A non-destructive identification method for cultivated phoebe shearling by combining odor and headspace analysis

Abstract

The application discloses a kind of cultivation of smell and headspace analysis combination's odd nan hand string nondestructive identification method, it is related to precious timber true and false identification and quality detection technical field, including following step A: sample pool data model establishment: a1, sample selection: collect the cultivation of different producing areas, gram weight, shape and size odd nan hand string genuine sample not less than 200;Collect by high pressure injection paste, press oil, bubble medicine, high throwing, immersion paste smearing and non-lignum aquilariae imitated fake sample not less than 200;A2, electronic nose data acquisition: using electronic nose to the sample carries out smell data acquisition, acquisition condition includes;By being provided with electronic nose data acquisition and key sensor screening, the whole identification process does not need to be destructively treated to sample any cutting, grinding etc., completely maintain the integrity and value of hand string, and electronic nose detection can be completed within 1 hour, headspace mass spectrum analysis as supplementary means can be completed within 3 hours, improve detection efficiency.

Description

Technical Field

[0001] This invention relates to the field of identification and quality testing technology for precious woods, specifically a non-destructive identification method for cultivated Qinan bracelets that combines odor and headspace analysis. Background Technology

[0002] Agarwood is a traditional and precious Chinese medicinal herb and a top-tier spice in the world, with consistently strong market demand. Among them, cultivated Qinan agarwood, due to its short resin formation cycle, high yield, and ability to emit fragrance even at room temperature, has become the mainstream category in the collectible bead market, with huge production and sales volumes in Guangdong and Hainan provinces. However, behind this market prosperity lies a serious problem of counterfeiting and substandard products. For example, high-pressure injection of resin, surface oiling, soaking in chemical fragrances, and "resin-coating" are just some of the increasingly sophisticated and easily concealed methods of counterfeiting.

[0003] Currently, the identification of cultivated agarwood bracelets faces severe challenges. Firstly, standards such as the *Pharmacopoeia of the People's Republic of China* and the forestry industry standard *Agarwood* (LY / T<2904-2017) primarily target traditional agarwood, which differs significantly from the chemical characteristics of cultivated agarwood, thus lacking applicability. Existing group standards related to cultivated agarwood are complex to implement, have long testing cycles, and are costly, often requiring destructive sampling (such as cutting and grinding), leading consumers to refuse testing due to the risk of sample damage. Secondly, existing non-destructive identification patent technologies, such as CN114112951A (a method for identifying insect-damaged agarwood using hyperspectral imaging) and CN105699343A (a non-destructive method for identifying genuine agarwood), are mainly designed for traditional agarwood and are difficult to adapt to the characteristics of cultivated agarwood. Methods such as CN119470424A (A rapid and non-destructive method for identifying genuine and fake agarwood based on microscopic imaging and electronic nose (201)) have limited effectiveness in identifying counterfeit methods such as "sap application" that do not change the wood structure, and its electronic nose (201) model is often tied to specific equipment and lacks versatility.

[0004] Therefore, there is a need for a dedicated identification method for cultivated agarwood bracelets that can balance non-destructiveness, high accuracy, good versatility, and high efficiency, in order to protect consumer rights and regulate market order. Summary of the Invention

[0005] The purpose of this invention is to provide a non-destructive identification method for cultivated agarwood bracelets that combines odor and headspace analysis, in order to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A non-destructive method for identifying cultivated agarwood bracelets by combining odor and headspace analysis includes the following steps:

[0008] Step A: Establishing the sample pool data model:

[0009] a1. Sample selection: Collect no less than 200 genuine cultivated Qinan bracelets of different origins, weights, shapes and sizes; collect no less than 200 counterfeit samples made by high-pressure injection, oil pressing, soaking in medicine, high polishing, sizing and coating, and imitations made of non-agarwood.

[0010] a2. Electronic nose data acquisition: Use an electronic nose to acquire odor data from the sample. The acquisition conditions include: placing the sample in a headspace vial and equilibrating at 45°C for 30 minutes, performing baseline purge, sample gas inhalation, and sensor rinsing.

