is input into the gas feature model to train the gas feature model, to obtain a trained gas feature model, which comprises:

[0042] Randomly selecting one collection cycle from the training set;

[0043] The gas types contained in the analysis results of the gas samples of the collection cycle are sorted according to the sizes of the gas concentrations;

[0044] Selecting the gas types with the top M gas concentrations; and taking the top M gas types as the reference gas of the collection cycle.

[0045] Returning to randomly selecting one collection cycle from the training set, until all the collection cycles in the training set are selected, to obtain the reference gas of each collection cycle.

[0046] Preferably, real-time gas information is collected, and the collected real-time gas information is input into the gas feature model to obtain the root rot fraction of the real-time gas information output by the gas feature model, which comprises:

[0047] Collecting real-time gas information;

[0048] Performing gas chromatography analysis on the real-time gas information to obtain the analysis results of the real-time gas information;

[0049] Inputting the real-time gas information and the gas chromatography analysis results into the trained gas feature model, and obtaining the prediction results output by the trained gas feature model;

[0050] Calculating the similarity between the prediction results and the reference gas, and selecting the collection cycle corresponding to the reference gas with the highest similarity.

[0051] Preferably, the real-time gas information and the gas chromatography analysis results are input into the trained gas feature model, and the prediction results output by the trained gas feature model are obtained, which comprises:

[0052] Obtaining the same gas types of the real-time gas information and the reference gas of the corresponding collection cycle;

[0053] Calculating the root rot fraction of the real-time gas information by Formula 2;

[0054]

[0055] wherein X is the root rot fraction of the real-time gas information, Y a is the a-th gas type, A a is the weight parameter corresponding to the a-th gas type, and m is the total number of the same gas types in the real-time gas information and the reference gas of the corresponding collection cycle.

[0056] Preferably, the collection period of the real-time gas information is determined by the root rot fraction of the real-time gas information, and a warning is issued according to the collection period, including:

[0057] A root rot deviation threshold is set;

[0058] The difference between the root rot fraction of the real-time gas information and the root rot index of the alternative period is calculated;

[0059] It is determined whether the difference between the root rot fraction of the real-time gas information and the root rot index of the alternative period is less than or equal to the root rot deviation threshold;

[0060] If the difference between the root rot fraction of the real-time gas information and the root rot index of the alternative period is less than or equal to the root rot deviation threshold, it is determined that the collection period of the real-time gas is the same as the alternative period in the same root rot period.

[0061] On the other hand, the application also provides a volatile substance dynamic acquisition and analysis system for early detection of tomato root rot, comprising:

[0062] An acquisition component is used to acquire gas data;

[0063] A processing component, which includes a processing unit and a gas data unit, is used to control the gas data unit to analyze the gas data and input the analysis result to the processing unit, and the processing unit is used to execute the volatile substance dynamic acquisition and analysis method for early detection of tomato root rot according to any one of the preceding methods.

[0064] Compared with the prior art, the above technical solutions of the application have the following beneficial technical effects:

[0065] By setting the acquisition parameters, the gas information of multiple targets is acquired according to the acquisition parameters, and the gas information is preprocessed to obtain preprocessed information. Then, the preprocessed information is analyzed by gas chromatography, a gas characteristic model is created in combination with the gas chromatography analysis result, the real-time gas information acquired is input into the gas characteristic model, the root rot fraction of the real-time gas information output by the gas characteristic model is obtained, the collection period of the real-time gas information is determined by the root rot fraction of the real-time gas information, and a warning is issued according to the collection period. According to the application, the gas types contained in the gas information are sorted according to the concentration, and a higher weight is set for the gas type with high concentration, so as to calculate the root rot index of the gas information, and the root rot period of the gas information is determined according to the root rot index. BRIEF DESCRIPTION OF DRAWINGS

[0066] Figure 1 A flowchart of a volatile substance dynamic acquisition and analysis method for early detection of tomato root rot according to the application is shown;

[0067] Figure 2 A structure schematic diagram of a volatile substance dynamic acquisition and analysis system for early detection of tomato root rot disease is provided in the application.

