NL2027699B1 - A Multi-dimensional Integration Identification Method for Fermentation Degree of Meixiang Salted Fish Based on Odour Visualization - Google Patents

Abstract

The invention discloses a multi-dimensional integration identification method for fermentation degree of Meixiang salted fish based on odour visualization. In this invention, an effective multi-dimensional identification model for fermentation degree was established by integrating electronic nose system (E-nose), gas chromatography -ion mass spectrometry (GC-IMS) and multi-dimensional data fusion analysis. Specifically, the method integrates data analysis methods such as neural network algorithm and PCA analysis to construct a quick identification model for the fermentation degree of Meixiang salted fish. With low identification cost and high visualization of results, the method is suitable for quality evaluation and fermentation degree differentiation of Meixiang salted fish. Compared with the traditional sensory analysis and single flavour analysis, the invention based on the multi- dimensional integration identification and prediction strategy, shows high sensitivity, short analysis time, convenient operation and high predictability, and meets the requirements of rapid detection and analysis of modern aquatic foods.

Description

A Multi-dimensional Integration Identification Method for Fermentation Degree of Meixiang Salted Fish Based on Odour Visualization

TECHNICAL FIELD The invention is a multi-dimensional integration identification method for fermentation degree of Meixiang salted fish based on odour visualization in the technical field of rapid analysis and detection of aquatic food.

BACKGROUND The traditional fermented and pickled dried fish products with Chinese characteristics mainly 10° include Guangdong Meixiang salted fish, Anhui salted mandarin fish, Miao and Dong sour fish and Shaoxing vinasse fish. Specifically, Meixiang salted fish, also known as “enzymatic salted fish”, is a traditional solid natural fermented fish product, and its production is mainly distributed in southeast coastal areas such as Guangdong and Fujian. According to statistics, in 2018, the total output of dried salted fish products in China reached 1.62 million tons, accounting for 7.51% of the total processed aquatic products (21.57 million tons) and 9.13% of the processed seawater aquatic products (17.75 million tons). Traditional Meixiang salted fish 1s favoured by consumers because of its long history, rich nutrition and unique flavour. However, due to the lack of systematic theoretical support system, the automation level of Meixiang salted fish production is low. Therefore, one of the key problems is to realize the targeted process control and quality monitoring of traditional Meixiang salted fish, which urgently need to be solved in aquatic products processing industry. The fermentation degree of Meixiang salted fish is the key process for its quality. There are great differences in the internal nutritional components of Meixiang fish with different degrees of fermentation. A series of complex material and energy exchanges between microorganisms and metabolites occur in its fermentation system. At present, the identification of the fermentation degree of Meixiang salted fish still depends on the artificial experience sensory discrimination such as colour, hardness and smell, and there are some problems such as strong subjectivity, low standardization and unstable product quality. The characteristic flavour is the “skeleton” of traditional Meixiang salted fish fermentation quality identification. The odour visualization multi-dimensional integration technology can perform multi-dimensional integration analysis on response signals, fingerprints and characteristic volatile organic compounds (VOC), and construct an identification model for the fermentation degree of traditional Meixiang salted fish, which is an effective strategy to realize Meixiang salted fish quality monitoring. Different from artificial sensory analysis and single volatile flavour analysis, multi-dimensional integration technology with odour 3 visualization can identify the fermentation degree of traditional Meixiang salted fish more objectively and efficiently. At present, there is no technical report on using odour visualization and multi-dimensional data fusion technology to identify the fermentation degree of traditional Meixiang salted fish.

SUMMARY 10° In order to solve the above problems, the present invention provides a multi-dimensional integration identification method for fermentation degree of Meixiang salted fish based on odour visualization. In this invention, an effective multi-dimensional identification model for fermentation degree 1s established by integrating E-nose system, GC-IMS and multi- dimensional data fusion analysis (neural network algorithm, PCA analysis and random forest) together. Compared with the traditional sensory analysis and single flavour analysis, the invention based on the multi-dimensional integration identification and prediction strategy has the advantages of high sensitivity, short analysis time, convenient operation and high predictability, and meets the requirements of rapid detection and analysis of modern aquatic foods.