[0011] a3. Key sensor screening: After preprocessing the sensor response data of the electronic nose (201), orthogonal partial least squares discriminant analysis (OPLS-DA) was used to model the data, and key sensors were screened by combining variable projection importance (VIP>1) and significance level (P<0.05).

[0012] a4. Identification Model Construction: Based on the selected key sensor data, support vector machine (SVM), logistic regression (LR) and AdaBoost algorithms are used to train and construct a true / false discrimination model.

[0013] Step B: Initial screening and identification using electronic nose:

[0014] b1. Following the conditions in step A2, perform electronic nose testing on the cultivated Qinan bracelet to be tested to obtain its odor data;

[0015] b2. Input the data to be detected into the SVM, LR and AdaBoost models established in step A4 for discrimination respectively;

[0016] b3. If the discrimination results of all three models are ≥90%, the bracelet to be tested is determined to be genuine; if the discrimination result of any model is lower than 90%, proceed to step C.

[0017] Step C: Secondary discrimination by headspace gas mass spectrometry:

[0018] c1. Take 1-5 beads from the bracelet to be tested, place them in a headspace vial, and analyze them using solid phase microextraction-gas chromatography-mass spectrometry.

[0019] c2. Analyze the total ion chromatogram and identify the compounds. If plasticizers, chemical fragrances or other foreign substances are detected, or if the total ion chromatogram lacks characteristic chromatographic peaks in the 28-39 minute retention time range, it is determined to be a counterfeit product; otherwise, it is determined to be a genuine product.

[0020] As a further aspect of the present invention: the specific parameters for data acquisition by the electronic nose (201) in step a2 are as follows: baseline purge time 5-15 seconds, sample inhalation time 25-35 seconds, sensor rinsing time 35-45 seconds; the gas flow rate during the baseline and sample inhalation stages is 100-140cc / min, and the gas flow rate during the sensor rinsing stage is 160-200cc / min.

[0021] As a further aspect of the present invention: in step a4, the training process of the SVM, LR and AdaBoost models adopts cross-validation to evaluate the model performance and grid search to optimize the model hyperparameters.

[0022] As a further embodiment of the present invention: wherein the electronic nose in step a2 is a portable gas analysis system, and the sensor array contains 32 sensors.

[0023] As a further embodiment of the present invention: a constant temperature water bath is provided below the electronic nose, the constant temperature water bath has a placement opening, the headspace bottle is placed inside the placement opening, a support platform is fixedly connected to the upper side of the outer shell of the constant temperature water bath, a cylinder is installed on the support platform, and the output end of the cylinder is detachably connected to the electronic nose.

[0024] As a further embodiment of the present invention: a protective tube is fixedly connected to one side of the output end of the cylinder, a pressing rod is slidably connected to the bottom side of the protective tube, a return spring is sleeved on the pressing rod, a spiral rod is spirally connected to the upper side of the protective tube, a pressing switch is installed at the bottom end of the spiral rod, and the pressing switch is slidably connected to the inner wall of the protective tube.

[0025] As a further embodiment of the present invention: a protective plate is installed on the bottom side of the inner wall of the constant temperature water bath, a fixed pipe is fixedly connected to the protective plate, a lifting pipe is slidably connected to the fixed pipe, and the lifting pipe is connected to the headspace bottle.

[0026] As a further embodiment of the present invention: a threaded rod is rotatably connected inside the protective plate, the threaded rod passes through the protective plate and is threadedly connected to a threaded tube, the threaded tube is fixedly connected to the lifting tube, an adjusting rod is rotatably connected to one side of the protective plate, and a bevel gear is engaged between the adjusting rod and the threaded rod.

[0027] Compared with the prior art, the beneficial effects of the present invention are:

[0028] 1. By incorporating electronic nose data acquisition and key sensor screening, the entire identification process eliminates the need for any destructive treatment of the sample, such as cutting or grinding, thus preserving the integrity and value of the bracelet. Furthermore, electronic nose detection can be completed within 1 hour, and headspace mass spectrometry analysis, as a supplementary method, can be completed within 3 hours, thereby improving detection efficiency.

[0029] 2. By incorporating an electronic nose and headspace mass spectrometry, the two-step discrimination system complements each other. The electronic nose can quickly screen out most low-quality counterfeits, while headspace mass spectrometry can accurately identify high-quality counterfeits with similar ingredients such as "extract paste".