[0068] Reference signs

[0069] 100, acquisition assembly; 200, processing assembly; 201, processing unit; 202, gas data unit. DETAILED DESCRIPTION

[0070] In an embodiment, as shown in the accompanying drawings, a volatile substance dynamic acquisition and analysis method for early detection of tomato root rot disease is provided in the application, which comprises: Figure 1

[0071] S100, setting acquisition parameters, respectively acquiring gas information of multiple targets according to the acquisition parameters, and preprocessing the gas information to remove invalid gas information and obtain preprocessed information;

[0072] S200, performing gas chromatography analysis on the preprocessed information, and creating a gas feature model in combination with the gas chromatography analysis result;

[0073] S300, acquiring real-time gas information, inputting the acquired real-time gas information into the gas feature model, and obtaining a root rot score of the real-time gas information output by the gas feature model;

[0074] S400, determining a collection cycle in which the real-time gas information is located through the root rot score of the real-time gas information, and issuing a warning according to the collection cycle.

[0075] In the application, the acquisition parameters are set, the gas information of multiple targets is respectively acquired according to the acquisition parameters, the gas information is preprocessed to obtain preprocessed information, the preprocessed information is then subjected to gas chromatography analysis, a gas feature model is created in combination with the gas chromatography analysis result, the acquired real-time gas information is then input into the gas feature model, a root rot score of the real-time gas information output by the gas feature model is obtained, the collection cycle in which the real-time gas information is located is determined through the root rot score of the real-time gas information, and a warning is issued according to the collection cycle. In the application, the gas types contained in the gas information are sorted according to the concentration, and a higher weight is set for the gas type with a higher concentration, so as to calculate a root rot index of the gas information, and thus the root rot cycle in which the gas information is located is determined according to the root rot index.

[0076] In an optional embodiment, the S100 comprises:

[0077] S110, creating a gas data table;

[0078] S120, setting a corresponding acquisition parameter for each acquisition target; the acquisition parameter comprises a collection cycle and a collection frequency.​

[0079] S130, collecting gas data and gas samples of each collection target in different collection periods according to the collection parameters, and putting all the collected gas data into a gas data table.

[0080] It should be noted that different collection periods correspond to different root rot periods, and the total period composed of all the collection periods should include the entire period of root rot, so that the subsequent gas feature model can make predictions for different root rot periods;

[0081] For example, the existing collection period A, collection period B and collection period C, collection period A corresponds to the early stage of root rot, collection period B corresponds to the middle stage of root rot, and collection period C corresponds to the late stage of root rot.

[0082] In an optional embodiment, the S100 further comprises:

[0083] S140, selecting a collection target;

[0084] S141, selecting a collection period of the collection target and all gas data under the collection period from the gas data table;

[0085] S142, setting a frequency threshold, and sequentially passing all gas data under the collection period through the frequency threshold to filter the gas data and obtain a filtered signal;

[0086] Specifically, the band-pass filtering method is a signal processing technique for allowing only signals within a specific frequency range to pass through from a group of frequencies, while attenuating signals of other frequencies to a very low level. The band-pass filter combines the characteristics of a high-pass filter and a low-pass filter, and realizes the frequency selection function by combining the two filters;

[0087] S143, calculating the standard deviation of the filtered signal, and setting a signal range based on the standard deviation;

[0088] Optionally, the maximum value of the signal range can be three times the standard deviation, that is, the signal range is created by the three-sigma method;

[0089] S144, determining whether the filtered signal is within the signal range;

[0090] S145, if the filtered signal is not within the signal range, deleting the filtered signal;

[0091] Specifically, when the filtered signal is deleted, the gas sample corresponding to the filtered signal is also deleted;

[0092] S146, if the filtered signal is within the signal range, recording the filtered signal as a preprocessed signal and putting it into the gas data table;

[0093] S147, return to select one collection period of the collection target and all gas data under the collection period from the gas data table until all collection periods of the collection target are selected and completed;

[0094] S148, return to select one collection target until all collection targets are selected and completed, to obtain multiple pre-processing signals of each collection target.