The invention provides a multi-dimensional integration identification method for fermentation degree of Meixiang salted fish based on odour visualization, which comprises the following steps; S1. Preparation of detection samples of Meixiang salted fish in different fermentation stages. Removing skin and bones from the detection samples of Meixiang salted fish. Then 2 em x 20m x 1 cm block samples is taken from the starting point to the tail direction with the middle line and the upper surface layer as the base point, the block samples are cooled with liquid nitrogen and ground for 5-8s with 3 repetitions to obtain a powdery sample of Meixiang salted fish; S2. the E-nose system is used to collect olfactory fingerprint spectrum of the powdery sample of Meixiang salted fish, then the characteristic response value of the gas sensor of the E-nose is extracted to construct an odour characteristic model,

S3. 2g of Meixiang salted fish powder sample is subjected to headspace incubation at 60 °C for 10 min, and then extracted by a GC-IMS device to obtain ion signal response peaks of characteristic VOC of the Meixiang salted fish powder sample, then based on NIST database and IMS database, qualitative analysis is carried out on the characteristic VOC through > corresponding peaks of ion signals to obtain the characteristics of VOC; S4. based on the characteristics of VOC, a three-dimensional map and a two-dimensional overhead map are acquired through a reporter plug-in, then according to the colour difference of signal intensity of VOC characteristic in the three-dimensional map and the two- dimensional overhead map, the characteristic model of VOC is constructed; g5 bBased on the odour characteristic model and the characteristic VOC model, a multi- dimensional identification model of the fermentation degree of the Meixiang salted fish is constructed, wherein the multi-dimensional identification model of the fermentation degree of the Meixiang salted fish comprises a determination coefficient and a root mean square error (RMSE). Specifically, the determination coefficient is large and the RMSE is small, the 135 multi-dimensional identification model of the fermentation degree of the Meixiang salted fish 1s used for identifying the actual fermentation degree of the Meixiang salted fish samples. Preferably, the detection samples of Meixiang salted fish include 5 samples fermented for 0, 3,6,9, 12, 15, 18, 21 and 24 days respectively. The block samples to be measured are 5 fish blocks with width of 2 cm.

preferably, the characteristics of VOC in the detection samples of Meixiang salted fish fermented for O day include butyraldehyde, heptanal, nonanal, 3-methyl butyraldehyde and 2, 3-butanedione. The VOC in the Meixiang salted fish samples fermented for 3-9 days include butanone, ethyl butyrate, n-butyl butyrate, propyl 2-methylbutyrate, ethyl 2- methylbutyrate, ethyl acetate and ethyl propanoate. VOC in the samples of Meixiang salted fish fermented for 10-18 days include n-propyl butyrate and 3-nonanone. The VOC in the samples of Meixiang salted fish fermented for 19-24 days include octanal, benzaldehyde, 1- octen-3-ketone, 2, 3, 5-trimethylpyrazine, cyclohexene -2-ketone and 2, 6-Dimethylpyrazine. Preferably, the E-nose comprises 18 gas sensors at least. The characteristic response value 1s the average value of the response signal and the integral value of the response message of the gas sensor;

Gas sensors at least include LY2/LG, P30/2, T40/2, LY2/G, LY2/AA, T30/1, PI0/2, LY2/GH, LY2/gCTL, LY2/gCT, P40/1, T70/2, PA/2, P30/1, P40/2, T40/1P10/1 and TA/2. Preferably, S2 further comprises the following steps that the E-nose should be cleaned with ultra-pure dry air with the flow rate of 300 mL/min and the equilibrium time of 10 min before using.

1.00 g of Meixiang salted fish powder sample is placed in a 15mL headspace bottle, sealed with a lid, incubated at 4 °C for 0.5 h. And then the headspace gas is sucked for detection and analysis. The detection flow rate is set at 150 mL/min, the sample collection time is 120 s, and the delay time is 10 min. 810 data points are obtained through 5 biological repetitions. Based on the data points, the characteristic response values are obtained. Preferably, the construction method of the multi-dimensional identification model of the fermentation degree of Meixiang salted fish at least includes PCA analysis, PLS-DA analysis, neural network algorithm, random forest algorithm and steepest climbing algorithm. The coefficient of determination and RMSE are calculated as follows: Rs JAAP LL Vy a ZX YE Wherein, R? represents the determination coefficient, RMSE represents the RMSE, Xi represents the actual fermentation degree of the i sample in the model building process, X represents the average value of the actual fermentation degree of all samples in the model building process, Y; represents the predicted fermentation degree of the ith sample in the model building process, Y represents the average value of the predicted fermentation degree of all samples in the model building process, and N represents the sample quantity with known fermentation time used in the model building process. The positive progress effect of the invention is described below. The method disclosed by the invention breaks through the shortcoming, such as strong subjectivity, low standardization degree, unstable product quality and so on, that exist in the existing artificial empirical sensory discrimination. Moreover, multi-dimensional integration odour visualization technology 1s used to identify the fermentation degree of Meixiang salted fish, which is time-consuming and easy to operate, and is suitable for targeted process control and quality monitoring in industrial production of Meixiang salted fish.

5 The method of the invention integrates neural network algorithm, PCA analysis and other data analysis methods to establish a quick identification model of the fermentation degree of the Meixiang salted fish, which is suitable for the quality evaluation and the differentiation of the fermentation degree of the Meixiang salted fish, and has low identification cost and high visualization degree of results.

