Method for identifying different base original chrysanthemums based on characteristic smell difference of electronic nose
Through the characteristic odor difference analysis method based on electronic nose and combined with multiple data analysis methods, the accuracy and efficiency of chrysanthemum quality evaluation in the existing technology has been successfully solved, and the rapid and accurate identification of chrysanthemums of different basic roots has been achieved.
Patent Information
- Application Number
- CN202411517779.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-29
- Publication Date
- 2025-02-14
AI Technical Summary
The prior art has problems in the evaluation of chrysanthemum quality, which is complicated to handle and low accuracy of results. In particular, the identification of human olfactory sense is greatly affected by subjective factors and it is difficult to achieve rapid and accurate odor identification.
The characteristic odor difference analysis method based on electronic nose is used to collect chrysanthemum odor sensing information through the electronic nose detection system, combined with cluster analysis, principal component analysis, OPLS-DA analysis, MLP analysis and Fisher discriminant analysis method, verify and screen out Difference markers to identify chrysanthemums of different bases.
It realizes rapid and accurate identification of chrysanthemum odor, can effectively distinguish chrysanthemums from different bases, improves the accuracy and efficiency of quality evaluation, and has non-destructive testing.
Smart Images
Figure CN119446320A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of analysis and identification of Chinese herbal medicine pieces, and particularly relates to a method for identifying chrysanthemums of different origins based on characteristic odor differences of an electronic nose. Background Art
[0002] There are many varieties of chrysanthemums, and their quality varies. The smell information of most flower-based Chinese medicinal materials is closely related to the variety and quality. The smell of chrysanthemums of different varieties, origins and processing methods can reflect the differences in their volatile components, which is one of the important criteria for quality evaluation. At present, the evaluation methods of chrysanthemum quality are mainly based on traditional appearance sensory methods, and the evaluation of floral fragrance quality is mainly completed by traditional experienced people, but human olfactory identification is greatly affected by subjective factors, and it is time-consuming and laborious, and the results are not accurate. Modern volatile substance analysis methods include GC, GC-MS, etc., but there are problems such as long time, cumbersome processing, and high cost. Therefore, it is particularly necessary to seek a fast, effective, accurate and feasible method to assist in the identification of the smell of chrysanthemums.
[0003] Electronic nose technology (E-nose) is an analytical technology that simulates the human olfactory system. It uses a large number of sensor arrays to detect and analyze volatile components, and then obtains the "fingerprint data" of the overall information. This method of odor information analysis has the advantages of being fast, convenient, and low-cost, and has high practical value. It has been widely used in food testing, agriculture, botany and other fields. It can be seen that electronic nose technology is reliable and trustworthy in identifying material components through odor analysis. Summary of the invention
[0004] The purpose of the present invention is to provide a fast, effective and highly accurate method for identifying the smell of chrysanthemums, and further to distinguish chrysanthemums of different origins.
[0005] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:
[0006] The present invention provides a method for identifying chrysanthemums of different origins based on the difference in characteristic odors of an electronic nose, comprising the following steps:
[0007] (1) Selecting a chrysanthemum sample to be tested, and using an electronic nose detection system to collect and analyze chrysanthemum odor sensing information;
[0008] (2) performing cluster analysis and principal component analysis on the chrysanthemum odor sensing information in sequence to obtain classification and grouping of the chrysanthemum samples to be tested;
[0009] (3) using OPLS-DA analysis, MLP analysis and Fisher discriminant analysis to verify the correctness of the grouping described in step (2);
[0010] (4) performing OPLS-DA analysis and / or Fisher discriminant analysis on chrysanthemum according to the grouping described in step (3) to obtain differential markers;
[0011] (5) Identifying chrysanthemums of different origins according to the content of the differential markers.
[0012] Preferably, the chrysanthemum sample to be tested is in powder form and passed through a 10-20 mesh sieve, and the sampling method during the detection is headspace sampling.
[0013] Preferably, the odor sensors of the electronic nose detection system are metal oxide sensors, and the number of the odor sensors is ≥10.