[0095] It should be noted that by setting the signal range of the signal, the abnormal values in the filtered signal are eliminated, so as to avoid affecting the training of the subsequent gas feature model, and at the same time, the number of samples can be reduced, thereby improving the training efficiency.

[0096] In an optional embodiment, the S200 comprises:

[0097] S210, selecting one pre-processing signal from the gas data table;

[0098] S220, obtaining the gas sample corresponding to the pre-processing signal;

[0099] S230, performing gas chromatography analysis on the gas sample to divide the gas sample into multiple sub-samples;

[0100] S240, respectively performing ion mobility spectrum analysis on each sub-sample to obtain the analysis result of the gas sample;

[0101] S250, return to select one pre-processing signal from the gas data table until all pre-processing signals in the gas data table are selected and completed, to obtain the analysis result of the gas sample corresponding to each pre-processing signal.

[0102] It should be noted that the present application detects the gas types contained in the gas sample by gas chromatography-ion mobility spectrum detection technology, and since different collection periods contain different gas types, the collection period corresponding to the gas sample can be determined according to the types of gas types, so as to further determine the root rot period corresponding to the gas sample.

[0103] The reason for using gas chromatography-ion mobility spectrum detection technology is that this technology combines the high separation efficiency of gas chromatography and the high sensitivity of ion mobility spectrum, and can accurately and timely determine volatile and semi-volatile organic compounds. Through the gas chromatography part, each component in the mixture is separated according to its chemical properties in the chromatographic column, and then these separated compounds enter the ion mobility spectrum part, and the ions move in the migration tube according to their respective mobility, thereby realizing secondary separation and detection.

[0104] In an optional embodiment, the S200 further comprises:

[0105] S260, creating a gas feature model;

[0106] S261, for each collection target, obtaining a plurality of pre-processing signals of each collection period of each collection target and an analysis result of a gas sample corresponding to each pre-processing signal, and creating a coupling relationship of collection period-pre-processing signal-gas sample-gas sample analysis result for each pre-processing signal, recording the coupling relationship of collection period-pre-processing signal-gas sample-gas sample analysis result as a training sample, thereby obtaining a plurality of training samples;

[0107] Specifically, since each collection period has a unique corresponding root rot period, after determining the collection period, the corresponding root rot period can be determined;

[0108] S262, dividing all training samples into a training set and a test set according to a random ratio;

[0109] Specifically, the proportion of the training set should be higher than that of the test set;

[0110] S263, inputting the training set into the gas feature model to train the gas feature model, and obtaining a trained gas feature model;

[0111] S264, inputting the test set into the trained gas feature model to verify whether the trained gas feature model is trained;

[0112] S265, for each collection period, setting a gas weight corresponding to a reference gas;

[0113] S266, combining the gas weight corresponding to the reference gas to calculate the root rot index of each collection period by a formula;

[0114]

[0115] Wherein, P i is the root rot score of the real-time gas information, Q i is the gas score of the i-th reference gas, A i is the weight parameter of the i-th reference gas, and M is the total number of reference gases;

[0116] Specifically, in order to ensure that the credibility of the root rot index for all root rot periods is the same, before calculating the root rot index by formula 1, the corresponding gas score of all gas types needs to be set;

[0117] When step S264 is performed, the reference gas of the collection period corresponding to each gas sample output by the gas feature model is obtained by inputting the gas samples in the test set and the preprocessed signals into the trained gas feature model, and the reference gas of the collection period corresponding to each gas sample output by the gas feature model is compared with the analysis result of the gas sample in the training set. If the results are consistent, it is proved that the gas feature model has been trained.

[0118] It should be noted that by creating and training the gas feature model, the gas feature model has the ability to output the collection period corresponding to the gas information according to the input gas data and gas sample. Then, the real-time gas information and the corresponding gas sample can be input into the gas feature model to obtain the collection period corresponding to the real-time gas information, so as to further determine the root rot period corresponding to the real-time gas information.