BRIEF DESCRIPTION OF THE FIGURES Figure 1 1s an E-nose flavour profile in the fermentation process of Meixiang salted fish according to the present invention. Figure 2 is a qualitative result of VOC three-dimensional fingerprint spectrum m the fermentation process of Meixiang salted fish according to the present invention. 13 Figure 3 is the VOC spectrogram (top view) in the fermentation process of Meixiang salted fish according to the present invention. Figure 4 1s a contrast difference spectrogram of VOC in the fermentation process of Meixiang salted fish according to the present invention. Figure 5 1s a GalleryPlot fingerprint spectrum in the fermentation process of Meixiang salted fish according to the present invention. Figure 6 is an E-nose PCA analysis in the fermentation process of Meixiang salted fish according to the present invention. Figure 7 1s a PCA analysis of gas migration mass spectrometry in the fermentation process of Meixiang salted fish according to the present invention.

DESCRIPTION OF THE INVENTION In order to make the purpose, technical scheme and advantages of the embodiments of this application clearer, the technical scheme of the embodiments of this application will be described clearly and completely with reference to the figures in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. The components of the embodiments of the present application generally described and illustrated in the figures, which may be arranged and designed in various different configurations.

Therefore, the following detailed description of the embodiments of the application provided in the figures is not intended to limit the scope of the claimed application, but only represents selected embodiments of the application.

For the 3 embodiment of the application, all other embodiments obtained by those skilled in the art without making creative work should belong to the protection scope of the application.

As shown in Figures 1-7, the present invention provides a multi-dimensional integration identification method for the fermentation degree of Meixiang salted fish based on odour visualization, which comprises the following steps. 10° $1. Preparation of detection samples of Meixiang salted fish in different fermentation stages.

Removing skin and bones from the detection samples of Meixiang salted fish.

Then 2 cm x 2 cm x 1 cm block samples was taken from the starting point to the tail direction with the middle line and the upper surface layer as the base point.

The block samples were cooled with liquid nitrogen and then grinding for 5-8s with 3 repetitions to obtain a powdery sample of 13 Meixiang salted fish.

S2. Based on the powdery sample of the Meixiang salted fish, The E-nose system was used to collect its olfactory fingerprint spectrum.

Then the characteristic response values of the gas sensor of the E-nose were extracted to construct an odour characteristic model.

S3. 2g of Meixiang salted fish powder sample is subjected to headspace incubation at 60 °C for 10 min, and then extracted by a GC-IMS device to obtain ion signal response peaks of characteristic VOC of the Meixiang salted fish powder sample.

Then based on NIST database and IMS database, qualitative analysis was carried out on the characteristic VOC through corresponding peaks of ion signals to obtain the characteristics of VOC.

S4. Based on the characteristics of VOC, a three-dimensional map and a two-dimensional overhead map were acquired through a reporter plug-in.

Then according to the colour difference of signal intensity of VOC characteristic in the three-dimensional map and the two-dimensional overhead map, the characteristic model of VOC is constructed.

S5. Based on the odour characteristic model and the characteristic VOC model, a multi- dimensional identification model of the fermentation degree of the Meixiang salted fish was constructed.

Wherein the multi-dimensional identification model of the fermentation degree of the Meixiang salted fish comprised a determination coefficient and a RMSE. Specifically, the determination coefficient was large and the RMSE was small. The multi-dimensional identification model of the fermentation degree of the Meixiang salted fish was used for identifying the actual fermentation degree of the Meixiang salted fish samples.

The detection samples of Meixiang salted fish include 5 samples fermented for 0, 3, 6, 9, 12, 15, 18, 21 and 24 days respectively. The block samples to be measured were 5 fish blocks with width of 2 cm.

The characteristics of VOC in the Meixiang salted fish samples fermented for 0 days included butyraldehyde, heptanal, nonanal, 3-methyl butyraldehyde and 2, 3-butanedione. The VOC 10° in the Meixiang salted fish samples fermented for 3-9 days included butanone, ethyl butyrate, n-butyl butyrate, propyl 2-methylbutyrate, ethyl 2-methylbutyrate, ethyl acetate and ethyl propanoate. VOC in the samples of Meixiang salted fish fermented for 10-18 days included n-propyl butyrate and 3-nonanone. The VOC in the samples of Meixiang salted fish fermented for 19-24 days included octanal, benzaldehyde, 1-octen-3-ketone, 2, 3, 5- 13 trimethylpyrazine, cyclohexene -2-ketone and 2, 6-Dimethylpyrazine. Preferably, the E-nose comprised at least 18 gas sensors. The characteristic response value was the average value of the response signal and the integral value of the response message of the gas sensor; Gas sensors at least included LY2/LG, P30/2, T40/2, LY2/G, LY2/AA, T30/1, P10/2, LY2/GH, LY2/gCTL, LY2/gCT, P40/1, T70/2, PA/2, P30/1, P40/2, T40/1P10/1 and TA/2. In Step 2, the E-nose should be further cleaned with ultra-pure dry air with an air flow rate of 300 mL/min and an equilibrium time of 10 min before using. Furthermore, in Step 2 1.00 g of Meixiang salted fish powder sample was placed ina 15mL headspace bottle, sealed with a lid, incubated at 4 °C for 0.5 h. And then the headspace gas is sucked for detection and analysis. The detection flow rate was set at 150 mL/min, the sample collection time was 120 s, and the delay time was 10 min. 810 data points were obtained through 5 biological repetitions. Based on the data points, the characteristic response values were obtained.