[0014] Preferably, when the chrysanthemum samples to be tested are Boju, Jiaxiang yellow chrysanthemum, golden silk chrysanthemum, Jiaxiang white chrysanthemum, Huaiju, Hangju and Gongju, the chrysanthemum samples to be tested are classified into: Class I Boju, Jiaxiang yellow chrysanthemum, golden silk chrysanthemum and Class II Jiaxiang white chrysanthemum, Huaiju, Hangju and Gongju, among which Class II can be divided into Class IIa Jiaxiang white chrysanthemum, Huaiju and Class IIb Hangju and Gongju, which are divided into three groups in total.
[0015] Preferably, the MLP analysis in step (3) uses chrysanthemum odor sensing information as an input layer and the grouping of the chrysanthemum samples to be tested as an output layer.
[0016] Preferably, the Fisher discriminant analysis method in step (3) is to establish a Fisher function discriminant model for grouping the chrysanthemum samples to be tested using the chrysanthemum odor sensing information;
[0017] The Fisher function is:
[0018] Y1=-1584.301W1C-5.249W5S+1340.096W3C+8152.033W6S+2637.098W5C+1 91.461W1S+19.499W1W-576.156W2S+19.278W2W+3062.758W3S-6742.131;
[0019] Y2=-1584.256W1C-4.183W5S+1385.556W3C+8495.005W6S+2436.899W5C+1 87.224W1S+21.442W1W-600.622W2S+13.943W2W+3251.015W3S-7085.571;
[0020] Y3=-2417.663W1C-8.188W5S+830.718W3C+8039.501W6S+4002.632W5C+1 68.963W1S+19.387W1W-511.744W2S+30.469W2W+3445.627W3S-7321.494.
[0021] Preferably, during the OPLS-DA analysis in step (4), components with VIP values > 1 are selected as differential markers.
[0022] Preferably, in step (4), the differential markers are selected from high to low importance during the Fisher discriminant analysis.
[0023] Preferably, the differential markers in step (4) are inorganic sulfides W1W and aromatic components and organic sulfides W2W.
[0024] The present invention also provides the use of the differential marker W2W to evaluate the quality of chrysanthemums of different origins. The higher the content of the differential marker W2W, the richer the odor information of the chrysanthemums in this group, and the better the quality of the chrysanthemums.
[0025] The present invention firstly uses a variety of analysis methods such as cluster heat map, PCA, OPLS-DA, MLP and Fisher discriminant to find that the volatile components of different varieties of chrysanthemums are different, and can be divided into three groups according to the odor information (information olfactory matrix data of the electronic nose sensor); and then uses analysis methods such as OPLS-DA and Fisher discriminant to screen the main differential substances as W1W and W2W. Among them, W1W is highly correlated with inorganic sulfur components, and W2W is highly correlated with aromatic components, and aromatic components are recognized as important components of rose volatile substances, indicating that from the perspective of aroma, the third group of Hangju and Gongju have richer odor information, followed by Huaiju and Jiaxiang Baiju.
[0026] The present invention uses electronic nose technology to analyze and compare the types and quantities of odor information of different chrysanthemums, and classifies and evaluates the quality of different chrysanthemums; the present invention can also judge the quality of chrysanthemums by the changes in the chrysanthemum odor to ensure the consistency and high quality of the product; the present invention finds that the electronic nose sensor array has a good response capability to the chrysanthemum aroma, and can realize non-destructive detection of the chrysanthemum aroma. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 This is the clustering heat map of different varieties of chrysanthemum in Example 1.
[0028] Figure 2 This is the PCA analysis diagram of different varieties of chrysanthemum in Example 1.
[0029] Figure 3This is the OPLS-DA analysis of different varieties of chrysanthemum in Example 1.
[0030] Figure 4 This is the OPLS-DA score diagram of different groups of chrysanthemum in Example 2.
[0031] Figure 5 This is a 200 permutation test diagram of the OPLS-DA analysis in Example 2.
[0032] Figure 6 This is the VIP diagram of different components of chrysanthemum in the OPLS-DA analysis in Example 2.
[0033] Figure 7 This is the importance diagram of the three groups of chrysanthemum independent variables in Example 2.
[0034] Figure 8 This is a bar graph of the three groups of chrysanthemum difference markers in Example 2. DETAILED DESCRIPTION
[0035] The technical solutions provided by the present invention are described in detail below in conjunction with the embodiments, but they should not be construed as limiting the protection scope of the present invention.