[0119] In an optional embodiment, S263 comprises:

[0120] K100, randomly selecting a collection period from the training set;

[0121] K101, sorting the gas types contained in the analysis result of the gas sample of the collection period according to the size of the gas concentration;

[0122] K102, selecting the gas types with the top M gas concentrations; and taking the top M gas types as the reference gas of the collection period;

[0123] K103, returning to step K102 until all the collection periods in the training set are selected, and the reference gas of each collection period is obtained;

[0124] It should be noted that different collection periods correspond to different root rot periods, and the types and concentrations of gas types contained in different root rot periods are different. The greater the concentration, the greater the proportion of the gas type in the corresponding root rot period, so the gas type with a large concentration is given a large weight, thereby reflecting the importance of the gas type in the corresponding root rot period.

[0125] After the weight corresponding to the reference gas is given, the root rot index of the collection period is calculated by combining the gas weight corresponding to the reference gas, and the root rot index is used as a judgment standard for identifying the period of the gas.

[0126] For example, the reference gases of the collection period A are gas 1, gas 2, gas 3, gas 4 and gas 5, the reference gases of the collection period B are gas 2, gas 4, gas 5, gas 6 and gas 9, and the reference gases of the collection period C are gas 1, gas 3, gas 6, gas 13 and gas 15. The root rot index of the collection period A, the root rot index of the collection period B and the root rot index of the collection period C are respectively calculated by formula 1 according to the gas scores of the gases 1 to 6 and the weights corresponding to each gas type

[0127] In an optional embodiment, the S300 comprises:

[0128] S310, collecting real-time gas information;

[0129] S320, performing gas chromatography analysis on the real-time gas information to obtain an analysis result of the real-time gas information;

[0130] S330, inputting the real-time gas information and the gas chromatography analysis result into the trained gas feature model and obtaining a prediction result output by the trained gas feature model;

[0131] S340, calculating the similarity between the prediction result and the reference gas, and selecting a collection period corresponding to a reference gas with the highest similarity; the collection period corresponding to the reference gas with the highest similarity is recorded as a candidate period.

[0132] It should be noted that, since the gas feature model does not include irrelevant gas types, even if the real-time gas information includes gas types irrelevant to the root rot period, these irrelevant gas types will not affect the root rot score calculated by the real-time gas information, so that the application can avoid the interference of irrelevant gas types to improve the accuracy of the root rot period determination.

[0133] After inputting the real-time gas information into the gas feature model, if the reference gas of the real-time gas information output by the gas feature model is gas 1, gas 2, gas 4, gas 5 and gas 9, since the reference gas of the real-time gas information has the highest similarity with the type of the reference gas of the collection period B. Then it can be preliminarily determined that the candidate period of the real-time gas information is the collection period B.

[0134] In an optional embodiment, the S300 further comprises:

[0135] S350, screening the gas types included in the real-time gas information which are the same as the reference gas of the candidate period;

[0136] S360, calculating the root rot score of the real-time gas information by formula 2;

[0137]

[0138] wherein X is the root rot score of the real-time gas information, Y a is the gas score of the ath gas type in the real-time gas information, A a is the weight parameter corresponding to the ath gas type, and m is the total number of gas types in the real-time gas information that are the same as the reference gas of the corresponding collection period;

[0139] Specifically, the weight parameter corresponding to each gas type is uniquely determined within the collection period, so after the preliminary confirmation of the candidate period of the real-time gas information by step S340, the weight parameter corresponding to the gas type of the candidate period can be substituted into the real-time gas information and calculated by formula 2.

[0140] It should be noted that since the candidate period of the real-time gas information has been preliminarily determined as the collection period B by the foregoing embodiment, the weight parameter corresponding to the gas type of the collection period B is substituted into formula 2, and the root rot score of the real-time gas information is calculated by formula 2.

[0141] The root rot score is a numerical value used to calculate the similarity of the root rot index of the collection period and the candidate period in which the real-time gas information is located. The closer the root rot score of the real-time gas information is to the root rot index of the collection period, the closer the collection period in which the real-time gas information is located is to the candidate period.