The construction method of the multi-dimensional identification model of the fermentation degree of Meixiang salted fish at least included PCA analysis, PLS-DA analysis, neural network algorithm, random forest algorithm and steepest climbing algorithm.

The coefficient of determination and RMSE were calculated as follows: de AEX ALY TV} o Hag var Fun RMSE- EEL (XT)? Wherein, R? represents the determination coefficient, RMSE represents the RMSE, X; represents the actual fermentation degree of the i sample in the model building process, X represents the average value of the actual fermentation degree of all samples in the model building process, Yi represents the predicted fermentation degree of the i™ sample in the model building process, Y represents the average value of the predicted fermentation degree of all samples in the model building process, and N represents the sample quantity with known fermentation time used in the model building process.

The specific implementation process of the invention will be further explained by specific 13 embodiments below.

Embodiment 1 The method proposed by the invention was applied to the prediction of the fermentation degree of the traditional Meixiang salted Trachinotus ovatus. The Trachinotus ovatus used in this embodiment were purchased from Hongkai Yufeng Co., Ltd. in Yangjiang City, and there were 5 samples with fermentation time of 0, 3, 6, 9, 12, 15, 18, 21 and 24 days respectively, totalling 45 samples. According to the method in step (1) of the invention patent, the Trachinotus ovatus were pre-treated. Specifically, 5 pieces of 2cm wide fish blocks were cut from the starting point to the tail of each fish, and then each fish block were carefully cut off the surface and midline fish. Then 2 cm * 2 cm x 1 cm (length x width x thickness) block 23 samples to be measured were taken from the starting point to the tail direction with the middle line and the upper surface layer as the base pont. The Trachinotus ovatus block samples were cooled with liquid nitrogen and then ground for 5-8s with 3 repetitions to obtain a powdery sample of Meixiang salted 7rachinotus ovatus.

The obtained samples were analysed by E-nose and mass spectrometry.

The E-nose system 1s Fox4000 of AlphaMOS company in France, wich includes 18 gas sensor arrays of LY2/LG, P30/2, T40/2, LY2/G, LY2/AA, T30/1, P10/2, LY2/GH, 3 LY2/gCTL, LY2/gCT, P40/1, T70/2, PA/2, P30/1, P40/2, T40/1P10/1 and TA/2. Besides, different gas sensor arrays had different sensitivity to VOC, so that the electrical signals collected by the signal processing system can make an overall evaluation of the overall volatile flavour of the sample through the pattern recognition system.

The main characteristics of the 18 sensors are shown in Table 1. 10 Table 1 Array Sequence Sensor Name Sensitive Materials Chlorine, Fluorine, Nitrogen Oxides, 1 LY2/LG Sulfides Ammonia, Amine compounds, Carbon 2 LY2/G oxides 3 LY2/AA Ethanol, Acetone, Ammonia 4 LY2/GH Ammonia, Amine compounds LY2/eCTL Hydrogen sulfide 6 LY2/gCT Propane, Butane 7 T30/1 Polar Compounds, Hydrogen chloride 8 PIO Non-Polar: Hydrocarbon, Ammonia, Chlorine 9 P10/2 Non-Polar: Methane, Ethane P40/1 Fluorine, Chlorine 11 T70/2 Toluene, Xylene, Carbon monoxide In PA Ethanol, Ammonia water, Amine compounds 13 P30/1 Hydrocarbons, Ammonia, Ethanol

