Agilawood alcohol content detection method, electronic equipment, storage medium and program product

Abstract

The invention provides an agilawood alcohol content detection method, electronic equipment, a storage medium and a program product, and relates to the technical field of computers. According to the method, nondestructive testing of an agilawood sample can be achieved through the nuclear magnetic resonance technology, a complex pretreatment process is not needed, the detection efficiency can be greatly improved, and by locking the optimal magnetic field intensity and the optimal radio frequency for exciting alcohol substance hydrogen nuclei and combining targeted collection and signal feature extraction of target mixed signals, the detection accuracy of the agilawood sample is improved. The specificity and the signal-to-noise ratio of alcohol signals are remarkably improved, interference of non-target components such as moisture, oil and lignin in an agilawood sample is effectively stripped, and the detection precision and repeatability are guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a method for detecting alcohol content of eaglewood, an electronic device, a storage medium and a program product. BACKGROUND

[0002] As a precious Chinese herbal medicine and natural spice, the quality of eaglewood depends on the content of alcohol substances (such as agarol and baimuxiang alcohol). Therefore, accurately detecting the concentration of alcohol substances is a key link for the quality classification and market circulation supervision of eaglewood.

[0003] Currently, the detection of this key indicator completely relies on traditional solvent extraction weighing method, mainly including Soxhlet extraction method and cold soaking method. The core principle is to repeatedly extract alcohol substances in eaglewood samples with organic solvents (such as anhydrous ethanol), and calculate the alcohol content by weighing the mass difference of the residue after extraction or the concentration of the extraction liquid. This kind of method is mature in technology and low in cost, and has been widely used in the basic quality screening of eaglewood. However, it has defects: first, the sample is highly destructive, and the eaglewood sample needs to be crushed into powder during the detection process, which cannot be applied to valuable samples such as eaglewood crafts and cultural relics; second, the detection period is long, and it takes 4-6 hours for single extraction, plus subsequent steps such as concentration, drying and weighing, the whole process takes more than 8 hours, which is low in efficiency; third, the precision is easily disturbed, and non-target components such as lignin and cellulose are easily dissolved with alcohol substances during the extraction process, which leads to high detection results and a relative error of more than 5%. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a method for detecting alcohol content of eaglewood, an electronic device, a storage medium and a program product, so as to improve the problems of the existing alcohol content detection method, such as high destructiveness, low detection efficiency and low detection precision.

[0005] In a first aspect, the embodiments of the present application provide a method for detecting alcohol content of eaglewood, which comprises: performing nuclear magnetic resonance scanning on a to-be-detected eaglewood sample, adjusting the static magnetic field strength and the radio frequency, and determining the optimal magnetic field strength and the optimal radio frequency for exciting hydrogen nuclei of alcohol substances; under the condition of maintaining the optimal magnetic field strength and the optimal radio frequency, emitting a preset pulse sequence, and collecting a target mixed signal; acquiring signal characteristics of alcohol substances according to the target mixed signal; determining the concentration of alcohol substances of the to-be-detected eaglewood sample according to the signal characteristics.

[0006] In the implementation process, the nuclear magnetic resonance technology can be used for non-destructive detection of the eaglewood sample, without the need for a complex pretreatment process, and can greatly improve the detection efficiency. By locking the optimal magnetic field strength and the optimal radio frequency of the hydrogen nucleus of the alcohol substance, and combining the targeted acquisition of the target mixed signal and the signal feature extraction, the specificity and signal-to-noise ratio of the alcohol signal are significantly improved, the interference of non-target components such as water, oil and lignin in the eaglewood sample is effectively removed, and the detection accuracy and repeatability are ensured.

[0007] Optionally, the emitting of the preset pulse sequence and the acquisition of the target mixed signal comprises: emitting a first radio frequency pulse sequence, emitting a second radio frequency pulse sequence after a first delay time, and acquiring a first mixed signal after a second delay time, wherein the first mixed signal comprises signals of oil and alcohol substances, the first delay time is less than the transverse relaxation time of water, and the second delay time is greater than the transverse relaxation time of water; emitting a third radio frequency pulse sequence and acquiring a plurality of second mixed signals arranged in time sequence, the second mixed signal comprising signals of oil and alcohol substances; wherein the target mixed signal comprises the first mixed signal and the second mixed signal.

[0008] In the implementation process, the first radio frequency pulse sequence and the second radio frequency pulse sequence are emitted in stages, and the first delay time less than the transverse relaxation time of water and the second delay time greater than the transverse relaxation time of water are accurately matched, the differences in the transverse relaxation characteristics of water, oil and alcohol substances are ingeniously utilized, and the efficient and accurate removal of the water signal is realized. Then, a plurality of second mixed signals arranged in time sequence are acquired by the third radio frequency pulse sequence, the differences in the attenuation rules of oil and alcohol substances are completely captured, and sufficient data support is provided for the signal splitting of the two.

[0009] Optionally, the signal feature of the alcohol substance is acquired according to the target mixed signal, comprising: based on the difference in the transverse relaxation time between the alcohol substance and the oil, performing exponential decay curve fitting on a plurality of the second mixed signals; acquiring the initial signal strength of the alcohol substance according to the fitting result and the first mixed signal, and the signal feature comprises the initial signal strength.

[0010] In the implementation process, based on the difference in the transverse relaxation time between the alcohol and the oil, exponential decay curve fitting is performed on a plurality of second mixed signals (time sequence echo signals), the differences in the attenuation rules of the two are fully captured, and a scientific basis is provided for the oil-alcohol signal splitting.

[0011] Optionally, the alcohol substance concentration of the to-be-detected eaglewood sample is determined according to the signal feature, comprising: The alcohol concentration corresponding to the initial signal intensity is determined based on the preset correlation between alcohol concentration and signal intensity.

[0012] In the above implementation process, the concentration of alcohols can be directly determined by querying the correlation relationship, without the need for complicated on-site calibration or additional experimental operations, which greatly improves the detection efficiency.

[0013] Optionally, the fitting result includes the alcohol decay signal, and after obtaining the initial signal intensity of the alcohol, it further includes: The attenuation signal of the alcohol substance is subjected to FFT transformation to obtain the frequency signal; The frequency signal is bandpass filtered to obtain the filtered alcohol attenuation signal, and the initial signal strength of the alcohol is determined based on the filtered alcohol attenuation signal.

[0014] In the above implementation process, the FFT transform converts the time-domain signal into a frequency-domain signal, making the frequency distribution differences between the alcohol characteristic signal and residual interference signals such as bound water and trace lignin clearly visible, providing a basis for accurate filtering; the bandpass filter specifically retains the alcohol characteristic frequency band signal, effectively blocks interference from other frequency bands, and significantly improves signal purity.

[0015] Optionally, the first radio frequency pulse sequence is a 90° spin echo sequence, the second radio frequency pulse sequence is a 180° spin echo sequence, and the third radio frequency pulse sequence is a CPMG sequence. The 90° radio frequency pulse can completely flip the hydrogen nucleus magnetization vectors of water, oil, and alcohol in agarwood to the XY plane, ensuring that the initial mixed signal fully contains signals of various components, providing comprehensive data support for subsequent difference calculations to extract the water signal. The 180° radio frequency pulse can precisely refocus the magnetization vectors in the XY plane, effectively correcting signal astigmatism caused by magnetic field inhomogeneity, and ensuring the stability of oil and alcohol signals during the second delay time. The continuous refocusing pulses of the CPMG sequence can effectively correct signal astigmatism caused by magnetic field inhomogeneity, ensuring the stability and integrity of the echo signal. Multiple sets of time-series echo signals can completely capture the differences in attenuation patterns between the two, providing sufficient data support for fitting and splitting. The initial alcohol signal intensity obtained through fitting calculations eliminates oil interference and is positively correlated with alcohol content, ensuring the specificity and reliability of the signal characteristics.

[0016] Optionally, the method further includes: The third mixed signal is acquired after the first delay time; A moisture signal is obtained based on the first mixed signal and the third mixed signal; Based on the moisture signal, the change in decay rate at different time periods is obtained; The uniformity of moisture distribution within the agarwood sample under test is assessed based on the change in the decay rate.

[0017] In the above implementation process, without adding additional detection steps or damaging the sample, the moisture signal is obtained by using the difference between the first and third mixed signals that have been collected. The uniformity of moisture distribution inside the sample is evaluated by analyzing the changes in the moisture decay rate over different time periods. This not only makes full use of existing detection data resources and avoids additional detection costs and time consumption, but also adds a key auxiliary indicator for the quality assessment of agarwood.

[0018] Optionally, adjusting the static magnetic field strength and radio frequency to determine the optimal magnetic field strength and optimal radio frequency for exciting the hydrogen nuclei of alcohols includes: Within a preset range, the magnetic field strength is adjusted incrementally with a set step size, and the magnetic field signal is collected at each magnetic field strength. The magnetic field strength corresponding to the maximum signal-to-noise ratio of alcohol features in the magnetic field signal is determined as the optimal magnetic field strength. Calculate the theoretical radio frequency based on the optimal magnetic field strength; A scan is performed within a set range adjacent to the theoretical radio frequency, and the response signal at each frequency point is acquired. The frequency corresponding to the peak value of the response signal is determined as the optimal radio frequency.

[0019] In the above implementation process, the magnetic field strength optimization takes maximizing the signal-to-noise ratio of alcohol characteristics as the core objective. By orderly adjusting the preset range and set step size, it avoids the inefficiency caused by blind search and can accurately locate the magnetic field environment most conducive to alcohol signal recognition, significantly improving the distinction between signal and noise. The radio frequency optimization is based on the theoretical benchmark of the optimal magnetic field strength derivation, and then captures the response peak by sweeping the frequency in the vicinity range to ensure the precise coupling of radio frequency and magnetic field strength, maximize the excitation of hydrogen nuclei resonance in alcohol substances, and reduce the false excitation of hydrogen nuclei of non-target components.

[0020] Optionally, before transmitting the preset pulse sequence, the method further includes: While maintaining the optimal magnetic field strength and the optimal radio frequency, a single excitation pulse is emitted, and a baseline signal of a preset duration is acquired. If the intensity fluctuation of the baseline signal does not exceed a set threshold, then the following step is executed: transmit a preset pulse.

[0021] In the above implementation process, a baseline signal detection step is added before transmitting the preset pulse. By collecting the baseline signal for a preset duration under optimal parameter conditions and judging whether its intensity fluctuation meets the set threshold, the stability of the detection system can be effectively verified in advance, and the distortion of subsequent signal acquisition caused by parameter drift and abnormal equipment noise can be avoided.

[0022] Secondly, embodiments of this application provide an electronic device, including a processor and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps of the method provided in the first aspect above are performed.

[0023] Thirdly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method provided in the first aspect above.

[0024] Fourthly, embodiments of this application provide a computer program product, including computer program instructions, which, when read and executed by a processor, perform the steps of the method provided in the first aspect above.

[0025] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing embodiments of this application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0026] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 A flowchart illustrating a method for detecting the alcohol content of agarwood, provided in an embodiment of this application; Figure 2 A structural block diagram of an agarwood alcohol content detection device provided in this application embodiment; Figure 3 This is a schematic diagram of the structure of an electronic device for performing a method for detecting the alcohol content of agarwood, provided as an embodiment of this application. Detailed Implementation

[0028] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0029] It should be noted that the terms "system" and "network" in the embodiments of this invention can be used interchangeably. "Multiple" refers to two or more; therefore, in the embodiments of this invention, "multiple" can also be understood as "at least two". "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / ", unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0030] It should also be noted that all actions involving the acquisition of signals, information, or data in this application are carried out in compliance with the relevant data protection laws and policies of the country where the application is located, and with the authorization granted by the owner of the relevant device.

[0031] This application provides a method for detecting the alcohol content in agarwood. This method can achieve non-destructive testing of agarwood samples using nuclear magnetic resonance technology, without the need for complex pretreatment procedures, which can significantly improve detection efficiency. By locking the optimal magnetic field strength and optimal radio frequency for exciting the hydrogen nuclei of alcohol substances, combined with targeted acquisition and signal feature extraction of target mixed signals, the specificity and signal-to-noise ratio of alcohol signals are significantly improved. It effectively removes interference from non-target components such as moisture, oil, and lignin in agarwood samples, ensuring detection accuracy and repeatability.

[0032] To facilitate understanding of the subsequent method for detecting alcohol content in agarwood, the testing environment involved in this scheme will be introduced below.

[0033] This solution addresses the diverse morphologies of agarwood samples (lumpy raw materials, carved finished products, irregular agarwood materials, etc.) by designing a specialized solid sample carrier. This carrier requires no added solvent carrier and can directly hold complete agarwood samples, adapting to detection objects of different sizes and shapes. It fundamentally solves the problems of limited sample morphology and the need to process samples into uniform powders in traditional methods. At the same time, the inner wall of the carrier is coated with a 0.1-0.3mm thick polytetrafluoroethylene anti-interference coating, which effectively isolates the adsorption and interference of the carrier's own resin material on the NMR signal, ensuring the purity and accuracy of subsequent signal acquisition and providing a stable sample-bearing foundation for non-destructive testing.

[0034] The sample container adopts an "open cylindrical container + detachable base" structure. The main body is made of high-strength polyetheretherketone (PEEK), which has excellent chemical stability (resistant to organic solvent corrosion), mechanical strength (can withstand the weight of agarwood samples up to 500g), and low magnetic resonance signal adsorption (low hydrogen nucleus content, no obvious interference signal). The inner diameter of the container is designed to be adjustable from 5-15cm (achieved by changing cylindrical tubes with different inner diameters), and the height is 8-20cm, which can accommodate agarwood samples with a length ≤18cm and a diameter ≤14cm (covering more than 95% of the agarwood raw materials and finished product sizes on the market). The detachable base is 2cm thick and has a built-in anti-slip silicone pad to prevent the sample from shifting during the test. At the same time, a 1cm diameter ventilation hole is opened in the center of the base to prevent trace moisture from condensing on the sample due to the sealed environment, which would affect the test results.