[0142] In an optional embodiment, the S400 comprises:

[0143] S410, setting a root rot deviation threshold;

[0144] S420, calculating the difference between the root rot score of the real-time gas information and the root rot index of the candidate period;

[0145] S430, determining whether the difference between the root rot score of the real-time gas information and the root rot index of the candidate period is less than or equal to the root rot deviation threshold;

[0146] S440, if the difference between the root rot score of the real-time gas information and the root rot index of the candidate period is less than or equal to the root rot deviation threshold, determining that the collection period in which the real-time gas is located and the candidate period are in the same root rot period;

[0147] Specifically, "the collection period in which the real-time gas is located and the candidate period are in the same root rot period" does not refer to the same period in the collection time, but refers to the unity of the tomato root rot period, i.e., if the difference between the root rot score of the real-time gas information and the root rot index of the candidate period is less than or equal to the root rot deviation threshold, and the candidate period is in the middle of the root rot period, then it can also be considered that the corresponding collection period of the real-time gas information is also in the root rot period.

[0148] Optionally, if the difference between the root rot fraction of the real-time gas information and the root rot index of the alternative period is greater than the root rot deviation threshold, it is determined that the collection period in which the real-time gas information is located and the alternative period are not in the same period, and step S310 is returned to.

[0149] Optionally, since there may be multiple alternative periods, in this case, the weight corresponding to the gas type contained in each alternative period needs to be respectively brought into formula 2, and the root rot fraction of the real-time gas information is calculated multiple times under different weights by formula 2, and the alternative period with the smallest difference value of the root rot fraction of the gas information is selected as the corresponding period.

[0150] For example, the existing alternative periods are collection period A and collection period D, and the weight parameters corresponding to the gas types contained in collection period A and collection period D are shown in Table-1 gas type weight table.

[0151] Table-1 gas type weight table

[0152]

[0153] Then, the weight corresponding to the gas type contained in collection period A and the weight corresponding to the gas type contained in collection period D are respectively brought into formula 2, and the root rot fraction of the real-time gas information in the case of the alternative period being collection period A and the root rot fraction of the real-time gas information in the case of the alternative period being collection period D are calculated respectively according to the gas types contained in the real-time gas information. The root rot fraction of the real-time gas information in the case of the alternative period being collection period A is recorded as root rot fraction A, and the root rot fraction of the real-time gas information in the case of the alternative period being collection period D is recorded as root rot fraction D.

[0154] Then, the difference value a between root rot fraction A and the root rot index of collection period A, and the difference value d between root rot fraction D and the root rot index of collection period D are calculated respectively. If the difference value a and the difference value d both satisfy the root rot deviation threshold, the collection period with the smaller difference value is selected as the corresponding period. If one of the difference value a and the difference value d does not satisfy the root rot deviation threshold, the collection period that satisfies the root rot deviation threshold is selected as the corresponding period. If the difference value a and the difference value d both do not satisfy the root rot deviation threshold, step S310 is returned to.

[0155] It should be noted that after the root rot fraction of the real-time gas information is calculated, the similarity between the collection period in which the real-time gas information is located and the alternative period is observed by judging the difference between the root rot fraction of the real-time gas information and the root rot index of the alternative period. The smaller the difference is, the higher the similarity is.

[0156] The application sets double detection, first, the similarity between the gas type contained in the real-time gas information and the gas type contained in all collection periods is checked to preliminarily screen out the candidate period which is close to the collection period where the real-time gas information is located, then the difference between the root rot fraction of the real-time gas information and the root rot index of the candidate period is calculated to determine the closest collection period, and the collection period where the real-time gas information is located and the candidate period are in the same period, through double detection, the accuracy and reliability of the detection of the tomato root rot period are improved.

[0157] The application also provides a volatile substance dynamic collection and analysis system for early detection of tomato root rot, comprising a collection assembly 100 and a processing assembly 200.

[0158] The gas data is collected by the collection assembly 100, the processing assembly 200 comprises a processing unit 201 and a gas data unit 202, the gas data unit 202 is controlled by the processing unit 201 to analyze the gas data, and the analysis result is input to the processing unit 201, and the processing unit 201 executes the volatile substance dynamic collection and analysis method for early detection of tomato root rot in embodiment one;

[0159] Specifically, the collection target in the application should contain the whole period of tomato root rot, so as to ensure that the collected gas information can cover the whole period of tomato root rot.