14 P40/2 Chlorine, Hydrogen sulfide, Fluoride 15 P30/2 Hydrogen sulfide, Ketone 16 T40/2 Chlorine 17 T40/1 Fluorine 18 TA/2 Ethanol Before using the E-nose system, the E-nose was cleaned with ultra-pure dry air with an air flow rate of 300 mL/min and an equilibrium time of 10 min. Then 1.00 g of Meixiang salted Trachinotus ovatus powder sample was placed in a 15mL headspace bottle, sealed with a lid, incubated at 4 °C for 0.5 h. And then the headspace gas is sucked for detection and analysis. The detection flow rate was set at 150 mL/min, the sample collection time was 120 s, and the delay time was 10 min. 810 data points (18 sensors x 9 samples x 5 biological replicates) were obtained through 5 biological repetitions. Based on the data points, the average and integral values of response signals of theses 18 gas sensors were extracted. Figure 1 is a radar chart of sensor response of E-nose to VOC during fermentation of traditional Meixiang salted Trachinotus ovatus. It can be seen from Figure 1 that it is quite different in the response intensity of different sensors to volatile flavor substances in samples. Combined with the representative sensitive substance types of the gas sensor array, it can be seen that with the extension of fermentation time, the components and contents of aldehydes, alcohols, ketones and nitrogen-containing compounds in the traditional Meixiang salted lS Trachinotus ovatus samples changed, which led to the change of volatile flavour, so that the E-nose can distinguish the volatile flavour of Meixiang salted 7rachinotus ovatus samples at different fermentation time points to a certain extent. The adopted GC-IMS was FlavourSpec of GAS Company in Germany, and the detection limit of this migration mass spectrometry is ppb level, without enrichment and concentration. The chromatographic column type was MXT-5 chromatographic column (15 m x 0.53 mm). Analysis time was 15 min. Column temperature was 60 °C. Carrier gas was Na. Injection needle temperature was 65 °C.

2.00 g of Meixiang salted Trachinotus ovatus powder sample was taken, incubated in headspace at 60 °C for 10 min, and then introduced samples by GC-IMS to extract the ion signal response peak of VOC.

Qualitative analysis of substances can be carried out according to NIST database and IMS database.

Shown in Figure 2, the three-dimensional spectrogram (retention time, migration time and peak intensity) of sample VOC was automatically generated by using the built-in plug-in of FlavourSpec and the two-dimensional overhead view was shown in Figure 3 (retention time and migration time). It can be intuitively seen from Figure 2 and Figure 3 that there were obvious differences in VOC in samples at different fermentation stages.

Meanwhile, in combination with the two-dimensional overhead view, the differences in VOC types and concentrations in different samples are relatively intuitive.

As shown in Figure 4, when 0 DAY samples as a reference, the concentrations of corresponding volatile substances in other groups of samples are clear at a glance.

The deeper the red is, the higher the concentration is.

Therefore, samples with different fermentation degrees can be visually distinguished.

With the help of GalleryPlot plug-in, the fingerprint spectrum of volatile substances was drawn, and the results were shown in Figure 5. The VOC differences between different 13 samples were compared intuitively and quantitatively.

The quick identification method constructed according to the Claims is characterized in identifying VOCs quickly and the characteristic VOCs of fresh Meixiang salted Trachinotus ovatus (0 day) included butyraldehyde, heptanal, nonanal, 3-methyl butyraldehyde and 2, 3-butanedione.

The VOC in the Meixiang salted Trachinotus ovatus samples with light fermentation (3-9 days) included butanone, ethyl butyrate, n-butyl butyrate, propyl 2-methylbutyrate, ethyl 2- methylbutyrate, ethyl acetate and ethyl propanoate.

VOC in the samples of Meixiang salted Trachinotus ovatus with moderate fermentation (10-18 days) included n-propyl butyrate and 3-nonanone.

The VOC in the samples of Meixiang salted Trachinotus ovatus with full fermentation (19-24 days) included octanal, benzaldehyde, 1-octen-3-ketone, 2, 3, 5- trimethylpyrazine, cyclohexene -2-ketone and 2, 6-Dimethylpyrazine.

Furthermore, PCA method was used to reduce the dimension of data of E-nose system and gas migration mass spectrometry as well as the integrated data.

The PCA results are shown in Figure 6 and Figure 7. Figure 6 showed the PCA analysis results of E-nose and Figure 7 showed the PCA analysis results of gas migration mass spectrometry.

The results showed that the volatile flavour of Meixiang salted Trachinotus ovatus with different fermentation time was distributed in the non-interference area, and the distance between areas reflected the volatile flavour differences among Meixiang salted Trachinotus ovatus samples. According to the fermentation time, the samples in PC1 direction showed quasi-linear distribution, and the samples with nine fermentation degree showed cluster phenomenon, and each cluster can be well distinguished.