[0036] Example 1
[0037] 1. Materials: Collect 7 kinds of chrysanthemums that are currently widely circulated in the market, namely Boju, Jiaxiang Huangju, Jinsihuangju, Jiaxiang Baiju, Huaiju, Hangju and Gongju, 3 batches of each variety, a total of 21 batches, and crush them through No. 1 sieve (10 mesh).
[0038] 2. Detection method: Use an electronic balance to weigh the crushed chrysanthemums, and 0.5 g of each batch is used as the chrysanthemum sample to be tested. Three parallel samples are placed in a 30 mL headspace bottle, and the bottle mouth is sealed with a sealing strip. Let it stand for 5 minutes, and an electronic nose is used to detect and collect the chrysanthemum odor sensing information - sensor information olfactory matrix data (electronic nose response value). The electronic nose contains an array of 10 metal oxide sensors.
[0039] 3. Data Analysis:
[0040] Origin 2018 software was used to perform cluster heat map analysis on the data; SIMCA 14.1 analysis software was used for principal component analysis (PCA) and orthogonal partial least-squares discrimination analysis (OPLS-DA); SPSS21.0 software was used for multilayer perceptron (MLP) and Fisher discriminant analysis; the above analysis software and data analysis methods were combined to analyze the odor of chrysanthemum, and then classified and evaluated its quality.
[0041] 3.1. Cluster heat map analysis:
[0042] The information olfactory matrix data of 10 sensors of chrysanthemum were normalized, and cluster heat map analysis was performed on the normalized data. The results are as follows Figure 1 As shown in the figure, chrysanthemum samples can be clustered into two categories, namely, Class I (Boju, Jiaxiang Huangju, Jinsihuangju) and Class II (Jiaxiang Baiju, Huaiju, Hangju, Gongju), and Class II can be clustered into Class IIa (Jiaxiang Baiju, Huaiju) and Class IIb (Hangju, Gongju). This shows that there are great differences in the volatile components of different varieties of chrysanthemums, and different chrysanthemums can be divided into three groups according to their odor.
[0043] 3.2. Principal component analysis (PCA analysis):
[0044] The data of 10 sensors of chrysanthemum were used as variables and imported into simca 14.1 software for PCA analysis. The results are as follows Figure 2 As shown in the figure, Jiaxiang Huangju, Jinsihuangju and Boju are clustered together and are not distinguished. Gongju, Hangju, Huaiju and Jiaxiang Baiju have a trend of distinction, and the horizontal coordinates of Jiaxiang Baiju and Huaiju are similar. This shows that there are differences between different varieties of chrysanthemums, but the differences between Jiaxiang Huangju, Jinsihuangju and Boju are not obvious, and the differences between Jiaxiang Baiju and Huaiju are small, which is similar to the results of cluster heat map analysis.
[0045] 3.3. Orthogonal partial least squares discriminant analysis (OPLS-DA analysis):
[0046] To further distinguish different varieties of chrysanthemum, chrysanthemum samples were subjected to OPLS-DA analysis. Figure 3As shown in the figure, Jiaxiang Huangju, Jinsihuangju and Boju are clustered together and are not distinguished, Gongju, Hangju, Huaiju and Jiaxiang Baiju are clustered together and are clearly distinguished, among which Jiaxiang Baiju and Huaiju have similar horizontal coordinates. This shows that there are differences among different varieties of chrysanthemums, and there is no obvious difference between Jiaxiang Huangju, Jinsihuangju and Boju, and the difference between Jiaxiang Baiju and Huaiju is small, which is consistent with the PCA analysis results.
[0047] 3.4.MLP analysis:
[0048] Through the above clustering heat map, PCA and OPLS-DA analysis, chrysanthemum can be divided into three groups, group 1 is Boju, Jiaxiang Huangju, Jinsihuangju, group 2 is Huaiju and Jiaxiang Baiju, and group 3 is Hangju and Gongju. MLP analysis based on backpropagation (BP) algorithm was adopted, with the information olfactory matrix data of 10 sensors of chrysanthemum as the input layer and the above three groups as the output layer. The original data set of the chrysanthemum samples to be tested obtained by electronic nose analysis was randomly divided into training set (70%) and test set (30%) for analysis. The results are shown in Table 1. Under this condition, the training accuracy and test accuracy of the model are both 100%, indicating that it is reasonable to divide chrysanthemum into these three groups. This model has a high accuracy rate in the rapid identification of the three groups of chrysanthemums, and can quickly, objectively and accurately distinguish the types of the three groups of chrysanthemums.