[0160] It should be noted that the gas information of the multiple collection targets is collected by the multiple collection assemblies 100 respectively, and the collected gas information is input to the processing assembly 200, the gas information is processed by the gas data unit 202 of the processing assembly 200, so as to calculate the root rot index of the collection target, and finally the processing unit 201 judges whether the early warning is needed according to the root rot index.

[0161] The embodiments of the application are described in detail above in combination with the drawings, but the application is not limited thereto, and various changes can be made within the knowledge range of those skilled in the art without departing from the purpose of the application.

Claims

1. A method for dynamic collection and analysis of volatile substances for early detection of tomato root rot, characterized in that, The application relates to a method for collecting and analyzing gas information of multiple targets, comprising the following steps: Setting acquisition parameters, collecting gas information of multiple targets according to the acquisition parameters, and preprocessing the gas information to remove invalid gas information and obtain preprocessed information; Performing gas chromatography analysis on the preprocessed information, and creating a gas feature model in combination with the gas chromatography analysis result; Collecting real-time gas information, inputting the collected real-time gas information into the gas feature model, and obtaining a root rot fraction of the real-time gas information output by the gas feature model; Judging the collection cycle in which the real-time gas information is located through the root rot fraction of the real-time gas information, and issuing a warning according to the collection cycle.

2. The method for dynamic collection and analysis of volatile substances for early detection of tomato root rot according to claim 1, characterized in that, Setting acquisition parameters, collecting gas information of multiple targets according to the acquisition parameters, and preprocessing the gas information to remove invalid gas information and obtain preprocessed information, comprising the following steps: Creating a gas data table; Setting corresponding acquisition parameters for each collection target; the acquisition parameters include a collection cycle and a collection frequency; Collecting gas data and gas samples of each collection target in different collection cycles according to the acquisition parameters, and placing all the collected gas data into the gas data table.

3. The method for dynamic collection and analysis of volatile substances for early detection of tomato brown root rot according to claim 2, characterized in that, Setting acquisition parameters, collecting gas information of multiple targets according to the acquisition parameters, and preprocessing the gas information to remove invalid gas information and obtain preprocessed information, further comprising the following steps: Selecting one collection target; Selecting one collection cycle of the collection target and all the gas data in the collection cycle from the gas data table; Setting a frequency threshold, and sequentially passing all the gas data in the collection cycle through the frequency threshold to filter the gas data and obtain a filtered signal; Calculating the standard deviation of the filtered signal, and setting a signal range based on the standard deviation; Judging whether the filtered signal is in the signal range; If the filtered signal is not in the signal range, deleting the filtered signal; If the filtered signal is in the signal range, recording the filtered signal as a preprocessed signal and placing the preprocessed signal into the gas data table; Returning to the step of selecting one collection cycle of the collection target and all the gas data in the collection cycle from the gas data table until all the collection cycles of the collection target are selected and completed; Returning to the step of selecting one collection target until all the collection targets are selected and completed, and obtaining multiple preprocessed signals of each collection target.

4. The method for dynamic collection and analysis of volatile substances for early detection of tomato brown root rot according to claim 3, characterized in that, Performing gas chromatography analysis on the preprocessed information, and creating a gas feature model in combination with the gas chromatography analysis result, comprising the following steps: Selecting one preprocessed signal from the gas data table; Obtaining a gas sample corresponding to the preprocessed signal; Performing gas chromatography analysis on the gas sample to divide the gas sample into multiple sub-samples; Performing ion mobility spectrum analysis on each sub-sample to obtain an analysis result of the gas sample; Returning to the step of selecting one preprocessed signal from the gas data table until all the preprocessed signals in the gas data table are selected and completed, and obtaining an analysis result of the gas sample corresponding to each preprocessed signal.