3 Based on multi-data fusion and BP neural network algorithm, the prediction model of Meixiang salted Trachinotus ovatus was constructed. In the process of modelling, the number of neurons in hidden layer was continuously optimized, and the model was trained based on the extracted eigenvalues. Moreover, the test samples of each fermentation degree were randomly divided into two data sets, namely, a calibration set and a verification set. The calibration set was used to build the model and the verification set was used to test the performance of the model. Among them, there were 20 samples in the calibration set and 10 samples in the verification set in each year. The model was optimized by using the determination coefficient (R?) and the RMSE as the evaluation indexes of the model, wherein the calculation formulas of the determination coefficient (R?) and the RMSE were as follows: iy (A EP LL IE YP RSE |, IN EF Based on BP neural network algorithm, the R? and RMSE of the calibration set and the verification set was 0.9982 and 0.0601, respectively. The larger R? and the smaller RMSE showed that the prediction model has better performance.

To sum up, the multi-dimensional identification method for the fermentation degree of the traditional Meixiang salted fish based on odour visualization provided by the invention has strong feasibility. Besides, it can accurately predict the fermentation degree of the traditional Meixiang salted fish. Further, there is high correlation between the predicted value and the real value, so that the method is worthy of wide popularization.

Finally, it should be noted that the above-mentioned embodiments are only specific embodiments of the present invention, which are used to illustrate the technical scheme of the present invention, but not to limit it. Although the present invention has been described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that person familiar with the technical field can still modify or easily think of changes to the technical scheme described in the above-mentioned embodiments within the technical scope disclosed by the present invention.

However, these modifications, changes or substitutions do not make the essence of the corresponding technical solutions 3 deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Instead, they should be covered within the protection scope of the present invention.

Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (6)