[0030] 3. By setting up a recognition model and selecting key sensors and establishing a machine learning model, the electronic nose discrimination model is not dependent on specific brands or models of equipment, which lowers the detection threshold and facilitates promotion.

[0031] 4. The above methods are specifically designed to address the chemical characteristics of cultivated agarwood bracelets and current mainstream counterfeiting techniques, effectively filling the technological gaps in existing standards and patents in this niche field. Attached Figure Description

[0032] Figure 1 This is a schematic diagram comparing genuine and counterfeit samples after headspace analysis in this invention (fake on top, real on the bottom).

[0033] Figure 2 This is a schematic diagram of genuine and counterfeit samples in this invention;

[0034] Figure 3 This is a schematic diagram of sensor analysis data in this invention;

[0035] Figure 4 This is a schematic diagram of the process structure in this invention;

[0036] Figure 5 This is a schematic diagram of the constant temperature water bath structure in this invention;

[0037] Figure 6 This is a schematic diagram of the internal structure of the protective tube in this invention;

[0038] Figure 7 This is a schematic diagram of the internal structure of the constant temperature water bath in this invention;

[0039] Figure 8 This is a schematic diagram of the protective plate structure in this invention;

[0040] The correspondence between the labels and component names in the attached figures is as follows:

[0041] 1. Constant temperature water bath; 101. Placement port; 102. Support platform; 2. Cylinder; 201. Electronic nose; 3. Protective tube; 301. Pressing rod; 302. Screw rod; 303. Press switch; 4. Protective plate; 401. Threaded rod; 402. Threaded tube; 403. Adjusting rod; 404. Fixing tube; 405. Lifting tube; 406. Bevel gear; 5. Headspace bottle. Detailed Implementation

[0042] Please see Figures 1-8 A non-destructive identification method for cultivated agarwood bracelets combining odor and headspace analysis, comprising the following steps:

[0043] Step A: Establishing the sample pool data model:

[0044] a1. Sample selection: Collect no less than 200 genuine cultivated Qinan bracelet samples from different origins (Guangdong, Hainan, Guangxi), weights, shapes and sizes; collect no less than 200 counterfeit samples made by high-pressure injection, oil pressing, soaking in medicine, high polishing, smearing with extract, and imitations made from non-agarwood.

[0045] a2. Data acquisition by electronic nose 201: Odor data is acquired from the sample using electronic nose 201. Acquisition conditions include: placing the sample in headspace vial 5 and equilibrating at 45°C for 30 minutes, performing baseline purging, sample gas inhalation, and sensor rinsing. In step a2, electronic nose 201 is a portable gas analysis system, and the sensor array contains 32 sensors.

[0046] Furthermore, the specific parameters for data acquisition by the electronic nose 201 in step a2 are as follows: baseline purge time 5-15 seconds, sample aspiration time 25-35 seconds, and sensor rinsing time 35-45 seconds; the gas flow rate during the baseline and sample aspiration stages is 100-140 cc / min, and the gas flow rate during the sensor rinsing stage is 160-200 cc / min. Different flow rates are suitable for different operating stages. A lower aspiration flow rate is conducive to the full interaction between the sensor and gas molecules; a higher rinsing flow rate can quickly remove residual molecules and improve detection efficiency.

[0047] Specifically, a headspace vial is a laboratory glassware specifically designed for headspace analysis. It typically consists of a glass vial, a special sealing cap (usually an aluminum cap), and a sealing septum. Its core function is to create a closed, controlled microenvironment for storing the sample to be tested and collecting the volatile odor molecules that are released naturally.

[0048] a3. Key sensor screening: After preprocessing the sensor response data of the electronic nose 201, orthogonal partial least squares discriminant analysis (OPLS-DA) was used to model the data, and key sensors were screened by combining variable projection importance (VIP>1) and significance level (P<0.05).

[0049] Furthermore, the key sensors selected in step a3 include S31, S5, S2, S23, S6, and S18. These specifically numbered sensors (such as S31, S5, S2, S23, S6, and S18) are the core sensing units for high-precision identification. These sensing units can distinguish the chemical substances most relevant to the authenticity of Qinan (such as specific terpenes and sesquiterpenes). They are the characteristic carriers of Qinan's "odor fingerprint" and form the basis of the general technical foundation of the above model.