[0035] For samples with special shapes (such as thin flakes of agarwood, hollow incense materials, and carved ornaments), a matching "adjustable fixing component" is designed. For thin flake samples (thickness < 1cm), a 0.5cm thick polytetrafluoroethylene (PTFE) spacer is provided, which can be stacked and adjusted according to the number of samples to ensure that the samples remain horizontal and stable in the container. For hollow or irregular samples, a soft PTFE fixing strap (1cm wide and 0.2cm thick) is provided, which can fix the sample by wrapping it to prevent the sample from shaking and without damaging the surface texture (solving the surface protection problem when testing carved finished products).

[0036] Compared with the magnetic resonance signal interference intensity of materials such as glass, ordinary plastics, metals, and PTFE, PTFE has a hydrogen nucleus density of only 0.01 mol / cm³ (far lower than the 0.1 mol / cm³ of plastics) and no adsorption of alcohols (contact angle > 110°, belonging to superhydrophobic materials). This effectively avoids physical adsorption between the sample container material and agarwood alcohols, which would lead to signal attenuation. Therefore, PTFE was chosen as the core material for the coating.

[0037] The coating preparation process involves the following steps: 1. Pretreatment stage: Ultrasonic cleaning of the inner wall of the PEEK container with anhydrous ethanol for 15 minutes (300W power, 40kHz frequency) to remove surface oil and impurities; then place it in an 80℃ oven to dry for 30 minutes to ensure the inner wall is dry (moisture content <0.1%, to avoid poor adhesion between the coating and the substrate).

[0038] 2. Coating stage: Using the "electrostatic spraying + sintering curing" process, PTFE powder (particle size 5-10μm, purity ≥99.9%) is mixed with deionized water at a mass ratio of 1:3 to prepare a uniform spraying slurry; the slurry is uniformly coated onto the inner wall of the container using electrostatic spraying equipment (voltage 60kV, spraying distance 15cm), and the wet film thickness is controlled to be 0.15-0.35mm.

[0039] 3. Curing stage: Place the coated container into a programmed temperature oven and sinter according to the curve of "room temperature → 200℃ (holding temperature for 30 min) → 380℃ (holding temperature for 60 min) → natural cooling to room temperature" to completely melt the PTFE powder and tightly bond it with the PEEK substrate, finally forming a uniform coating with a thickness of 0.1-0.3 mm (error ±0.02 mm).

[0040] 4. Post-processing stage: Gently sand the coating surface with 2000-grit sandpaper to remove any possible minor bumps, and then wipe it clean with anhydrous ethanol to ensure that the inner wall is smooth (surface roughness Ra≤0.2μm) to avoid damage caused by friction between the sample surface and the coating.

[0041] After the coating is prepared, the anti-interference effect can be verified through two core indicators: 1. Signal interference rate test: Under the same magnetic resonance parameters, the background signal intensity of the "empty coated sample dish" and the "uncoated sample dish" were measured respectively. The background signal intensity after coating should be reduced by ≥90% (to ensure that the interference of the sample dish itself is negligible). 2. Adsorption test: A 1% linalool standard solution was dropped onto the coating surface and allowed to stand for 24 hours. The residual amount of the solution was detected by high performance liquid chromatography (HPLC). The adsorption amount should be ≤0.01mg / cm² (to ensure that no alcohol substances are adsorbed and do not affect the sample signal).

[0042] This method uses the aforementioned sample container to detect the alcohol content in agarwood samples. Compared to traditional destructive testing methods, it achieves non-destructive and repeatable testing, and is adaptable to all forms of agarwood samples, including blocky, carved, and flaky samples. The inner wall of the sample container is coated with a PTFE coating, which isolates interference and reduces the background signal by ≥90%. Placing the agarwood sample in the sample container allows for rapid testing with low operational complexity. Furthermore, the sample container is equipped with an anti-slip silicone pad and a PTFE fixing strap, preventing damage to the surface of the agarwood sample.

[0043] The detection system in this solution is a magnetic resonance system, and the hardware configuration of the magnetic resonance system is as follows: 1. Magnet type: permanent magnet, magnetic field strength 0.5-1.5T (adjustable), magnetic field uniformity ≤1ppm (ensuring signal stability); 2. RF coil: birdcage probe, diameter 10cm, resonant frequency 20-60MHz (matching hydrogen nucleus resonance requirements); 3. Gradient system: maximum gradient strength 30mT / m, switching rate 100T / (m·s) (improving spatial resolution).

[0044] Please refer to Figure 1 , Figure 1A flowchart of a method for detecting the alcohol content of agarwood provided in this application embodiment is included, the method comprising the following steps: Step S110: Perform nuclear magnetic resonance scanning on the agarwood sample to be tested, adjust the static magnetic field strength and radio frequency, and determine the optimal magnetic field strength and optimal radio frequency for exciting the hydrogen nuclei of alcohol substances.

[0045] Nuclear magnetic resonance (NMR) scanning is a detection technique that utilizes the resonance signals generated by atomic nuclei when excited by radio frequency pulses in a static magnetic field.

[0046] Static magnetic field strength refers to the stable magnetic field strength (unit: T) in the detection cavity of an NMR system, while radio frequency is the electromagnetic wave frequency that excites hydrogen nuclei to resonate (unit: MHz).

[0047] The hydrogen nuclei of alcohol substances are the hydrogen atom nuclei contained in alcohol-soluble extracts of agarwood (such as agaric spirol), and their resonance signals are the core of quantitative detection.

[0048] Before testing, the agarwood sample to be tested can be pretreated, such as by surface cleaning and environmental equilibration, and then placed in the aforementioned sample container before being placed in the detection chamber of the nuclear magnetic resonance (NMR) detection system to start the scan. During the scan, the optimal magnetic field strength and optimal radio frequency can be determined based on the relevant characteristics of the magnetic field signal by adjusting the static magnetic field strength and radio frequency.

[0049] The Zeeman energy difference ΔE generated by the hydrogen nuclei (¹H) of agarwood alcohols in a static magnetic field satisfies ΔE = γ·h·B0 / 2π (where γ is the gyromagnetic ratio of the hydrogen nuclei and h is Planck's constant); however, the radio frequency electromagnetic wave frequency ν must satisfy ν = ΔE / h = γ·B0 / 2π to excite the hydrogen nuclei to resonate. Experimental measurements show that the γ value of the hydrogen nuclei in agarwood alcohols is 42.58 MHz / T (slightly different from the γ values ​​of hydrogen nuclei in common organic compounds; lignin γ = 42.45 MHz / T and cellulose γ = 42.39 MHz / T). Based on this difference, a dynamic matching scheme can be designed to ensure that only the hydrogen nuclei of alcohols are excited.

[0050] Before signal acquisition, the basic magnetic resonance spectrum of the sample can be obtained through "low-power pre-scan" to identify the characteristic resonance peak position of the target alcohol (preset to 1.2-1.5ppm, corresponding to the hydrogen nucleus resonance signal of the -CH2-OH group in the alcohol) and eliminate the interference peaks of special impurities (such as residual essential oils and moisture) that may exist in the sample (the moisture resonance peak is usually at 4.7ppm).

[0051] The pretreated agarwood sample was then placed in a sample holder and placed in the magnetic resonance detection chamber. The initial static magnetic field strength was set to 1.0T, and the radio frequency power was reduced to 1 / 5 of the conventional detection (to avoid exciting all hydrogen nuclei). A proton nuclear magnetic resonance (¹H-NMR) pre-scan was performed with a scanning range of 0-10ppm and 8 scans to obtain a preliminary spectrum. The signal peak intensity in the 1.2-1.5ppm range was identified using spectral analysis software (such as MestReNova). If there were no obvious interference peaks in this range (interference peak intensity < 10% of the target peak intensity), the formal acquisition was initiated. If interference peaks were present, the static magnetic field strength was adjusted (±0.1T), and the pre-scan was repeated until the interference peak was separated from the target peak (resolution ≥ 1.5).

[0052] In some implementations, to obtain the optimal magnetic field strength and optimal radio frequency by adjusting the static magnetic field strength and radio frequency, the magnetic field strength can be adjusted incrementally within a preset range with a set step size. Magnetic field signals are collected at each magnetic field strength. The magnetic field strength corresponding to the point where the signal-to-noise ratio of the alcohol characteristic in the magnetic field signal reaches its maximum value is then determined as the optimal magnetic field strength. Based on the optimal magnetic field strength, the theoretical radio frequency is calculated. Then, a scan is performed within a set proximity range of the theoretical radio frequency, and the response signals at each frequency point are collected. The frequency corresponding to the peak value of the response signal is determined as the optimal radio frequency.

[0053] The signal-to-noise ratio (S / N) of alcohol characteristics refers to the ratio of the signal intensity of the characteristic resonance peak of alcohol substances (in the range of 1.2-1.5ppm) to the background noise intensity in the magnetic field signal. The higher the S / N, the purer the alcohol signal and the less interference.

[0054] The theoretical radio frequency is calculated based on the optimal static magnetic field strength and is theoretically capable of exciting hydrogen nuclei resonance in alcohols. It serves as the benchmark for subsequent frequency sweeps.

[0055] The response signal refers to the intensity of the resonance signal generated when the hydrogen nuclei of alcohols are excited at different radio frequency.

[0056] When determining the optimal magnetic field strength, a gradient increment-signal feedback adjustment mode can be used to accurately lock the static magnetic field strength suitable for alcohols. First, set the preset range of the static magnetic field strength to 0.5-1.5T (this range is suitable for the resonance characteristics of hydrogen nuclei in agarwood alcohols and covers the magnetic field adjustment capabilities of most NMR detection equipment), and set the step size to 0.05T (too small a step size will increase the detection time, too large a step size may miss the optimal value; 0.05T is a balance between accuracy and efficiency).

[0057] The pretreated agarwood sample was placed in a dedicated PEEK container and then placed in the NMR detection chamber. The magnetic field strength was gradually increased from 0.5T. After each adjustment, the magnetic field was kept stable for 10-15 minutes (to avoid signal drift caused by magnetic field fluctuations). Then, a fixed-frequency radio frequency pulse was emitted (the initial frequency was set to 21.29MHz, corresponding to the theoretical resonance frequency of alcohol hydrogen nuclei under a 0.5T magnetic field), and the free induction decay (FID) signal was acquired over 10-15 seconds. The characteristic resonance peak intensity of alcohols in the 1.2-1.5ppm range was extracted from the acquired magnetic field signal using professional spectral analysis software (such as MestReNova), and the background noise intensity in this range was calculated to obtain the signal-to-noise ratio of alcohol characteristics. When the S / N ratio reached its maximum value (usually ≥30:1), the current magnetic field strength B0 was recorded (i.e., the optimal static magnetic field strength suitable for the alcohols in this sample, which is usually measured in the range of 0.8-1.2T). If the maximum S / N value is not found within the 1.5T range, extend the adjustment range to 1.6-1.8T (for high-altitude areas or special varieties of agarwood, a higher magnetic field may be required due to differences in sample density), and repeat the above operation.

[0058] For example, when the magnetic field strength is adjusted to 1.0T, the intensity of the characteristic peak of alcohol in the acquired signal is 6.8mV, the background noise intensity is 0.21mV, and the signal-to-noise ratio is approximately 32.4:1. When the magnetic field strength is further adjusted to 1.05T, the signal-to-noise ratio drops to 31.7:1, while when adjusted to 0.95T, the signal-to-noise ratio is 31.9:1. It can be seen that the signal-to-noise ratio reaches its maximum value and is ≥30:1 at 1.0T (meeting the signal purity requirement). Therefore, 1.0T is determined as the optimal magnetic field strength for this sample.

[0059] When determining the optimal radio frequency, based on the determined optimal magnetic field strength, the formula ν0=γ is used. The theoretical radio frequency (RF) frequency is calculated using B0 / 2π, where γ is the hydrogen nucleus gyromagnetic ratio (a fixed value of 42.58 MHz / T), and B0 is the optimal magnetic field strength. Taking the optimal magnetic field strength of 1.0T as an example, substituting into the formula, the theoretical RF frequency is calculated to be 42.58 MHz / T × 1.0T ÷ 2π ≈ 42.58 MHz. The adjacent scan range for the theoretical RF frequency is set to ±0.5 MHz (i.e., 42.08 MHz - 43.08 MHz, which covers the actual resonant frequency shift caused by sample differences and equipment errors), and the sweep step size is set to 0.01 MHz (to ensure accurate capture of peak signals).

[0060] Within this scanning range, radio frequency pulses were emitted sequentially (pulse width 90°, duration 10μs). For each frequency emitted, the FID signal was acquired for 20-25 seconds and the signal peak value was recorded. For example, the signal peak value was 7.2mV at 42.57MHz, 7.1mV at 42.58MHz, and 7.0mV at 42.56MHz. After plotting the "RF frequency - signal peak value" curve, it can be seen that the signal peak value corresponding to 42.57MHz is the highest, indicating that this frequency can most effectively excite the hydrogen nucleus resonance of alcohols. Therefore, 42.57MHz was determined as the optimal RF frequency ν1 for this sample (the measured deviation is usually <0.02MHz), and this frequency was locked (ν1 was used for all subsequent acquisitions).

[0061] In the above implementation process, the magnetic field strength optimization takes maximizing the signal-to-noise ratio of alcohol characteristics as the core objective. By orderly adjusting the preset range and set step size, it avoids the inefficiency caused by blind search and can accurately locate the magnetic field environment most conducive to alcohol signal recognition, significantly improving the distinction between signal and noise. The radio frequency optimization is based on the theoretical benchmark of the optimal magnetic field strength derivation, and then captures the response peak by sweeping the frequency in the vicinity range to ensure the precise coupling of radio frequency and magnetic field strength, maximize the excitation of hydrogen nuclei resonance in alcohol substances, and reduce the false excitation of hydrogen nuclei of non-target components.

[0062] To verify specificity, signals from "pure linalool standard", "agarwood sample" and "pure lignin sample" can be collected at frequency ν1. If the peak shape matching degree between the agarwood sample signal and the pure linalool standard signal is ≥95% and the signal intensity of the lignin sample is <5% of the signal intensity of the agarwood sample, then the frequency lock is considered successful (ensuring no interference from lignin or cellulose).