ConclusiesConclusions 1. Een multidimensionale integratie identificatiemethode voor de fermentatiegraad van Meixiang gezouten vis op basis van geurvisualisatie, gekenmerkt door de volgende stappen; S1. het bereiden van detectiemonsters van Meixiang gezouten vis in verschillende fasen van fermentatie, het verwijderen van huid en graten van de detectiemonsters van Meixiang gezouten vis, vervolgens 2 cm x 2 cm x 1 cm blokmonsters afnemen vanaf het startpunt naar de richting van de staart met de middelste lijn en de povenste oppervlaktelaag als uitgangspunt, de blokmonsters koelen met vioeibare stikstof en vermalen gedurende 5-8 seconden met 3 herhalingen om een poederachtig monster van Meixiang gezouten vis te verkrijgen; S2. het op basis van het poederachtige monster van de Meixiang gezouten vis verzamelen van de olfactorische vingerafdruk spectrum via het E-neussysteem, het extraheren van de karakteristieke responswaarde van de gassensor van de E-neus om een geurkarakteristiek model te bouwen, S3. 2g poedermonster van de Meixiang gezouten vis onderwerpen aan incubatie in de kopruimte bij 60 °C gedurende 10 min, en vervolgens geëxtraheerd door een GC- IMS-apparaat om ionen signaal responspieken van karakteristieke VOC van het poedermonster van de Meixiang gezouten vis te verkrijgen, op basis van NIST- database en IMS-database wordt een kwalitatieve analyse uitgevoerd op de karakteristieke VOC door overeenkomstige pieken van ionsignalen om de karakteristieken van VOC te verkrijgen; S4. op basis van de karakteristieken van de VOC, het verkrijgen van een driedimensionale kaart en een tweedimensionale overheadkaart door middel van een reporter-plug-in, het op basis van het kleurverschil van signaalintensiteit van VOC-karakteristiek in de driedimensionale kaart en de tweedimensionale overhead kaart, het vervaardigen van het karakteristieke model van VOC; S5. op basis van het geurkenmerkende model en het kenmerkende VOC-model, het vervaardigen van een miltudimensionaal identificatiemodel van de fermentatiegraad van de Meixiang gezouten vis, waarin het multidimensionale identificatiemodel van de fermentatiegraad van de Meixiang gezouten vis een bepalingscoéfiiciént en een RMSE omvat, in het bijzonder, is de bepalingscoéfficiént groot en de RMSE klein, waarbij het multidimensionale identificatiemodel van de fermentatiegraad van de1. A multidimensional integration identification method for the fermentation degree of Meixiang salted fish based on odor visualization, characterized by the following steps; S1. preparing detection samples of Meixiang salted fish at different stages of fermentation, removing skin and bones from the detection samples of Meixiang salted fish, then taking 2 cm x 2 cm x 1 cm block samples from the starting point to the direction of the tail with the middle line and the lowermost surface layer as a starting point, cool the block samples with liquid nitrogen and grind for 5-8 seconds with 3 repetitions to obtain a powdery sample of Meixiang salted fish; S2. based on the powdery sample of the Meixiang salted fish, collecting the olfactory fingerprint spectrum through the E nose system, extracting the characteristic response value of the gas sensor of the E nose to build a odor characteristic model, S3. Subject 2g powder sample of the Meixiang salted fish to headspace incubation at 60 °C for 10 min, and then extracted by a GC-IMS apparatus to obtain ion signal response peaks of characteristic VOC of the powder sample of the Meixiang salted fish, on based on NIST database and IMS database, a qualitative analysis is performed on the characteristic VOC by corresponding peaks of ion signals to obtain the characteristics of VOC; S4. based on the characteristics of the VOC, obtaining a three-dimensional map and a two-dimensional overhead map by means of a reporter plugin, the based on the color difference of signal intensity of VOC characteristic in the three-dimensional map and the two-dimensional overhead map , manufacturing the characteristic model of VOC; S5. on the basis of the odor characteristic model and the VOC characteristic model, the preparation of a spleen-dimensional identification model of the degree of fermentation of the Meixiang salted fish, in which the multidimensional identification model of the degree of fermentation of the Meixiang salted fish includes a determination coefficient and an RMSE, in particular , the determination coefficient is large and the RMSE is small, where the multidimensional identification model of the degree of fermentation of the Meixiang-gezouten vis wordt gebruikt om de werkelijke fermentatiegraad van de monsters van de Meixiang gezouten vis te identificeren.Meixiang salted fish is used to identify the actual degree of fermentation of the Meixiang salted fish samples. 2. Het multidimensionale identificatiemodel van de fermentatiegraad van de Meixiang-gezouten vis gebaseerd op geurvisualisatie volgens conclusie 1, met het kenmerk dat de detectiemonsters van Meixiang gezouten vis 5 monsters omvatten waarvoor respectievelijk 0, 3, 6, 9, 12, 15, 18, 21 en 24 dagen is gefermenteerd, waarbijde blokmonsters die moeten worden gemeten 5 visblokken zijn met een breedte van 2 cm.The multidimensional identification model of the degree of fermentation of the Meixiang salted fish based on odor visualization according to claim 1, characterized in that the detection samples of Meixiang salted fish comprise 5 samples for which 0, 3, 6, 9, 12, 15, 18, respectively 21 and 24 days was fermented in which the block samples to be measured are 5 fish blocks with a width of 2 cm. 3. Het multidimensionale identificatiemodel van de fermentatiegraad van de Meixiang-gezouten vis gebaseerd op geurvisualisatie volgens een der voorgaande conclusies, met het kenmerk dat VOC-karakteristieken in de detectiemonsters van Meixiang gezouten vis die gedurende 0 dagen zijn gefermenteerd, butyraldehyde, heptanal, nonanal, 3-methylbutyraldehyde en 2,3-butaandion omvatten, de VOC in het monster van de Meixiang gezouten vis die 3-9 dagen is gefermenteerd, omvatten butanon, ethylbutyraat, n-butylbutyraat, propyl-2-methylbutyraat, ethyl-2- methylbutyraat, ethylacetaat en ethylpropanoaat, waarbij VOC in de monsters van Meixiang gezouten vis die 10-18 dagen zijn gefermenteerd, omvatten n- propylbutyraat