[0050] Specifically, the electronic nose 201 includes sampling, a sensor array, and a pattern recognition algorithm.

[0051] Sampling: The gas in the headspace bottle 5 of the sample to be tested (such as a Qinan bracelet) is "drawn" into the instrument through the gas path system.

[0052] Sensor array: An array of multiple sensors that are cross-sensitive to different chemical substances. It reacts more strongly to specific types of volatile organic compounds (such as alcohols, aldehydes, ketones, alkanes, aromatic compounds, etc.). When a complex odor (such as the fragrance of Qinan) passes by, all sensors will respond simultaneously, forming a unique combination of signals.

[0053] Pattern recognition algorithm: The electrical signals generated by the sensor are received by the computer. Through machine learning algorithms (SVM, LR, AdaBoost), the computer learns and remembers the odor characteristics of different samples (genuine Qinan and various counterfeit products).

[0054] Then, by using a large number of known real and fake samples, a discrimination model is established. When an unknown sample is encountered, the electronic nose 201 collects its odor characteristics and the model determines whether it is closest to the real product or a counterfeit product.

[0055] a4. Identification Model Construction: Based on the selected key sensor data, support vector machine (SVM), logistic regression (LR) and AdaBoost algorithms are used to train and construct a true / false discrimination model.

[0056] Specifically, in step a4, the training process of SVM, LR and AdaBoost models uses cross-validation to evaluate model performance and grid search to optimize model hyperparameters. Through cross-validation, the dataset is divided into multiple parts, and one part is used as the test set and the rest as the training set in turn. This can more objectively evaluate the model's performance on unknown data, thereby selecting the best model parameters.

[0057] Step B: Initial screening and identification using the electronic nose 201:

[0058] b1. Following the conditions in step A2, the cultivated Qinan bracelet to be tested is subjected to electronic nose 201 testing to obtain its odor data;

[0059] b2. Input the data to be detected into the SVM, LR and AdaBoost models established in step A4 for discrimination respectively;

[0060] b3. If the discrimination results of all three models are ≥90%, the bracelet to be tested is determined to be genuine; if the discrimination result of any model is lower than 90%, proceed to step C. The judgment to proceed to step C in step b3 is automatically triggered. Once the model output result of the electronic nose 201 does not meet the preset confidence threshold (<90%), the system will unconditionally and automatically start the next analysis, ensuring the rigor and efficiency of the identification process.

[0061] Step C: Secondary discrimination by headspace gas mass spectrometry:

[0062] c1. Take 1-5 beads from the bracelet to be tested, place them in headspace vial 5, and analyze them using solid phase microextraction-gas chromatography-mass spectrometry.

[0063] c2. Analyze the total ion chromatogram and identify the compounds. If plasticizers, chemical fragrances or other foreign substances are detected, or if the total ion chromatogram lacks characteristic chromatographic peaks in the 28-39 minute retention time range, it is determined to be a counterfeit product; otherwise, it is determined to be a genuine product. In step c2, the compound identification is completed by calculating the retention index of volatile organic compounds using C7-C40 n-alkane standards and comparing it with the NIST17 mass spectrometry library.

[0064] Specifically, the conditions for solid-phase microextraction-gas chromatography-mass spectrometry in step c1 include: using an HP-5MS column, and the gas chromatography temperature program is as follows: hold at 50℃ for 1 minute, increase to 140℃ at 3℃ / min, then increase to 160℃ at 1℃ / min, and then increase to 280℃ at 4℃ / min and hold for 10 minutes. The different heating rates (3℃ / min, 1℃ / min, 4℃ / min) are designed to address the complexity of the aroma components of Qinan. In the critical range (such as 140℃-160℃), a slow heating rate (1℃ / min) is used to allow key aroma components with similar boiling points and polarities sufficient time to separate on the chromatographic column, forming identifiable independent chromatographic peaks, ensuring that the target compounds (especially the key components that elute at 28-39 minutes) can be clearly and accurately detected.