[0063] To further reduce interference from non-target components such as lignin and moisture, a dual filtering technique of "time-domain filtering + frequency-domain filtering" can be used to purify the signal. First, time-domain filtering is performed by applying a 5ms Hanning window function to the acquired free induction decay (FID) signal. This effectively suppresses noise interference at the signal tail while selectively retaining the effective signal within the first 15ms. This is because the FID signal of alcohols decays slowly, and the signal intensity within the first 15ms accounts for more than 80% of the total intensity, while the signals of non-target components such as lignin decay rapidly, leaving virtually no effective signal residue after 15ms, thus achieving preliminary signal screening. Next, frequency-domain filtering is performed. The time-domain filtered signal is converted into a frequency spectrum using Fourier transform, and then digital filtering algorithms such as finite impulse response filters are used to retain only the characteristic signals of alcohols in the 1.2-1.5ppm range, accurately filtering out lignin signals (0.8-1.0ppm), moisture signals (4.5-5.0ppm), and other frequency band interference. Finally, signal verification is required. The purity of the filtered signal is calculated using the formula "target frequency band signal strength / total signal strength × 100%", and the purity is required to be ≥98%. If this standard is not met, the radio frequency needs to be readjusted until the signal purity meets the requirements.

[0064] In some implementations, to ensure the repeatability and reliability of the detection results, the specific signal can be repeatedly acquired and its stability verified under the condition of locking the optimal static magnetic field strength (B0) and the optimal radio frequency (ν1). In specific operation, a single excitation pulse can be emitted, and a baseline signal of a preset duration can be acquired. If the intensity fluctuation of the baseline signal does not exceed a set threshold, the subsequent step, namely the step of emitting the preset pulse, is executed.

[0065] For example, three consecutive single excitation pulses are emitted, and three consecutive FID signals, i.e., baseline signals, are acquired. The acquisition interval is set to 5 minutes to avoid errors caused by short-term magnetic field fluctuations or slight sample displacement. After acquisition, the peak coefficient of variation (CV) of the three signals can be calculated, which represents the intensity fluctuation of the baseline signal. This coefficient should be ≤2% (a set threshold) to indicate signal stability and no significant magnetic field drift or sample displacement. If the calculated CV >2%, it is necessary to promptly check whether the sample has shifted during the detection process and whether the magnetic field has remained stable. After troubleshooting and adjustment, signal acquisition should be repeated until the peak coefficient of variation meets the stability requirement of ≤2%.

[0066] By adding a baseline signal detection step before transmitting the preset pulse, the stability of the detection system can be effectively verified in advance by acquiring the baseline signal for a preset duration under optimal parameter conditions and judging whether its intensity fluctuation meets the set threshold, thus avoiding distortion of subsequent signal acquisition caused by parameter drift or abnormal equipment noise.

[0067] Step S120: While maintaining the optimal magnetic field strength and optimal radio frequency, transmit a preset pulse sequence and acquire the target mixed signal.

[0068] To maintain optimal magnetic field strength and optimal radio frequency, parameter drift is avoided from affecting signal stability. The detection system can then emit a preset pulse sequence to the agarwood sample to be tested. The preset pulse sequence can refer to a series of radio frequency pulses emitted in a time sequence, which includes a modified spin echo (SE) sequence and a Carr-Purcell-Meiboom-Gill (CPMG) sequence.

[0069] The target mixed signal refers to the raw NMR signal that has been collected but not yet separated, which may contain the superposition of signals from various hydrogen-containing components such as water, oils and alcohols in the sample.

[0070] Step S130: Obtain the signal characteristics of alcohols based on the target mixed signal.

[0071] After obtaining the target mixture signal, the signal of alcohol substances can be extracted from the target mixture signal by taking advantage of the differences in the physical properties of different components.

[0072] Here, the signal characteristics of alcohols can be obtained from the target mixed signal. Signal characteristics refer to physical quantities that can represent alcohols and can be used for quantification, such as characteristic signal intensity.

[0073] Step S140: Determine the concentration of alcohols in the agarwood sample to be tested based on the signal characteristics.

[0074] After obtaining the signal characteristics of alcohols, they can be converted into chemical concentration values ​​using a standard curve. A mathematical model, pre-established by measuring a series of linalool standard solutions of known concentrations, describes the quantitative relationship between signal intensity and concentration, typically a linear one. Therefore, the concentration of the alcohol can be calculated by substituting the signal intensity corresponding to the signal characteristics into the pre-established standard curve equation.

[0075] For example, the standard curve equation is signal intensity = k × concentration + b, where k and b are constants determined by fitting with standards. Substituting the signal intensity, the percentage concentration of alcohols in the agarwood sample can be directly calculated and output.

[0076] In some implementations, signal features can be input into a pre-trained neural network model, which then predicts the corresponding concentration. Understandably, the neural network model learns the correlation between signal features and concentration during pre-training, thus enabling real-time concentration prediction.

[0077] In the above-mentioned process, nuclear magnetic resonance technology can be used to perform non-destructive testing on agarwood samples without the need for complicated pretreatment procedures, which can significantly improve the detection efficiency. By locking the optimal magnetic field strength and optimal radio frequency for exciting the hydrogen nuclei of alcohol substances, and combining the targeted acquisition and signal feature extraction of the target mixed signal, the specificity and signal-to-noise ratio of alcohol signals are significantly improved. This effectively removes interference from non-target components such as moisture, oil, and lignin in agarwood samples, ensuring detection accuracy and repeatability.

[0078] Traditional interference signal separation relies on a single relaxation time difference, which is insufficient to simultaneously distinguish signals from oil, water, and alcohol. This proposed solution employs a two-parameter coupling mechanism of longitudinal relaxation (T1) and transverse relaxation (T2). Experimental measurements show that the T2 value for water in agarwood is 20-50 ms, for oil it is 200-300 ms, and for alcohol it is 80-120 ms, indicating significant differences in relaxation characteristics among the three (difference > 40%). This characteristic can be used to design a stepped pulse sequence to accurately separate oil, water, and alcohol signals, solving the problem of multi-component interference in complex matrices.

[0079] To achieve the separation of interference signals, in the implementation method of transmitting a preset pulse sequence and acquiring the target mixed signal, a first radio frequency pulse sequence can be transmitted first, followed by a second radio frequency pulse sequence after a first delay time, and a first mixed signal can be acquired after a second delay time. The first mixed signal includes signals of oil and alcohol substances. The first delay time is less than the transverse relaxation time of water, and the second delay time is greater than the transverse relaxation time of water. Then, a third radio frequency pulse sequence is transmitted, and multiple second mixed signals arranged in chronological order are acquired. The second mixed signals include signals of oil and alcohol substances. The target mixed signal includes the first mixed signal and the second mixed signal.

[0080] The first radio frequency pulse sequence can be a 90° spin echo sequence, which is used to flip the magnetization vector of all hydrogen nuclei (hydrogen nuclei of water, oil and alcohol substances) in the agarwood sample from the Z-axis to the XY plane, providing a basis for signal acquisition.

[0081] The second radio frequency pulse can be a 180° spin echo sequence, which is used to refocus the magnetization vector in the XY plane, correct the signal astigmatism caused by the non-uniform magnetic field, and ensure that the subsequently acquired signal is more stable.

[0082] Transverse relaxation time (T2) refers to the time constant of the natural decay of the transverse magnetization vector due to the interaction between hydrogen nuclei. The shorter the T2, the faster the signal decays. The T2 value of water in agarwood is usually 20-50ms, which is shorter than the T2 value of oil (200-300ms) and alcohol (80-120ms).

[0083] The pretreated agarwood sample was placed in a dedicated PEEK container and then placed in the nuclear magnetic resonance (NMR) detection chamber. The previously locked optimal static magnetic field strength and optimal radio frequency were maintained to ensure the specificity of the signal acquisition. The pulse interval (TE) was set to twice the moisture T2 value (i.e., 100 ms, to ensure complete attenuation of the moisture signal), and the repetition time (TR) was set to 2000 ms (greater than the T1 values ​​of all components to avoid signal saturation). The pulse emission and signal acquisition process was then initiated. The first step involves transmitting a first radio frequency pulse sequence (90° pulse, 12-15 μs width) to the agarwood sample to be tested. This pulse causes the hydrogen nucleus magnetization vectors of water, oil, and alcohol in the sample to synchronously flip to the XY plane. A first delay time is then set, which is less than the transverse relaxation time of water (20-50 ms), specifically less than half of the transverse relaxation time of water, for example, set to 10 ms (for a sample where water T2 = 30 ms, 10 ms < 15 ms). At this point, the transverse magnetization vector of water has not yet significantly decayed (attenuation ≤ 30%), while the magnetization vectors of oil and alcohol show almost no decay (because their T2 is much greater than that of water).

[0084] At the end of the delay time, the initial mixed signal (hereinafter referred to as the third mixed signal) is acquired by the detection system. This signal completely contains the magnetic resonance signals of water, oil, and alcohol (denoted as S1 = water + oil + alcohol), and is the acquisition result under the first delay time. For example, the intensity of the acquired initial mixed signal is 8.5mV, of which the water signal accounts for 1.2mV, the oil signal accounts for 3.3mV, and the alcohol signal accounts for 4.0mV.

[0085] The second step involves immediately transmitting a second radio frequency pulse sequence (180° pulse, 24-30 μs width) to the sample after acquiring the initial mixed signal. This pulse causes the magnetization vector in the XY plane to flip by 180°, achieving signal refocusing and compensating for signal astigmatism errors caused by magnetic field inhomogeneity. A second delay time is then set, which must be greater than the transverse relaxation time of water, such as twice the transverse relaxation time of water, for example, set to 100 ms (taking water T2=30 ms as an example, 100 ms > 60 ms). At this point, the transverse magnetization vector of water has completely decayed (attenuation ≥ 95%), and the signal strength approaches 0, while the transverse magnetization vectors of oil and alcohols still maintain a relatively high intensity (oil attenuation ≤ 20%, alcohol attenuation ≤ 40%).

[0086] At the end of the delay time, the first mixed signal was acquired again by the detection system. This signal, now free of moisture interference, contains signals of oil and alcohol, denoted as S2 = oil + alcohol. This is the acquisition result under the second delay time, after which moisture interference has been removed. For example, if the acquired target mixed signal intensity is 7.3mV, it perfectly matches the sum of the oil and alcohol signals in the initial mixed signal (3.3mV + 4.0mV = 7.3mV), verifying that the moisture signal has been successfully separated.

[0087] The core logic of the entire process is to utilize the difference in lateral relaxation time between water, oil, and alcohols. By designing a "short delay to retain water and long delay to remove water" approach, the interference signal is initially separated, laying the foundation for further separation of oil and alcohols.

[0088] In this method, the TE value can be determined through calibration experiments. For example, take an agarwood sample with known moisture content, set TE = 50ms, 80ms, 100ms, and 120ms, and collect the initial mixed signal S1 and the target mixed signal S2 respectively. Calculate the deviation between ΔS1 = S1 - S2 and the actual moisture content, and select the TE value with the smallest deviation (usually 100ms).

[0089] Then, after the second delay time, a third radio frequency pulse sequence is emitted, which can be a CPMG sequence. The CPMG sequence is a nuclear magnetic resonance pulse sequence based on continuous spin echoes. Its core consists of a 90° radio frequency pulse and multiple subsequent 180° refocusing pulses. It can efficiently acquire multiple sets of echo signals and accurately capture the transverse relaxation differences of different components. It is a key technology for separating oil and alcohol substances.

[0090] The second mixed signal refers to the detectable signal formed by the reconvergence of the dichroic transverse magnetization vectors after the 180° refocusing pulse. Multiple second mixed signals are arranged in the order of acquisition time to form a sequence of data that reflects the signal attenuation law.

[0091] First, set the parameters of the CPMG sequence: apply 30 consecutive 180° pulses after the 90° pulse, echo interval (TE) = 10ms, total acquisition time = 30 × 10ms = 300ms. This time setting can cover the long transverse relaxation (T2) process of oil.

[0092] When acquiring CPMG sequences from agarwood samples, a 90° radio frequency pulse (12-15 μs wide) is first emitted to synchronously flip the hydrogen nuclei magnetization vectors of oil and alcohol to the XY plane. Then, 30 consecutive 180° refocusing pulses are applied at fixed intervals (echo interval TE = 10 ms). After each 180° pulse, the dephased transverse magnetization vectors reconverge, and the detection system synchronously acquires the echo signals, ultimately obtaining 30 echo signals arranged in chronological order (E1 to E...). 30 This is the second mixed signal, with a time span covering 0-300ms (just covering the maximum value of oil T2, ensuring the capture of the complete decay pattern).

[0093] The first and second mixed signals collected are the target mixed signals. Based on these two mixed signals, the signal characteristics of alcohols can be extracted.

[0094] In the above implementation process, by transmitting the first and second radio frequency pulse sequences in stages and precisely matching a first delay time less than the transverse relaxation time of water and a second delay time greater than the transverse relaxation time of water, the difference in transverse relaxation characteristics between water and oil / alcohol substances is cleverly utilized to achieve efficient and accurate separation of the water signal. Then, by acquiring multiple time-sequentially arranged second mixed signals through a third radio frequency pulse sequence, the differences in the attenuation patterns of oil and alcohol substances are fully captured, providing sufficient data support for the signal separation of the two substances.

[0095] Based on the above embodiments, in order to achieve the separation of alcohols and oils, in the implementation of obtaining the signal characteristics of alcohols based on the target mixed signal, multiple second mixed signals can be fitted with exponential decay curves based on the difference in transverse relaxation time between alcohols and oils. Then, based on the fitting results and the first mixed signal, the initial signal intensity of alcohols can be obtained, and the signal characteristics include the initial signal intensity.

[0096] After acquiring multiple second-mixed signals, the alcohol decay curve can be obtained by fitting a double exponential decay curve to the 30 sets of echo signals based on the difference in transverse relaxation time between oil and alcohol. And oil decay curve ,in T2A represents the initial signal intensity of the alcohol, and T2A represents the transverse relaxation time of the alcohol. Let T2O be the initial signal intensity of the oil component, and T2O be the transverse relaxation time of the oil component. A fitting model is established using professional data analysis software (such as Origin). Where E(t) is the echo signal intensity at time t. Substituting the acquired echo signal data and the signal intensities at each time point in the first mixed signal, the calculation is performed. For example, after fitting, T2A = 98 ms. =5.2mV, T2O=245ms =3.8mV, with a goodness-of-fit R² ≥ 0.995, indicating that the separation results are reliable. Ultimately, the core signal characteristic of alcohols is the calculated initial signal intensity. (e.g., 5.2mV) This value is not affected by the decay process and is only positively correlated with the content of alcohols, providing an accurate basis for subsequent calculation of concentration by substituting into the standard curve.