en 3-nonanon, de VOC in de monsters van Meixiang gezouten vis die 19-24 dagen zijn gefermenteerd, omvatten octanal, benzaldehyde, 1-octeen-3- keton, 2,3,5-trimethylpyrazine, cyclohexeen-2-keton en 2,6-dimethylpyrazine.The multidimensional identification model of the degree of fermentation of the Meixiang salted fish based on odor visualization according to any one of the preceding claims, characterized in that VOC characteristics in the detection samples of Meixiang salted fish fermented for 0 days, butyraldehyde, heptanal, nonanal, 3-methylbutyraldehyde and 2,3-butanedione, the VOC in the sample of the Meixiang salted fish fermented for 3-9 days include butanone, ethyl butyrate, n-butyl butyrate, propyl-2-methylbutyrate, ethyl-2-methylbutyrate, ethyl acetate and ethyl propanoate, where VOC in the samples of Meixiang salted fish fermented for 10-18 days include n-propyl butyrate and 3-nonanone, the VOC in the samples of Meixiang salted fish fermented for 19-24 days include octanal, benzaldehyde, 1-octene-3-ketone, 2,3,5-trimethylpyrazine, cyclohexene-2-ketone and 2,6-dimethylpyrazine. 4. Het multidimensionale identificatiemodel van de fermentatiegraad van de Meixiang-gezouten vis gebaseerd op geurvisualisatie volgens een der voorgaande conclusies, met het kenmerk dat de E-neus tenminste 18 gassensoren omvat, de karakteristieke responswaarde is de gemiddelde waarde van het reactiesignaal en de integrale waarde van het responsbericht van de gassensor, waarbij de gassensoren tenminste LY2/LG, P30/2, T40/2, LY2/G, LY2/AA, T30/1, P10/2, LY2/GH, LY2/gCTL, LY2/gCT, P40/1, T70/2, PA/2, P30/1, P40/2, T40/1P10/1 en TA/2 omvatten.The multidimensional identification model of the degree of fermentation of the Meixiang salted fish based on odor visualization according to any one of the preceding claims, characterized in that the E nose comprises at least 18 gas sensors, the characteristic response value is the mean value of the response signal and the integral value of the response message from the gas sensor, wherein the gas sensors are at least LY2/LG, P30/2, T40/2, LY2/G, LY2/AA, T30/1, P10/2, LY2/GH, LY2/gCTL, LY2/gCT , P40/1, T70/2, PA/2, P30/1, P40/2, T40/1P10/1 and TA/2. 5. Het multidimensionale identificatiemodel van de fermentatiegraad van de Meixiang-gezouten vis gebaseerd op geurvisualisatie volgens een der voorgaande conclusies, met het kenmerk dat de stap S2 verder de volgende stappen omvat dat de E-neus vóór gebruik moet worden gereinigd met ultra zuivere droge lucht met een luchtstroomsnelheid van 300 ml/min en een evenwichtstijd van 10 minuten, het plaatsen van 1,00g poedermonster van Meixiang gezouten vis in een hoofdruimte 9 flesvan 15 mi, verzegeld met een deksel, laten staan op 4 °C gedurende 0,5 uur, het aanzuigen van het gas uit de kopruimte voor detectie en analyse, het detectiedebiet is ingesteld op 150 ml/min, de tijd voor het verzamelen van monsters is 120 seconden, en de vertragingstijd is 10 min 810 gegevenspunten worden verkregen door middel van 5 biologische herhalingen, op basis van de gegevenspunten worden de karakteristieken responswaarden verkregen.The multidimensional identification model of the degree of fermentation of the Meixiang salted fish based on odor visualization according to any one of the preceding claims, characterized in that the step S2 further comprises the steps of cleaning the E-nose with ultra-pure dry air before use. with an airflow rate of 300 ml/min and an equilibration time of 10 minutes, placing 1.00g powder sample of Meixiang salted fish in a headspace 9 bottle of 15 ml, sealed with a lid, let stand at 4°C for 0.5 hours , sucking the gas from the headspace for detection and analysis, the detection flow rate is set to 150 ml/min, the sample collection time is 120 seconds, and the delay time is 10 min 810 data points are obtained through 5 biological repetitions, based on the data points, the characteristics response values are obtained. 6. Het multidimensionale identificatiemodel van de fermentatiegraad van de Meixiang-gezouten vis gebaseerd op geurvisualisatie volgens een der voorgaande conclusies, met het kenmerk dat de constructiemethode van het multidimensionale identificatiemodel van de fermentatiegraad van Meixiang gezouten vis op zijn minst PCA-analyse, PLS-DA-analyse, neuraal netwerkalgoritme, willekeurig bosalgoritme en steilste klimalgoritme omvat, de bepalingscoëfficiënt en RMSE worden als volgt berekend:The multidimensional identification model of the degree of fermentation of Meixiang salted fish based on odor visualization according to any one of the preceding claims, characterized in that the construction method of the multidimensional identification model of the degree of fermentation of Meixiang salted fish is at least PCA analysis, PLS-DA analysis, includes neural network algorithm, random forest algorithm and steepest climbing algorithm, the determination coefficient and RMSE are calculated as follows: ERAS VES ERI AE Ruse ISN (x, 2 Aa TER R2 vertegenwoordigt de bepalingscoëfficiënt, RMSE vertegenwoordigd Wortel- gemiddelde-kwadraat-deviatie, Xi vertegenwoordigt de feitelijke fermentatiegraad van het 1° monster in het bouwproces van het model, X vertegenwoordigt de gemiddelde waarde van de werkelijke fermentatiegraad van alle monsters in het bouwproces van het model, Yi vertegenwoordigt de voorspelde fermentatiegraad van het 1° monster in het bouwproces van het model, Y vertegenwoordigt de gemiddelde waarde van de voorspelde fermentatiegraad van alle monsters in het bouwproces van het model, en N vertegenwoordigt de hoeveelheid van het monster met bekende fermentatietijd die wordt gebruikt in het bouwproces van het model.ERAS VES ERI AE Ruse ISN (x, 2 Aa TER R2 represents the coefficient of determination, RMSE represents Root mean square deviation, Xi represents the actual degree of fermentation of the 1° sample in the model building process, X represents the mean value of the actual fermentation rate of all samples in the model build process, Yi represents the predicted fermentation rate of the 1° sample in the model build process, Y represents the mean value of the predicted fermentation rate of all samples in the model build process, and N represents the amount of the sample with known fermentation time used in the model building process.