[0065] Furthermore, the electronic nose 201 is based on overall odor fingerprint recognition, which is good at capturing overall pattern differences in odors. However, it may fail to counterfeit products with extremely similar ingredients. The set-up qualitative and quantitative chemical composition analysis can accurately identify the absence of specific foreign molecules or key components. The two complement each other perfectly in terms of technical principle. Moreover, the non-destructive nature of the whole process is based on "headspace analysis" technology. The sample only needs to be heated in a sealed headspace vial 5, and its naturally released volatile molecules are detected without any physical or chemical damage. This improves detection efficiency while ensuring the integrity of the sample.

[0066] like Figure 5 As shown, a constant temperature water bath 1 is provided below the electronic nose 201. The constant temperature water bath 1 has a placement port 101, and a headspace vial 5 is placed inside the placement port 101. A support platform 102 is fixedly connected to the upper side of the outer shell of the constant temperature water bath 1. A cylinder 2 is installed on the support platform 102. The output end of the cylinder 2 is detachably connected to the electronic nose 201. In use, the sample is placed in the headspace vial 5, and the sealing cap of the headspace vial 5 is sealed. Then, the headspace vial 5 is placed inside the placement port 101, so that the placement port 101 is aligned with the headspace vial 5. The headspace vial 5 is locked in place. Then, the constant temperature water bath 1 is operated to raise the water temperature inside the constant temperature water bath 1 to 45°C and maintain it. The water inside the constant temperature water bath 1 heats the headspace vial 5 until the temperature of the headspace vial 5 is the same as that of the water. Then, it is left to stand for 30 minutes. Then, the output of the cylinder 2 controls the electronic nose 201 to move down, so that the probe of the electronic nose 201 pierces the rubber ring of the sealing cap on the headspace vial 5 and inserts into the top side inside the headspace vial 5. The probe extracts gas and analyzes it to complete the detection of the sample.

[0067] like Figure 6 As shown, a protective tube 3 is fixedly connected to one side of the output end of cylinder 2. A pressing rod 301 is slidably connected to the bottom side of the protective tube 3. A return spring is sleeved on the pressing rod 301. A spiral rod 302 is spirally connected to the upper side of the protective tube 3. A pressing switch 303 is installed at the bottom end of the spiral rod 302. The pressing switch 303 is slidably connected to the inner wall of the protective tube 3. When cylinder 2 moves the electronic nose 201 downward, it also moves the protective tube 3 downward. When the protective tube 3 moves downward, it moves the pressing rod 301 downward. When the pressing rod 301 moves downward and contacts the constant temperature water bath... On the outer wall, the pressing rod 301 moves inside the protective tube 3 under the reaction force until the top of the pressing rod 301 contacts the pressing switch 303. The pressing switch 303 controls the operation of the cylinder 2. When the pressing switch 303 is pressed, the cylinder 2 controls the electronic nose 201 to stop moving down. The position of the pressing switch 303 can be adjusted by rotating the screw rod 302. This setting can adjust the downward movement distance of the electronic nose 201 according to the height of the headspace vial 5 to avoid the electronic nose 201 from contacting the sample inside the headspace vial 5 too much.

[0068] like Figure 7 and Figure 8 As shown, a protective plate 4 is installed on the bottom side of the inner wall of the constant temperature water bath 1. A fixed tube 404 is fixedly connected to the protective plate 4, and a lifting tube 405 is slidably connected to the fixed tube 404. The lifting tube 405 is connected to the headspace bottle 5. A threaded rod 401 is rotatably connected inside the protective plate 4. The threaded rod 401 passes through the protective plate 4 and is threadedly connected to a threaded tube 402. The threaded tube 402 is fixedly connected to the lifting tube 405. An adjusting rod 403 is rotatably connected to one side of the protective plate 4. A bevel gear 406 meshes between the adjusting rod 403 and the threaded rod 401. When the headspace bottle 5 is detected by the electronic nose 201, the force generated when the probe pierces the bottle cap may cause the headspace bottle 5 to shift. Therefore, a special... The lifting tube 405 is installed to fix the position of the headspace bottle 5, preventing it from shifting downwards. It can also adjust the position of the headspace bottle 5 inside the constant temperature water bath 1 according to its size. Specifically, the adjusting rod 403 is rotated, which, through the bevel gear 406, causes the threaded rod 401 to rotate. When the threaded rod 401 rotates, because the lifting tube 405 is fixedly connected to the threaded tube 402, the lifting tube 405 can only move vertically up and down inside the fixed tube 404 (a limit block is installed between the lifting tube 405 and the fixed tube 404, ensuring that the lifting tube 405 can only move vertically). This allows for adjustment of the position of the headspace bottle 5 inside the constant temperature water bath 1.