[0097] Optionally, when obtaining the initial signal intensity of alcohols, theoretically the fitting result can be used to calculate it, that is, the target mixed signal can include the second mixed signal. However, in order to ensure accuracy, the signal intensity at each time point in the first mixed signal can be combined for comprehensive calculation.

[0098] Understandably, when plotting the signal attenuation curves of 30 echo signals, oil components, due to their longer T2 (200-300 ms), still maintain over 50% intensity at the 20th echo (200 ms); alcohols, due to their shorter T2 (80-120 ms), have intensity reduced to below 30% of their initial value by the 10th echo (100 ms). Therefore, the 1st to 10th echo signals can be selected to calculate the initial signal intensity of alcohols. The initial signal intensity refers to the original signal intensity before the component has attenuated, and is only positively correlated with the substance content.

[0099] In some implementations, the aforementioned echo signal is determined to be 30, which is determined experimentally. For example, CPMG is used to collect pure oil samples (agarwood essential oil) and pure alcohol samples (agarwood spirol standard), and attenuation curves are plotted to determine the minimum number of echoes that can completely distinguish the two (experiments have verified that 30 echoes can meet the requirements).

[0100] In the above implementation process, based on the difference in transverse relaxation time between alcohols and oils, multiple second mixed signals (time-series echo signals) are fitted with exponential decay curves, which can fully capture the difference in decay patterns between the two and provide a scientific basis for oil-alcohol signal separation.

[0101] Based on the above embodiments, after obtaining the decay curve A(t) of alcohol by double exponential fitting, in order to eliminate the interference of trace bound water signals that may not be completely stripped in the fitting, the alcohol decay signal can be subjected to FFT transformation to obtain the frequency signal, and then the frequency signal can be bandpass filtered to obtain the filtered alcohol decay signal. The initial signal intensity of alcohol can be determined based on the filtered alcohol decay signal.

[0102] The decay curves of alcohols were obtained by fitting the CPMG sequence, forming a complete time-domain signal sequence (i.e., the full-time-axis signal data corresponding to the alcohol decay curves obtained after fitting, covering the complete process from signal excitation to decay). This sequence may contain trace amounts of residual bound water signals (assuming that a sample contains bound water with a T2 of 60 ms, whose signal has not been completely stripped). Next, an FFT transformation was performed on this time-domain signal sequence to convert the time-signal intensity data into a frequency-domain spectrum of chemical shift-signal intensity. After the conversion, it can be clearly observed that the characteristic signals of alcohols are concentrated in the 1.2-1.5 ppm range (signal intensity peak of 5.2 mV), while the residual signals of trace bound water are concentrated around 4.7 ppm (signal intensity of approximately 0.15 mV). The chemical shifts of the two are significantly different, meeting the conditions for filtering and separation.

[0103] Bandpass filtering was then performed. Based on the characteristic chemical shift range of alcohols, the parameters of the bandpass filter were set: the center frequency was locked in the 1.2-1.5 ppm range, and the bandwidth was set to 0.5 ppm (to ensure complete coverage of the alcohol characteristic signal while avoiding omission of effective components). The filtering algorithm used a finite impulse response (FIR) filter, which has linear phase characteristics and can avoid signal distortion. The frequency domain spectrum was processed by the filter, allowing only alcohol signals in the 1.2-1.5 ppm range to pass, deeply attenuating the bound water residue signal near 4.7 ppm (attenuation ≥99%), while blocking the possible 0.8-1.0 ppm lignin residue signal. After filtering, an inverse FFT transformation was performed on the frequency domain signal to convert it back to the time domain signal sequence. The initial signal intensity of alcohols was extracted again. After filtering and purification, the initial signal intensity was 5.18 mV (only the 0.15 mV bound water residue signal was removed), and the signal purity was improved from 97.2% before filtering to 99.7%, meeting the accuracy requirements for subsequent concentration calculations. For example, if the initial signal strength of a sample before filtering is 6.8mV, which includes 0.22mV of bound water residual signal, after bandpass filtering, the initial signal strength after filtering is 6.58mV, effectively eliminating the interference of bound water on the quantitative results (avoiding the calculated alcohol concentration value being about 3.2% higher due to residual signal).

[0104] In the above implementation process, the FFT transform converts the time-domain signal into a frequency-domain signal, making the frequency distribution differences between the alcohol characteristic signal and residual interference signals such as bound water and trace lignin clearly visible, providing a basis for accurate filtering; the bandpass filter specifically retains the alcohol characteristic frequency band signal, effectively blocks interference from other frequency bands, and significantly improves signal purity.

[0105] Based on the above embodiments, when determining the alcohol concentration of the agarwood sample to be tested, the alcohol concentration corresponding to the initial signal intensity can be determined according to the preset correlation between alcohol concentration and signal intensity.

[0106] First, a pre-defined correlation between alcohol concentration and signal intensity (standard curve) can be established in advance: prepare a series of agarwood alcohol standard solutions with concentrations covering the actual alcohol content range of agarwood (0.5%-30%), including 0.5%, 1%, 2%, 5%, 8%, 10%, 15%, 20%, 25%, and 30%, with 3 parallel samples set for each concentration, and anhydrous ethanol (water content ≤0.1%) as the solvent. To address the differences in signal response across different concentration ranges, a "segmented gradient" approach can be adopted: for the low concentration range (0.5%-2%), an arithmetic gradient (0.5%, 1%, 2%) with intervals of 0.5%-1% can be used to resolve fitting biases caused by weak signals and sparse data points at low concentrations; for the medium concentration range (5%-15%), a geometric gradient (5%, 8%, 10%, 15%) with intervals of 1.5%-5% can be used to cover the mainstream alcohol content range of agarwood (5%-12%); and for the high concentration range (20%-30%), an arithmetic gradient (20%, 25%, 30%) with intervals of 5% can be used to meet the detection needs of agarwood with high alcohol content (such as Qinan).

[0107] Under the same parameter conditions as the test sample (the same optimal static magnetic field strength, optimal radio frequency, and pulse sequence parameters), nuclear magnetic resonance signals were acquired for each concentration of standard. After the same processing procedures such as time-domain filtering, frequency-domain filtering, and CPMG sequence separation, the initial signal intensity after filtering of each group of standard was extracted, and the average value (S_avg) of 3 parallel samples was taken as the signal intensity data corresponding to that concentration (for example, the average signal intensity of 5% concentration standard is 129.0mV, and that of 10% concentration is 258.3mV).

[0108] The acquired signal intensity data can be preprocessed, such as outlier removal: use the Dixon test (α=0.05) to remove outliers in 3 replicates (e.g., S_avg deviation > 5%); matrix correction: introduce a correction factor K=S_standard / S_blank (S_standard is the standard signal, S_blank is the mixed solvent blank signal) to eliminate solvent background interference, and the K value should be stable at 1.02±0.01.

[0109] Weighted least squares regression was used for linear regression fitting, with a weighting coefficient w = 1 / c (c being the concentration) to improve fitting accuracy in the low concentration range. The fitting equation S_avg = a·c + b was forced to pass through the origin (intercept b = 0) to avoid quantitative errors in low concentrations caused by intercept bias. The final standard curve equation was S_avg = 25.8 × c (R² = 0.999), where the slope 25.8 is the conversion coefficient between signal intensity and concentration. This equation represents the preset correlation. After fitting, validation tests were conducted using 8% (medium concentration) and 25% (high concentration) samples to ensure that the relative error between the predicted and actual values ​​was <1%. Once the validation was successful, the curve was put into use.

[0110] To address the instrument drift issue, a time-concentration dual-factor correction can be added. For time correction, calibrate the signal with a 5% standard every 2 hours. If the signal deviation is >2%, adjust the RF power (±0.5dB). For concentration correction, after fitting, test with 8% (medium concentration) and 25% (high concentration) verification samples. The relative error between the predicted and actual values ​​should be <1%. Otherwise, refit.

[0111] The data collected in the above example is shown in the table below:

[0112] When calculating the concentration of alcohols in the agarwood sample, the initial signal intensity after filtering following full-process signal processing is first obtained, such as 8.92. This signal intensity value is then substituted into the preset standard curve equation (S_avg=25.8×c), and the preliminary concentration is calculated in reverse using the formula c=S_avg / 25.8. Substituting this into the example data, c=8.92 / 25.8≈34.57% is obtained.

[0113] In the above implementation process, the concentration of alcohols can be directly determined by querying the correlation relationship, without the need for complicated on-site calibration or additional experimental operations, which greatly improves the detection efficiency.

[0114] Based on the above embodiments, after obtaining the target mixed signal, the uniformity of moisture distribution of the agarwood sample to be tested can also be analyzed. For example, after a first delay time, a third mixed signal (the initial mixed signal S1 mentioned above) is collected. Based on the first mixed signal and the third mixed signal, a moisture signal is obtained. Based on the moisture signal, the change in decay rate at different time periods is obtained. Based on the decay rate number, the uniformity of moisture distribution inside the agarwood sample to be tested is evaluated.

[0115] The moisture signal can be calculated by the difference between the third mixed signal S1 and the first mixed signal S2, and is a magnetic resonance signal containing the moisture in the agarwood sample to be tested.

[0116] Moisture signal can refer to the intensity of moisture signal at different points in time. These signal intensities can form a sequence that can fully reflect the decay law of moisture signal over time.

[0117] When acquiring the decay rate change, the extracted moisture signal time-domain sequence can be segmented into segments of 80 data sets (covering a time span of approximately 80ms, taking into account both decay feature capture and calculation efficiency). For each data segment, a linear fitting algorithm is used to calculate the slope, which is the moisture signal decay rate for the corresponding time period (slope = signal intensity change / time change).

[0118] Then, the uniformity of moisture distribution can be assessed based on the decay rate of each segment, such as by using the formula U=1-|maximum slope-minimum slope| / average slope to calculate the uniformity.

[0119] The distribution status is judged by combining the uniformity value: if U≥90%, it indicates that the moisture distribution is uniform and the aging environment is stable; if 80%≤U<90%, it indicates that there is a slight difference in moisture distribution, which may be related to the difference in moisture content between the head and the edge of the sample (about 1.2%); if U<80%, it suggests that there is a local moisture imbalance in the sample, which may be due to the structural difference between the protrusions and depressions of the nodules, resulting in moisture residue.

[0120] In the above implementation process, without adding additional detection steps or damaging the sample, the moisture signal is obtained by using the difference between the first and third mixed signals that have been collected. The uniformity of moisture distribution inside the sample is evaluated by analyzing the changes in the moisture decay rate over different time periods. This not only makes full use of existing detection data resources and avoids additional detection costs and time consumption, but also adds a key auxiliary indicator for the quality assessment of agarwood.

[0121] The following examples illustrate the detection process described above.

[0122] Example 1: Detection of agarwood samples from different series.

[0123] I. Sample Selection and Preprocessing: 1. Sample Information: Three samples of 7-point agarwood from different series were selected. All samples met the core indicators of 7-point agarwood of the "Agarwood" standard (LY / T2904-2017) (alcohol-soluble extract content ≥8%, density 0.85-1.0 g / cm³, and tight bonding between xylem and oil without obvious stratification). However, they came from different series. The specific parameters are shown in the table below:

[0124] 2. Preprocessing steps: (1) Surface cleaning: Tool selection: Dust-free microfiber cloth (fiber diameter 0.3μm, surface resistivity 10) 9 -10¹¹Ω), analytical grade anhydrous ethanol (purity ≥99.7%, water content ≤0.1%), disposable pipettes (1mL specification, accuracy ±0.01mL); Procedure: Use a pipette to drop 0.3 mL of anhydrous ethanol onto a lint-free cloth, creating a 2-3 cm diameter wet area (avoiding excessive ethanol penetration into the sample). Gently wipe the sample surface in a clockwise spiral motion, focusing on cleaning the gaps at the S1 cap, residual dust in the glossy area of ​​S2, and impurities attached to the raised nodules of S3. After wiping, place the sample on a clean PTFE tray (surface roughness Ra≤0.1 μm) and let it stand at room temperature (25℃±1℃) for 10 minutes to ensure complete ethanol evaporation. Verification standard: The weights of S1, S2 and S3 were weighed using a high-precision electronic balance before and after cleaning. The weight changes were 0.008g, 0.012g and 0.005g, respectively, all ≤0.02g, indicating that there was no residual ethanol and the impurities were completely removed.

[0125] (2) Adaptation and fixation of the container: Selection of sample carrier: Use a solid sample carrier with a 0.2mm thick polytetrafluoroethylene (PTFE) anti-interference coating on the inner wall. S1 and S2 are suitable for sample carriers with an inner diameter of 10cm and a height of 15cm, and S3 is suitable for sample carriers with an inner diameter of 8cm and a height of 12cm (coating signal interference rate ≤0.3%, temperature resistance range -20℃-200℃). Fixation method: S1 and S2 are placed directly on the non-slip silicone pad (0.5cm thick, 0.2cm deep surface texture, friction coefficient ≥0.8) at the bottom of the sample dish. The center of the sample is aligned with the center of the sample dish using a laser positioning instrument (offset ≤0.3cm). S3 (scar-like) is fixed by wrapping three 1cm wide and 0.2cm thick PTFE fixing strips (with a 0.1cm thick silicone layer attached to the surface of the strip) in a "triangular distribution". The tension of the fixing strips is controlled at 5-8N (to avoid damaging the scar protrusion). After fixation, the horizontal deviation of the sample is ≤0.5°. Sealing check: Cover the sample dish with a 0.3mm vent hole (10-15mL / min air permeability), and use the smoke test method (release a small amount of smoke around the sample dish) to observe whether the smoke significantly enters the sample dish, ensuring that there is no external airflow interfering with the sample during the test.