[0049] Table 1 MLP discriminant analysis results
[0050]
[0051] 3.5.Fisher Discriminant Analysis:
[0052] Fisher discriminant, also known as canonical discriminant, projects the independent variable combination of high-dimensional space into a space with lower dimensions, and then classifies it in the low-dimensional space. The Fisher function discriminant model and Fisher discriminant function were established using the three groups of chrysanthemums divided by the electronic nose response value pairs. The results of Fisher discriminant analysis are shown in Table 2:
[0053] The Fisher function is:
[0054] Y1=-1584.301W1C-5.249W5S+1340.096W3C+8152.033W6S+2637.098W5C+1 91.461W1S+19.499W1W-576.156W2S+19.278W2W+3062.758W3S-6742.131;
[0055] Y2=-1584.256W1C-4.183W5S+1385.556W3C+8495.005W6S+2436.899W5C+1 87.224W1S+21.442W1W-600.622W2S+13.943W2W+3251.015W3S-7085.571;
[0056] Y3=-2417.663W1C-8.188W5S+830.718W3C+8039.501W6S+4002.632W5C+1 68.963W1S+19.387W1W-511.744W2S+30.469W2W+3445.627W3S-7321.494;
[0057] Table 2 Fisher discriminant analysis results
[0058]
[0059] As shown in Table 2, the discrimination rates of the model's initial test and cross-validation were both 100%, indicating that the three groups of chrysanthemums could be accurately distinguished, with good discrimination effect.
[0060] In summary, through cluster heat map, PCA, OPLS-DA, MLP and Fisher discriminant analysis methods, it was repeatedly verified from multiple angles that there are differences in the volatile components of different varieties of chrysanthemum, and they can be divided into three groups according to the odor information. The first group is Boju, Jiaxiang Huangju, and Jinsihuangju, the second group is Huaiju and Jiaxiang Baiju, and the third group is Hangju and Gongju.
[0061] Example 2 Analysis of different components of chrysanthemum varieties
[0062] OPLS-DA analysis was performed on the chrysanthemums according to the chrysanthemum grouping obtained in Example 1. The results are as follows: Figures 4 to 6 As shown, the three groups of chrysanthemums were clearly distinguished. According to VIP>1, two differential markers were screened, namely W1W and W2W, namely inorganic sulfide (W1W), aromatic components and organic sulfide (W2W) ( Figure 4 is the OPLS-DA score graph; Figure 5 is the 200 permutation test graph; Figure 6 VIP diagram for different components).
[0063] To verify the rationality of the selected differential components, chrysanthemum was subjected to Fisher discriminant analysis by classification. The results are as follows Figure 7As shown, the two most important difference markers W1W and W2W were selected. The importance of W1W was 100% and W2W was 87.5%, which was consistent with the results of OPLS-DA analysis. It showed that the difference markers among the three groups of chrysanthemums were mainly W1W and W2W. Among them, the content of inorganic sulfide in group 2 was the highest, followed by group 3, and group 1 was the lowest; the aromatic substance components in group 3 were the richest, followed by group 2, and group 1 was the lowest. For details, see Figure 8 Among them, aromatic components are recognized as important components of rose volatile substances. Therefore, from the perspective of aroma, the odor information of the third group of Hangju and Gongju is relatively rich, followed by Huaiju and Jiaxiang Baiju, which have relatively rich odor information, between the first and third groups.