[0069] Example 1: Establishment of the sample pool and electronic nose 201 discriminant model:

[0070] Sample preparation: Collect 210 cultivated Qinan bracelets confirmed by experts (70 from Maoming, Guangdong; 70 from Haikou, Hainan; and 70 from Yulin, Guangxi), weighing 8-25g and with bead diameters of 10-18mm. Collect 200 counterfeit bracelets (40 from high-pressure injection, 50 from oiling, 40 from medicinal soaking, 30 from high-polishing, 30 from extracting and coating, and 10 from non-agarwood).

[0071] Data acquisition using the electronic nose 201: The Cyranose® 320 electronic nose 201 was used with the following parameters: equilibration at 45°C for 30 min, baseline purge for 10 s, sample aspiration for 30 s, sensor flushing for 40 s, and flow rates of 120 cc / min (baseline / aspiration) and 180 cc / min (flushing).

[0072] Key sensor selection and model training: After preprocessing the collected data, OPLS-DA analysis was performed to select six key sensors (S31, S5, S2, S23, S6, and S18) with VIP>1 and P<0.05. The SVM, LR, and AdaBoost algorithms were used to train the models respectively. After grid search optimization and 5-fold cross-validation, the accuracy of the SVM model (C=100, gamma=0.01) reached 95.0% and AUC=0.98; the accuracy of the LR model (C=10) reached 92.5% and AUC=0.96; and the accuracy of the AdaBoost model (n_estimators=100) reached 90.0% and AUC=0.95.

[0073] Example 2: Sample Identification Example 1 (Genuine Product Determination):

[0074] A sample of cultivated agarwood bracelet (sample A) was taken for testing. After detection by the electronic nose 201, the data was input into the established model. The discrimination results were: SVM model 94% (probability of authenticity), LR model 93%, and AdaBoost model 91%. Since the results of the three models were all ≥90%, according to the discrimination rules, it was initially determined to be authentic. To be on the safe side, headspace GC-MS verification was performed. Its total ion chromatogram showed obvious characteristic peaks in 28-39 min, and no plasticizers or chemical foreign matter were detected. Finally, sample A was determined to be authentic.

[0075] Example 3: Sample Identification Example 2 (Counterfeit Product Determination):

[0076] Another bracelet to be tested (sample B) was used. The initial screening results of the electronic nose 201 were: SVM model 88%, LR model 91%, and AdaBoost model 89%. Since the results of SVM and AdaBoost models were below 90%, a secondary headspace GC-MS discrimination was automatically triggered. The total ion chromatogram showed that the characteristic peak group of 28-39 min was completely missing, and trace amounts of dibutyl phthalate were detected in the mass spectrometry analysis. Based on this, it was determined that sample B was an extract-coated counterfeit.