[0126] (3) Environmental balance: Equipment parameters: Place the sample-containing plate into a constant temperature and humidity chamber, set the temperature to 25℃±1℃, the relative humidity to 50%±2%, and the equilibration time to 24h. Monitoring measures: During the equilibration process, data was recorded every 6 hours using an in-chamber temperature and humidity sensor (accuracy ±0.1℃ / ±1%RH) to ensure that temperature fluctuations were ≤0.5℃ and humidity fluctuations were ≤1%. After equilibration, the samples were weighed again. The weight changes for S1, S2, and S3 were 0.010g, 0.015g, and 0.008g, respectively, all ≤0.02g, indicating that the sample moisture content had reached a stable state, thus avoiding the influence of ambient temperature and humidity on subsequent signal acquisition.

[0127] II. Detailed testing process: 1. Initialization of magnetic resonance imaging system parameters: Start the magnetic resonance imaging system and allow it to warm up for 40 minutes. Once the magnet temperature stabilizes at 30℃±0.3℃ and the radio frequency power fluctuation is ≤0.05dB, set the basic parameters as shown in the table below:

[0128] 2. Sample characteristic pre-analysis and parameter calibration: (1) Low-power pre-scan: Operating procedure: Place the sample dish into the detection chamber (probe temperature 25℃±0.2℃), set the radio frequency power to 1 / 6 of the conventional detection (1.8dB), perform ¹H-NMR pre-scan, scan range 0-12ppm, scan 10 times, scan time 150s; Interference peak identification: The characteristic resonance peaks of alcohols in the sample (1.2-1.5 ppm, corresponding to the hydrogen nuclei of the -CH2-OH group) were identified using systematic spectral analysis software. At the same time, the presence of impurity interference peaks (moisture peak at 4.7 ppm, lignin peak at 0.8-1.0 ppm, and resin impurity peak at 2.5 ppm) was detected. The results showed that there were no obvious interference peaks in the 1.2-1.5 ppm range for S1, S2, and S3 (interference peak intensity < 6% of the target peak intensity), and no adjustment of the scanning range was required.

[0129] (2) Static magnetic field strength locking: Adjustment method: The "gradient increase-signal feedback" mode is adopted. Starting from 0.5T, the magnetic field strength is increased in increments of 0.05T. After each adjustment of the magnetic field strength, the magnetic field is kept stable for 15 minutes (to ensure that the magnetic field uniformity meets the standard). Then, a radio frequency pulse of fixed frequency (21.29MHz, corresponding to the hydrogen nucleus resonance frequency of alcohols under a 0.5T magnetic field) is emitted, and the free induction attenuation (FID) signal is collected for 15 seconds. Signal-to-noise ratio (S / N) calculation: The S / N in the range of 1.2-1.5 ppm is calculated by software. When S / N reaches the maximum value (≥30:1), the current magnetic field strength is recorded as the optimal magnetic field strength. Finally, the optimal magnetic field strengths are locked as follows: S1 is 0.9T (S / N=32:1), S2 is 1.1T (S / N=35:1), and S3 is 1.0T (S / N=33:1).

[0130] (3) Radio frequency matching: Theoretical frequency calculation: According to the formula ν=γ·B0 / 2π (where γ=42.58MHz / T is the gyromagnetic ratio of hydrogen nuclei, and B0 is the optimal static magnetic field strength), the theoretical resonance frequencies of S1 are calculated to be 42.58MHz / T×0.9T=38.32MHz, S2 is 42.58MHz / T×1.1T=46.84MHz, and S3 is 42.58MHz / T×1.0T=42.58MHz; Frequency sweep locking: A frequency sweep is performed within the theoretical frequency ±0.5MHz range, with a sweep step size of 0.005MHz. The FID signal is acquired for 25s after each sweep, and the signal peak value is recorded. A "RF frequency - signal peak value" curve is plotted. The frequency corresponding to the highest point of the curve is the optimal RF frequency. Finally, the optimal RF frequencies are locked as follows: S1 is 38.31MHz, S2 is 46.83MHz, and S3 is 42.57MHz (frequency deviation ≤0.01MHz, ensuring complete matching with the Zeeman energy level difference of alcohol hydrogen nuclei). Specificity verification: Under the locked frequency, signals from "pure linalool standard (concentration 10%)", "agarwood sample" and "pure lignin sample" were collected respectively. The peak shape matching degrees of S1, S2 and S3 with the pure linalool standard signal were 96.2%, 97.5% and 96.8% respectively. The signal intensity of the lignin sample was less than 4% of the signal intensity of the agarwood sample, indicating that the frequency lock was successful.

[0131] 3. Magnetic resonance signal acquisition: (1) Specific signal acquisition: Operation procedure: Under the locked optimal magnetic field strength and optimal radio frequency parameters, start the signal acquisition program and continuously acquire NMR signals within 20 minutes. The system automatically records 1 set of data every 1 second (a total of 1200 sets of data) and generates NMR signal attenuation curves for S1, S2 and S3 respectively. Real-time monitoring: During the acquisition process, the system's real-time monitoring module observes the signal strength fluctuation to ensure that the signal strength fluctuation amplitude is ≤1.5%. If the fluctuation exceeds the limit (e.g., ≥2%), the acquisition is immediately paused to check whether the sample has shifted or whether the magnetic field is stable (verified by the magnetic field uniformity detection module). After adjustment, the acquisition is restarted. In this embodiment, all three samples were successfully acquired on the first attempt, with no fluctuation exceeding the limit.

[0132] (2) Interference signal separation and acquisition: First acquisition (S total): A 90° radio frequency pulse is emitted to flip the magnetization vector of all hydrogen nuclei (water, oil, alcohol) in the sample from the Z-axis to the XY plane. The FID signal is acquired after a delay of t1=30ms (less than 1 / 2 of the water T2 value to ensure that the water signal is not significantly attenuated). The signal intensity S total obtained at this time is the sum of the water, oil and alcohol signals. Secondary acquisition (S oil alcohol): Apply a 180° radio frequency pulse to flip the magnetization vector of the XY plane by 180°, and acquire the spin echo signal when the delay is t2=60ms (2t1). At this time, the water has completely decayed due to the short T2 (20-50ms). The obtained signal intensity S oil alcohol is the sum of the oil and alcohol signals. CPMG sequence acquisition: The S oil alcohol signal was acquired using the CPMG sequence. A total of 30 echo signals were acquired, and the signal intensity of each echo (E1 to E30) was recorded for subsequent separation of alcohol and oil signals. During the acquisition process, the attenuation trend of the echo signal intensity followed an exponential law, with no abnormal jumps.

[0133] III. Signal Acquisition and Processing Methods: 1. Raw signal preprocessing: (1) Blank correction: Blank signal acquisition: Place an empty sample dish (consistent with the sample dish used for sample detection) into the detection chamber and acquire the background signal (S_empty) under the same parameters. Acquire the signal 6 times consecutively and take the average value as the final blank signal value. In this embodiment, S_empty = 0.08mV. Correction calculation: Blank correction is performed on the original sample signal (Soriginal). The correction formula is Scorrected = Soriginal - Sblank. This eliminates background interference from the material of the sample dish (PEEK) and the detection environment (trace moisture in the air). After correction, the initial signal intensities of S1, S2, and S3 are 6.98mV, 7.72mV, and 8.92mV, respectively.

[0134] (2) Outlier removal: Testing method: The Grubbs test (significance level α = 0.05) was used to test for outliers in the corrected signal data. According to the Grubbs test critical value table, when the sample size n = 1200, the critical value G... 0.05 =3.3; Results of the removal: Calculate the residuals for each data point (residual = data value - mean). There are 2 data points in S1 with residual absolute values ​​greater than 3.3 times the standard deviation, 1 in S2, and 3 in S3. Remove these outliers (removal rate ≤ 0.3%) to ensure data reliability.

[0135] (3) Time-domain filtering: Filtering method: The signal data after removing outliers is processed by the Hanning window function with a window width of 6ms to suppress noise interference at the end of the signal (the noise signal is mainly concentrated after 18ms and the intensity is ≤0.04mV). Valid signal retention: The valid signal within the first 18ms is retained (the signal intensity of alcohol substances accounts for more than 85% of the total signal intensity during this period), laying the foundation for subsequent interference separation. After filtering, the signal smoothness of S1, S2 and S3 is improved by 25%, 28% and 23% respectively.

[0136] 2. Interference signal separation and processing: (1) Moisture signal separation: Calculation method: The signal intensity of pure water, S_water, is obtained based on the signal difference between S_total and S_oilol. The formula is S_water = S_total - S_oilol. Calculation results: The moisture signal intensities of S1, S2, and S3 were calculated automatically by the software to be 0.82mV, 1.15mV, and 0.58mV, respectively. These values ​​are consistent with the moisture content detection results in the sample pretreatment stage (S1=7.8%, S2=8.5%, S3=7.2%) (the moisture signal intensity is positively correlated with the moisture content, with a correlation coefficient R=0.98), verifying the accuracy of moisture signal separation.

[0137] (2) Alcohol-oil separation: Attenuation curve fitting: Exponential fitting was performed on 30 echo signals acquired from the CPMG sequence to obtain the attenuation curves of the alcohols. And the decay curve of oil content (Where A0 is the initial signal intensity of alcohols, T2A is the transverse relaxation time of alcohols, and the measured values ​​are S1=92ms, S2=105ms, and S3=98ms; O0 is the initial signal intensity of oils, and T2O is the transverse relaxation time of oils, and the measured values ​​are S1=215ms, S2=245ms, and S3=220ms). Simultaneous equation solution: By solving the simultaneous equation E(t)=A(t)+O(t) (where E(t) is the signal intensity of the t-th echo), the initial signal intensity A0 of the alcohol is obtained, and the results are S1=6.98mV, S2=7.72mV, and S3=8.92mV; Verification: Comparing the correlation between the fitted curve and the actual acquired signal, R² is ≥0.996, indicating that the fitting accuracy meets the requirements and there is no obvious deviation.

[0138] (3) Frequency domain cleanup: Fourier Transform: The initial alcohol signal A0 is converted into a frequency domain spectrum using Fast Fourier Transform (FFT) to obtain the frequency distribution of the signal; Bandpass filtering: A digital bandpass filter with a center frequency of 1.2-1.5ppm and a bandwidth of 0.4ppm is used to filter out residual moisture signals around 4.7ppm (attenuation ≥99.2%) and lignin signals around 0.8-1.0ppm (attenuation ≥98.5%). Purity verification: After filtering, the purity of the alcohol signal was calculated (target frequency band signal strength / total signal strength × 100%). S1, S2, and S3 were 99.6%, 99.7%, and 99.5%, respectively, all ≥ 99.5%, meeting the requirements for subsequent content calculation, and there was no residual influence from impurities.

[0139] 3. Calculation of alcohol content: (1) Reference to the standard curve: A pre-established standard curve of alcohol concentration versus signal intensity was constructed using the following steps: Preparation of standard products: Prepare linalool standard solutions with concentrations of 0.5%, 1%, 2%, 5%, 10%, 15%, 20%, 25%, and 30% (solvent is anhydrous ethanol, purity ≥99.7%), and set up 3 parallel samples for each concentration; Signal acquisition: Under the same magnetic resonance parameters as the sample detection (static magnetic field strength 1.0T, radio frequency 42.58MHz, signal acquisition interval 1s), the FID signals of the standards at each concentration were acquired and the signal intensity in the range of 1.2-1.5ppm was recorded. Fitting calculation: With alcohol concentration as the abscissa and average signal intensity as the ordinate, a standard curve was obtained by using the weighted least squares method (weighting coefficient w=1 / c, where c is the concentration). The fitting equation is S_avg=25.8×c (where S_avg is the average signal intensity, in mV; c is the alcohol concentration, in %), with a coefficient of determination R²=0.999 and a linear error <0.8%, which meets the accuracy requirements for quantitative analysis.

[0140] (2) Content calculation: Substitute the initial signal intensity A0 of the separated pure alcohols into the standard curve equation to calculate the alcohol content c in the sample, using the formula c = A0 / 25.8. S1 (Hainan Guoxiang series): c1=6.98 / 25.8≈27.05%. Considering the slightly higher oil density in the sample's top part, the final content after correction is 28% (reason for correction: the signal intensity in the top part accounts for about 55%, and the edge accounts for 45%. After adjustment according to the regional weight, the error is <1%). S2 (Kalimantan Xingzhou system): c2=7.72 / 25.8≈29.92%. Due to the uniform distribution of oil in the sample (dense dotted distribution), no additional correction is needed, and the final content is 30%. S3 (Thailand Huian series): c3=8.92 / 25.8≈34.57%. Considering the signal superposition effect of the annular oil grid of the nodule (the signal intensity in the annular region is about 2% higher than that in the center), the final content after correction is 35%.

[0141] 4. Moisture distribution uniformity analysis: (1) Slope calculation: Linear piecewise fitting was performed on the moisture signal attenuation curve (15 segments in total, each consisting of 80 data points), and the slope of the curve for each segment was calculated (slope = signal intensity change / time change, unit: mV / s). The results are shown in the table below:

[0142] (2) Uniformity calculation: The uniformity of moisture distribution (U) is calculated based on the slope value. The formula is U = 1 - |maximum slope value - minimum slope value| / average slope value. The results are as follows: S1: U1 = 1 - |0.018 - 0.015| / 0.0165 ≈ 0.818, which is 81.8% (the uniformity is slightly lower due to the difference of about 1.2% in moisture content between the head and the edge). S2: U2=1-|0.022-0.020| / 0.0210≈0.905, that is, 90.5% (the water penetration is uniform during the formation of water-sinking, the difference in moisture content between the oil and wood is <0.5%, and the uniformity is high). S3: U3 = 1 - |0.017 - 0.013| / 0.0150 ≈ 0.733, or 73.3% (the difference in water content between the raised and depressed parts of the nodule is about 1.8%, and the special structure leads to the lowest uniformity).

[0143] IV. Data Analysis and Result Verification: 1. Comparison of alcohol content: The order of agarwood alcohol content among different series is as follows: Thai Hoi An series (S3, 35%) > Kalimantan Xingzhou series (S2, 30%) > Hainan Guoxiang series (S1, 28%). The reasons for the differences are as follows: S3 (Thailand Hoi An series): The burl structure leads to more concentrated resin secretion during resin formation, with an oil coverage rate of 55%, and the ring-shaped distribution of the burl promotes the accumulation of alcohols, hence the highest content. S2 (Kalimantan Xingzhou series): Water-sinking black oil-coated quartz belongs to the high-quality category of Xingzhou series. Long-term underwater immersion accelerates the degradation of the xylem, and oil and alcohol substances are more easily deposited. The oil coverage rate is 65%, and the content is second to last. S1 (Hainan Guoxiang Series): Although wild Baotou material is a typical category of Hainan Guoxiang series, the Baotou part only accounts for 30% of the sample volume, and the edge oil density is low (coverage rate 40%), resulting in a slightly lower overall content, which is in line with the category characteristics of "sweet fragrance" of Hainan Guoxiang series.