[0064] The present invention selects 7 mainstream varieties of chrysanthemum currently circulating on the market, analyzes the differences in volatile components between different varieties of chrysanthemums through electronic nose technology, and classifies and evaluates their quality. First, through cluster heat map, PCA, OPLS-DA, MLP and Fisher discriminant analysis methods, it is found that there are differences in the volatile components of different varieties of chrysanthemums, and they can be divided into three groups according to the odor information, the first group is Boju, Jiaxiang Huangju, Jinsihuangju, the second group is Huaiju and Jiaxiang Baiju, and the third group is Hangju and Gongju; then through OPLS-DA, Fisher discriminant and other analysis methods, the main differential substances are screened as W1W and W2W, among which W1W is highly correlated with inorganic sulfur components, and W2W is highly correlated with aromatic substances. It shows that from the perspective of aroma, the odor information of the third group of Hangju and Gongju is richer, which is also consistent with the evaluation of chrysanthemum quality by traditional sensory methods, followed by Huaiju and Jiaxiang Baiju, whose odor information is richer, between the first group and the third group. Therefore, electronic nose technology can be used to analyze and classify the differences of different varieties of chrysanthemums and conduct effective quality evaluation of chrysanthemums.
[0065] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for identifying different origin chrysanthemums based on the difference in characteristic odor of an electronic nose, characterized in that: The following steps are involved: (1) Selecting a chrysanthemum sample to be tested, and using an electronic nose detection system to collect and analyze chrysanthemum odor sensing information; (2) performing cluster analysis and principal component analysis on the chrysanthemum odor sensing information in sequence to obtain classification and grouping of the chrysanthemum samples to be tested; (3) using OPLS-DA analysis, MLP analysis and Fisher discriminant analysis to verify the correctness of the grouping described in step (2); (4) performing OPLS-DA analysis and / or Fisher discriminant analysis on chrysanthemum according to the grouping described in step (3) to obtain differential markers; (5) Identifying chrysanthemums of different origins according to the content of the differential markers.
2. The method according to claim 1, characterized in that: The chrysanthemum sample to be tested is in powder form and is sieved through a 10-20 mesh sieve. The sampling method during the test is headspace sampling.
3. The method according to claim 2, characterized in that: The odor sensors of the electronic nose detection system are metal oxide sensors, and the number of the odor sensors is ≥10.
4. The method according to claim 3, characterized in that: When the chrysanthemum samples to be tested are Boju, Jiaxiang yellow chrysanthemum, golden silk chrysanthemum, Jiaxiang white chrysanthemum, Huaiju, Hangju and Gongju, the chrysanthemum samples to be tested are classified into: Class I Boju, Jiaxiang yellow chrysanthemum, golden silk chrysanthemum and Class II Jiaxiang white chrysanthemum, Huaiju, Hangju and Gongju, among which Class II can be divided into Class IIa Jiaxiang white chrysanthemum, Huaiju and Class IIb Hangju and Gongju, which are divided into three groups in total.
5. The method according to claim 4, characterized in that: The MLP analysis in step (3) uses the chrysanthemum odor sensing information as the input layer and the grouping of the chrysanthemum samples to be tested as the output layer.
6. The method according to claim 4, characterized in that: The Fisher discriminant analysis method in step (3) is to use the chrysanthemum odor sensing information to group the chrysanthemum samples to be tested to establish a Fisher function discriminant model, and the Fisher function is: Y1=-1584.301W1C-5.249W5S+1340.096W3C+8152.033W6S+2637.098W 5C+191.461W1S+19.499W1W-576.156W2S+19.278W2W+3062.758W3S-6742.131; Y2=-1584.256W1C-4.183W5S+1385.556W3C+8495.005W6S+2436.899W 5C+187.224W1S+21.442W1W-600.622W2S+13.943W2W+3251.015W3S-7085.571; Y3=-2417.663W1C-8.188W5S+830.718W3C+8039.501W6S+4002.632W5 C+168.963W1S+19.387W1W-511.744W2S+30.469W2W+3445.627W3S-7321.
494.
7. The method according to claim 6, characterized in that During the OPLS-DA analysis in step (4), components with VIP values > 1 were selected as differential markers.
8. The method according to claim 7, characterized in that: In step (4), the Fisher discriminant analysis is performed to select differential markers based on their importance from high to low.
9. The method according to claim 8, characterized in that The differential markers in step (4) are inorganic sulfide W1W and aromatic components and organic sulfide W2W.
10. Use of the differential marker W2W in the method of claim 9 to evaluate the quality of chrysanthemums of different origins, characterized in that: The higher the content of the differential marker W2W, the richer the odor information of the chrysanthemums in this group and the better the quality of the chrysanthemums.