[0077] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A non-destructive method for identifying cultivated agarwood bracelets by combining odor and headspace analysis, characterized in that, Includes the following steps: Step A: Establishing the sample pool data model: a1. Sample selection: Collect no fewer than 200 genuine cultivated Qinan bracelet samples of different origins, weights, shapes and sizes; Collect no fewer than 200 samples of counterfeit products made through high-pressure injection, oil pressing, soaking in medicine, high polishing, sizing, and imitation of non-agarwood. a2. Electronic nose (201) data acquisition: The electronic nose (201) is used to acquire odor data of the sample. The acquisition conditions include: placing the sample in a headspace vial (5) and equilibrating at 45°C for 30 minutes, performing baseline purging, sample gas inhalation and sensor rinsing. a3. Key sensor screening: After preprocessing the sensor response data of the electronic nose (201), orthogonal partial least squares discriminant analysis (OPLS-DA) was used to model the data, and key sensors were screened by combining the importance and significance level of variable projection. a4. Identification Model Construction: Based on the selected key sensor data, support vector machine (SVM), logistic regression, and AdaBoost algorithms are used to train the model to distinguish between true and false data. Step B: Initial screening and judgment using the electronic nose (201): b1. According to the conditions of step a2, the cultivated Qinan bracelet to be tested is tested by electronic nose (201) to obtain its odor data; b2. Input the data to be detected into the SVM, LR and AdaBoost models established in step a4 for discrimination respectively; b3. If the discrimination results of all three models are ≥90%, the bracelet to be tested is determined to be genuine; if the discrimination result of any model is lower than 90%, proceed to step C. Step C: Secondary discrimination by headspace gas mass spectrometry: c1. Take 1-5 beads of the bracelet to be tested and place them in the headspace vial (5). Analyze them using solid-phase microextraction-gas chromatography-mass spectrometry. c2. Analyze the total ion chromatogram and identify the compounds. If plasticizers, chemical fragrances or other foreign substances are detected, or if the total ion chromatogram lacks characteristic chromatographic peaks in the 28-39 minute retention time range, it is determined to be a counterfeit product; otherwise, it is determined to be a genuine product.

2. The non-destructive identification method for cultivated agarwood bracelets combining odor and headspace analysis according to claim 1, characterized in that, The specific parameters for data acquisition by the electronic nose (201) in step a2 are as follows: baseline purge time 5-15 seconds, sample inhalation time 25-35 seconds, sensor rinsing time 35-45 seconds; gas flow rate during baseline and sample inhalation phases is 100-140cc / min, and gas flow rate during sensor rinsing phases is 160-200cc / min.

3. The non-destructive identification method for cultivated agarwood bracelets combining odor and headspace analysis according to claim 1, characterized in that, In step a4, the training process of the SVM, LR, and AdaBoost models uses cross-validation to evaluate model performance and grid search to optimize model hyperparameters.

4. The non-destructive identification method for cultivated agarwood bracelets combining odor and headspace analysis according to claim 1, characterized in that, The electronic nose mentioned in step a2 is a portable gas analysis system, and the sensor array contains 32 sensors.

5. The non-destructive identification method for cultivated agarwood bracelets combining odor and headspace analysis according to claim 1, characterized in that, Below the electronic nose (201) is a constant temperature water bath (1), and the constant temperature water bath (1) has a placement port (101). The headspace bottle (5) is placed inside the placement port (101). A support platform (102) is fixedly connected to the upper side of the outer shell of the constant temperature water bath (1). A cylinder (2) is installed on the support platform (102). The output end of the cylinder (2) is detachably connected to the electronic nose (201).

6. The non-destructive identification method for cultivated agarwood bracelets combining odor and headspace analysis according to claim 5, characterized in that, A protective tube (3) is fixedly connected to one side of the output end of the cylinder (2). A pressing rod (301) is slidably connected to the bottom side of the protective tube (3). A reset spring is sleeved on the pressing rod (301). A spiral rod (302) is spirally connected to the upper side of the protective tube (3). A pressing switch (303) is installed at the bottom end of the spiral rod (302). The pressing switch (303) is slidably connected to the inner wall of the protective tube (3).

7. The non-destructive identification method for cultivated agarwood bracelets combining odor and headspace analysis according to claim 6, characterized in that, A protective plate (4) is installed on the bottom side of the inner wall of the constant temperature water bath (1). A fixed pipe (404) is fixedly connected to the protective plate (4). A lifting pipe (405) is slidably connected to the fixed pipe (404). The lifting pipe (405) is connected to the headspace bottle (5).

8. The non-destructive identification method for cultivated agarwood bracelets combining odor and headspace analysis according to claim 7, characterized in that, The protective plate (4) is rotatably connected to a threaded rod (401). The threaded rod (401) passes through the protective plate (4) and is threadedly connected to a threaded tube (402). The threaded tube (402) is fixedly connected to the lifting tube (405). An adjusting rod (403) is rotatably connected to one side of the protective plate (4). A bevel gear (406) meshes between the adjusting rod (403) and the threaded rod (401).

Images