[0144] 2. Comparison of signal stability: The coefficient of variation (CV = standard deviation / mean × 100%) of the pure alcohol signal intensity was calculated over 20 minutes, and the results are as follows: S2 (CV=0.78%)<S1 (CV=1.12%)<S3 (CV=1.45%); Reasons for the differences: S2 is a near-cubic block with a regular shape and uniform oil grid distribution, and there is no obvious interference in signal acquisition; S1 is an irregular block with slight differences in signal at the edge and the head; S3 is a scab-like structure with different signal propagation paths at the protrusion and depression, resulting in slightly larger fluctuations, but the CV of all samples is ≤2%, which meets the requirements for detection stability.

[0145] 3. Correlation analysis of moisture distribution: Moisture distribution uniformity is highly correlated with lineage characteristics: S2 (U2=90.5%): The water-sinking formation environment (long-term underwater) allows for uniform water penetration, and the difference in moisture content between the oil grain and the wood is small, which is consistent with the "warm and uniform" characteristics of Singapore-type water-sinking agarwood. S1 (U1=81.8%): The alternating wet and dry environment in the mountainous areas of Hainan caused the moisture content of the top part (near the center of the trunk) to be slightly higher than that of the edge, which is consistent with the growth characteristics of the Guoxiang series of trees with "high oil content in the center". S3 (U3=73.3%): The structure of the burl is a resin secretion channel, and the depression is prone to retain trace amounts of moisture, resulting in the lowest uniformity, which is consistent with the characteristic of Huian series burl materials that "structure affects moisture distribution".

[0146] 4. Result Verification: (1) Accuracy verification: The method for determining alcohol-soluble extracts (high performance liquid chromatography, HPLC) as specified in the "Agarwood" standard (LY / T2904-2017) was used for comparative verification. HPLC detection conditions: C18 column (250 mm × 4.6 mm, 5 μm), mobile phase methanol-water (70:30, v / v), flow rate 1.0 mL / min, detection wavelength 280 nm, column temperature 30 ℃; Results comparison: The relative errors (RE = |value of this method - value of HPLC| / value of HPLC × 100%) of S1 (HPLC value 27.8%), S2 (HPLC value 29.9%) and S3 (HPLC value 34.9%) with the values ​​calculated by this method were 0.72%, 0.33%, and 0.29%, respectively, all < 1.5%, which meets the accuracy requirements for quantitative analysis and proves that the method of this invention can replace traditional destructive detection methods.

[0147] (2) Level determination: According to the core requirements of the "Agarwood" standard (LY / T2904-2017) for agarwood of the "Stacked Agarwood" grade (alcohol-soluble extract content ≥8%, density of 7-point agarwood 0.85-1.0g / cm³): The alcohol content of S1, S2 and S3 is much higher than 8%, and their densities are 0.92 g / cm³, 0.98 g / cm³ and 1.01 g / cm³, respectively (S3 has a slightly higher density due to its burl structure, but it still meets the characteristics of Huian burl material). Final judgment: All three samples were qualified grade 7 agarwood. Among them, S3 (35%) and S2 (30%) were classified as high-quality grade agarwood (alcohol content > 25%), and S1 (28%) was classified as good grade agarwood (alcohol content 20%-25%).

[0148] In this embodiment, three samples cover different series (Guoxiang series, Xingzhou series, and Huian series) and different shapes (irregular blocky, near-cubic, and burl-like). Through the adaptation of special sample carriers (anti-slip silicone pads and PTFE fixing straps) and dynamic adjustment of parameters (magnetic field strength and radio frequency), stable detection was achieved in all samples, proving the wide applicability of the method of this invention to different series of agarwood. At the same time, considering the structural characteristics of each series of samples (such as S1 head material and S3 burl material), signal correction and uniformity analysis were used to further verify the anti-interference ability and comprehensive evaluation value of the method, which can meet the needs of the agarwood detection field for "non-destructive, high precision, and multi-dimensional" methods.

[0149] Example 2: Detection of agarwood samples from different sub-regions within the same major production area.

[0150] I. Sample Selection and Preprocessing: 1. Sample Information: Yellow-ripe agarwood samples from three different origins in Kalimantan were selected. The specific parameters of the samples are shown in the table below. All samples meet the basic appearance requirements for yellow-ripe agarwood in the "Agarwood" standard (LY / T2904-2017).

[0151]

[0152] 2. Preprocessing steps: (1) Surface cleaning: Tools selected: dust-free microfiber cloth (fiber diameter 0.5μm), analytical grade anhydrous ethanol (purity ≥99.7%), disposable dropper (1mL specification); Operating procedure: Use a dropper to apply 0.5 mL of anhydrous ethanol to a lint-free cloth, ensuring the cloth is evenly moistened (moisture content ≤10%). Gently wipe the sample surface in a clockwise direction, focusing on cleaning the dust in the weathering grooves of S4 and the impurities attached to the oil glands on the surface of S5. After wiping, place the sample on a clean PTFE tray and let it stand at room temperature (25℃±1℃) for 8 minutes to ensure complete evaporation of the ethanol. Verification standard: After cleaning, the sample is weighed using a high-precision electronic balance. If the weight change is ≤0.01g, it indicates that there is no residual ethanol or impurities and the removal is complete.

[0153] (2) Adaptation and fixation of the container: Selection of sample carriers: Use solid sample carriers made of polyetheretherketone (PEEK) with a 0.2mm thick polytetrafluoroethylene (PTFE) anti-interference coating on the inner wall. S4 and S5 are suitable for sample carriers with an inner diameter of 12cm and a height of 18cm, and S6 is suitable for sample carriers with an inner diameter of 10cm and a height of 15cm (coating surface roughness Ra≤0.2μm, signal interference rate≤0.5%). Fixing method: S4 and S5 are placed directly on the anti-slip silicone pad (0.5cm thick, 0.1cm texture depth) at the bottom of the sample dish, ensuring that the center of the sample is aligned with the center of the sample dish (offset ≤0.5cm); S6 (thin sheet) is fixed by clamping with two PTFE spacers (0.5cm thick, 8cm diameter), with a 0.1cm thick silicone layer pasted at the contact point between the spacers and the sample to avoid damaging the sample edges. After fixing, the sample level deviation is ≤1°. Sealing check: Cover the sample container with the lid (with a vent hole, 0.5mm in diameter) to ensure that there is no obvious airflow disturbance inside the sample container and to avoid sample displacement during the detection process.

[0154] (3) Environmental balance: Equipment parameters: Place the sample-containing plate into a constant temperature and humidity chamber, set the temperature to 25℃±1℃, the relative humidity to 50%±2%, and the equilibration time to 24h. Monitoring measures: Temperature and humidity data were recorded every 8 hours during the equilibration process to ensure that temperature fluctuations were ≤0.5℃ and humidity fluctuations were ≤1%; after equilibration, the samples were weighed again. The weight changes for S4, S5, and S6 were 0.012g, 0.008g, and 0.005g, respectively, all ≤0.02g, indicating that the moisture content of the samples had reached a stable state.

[0155] II. Testing process: 1. Initialization of magnetic resonance imaging system parameters: Start the magnetic resonance imaging system and allow it to warm up for 30 minutes. Once the magnet temperature stabilizes at 32℃±0.5℃ and the radio frequency power fluctuation is ≤0.1dB, set the basic parameters as follows:

[0156] 2. Sample characteristic pre-analysis and parameter calibration: (1) Low-power pre-scan: Operating procedure: Place the sample dish into the detection chamber, set the radio frequency power to 1 / 5 (2dB) of the conventional detection, perform a ¹H-NMR pre-scan, scan range 0-10ppm, scan 8 times, scan time 120s; Interference peak identification: The characteristic resonance peaks of alcohols in the sample (1.2-1.5 ppm, corresponding to the hydrogen nuclei of the -CH2-OH group) were identified using systematic spectral analysis software. At the same time, the presence of impurity interference peaks (such as the moisture peak at 4.7 ppm and the lignin peak at 0.8-1.0 ppm) was checked. The results showed that there were no obvious interference peaks in the 1.2-1.5 ppm range for S4, S5, and S6 (interference peak intensity < 8% of the target peak intensity), and no adjustment of the scanning range was required.

[0157] (2) Static magnetic field strength locking: Adjustment method: The "gradient increment-signal feedback" mode is adopted. Starting from 0.5T, the magnetic field strength is increased in increments of 0.05T. After each adjustment of the magnetic field strength, the magnetic field is kept stable for 10 minutes (to ensure that the magnetic field uniformity meets the standard). Then, the free induction decay (FID) signal is collected for 10 seconds. Signal-to-noise ratio (S / N) calculation: The software calculates the S / N in the range of 1.2-1.5 ppm. When S / N reaches its maximum value, the current magnetic field strength is recorded as the optimal magnetic field strength. Finally, the optimal magnetic field strengths are locked as follows: S4 is 0.95T (S / N=34:1), S5 is 1.05T (S / N=37:1), and S6 is 1.0T (S / N=32:1).

[0158] (3) Radio frequency matching: Theoretical frequency calculation: According to the formula ν=γ·B0 / 2π (where γ=42.58MHz / T is the gyromagnetic ratio of hydrogen nuclei, and B0 is the optimal static magnetic field strength), the theoretical resonance frequencies of S4 are calculated to be 42.58MHz / T×0.95T=40.45MHz, S5 is 42.58MHz / T×1.05T=44.71MHz, and S6 is 42.58MHz / T×1.0T=42.58MHz; Frequency sweep locking: Sweep the frequency within the theoretical frequency ±0.5MHz range with a sweep step size of 0.01MHz. Acquire the FID signal for 20s after each sweep and record the signal peak value. Plot the "RF frequency - signal peak value" curve. The frequency corresponding to the highest point of the curve is the optimal RF frequency. Finally, the optimal RF frequencies are locked as follows: S4 40.44MHz, S5 44.72MHz, and S6 42.57MHz (frequency deviation ≤0.02MHz, ensuring complete matching with the Zeeman energy level difference of alcohol hydrogen nuclei).

[0159] 3. Magnetic resonance signal acquisition: (1) Specific signal acquisition: Operation procedure: Under the locked optimal magnetic field strength and radio frequency parameters, start the signal acquisition program and continuously acquire NMR signals within 25 minutes. The system automatically records 1 set of data every 1 second (a total of 1500 sets of data) and generates NMR signal attenuation curves for S4, S5 and S6 respectively. Real-time monitoring: During the acquisition process, the signal strength fluctuation is monitored in real time. The system’s built-in signal stability module ensures that the signal strength fluctuation is ≤2%. If the fluctuation exceeds the limit (e.g., ≥3%), the acquisition is immediately paused, and the sample is checked for displacement or magnetic field instability. After adjustment, the acquisition is restarted.

[0160] (2) Interference signal separation and acquisition: First acquisition (S total): A 90° radio frequency pulse is emitted to flip the magnetization vector of all hydrogen nuclei (water, oil, alcohol) in the sample from the Z-axis to the XY plane. The FID signal is acquired after a delay of t1=25ms (less than 1 / 2 of the water T2 value to ensure that the water signal is not significantly attenuated). The signal intensity S total obtained at this time is the sum of the water, oil and alcohol signals. Secondary acquisition (S oil alcohol): Apply a 180° radio frequency pulse to flip the magnetization vector of the XY plane by 180°, and acquire the spin echo signal when the delay is t2=50ms (2t1). At this time, the water has completely decayed due to the short T2 (20-50ms). The obtained signal intensity S oil alcohol is the sum of the oil and alcohol signals. CPMG sequence acquisition: The S-oil alcohol signal was acquired using the CPMG sequence. A total of 35 echo signals were acquired, and the signal intensity of each echo (E1 to E35) was recorded for subsequent separation of alcohol and oil signals.

[0161] III. Signal Acquisition and Processing Methods: 1. Raw signal preprocessing: (1) Blank correction: Blank signal acquisition: Place an empty sample dish (consistent with the sample dish used for sample detection) into the detection chamber and acquire the background signal (S_empty) under the same parameters. Acquire the signal 5 times consecutively and take the average value as the final blank signal value. In this embodiment, S_empty = 0.09mV. Correction calculation: Blank correction is performed on the original sample signal (Soriginal). The correction formula is Scorrected = Soriginal - Sblank, which eliminates background interference from the material of the sample dish itself and the detection environment.

[0162] (2) Outlier removal: Testing method: The Grubbs test (significance level α = 0.05) was used to test for outliers in the corrected signal data. According to the Grubbs test critical value table, when the sample size n = 1500, the critical value G... 0.05 =3.2; Results of the removal: Calculate the residuals for each data point (residual = data value - mean). There are 2 data points in S4 with residual absolute values ​​greater than 3.2 times the standard deviation, 1 in S5, and 2 in S6. Remove these outliers to ensure data reliability.

[0163] (3) Time-domain filtering: Filtering method: The signal data after outlier removal is processed by the Hanning window function with a window width of 5ms to suppress noise interference at the end of the signal (the noise signal is mainly concentrated after 15ms and the intensity is ≤0.05mV). Valid signal retention: The valid signal within the first 15ms is retained (during this period, the signal intensity of alcohol substances accounts for more than 82% of the total signal intensity), laying the foundation for subsequent interference separation.

[0164] 2. Interference signal separation and processing: (1) Moisture signal separation: Calculation method: The signal intensity of pure water, S_water, is obtained based on the signal difference between S_total and S_oilol. The formula is S_water = S_total - S_oilol. Calculation results: The moisture signal intensities of S4, S5 and S6 were 0.95mV, 1.20mV and 0.78mV respectively, which are consistent with the moisture content detection results in the sample pretreatment stage (S4=8.2%, S5=7.5% and S6=6.8%), verifying the accuracy of moisture signal separation.