5. The method for dynamic collection and analysis of volatile substances for early detection of tomato brown root rot according to claim 4, characterized in that, Performing gas chromatography analysis on the preprocessed information, and creating a gas feature model in combination with the gas chromatography analysis result, further comprising the following steps: Creating a gas feature model; For each acquisition target, the analysis result of each pre-processing signal of each acquisition cycle of each acquisition target and the analysis result of the corresponding gas sample are obtained, and a coupling relationship of acquisition cycle-pre-processing signal-gas sample-analysis result of the gas sample is created for each pre-processing signal, and the coupling relationship of acquisition cycle-pre-processing signal-gas sample-analysis result of the gas sample is recorded as a training sample, thereby obtaining a plurality of training samples; All training samples are divided into a training set and a test set according to a random ratio; The training set is input into the gas feature model to train the gas feature model, and a trained gas feature model is obtained; The test set is input into the trained gas feature model to verify whether the trained gas feature model is trained; For each acquisition cycle, the gas weight corresponding to the reference gas is set respectively; The root rot index of each acquisition cycle is calculated by combining the gas weight corresponding to the reference gas through a formula; where P i is the root rot fraction of the real-time gas information, Q i is the gas score of the i-th reference gas, A i is the weight parameter of the i-th reference gas, and M is the total number of reference gases.

6. The method for dynamic collection and analysis of volatile substances for early detection of tomato brown root rot according to claim 5, characterized in that, The training set is input into the gas feature model to train the gas feature model, and a trained gas feature model is obtained, including: A acquisition cycle is randomly selected from the training set; The gas types contained in the analysis result of the gas sample of the acquisition cycle are sorted according to the size of the gas concentration; The gas types with the top M gas concentrations are selected; the top M gas types are recorded as the reference gas of the acquisition cycle; The acquisition cycle is randomly selected from the training set until all acquisition cycles in the training set are selected, and the reference gas of each acquisition cycle is obtained.

7. The method according to claim 6, wherein the method is characterized by, Real-time gas information is collected, and the collected real-time gas information is input into the gas feature model to obtain the root rot fraction of the real-time gas information output by the gas feature model, including: Collecting real-time gas information; Performing gas chromatography analysis on the real-time gas information to obtain the analysis result of the real-time gas information; The real-time gas information and the gas chromatography analysis result are input into the trained gas feature model, and a prediction result output by the trained gas feature model is obtained; The similarity between the prediction result and the reference gas is calculated, and the acquisition cycle corresponding to the reference gas with the highest similarity is selected.

8. The method for dynamic collection and analysis of volatile substances for early detection of tomato brown root rot according to claim 7, characterized in that, The real-time gas information and the gas chromatography analysis result are input into the trained gas feature model, and a prediction result output by the trained gas feature model is obtained, including: The same gas types as the reference gas of the corresponding acquisition cycle are obtained; The root rot fraction of the real-time gas information is calculated through formula 2; Wherein, X is the root rot percentage of the real-time gas information, Y a is the a-th gas type, A a is the weight parameter corresponding to the a-th gas type, and m is the total number of the gas types in the real-time gas information which are the same as the reference gas of the corresponding collection period.

9. The method for dynamic collection and analysis of volatile substances for early detection of tomato brown root rot according to claim 8, characterized in that, The acquisition cycle of the real-time gas information is determined through the root rot fraction of the real-time gas information, and a warning is issued according to the acquisition cycle, including: Setting a root rot deviation threshold; Calculating the difference between the root rot fraction of the real-time gas information and the root rot index of the candidate cycle; Determining whether the difference between the root rot fraction of the real-time gas information and the root rot index of the candidate cycle is less than or equal to the root rot deviation threshold; If the difference between the root rot fraction of the real-time gas information and the root rot index of the candidate cycle is less than or equal to the root rot deviation threshold, it is determined that the acquisition cycle of the real-time gas and the candidate cycle are in the same root rot cycle.

10. A volatile substance dynamic collection and analysis system for early detection of tomato foot rot, characterized in that, including: A collection component, through which gas data is collected; The processing assembly comprises a processing unit and a gas data unit, the gas data unit analyzes the gas data under the control of the processing unit, and inputs the analysis result to the processing unit, and the processing unit executes the volatile substance dynamic collection and analysis method for early detection of tomato root rot according to any one of claims 1-9.

Images