[0165] (2) Alcohol-oil separation: Attenuation curve fitting: Exponential fitting was performed on 35 echo signals acquired from the CPMG sequence to obtain the attenuation curves of the alcohols. And the decay curve of oil content (Where A0 is the initial signal intensity of alcohols, T2A is the transverse relaxation time of alcohols, and the measured values ​​are S4=95ms, S5=102ms, and S6=88ms; O0 is the initial signal intensity of oils, and T2O is the transverse relaxation time of oils, and the measured values ​​are S4=220ms, S5=240ms, and S6=210ms). Simultaneous equation solution: By solving the simultaneous equation E(t)=A(t)+O(t) (where E(t) is the signal intensity of the t-th echo), the initial signal intensity A0 of the alcohol is obtained, and the results are S4=5.93mV, S5=4.99mV, and S6=4.64mV; Verification: Comparing the correlation between the fitted curve and the actual acquired signal, R² is ≥0.995, indicating that the fitting accuracy meets the requirements.

[0166] (3) Frequency domain cleanup: Fourier Transform: The initial alcohol signal A0 is converted into a frequency domain spectrum using Fast Fourier Transform (FFT) to obtain the frequency distribution of the signal; Bandpass filtering: A digital bandpass filter with a center frequency of 1.2-1.5ppm and a bandwidth of 0.5ppm is used to filter out residual moisture signals around 4.7ppm (attenuation ≥99%) and lignin signals around 0.8-1.0ppm (attenuation ≥98%). Purity verification: After filtering, the purity of the alcohol signal was calculated as (target frequency band signal strength / total signal strength × 100%). S4, S5, and S6 were 99.6%, 99.7%, and 99.5%, respectively, all ≥ 99.5%, which meets the requirements for subsequent content calculation.

[0167] 3. Calculation of alcohol content: (1) Reference of standard curve: The standard curve of alcohol concentration-signal intensity was established in advance. This curve was obtained by preparing 0.5%, 1%, 2%, 5%, 10%, 15%, 20%, 25%, and 30% linalool standard solutions (solvent is anhydrous ethanol) and collecting magnetic resonance signals respectively. The alcohol concentration was used as the abscissa and the signal intensity was used as the ordinate. The weighted least squares method was used to fit the curve. The fitting equation was S_avg=25.8×c (where S_avg is the average signal intensity and c is the alcohol concentration), the coefficient of determination R²=0.999, and the linear error <0.8%.

[0168] (2) Content Calculation: Substitute the initial signal intensity A0 of the separated pure alcohols into the standard curve equation to calculate the alcohol content c in the sample. The formula is c = A0 / 25.8. S4 (North Brunei, Kalimantan): c4 = 5.93 / 25.8 ≈ 23.0%; S5 (Kalimantan Darakan): c5=4.99 / 25.8≈19.3% (Note: Based on the actual characteristics of the sample, the corrected value is 17.0%. The reason for the correction is that the oil spots in S5 contain a small amount of non-alcoholic oily components, which leads to a slightly higher initial signal. The correction coefficient is 0.88). S6 (Kalimantan Manilao): c6=4.64 / 25.8≈18.0%.

[0169] 4. Moisture distribution uniformity analysis: (1) Slope calculation: Perform linear segmented fitting on the moisture signal attenuation curve (each 100 data sets is a segment, a total of 15 segments), and calculate the slope of each segment (slope = signal intensity change / time change). (2) Calculation of uniformity: The uniformity of water distribution (U) is calculated based on the slope value. The calculation formula is U=1-|maximum slope value-minimum slope value| / average slope value: S4: Maximum slope = 0.019mV / s, minimum slope = 0.016mV / s, average slope = 0.0175mV / s, U4 = 1 - |0.019 - 0.016| / 0.0175 ≈ 0.83 (due to slightly uneven moisture distribution caused by surface weathering gullies). S5: Maximum slope = 0.021mV / s, minimum slope = 0.020mV / s, average slope = 0.0205mV / s, U5 = 1 - |0.021 - 0.020| / 0.0205 ≈ 0.95 (uniform texture, relatively uniform moisture distribution). S6: Maximum slope = 0.018mV / s, minimum slope = 0.015mV / s, average slope = 0.0165mV / s, U6 = 1 - |0.018 - 0.015| / 0.0165 ≈ 0.82 (the difference in moisture content between the edge and center of the thin sheet results in low uniformity).

[0170] IV. Data Analysis and Result Verification: 1. Alcohol content comparison: Kalimantan North Brunei (S4, 23.0%) > Manilao (S6, 18.0%) > Tarakan (S5, 17.0%), consistent with the sample characteristics of S4 having uniform oil distribution in the yellow-brown hue and S5 having non-alcoholic oil components in the oil flower structure, reflecting the regional differences in alcohol content of yellow-ripe agarwood from different production areas in Kalimantan; 2. Signal stability comparison: The coefficient of variation (CV) of signal intensity was calculated within 25 minutes. S5 (CV=0.72%) < S4 (CV=1.15%) < S6 (CV=1.38%). S6 had slightly larger signal fluctuations due to its sheet-like shape, but all met the stability requirement of CV≤2%. 3. Correlation analysis of moisture distribution: S5 (U5=0.95) has the most uniform moisture distribution, which is consistent with its formation environment (depth) as a water-sinking agarwood; S4 and S6 have lower uniformity of moisture distribution due to weathering or morphological reasons, which is consistent with the natural formation process of the samples.

[0171] 4. Result Verification: (1) Repeatability verification: Each sample was tested 5 times. The relative standard deviations (RSDs) of the alcohol content calculation results of S4, S5 and S6 were 1.23%, 0.98% and 1.45% respectively, all <2%, which proves that the method has good repeatability; (2) Accuracy verification: The alcohol-soluble extract determination method (HPLC) specified in "Agarwood" (LY / T2904-2017) was used as a reference. The relative errors (RE) between S4 (HPLC value 22.8%), S5 (HPLC value 16.9%), and S6 (HPLC value 17.9%) and the values ​​calculated by this method were 0.88%, 0.59%, and 0.56%, respectively, all <1.5%, indicating that the accuracy is reliable. (3) Grade determination: According to LY / T2904-2017, the alcohol-soluble extract content of Huangshuxiang grade agarwood must be ≥10%. S4, S5 and S6 all meet the requirements and are judged as qualified Huangshuxiang grade agarwood. Among them, S4 alcohol content >20% can be classified as high-quality Huangshuxiang.

[0172] In this embodiment, the three samples showed significant differences in morphology (irregular block, near cube, and thin sheet). Stable detection was achieved in all samples through the adaptation design of a special sample carrier, proving that the method of the present invention has wide applicability to different forms of Kalimantan agarwood. At the same time, for the non-alcoholic oil components in the S5 oil structure, accurate quantification was achieved through signal correction, further verifying the anti-interference ability of the method.

[0173] Example 3: Detection of Qinan agarwood samples from different sampling locations and with varying degrees of resin formation.

[0174] I. Sample Selection and Preprocessing: 1. Sample Information: Three agarwood samples from the Hoi An agarwood lineage in Nha Trang, Vietnam, were selected based on different harvesting locations and resin formation levels. All samples originated from the same wild agarwood tree approximately 50 years old, exhibiting high lineage consistency. The differences were solely due to the harvesting location (outer bark, core) and resin formation level (yellow mature agarwood, 7-point agarwood, and 9-point agarwood). All samples met the core indicators (alcohol-soluble extract content, density, and oil coverage) of the corresponding grade in the "Agarwood" standard (LY / T2904-2017). The specific parameters are shown in the table below.

[0175] 2. Preprocessing steps: (1) Surface cleaning: Tools selected: dust-free microfiber cloth (fiber diameter 0.2μm, water absorption ≤5g / m²), analytical grade anhydrous ethanol (purity ≥99.7%, evaporation residue ≤0.001%), disposable microdropper (0.5mL specification, accuracy ±0.005mL); Operating Procedure: Different cleaning methods are used for samples of different morphologies—S7 (outer bark, shallow texture) is wiped unidirectionally with a cloth moistened with 0.2 mL of anhydrous ethanol (to avoid embedding impurities into the texture); S8 (core material, smooth surface) is wiped circumferentially with a cloth moistened with 0.3 mL of anhydrous ethanol; S9 (burl material, many protrusions) is wiped locally with a cloth moistened with 0.2 mL of anhydrous ethanol (focusing on cleaning the recessed areas of the burls). After wiping, the sample is placed on a polytetrafluoroethylene tray (roughness Ra≤0.1μm) and left to stand at room temperature (25℃±1℃) for 12 minutes to ensure complete evaporation of the ethanol. Verification standard: Weigh the S7, S8 and S9 samples before and after cleaning using a high-precision electronic balance. The weight changes are 0.015g, 0.009g and 0.006g respectively, all ≤0.02g, with no ethanol residue or impurity residue.

[0176] (2) Adaptation and fixation of the container: Selection of sample carriers: Use polyetheretherketone (PEEK) solid sample carriers with a 0.2mm thick polytetrafluoroethylene (PTFE) anti-interference coating on the inner wall. S7 is suitable for 12cm inner diameter × 16cm height (suitable for sheet-shaped samples), S8 is suitable for 8cm inner diameter × 10cm height (suitable for cylindrical samples), and S9 is suitable for 6cm inner diameter × 8cm height (suitable for small nodule samples). The coating signal interference rate is ≤0.2%. Fixation method: S7 (sheet-shaped) is clamped and fixed with two PTFE spacers (0.5cm thick, 10cm diameter), and a 0.1cm thick silicone layer is attached to the contact area between the spacer and the sample (to avoid edge damage). After fixation, the horizontal deviation is ≤0.5°; S8 (cylindrical) is placed in a V-shaped silicone slot (60° angle, 1cm depth), ensuring that the cylinder axis is aligned with the center of the sample dish; S9 (scar-like) is fixed with three PTFE elastic fixing bands (elastic modulus 1.2GPa) with a "triangular support" of 0.8cm width, with the tension controlled at 3-5N (to avoid compressing the scar protrusion); Sealing check: Cover with the top cover with a 0.4mm diameter vent hole (air permeability 8-12mL / min), and test by introducing 0.01MPa nitrogen gas. The pressure drop should be ≤0.001MPa within 30s to ensure no air leakage.

[0177] (3) Environmental balance: Equipment parameters: Place the sample container in a constant temperature and humidity chamber, set the temperature to 25℃±1℃ and the relative humidity to 50%±2%, and allow it to equilibrate for 24 hours; Monitoring measures: Data were recorded every 4 hours using a temperature and humidity recorder (accuracy ±0.05℃ / ±0.5%RH), with temperature fluctuation ≤0.3℃ and humidity fluctuation ≤0.8%; after balancing, the samples were weighed, and the weight changes of S7, S8 and S9 were 0.018g, 0.011g and 0.007g respectively, all ≤0.02g, indicating that the moisture content of the samples was stable.

[0178] II. Testing process: 1. Initialization of magnetic resonance imaging system parameters: Start the magnetic resonance imaging system and preheat for 45 minutes. Once the magnet temperature stabilizes at 31℃±0.3℃ and the radio frequency power fluctuation is ≤0.05dB, set the basic parameters as shown in the table below:

[0179] 2. Sample characteristic pre-analysis and parameter calibration: (1) Low-power pre-scan: Operating procedure: Place the sample dish into the detection chamber (probe temperature 25℃±0.2℃), set the radio frequency power to 1 / 5 (2dB) of the conventional detection, and perform a ¹H-NMR pre-scan (range 0-11ppm, 12 scans, time 180s). Interference peak identification: The characteristic peaks of alcohols (1.2-1.5ppm) and impurity peaks (4.7ppm moisture peak and 0.9ppm lignin peak) were identified by TopSpin 4.1.4 software. The intensity of these peaks was less than 5% of the target peak, so no adjustment of the scanning range was required.

[0180] (2) Static magnetic field strength locking: Adjustment method: Start from 0.5T and increase in increments of 0.05T. After each step is stable for 12 minutes, acquire the FID signal for 12 seconds and calculate the signal-to-noise ratio (S / N) in the 1.2-1.5ppm range. The results show that the optimal magnetic field strength is 0.92T for S7 (S / N=31:1), 1.08T for S8 (S / N=36:1), and 1.05T for S9 (S / N=34:1), all falling within the range of 0.5-1.5T.

[0181] (3) Radio frequency matching: Theoretical frequency calculation: According to ν=γ·B0 / 2π (γ=42.58MHz / T), S7=39.17MHz, S8=45.99MHz, S9=44.71MHz; Frequency sweep lock: Sweep the frequency within the theoretical frequency range of ±0.5MHz (step size 0.008MHz), collect 22s FID signal in each step, plot the "frequency-peak" curve, and lock S7=39.16MHz, S8=45.98MHz, S9=44.70MHz (deviation ≤0.01MHz). Specificity verification: The peak shape matching degree between the pure linalool standard and the sample signal is ≥95.8%, and the lignin signal intensity is <3%, thus verifying the specificity.

[0182] 3. Magnetic resonance signal acquisition: (1) Specific signal acquisition: Operation procedure: Under locked parameters, continuously acquire signals for 20 minutes, record one set of data every 0.8 seconds, and generate NMR signal attenuation curves for S7, S8, and S9; Real-time monitoring: Signal strength fluctuation amplitude S7=1.3%, S8=0.9%, S9=1.1%, all ≤1.5%, with no abnormal fluctuations.

[0183] (2) Interference signal separation and acquisition: Initial sampling (S total): A 90° pulse is emitted, and the FID signal (S total = moisture + oil + alcohol) is collected after a delay of t1 = 28ms. Secondary acquisition (S oil alcohol): Apply a 180° pulse and acquire the spin echo signal after a delay of t2=56ms (2t1) (S oil alcohol = oil + alcohol, water completely attenuated). CPMG acquisition: 32 echoes of the S oil alcohol signal were acquired, and the signal intensities of E1-E32 were recorded for alcohol-oil separation.

[0184] III. Signal Acquisition and Processing Methods: 1. Raw signal preprocessing: (1) Blank correction: Blank signal acquisition: The background signal of the empty sample dish (S_empty) was acquired and averaged 6 times. S_empty = 0.07mV; Correction calculation: Scorrected = Soriginal - Sempty, after correction S7 = 3.87mV, S8 = 6.20mV, S9 = 10.09mV.

[0185] (2) Outlier removal: Test method: Grubbs' test (α=0.05, n=1500, critical value G) 0.05 =3.3); Outlier removal results: 2 outliers were removed from S7, 1 outlier was removed from S8, and 2 outliers were removed from S9, with a removal rate of ≤0.2%.

[0186] (3) Time-domain filtering: Filtering method: Hanning window function processing (window width 5.5ms) to suppress noise after 16ms (intensity ≤0.03mV); Valid signal retention: Retain the valid signal within the first 16ms (accounting for more than 83%), and improve the smoothness of the filtered signal by 26%-30%.

[0187] 2. Interference signal separation and processing: (1) Moisture signal separation: Calculation results: Swater = Stotal - Soil alcohol, S7 = 1.12mV, S8 = 0.95mV, S9 = 0.82mV, which are positively correlated with water content (S7 = 9.2%, S8 = 8.0%, S9 = 7.5%) (R = 0.97), verifying the accuracy.

[0188] (2) Alcohol-oil separation: Decay curve fitting: Alcohols T2: S7=88ms, S8=98ms, S9=105ms; Oils T2: S7=210ms, S8=235ms, S9=250ms; Simultaneous calculation: Initial signal intensity A0 of alcohols: S7=3.87mV, S8=6.20mV, S9=10.09mV, and the fitting correlation R²≥0.996.

[0189] (3) Frequency domain cleanup: Bandpass filter: 1.2-1.5ppm bandwidth, 0.45ppm, moisture residue attenuation ≥99.3%, lignin attenuation ≥98.7%; Purity verification: The purity of alcohol signals S7=99.5%, S8=99.7%, and S9=99.6%, all ≥99.5%.

[0190] 3. Calculation of alcohol content: (1) Reference to the standard curve: The standard curve established in Example 1 (fitting equation S_avg=25.8×c, R²=0.999, linear error <0.8%) was used. The system parameters were consistent, so there was no need to refit.

[0191] (2) Content calculation: S7 (yellow skin, ripe and fragrant): c7=3.87 / 25.8≈15.0% (no correction, oil distribution is uniform); S8 (7 points of core sedimentation): c8=6.20 / 25.8≈24.0% (no correction, stable radial oil grid signal); S9 (9 points of blister deposit): c9=10.09 / 25.8≈39.1%, corrected to 39% (correction reason: superposition of blister annular oil grid signal, correction coefficient 0.997).

[0192] 4. Moisture distribution uniformity analysis: (1) Slope calculation: Piecewise fitting (15 segments for every 100 data points), the slope results are shown in the table below:

[0193] (2) Uniformity calculation: S7: U7=1-|0.021-0.017| / 0.0190≈78.9% (unevenness caused by the texture of the outer skin); S8: U8 = 1 - |0.018 - 0.016| / 0.0170 ≈ 88.2% (the core material has a uniform structure); S9: U9=1-|0.016-0.014| / 0.0150≈86.7% (slight impact from lumpy scar protrusion).

[0194] IV. Data Analysis and Result Verification: 1. Analysis of differences in material sourcing location and agarwood formation degree: (1) Reasons for the difference in alcohol content: Material sampling location: core material (S8, 24%) > bark material (S7, 15%), because the xylem of the core material is dense and the resin channels are concentrated, which is conducive to alcohol deposition; the bark material is exposed for a long time, and some alcohols are lost with weathering. Agarwood formation degree: 9 points of agarwood (S9, 39%) > 7 points of agarwood (S8, 24%) > yellow agarwood (S7, 15%). The longer the agarwood formation time, the higher the oil coverage (S9=75%>S8=45%>S7=20%), and the more significant the enrichment effect of alcohols, which is consistent with the characteristic of Huian agarwood that "the longer the agarwood formation time, the better the quality".

[0195] (2) Differences in signal stability: Calculate the coefficient of variation (CV) of signal intensity over 20 minutes: S8 (CV=0.85%)<S9 (CV=1.02%)<S7 (CV=1.32%); Reasons for the differences: S8 is cylindrical core material with a regular shape and uniform radial distribution of oil cells, ensuring unobstructed signal propagation; S9, although burl material, has dense and symmetrically distributed oil cells in a ring shape, with relatively small fluctuations; S7 is outer bark material with uneven surface weathering and sparse oil cells, resulting in slightly poorer stability, but all samples have a CV of ≤1.5%, meeting the testing requirements.

[0196] 2. Correlation analysis of moisture distribution uniformity: Core material (S8, 88.2%) > Burl material (S9, 86.7%) > Outer bark material (S7, 78.9%). Related logic: The core part has a stable growth environment and uniform moisture penetration; the outer bark is greatly affected by changes in external temperature and humidity, with a moisture content difference of about 1.8% between the edge and the center; the burl material has a slightly enriched local moisture due to its raised structure, but the dense oil grid can reduce fluctuations, so its uniformity is better than that of the outer bark material.

[0197] 3. Accuracy verification: The method for determining alcohol-soluble extracts (Soxhlet extraction-HPLC) as specified in the "Agarwood" standard (LY / T2904-2017) was used as a reference: Extraction conditions: 95% ethanol as solvent, Soxhlet extraction for 6 hours, and the extract was concentrated to a final volume of 10 mL; HPLC detection: C18 column (250 mm × 4.6 mm, 5 μm), mobile phase methanol-water (65:35, v / v), flow rate 1.0 mL / min, detection wavelength 280 nm; Results comparison: The relative errors (RE) of S7 (HPLC=14.9%), S8 (HPLC=23.8%), and S9 (HPLC=38.9%) compared with this method were 0.67%, 0.84%, and 0.26%, respectively, all <1.5%, and the accuracy met the quantitative requirements.

[0198] 4. Level Determination: According to the corresponding grading indicators in "Agarwood" (LY / T2904-2017): S7 (Yellow-ripe agarwood): Alcohol content ≥ 10% (minimum requirement for yellow-ripe agarwood grade), judged as qualified yellow-ripe agarwood grade; S8 (7-point agarwood): Alcohol content ≥20%, density 0.92g / cm³ (meets the 7-point agarwood range), judged as qualified 7-point agarwood of the stacked fragrance grade; S9 (Burl 9-point sinking): Alcohol content ≥35% and density 1.08g / cm³ (meets the 9-point sinking range), judged as high-quality stacked incense grade 9-point sinking agarwood.

[0199] This embodiment of the sample covers different sampling locations (outer bark, core, burl) and resin formation levels (low, medium, high), with significant morphological differences (flaky, cylindrical, burl-like). Through the adaptation design of the sample carrier (V-shaped groove, elastic fixing band) and dynamic parameter adjustment, stable detection is achieved in all cases. This proves that the method has wide applicability to agarwood of "the same series but different sampling / resin formation levels" and can accurately distinguish differences in alcohol content.

[0200] Material sampling guidance: By testing the alcohol content at different sampling locations, a basis for processing can be provided (e.g., using heartwood to make high-end incense, and outer bark to make entry-level products), thereby improving the utilization rate of raw materials; Agarwood formation assessment: A model of "signal intensity-alcohol content-agarwood formation degree" can be established to quickly determine the agarwood formation stage and provide data support for the selection of artificial agarwood formation timing; Quality grading: Combining alcohol content and moisture uniformity, we can accurately classify the grades of Huian agarwood, avoiding the errors of traditional experience-based grading.

[0201] Example 4: Detection of white Qinan agarwood samples with different resin formation methods.

[0202] I. Sample Selection and Preprocessing: 1. Sample Information: Three samples of grade 90% white agarwood with different resin formation methods were selected, covering natural wild insect-damaged, artificially induced by fungi, and artificially physically induced resin formation types. All samples have authoritative museum or auction certifications. Specific parameters are shown in the table below:

[0203] 2. Preprocessing steps: (1) Surface cleaning: Tools selected: dust-free microfiber cloth (fiber diameter 0.2μm, antistatic treatment), analytical grade anhydrous ethanol (purity ≥99.7%, evaporation residue ≤0.001%), disposable microdropper (0.2mL specification, accuracy ±0.002mL), soft brush (bristle diameter 0.05mm, to avoid scratching the oil filter). Operational procedure: Differentiated cleaning based on the characteristics of different samples.

[0204] S10 (Insect-damaged material): First, use a soft brush to gently brush the inside and surface of the insect holes to remove residual dust; then, use a dropper to take 0.1mL of anhydrous ethanol and drop it onto a corner of the fiber cloth to make a "moist cotton swab" shape, and gently wipe it into the insect holes (avoid ethanol seeping into the oil filter and causing signal deviation). Use a "spiral wiping" method on the surface (from the center to the edge, covering all areas). S11 (Fungal Inducing Material): Take 0.15 mL of anhydrous ethanol to moisten the fiber cloth, and use "one-way wiping" (along the edge of the cube) to avoid repeated friction damage to the surface oil grid; S12 (physical induction material): Take 0.1 mL of anhydrous ethanol to moisten the fiber cloth, wipe along the direction of physical damage texture, and focus on cleaning the impurities in the texture grooves; Drying verification: After cleaning, the three samples were placed on a polytetrafluoroethylene tray and left to stand at room temperature (25℃±1℃) for 10 minutes. They were then weighed using a high-precision electronic balance (Sartorius CPA225D, accuracy 0.0001g). The weight changes of S10, S11, and S12 were 0.004g, 0.003g, and 0.005g, respectively, all ≤0.006g, confirming that the anhydrous ethanol had completely evaporated and that there was no oil loss.

[0205] (2) Adaptation and fixation of the container: S10. Selection of sample carrier: Based on the size and shape of the sample, a polyetheretherketone (PEEK) sample carrier with a 0.2mm thick polytetrafluoroethylene (PTFE) anti-interference coating on the inner wall should be selected. S11: Select a sample dish with an inner diameter of 6cm and a height of 8cm (suitable for small block samples). S12: Use a sample dish with an inner diameter of 8cm and a height of 10cm (suitable for cuboid samples); the bottom of the sample dish has a built-in anti-slip silicone pad (thickness 0.3cm, coefficient of friction ≥0.8), and a 1cm diameter ventilation hole is opened in the center of the base (to prevent moisture condensation in the sealed environment).

[0206] Fixing method: S10 (Irregular Insect-Leaked Sample): Three 0.5cm wide PTFE elastic fixing strips (elastic modulus 1.2GPa, elongation ≤5%) are used to fix the sample in a "triangular support" shape. The fixing strips avoid insect holes and areas with dense oil grids (identified by visual observation of the color difference of the oil grids; the milky yellow area is the area with dense oil grids). S11 (cube material): Place directly on the anti-slip pad in the center of the container without additional fixation (the cube structure has good stability and no risk of displacement). S12 (cubic prism material): Use two PTFE spacers (0.5cm thick, fitting against the inner wall of the sample dish) to hold the two ends of the cuboid, ensuring that the sample is placed horizontally (calibrated with a laser level, tilt ≤0.5°).

[0207] Position calibration: The laser positioning module built into the magnetic resonance detection system is used to ensure that the geometric center of the three samples coincides with the center of the detection cavity (deviation ≤ 0.2cm), so as to avoid the signal acquisition being affected by magnetic field inhomogeneity due to position offset.

[0208] (3) Environmental balance: Equipment parameters: Place the sample-containing plates into the constant temperature and humidity chamber (BinderKB240) simultaneously, set the temperature to 25℃±1℃ and the relative humidity to 50%±2%, and allow it to equilibrate for 24 hours; Monitoring measures: Temperature and humidity inside the chamber were recorded every 6 hours (accuracy ±0.05℃ / ±0.5%RH). Temperature fluctuation within 24 hours was ≤0.3℃ and humidity fluctuation was ≤0.8%. After balancing, the samples were weighed again. The weight changes of S10, S11, and S12 were 0.005g, 0.004g, and 0.006g, respectively, all ≤0.01g, confirming that the moisture content of the samples was stable and there was no interference from environmental factors.

[0209] II. Testing process: 1. Initialization of magnetic resonance imaging system parameters: Start the magnetic resonance imaging system and preheat for 45 minutes (ensuring the magnet temperature stabilizes at 30℃±0.3℃ and the RF power fluctuation is ≤0.05dB). Set the basic parameters as shown in the table below:

[0210] 2. Sample characteristic pre-analysis and parameter calibration: (1) Low-power pre-scan: Operating procedure: Place the sample dish into the detection chamber in the order of S10→S11→S12, close the detection door and stabilize for 5 minutes (to avoid mechanical vibration affecting the magnetic field); set the radio frequency power to 1 / 5 of the conventional detection (2dB, to avoid exciting non-target hydrogen nuclei), and perform a proton nuclear magnetic resonance (¹H-NMR) pre-scan, with a scanning range of 0-10ppm, 8 scans, and a scanning time of 180s; Interference peak identification: Using MestReNova 14.0 spectral analysis software, characteristic resonance peaks of alcohols in the sample were identified (preset range 1.2-1.5 ppm, corresponding to the hydrogen nucleus signal of the -CH2-OH group in the alcohol): S10: A strong signal peak (peak height 2.8 mV) appears in the 1.2-1.5 ppm range, with only a weak impurity peak (peak height 0.25 mV, intensity < 10% of the target peak) at 2.3 ppm, with no obvious interference; S11: The peak-to-peak height of the signal in the 1.2-1.5ppm range is 2.6mV, with no impurity peak interference; S12: The peak height of the signal in the 1.2-1.5ppm range is 2.0mV, with no impurity peak interference; all three samples meet the acquisition condition of "interference peak intensity < target peak intensity 10%", and no adjustment of the scanning range is required.

[0211] (2) Static magnetic field strength locking: Using a "gradient increment-signal feedback" model, the optimal static magnetic field strength for each alcohol sample was precisely determined. The steps are as follows: Set the static magnetic field strength adjustment range to 0.5-1.5T, with an adjustment step of 0.05T, starting from 0.5T and gradually increasing. After each adjustment of the magnetic field strength, the magnetic field was kept stable for 10 minutes (to avoid signal drift caused by magnetic field fluctuations). Then, a radio frequency pulse of a fixed frequency (21.29MHz, corresponding to the hydrogen nucleus resonance frequency of alcohols under a 0.5T magnetic field) was emitted, and the free induction attenuation (FID) signal was collected within 10 seconds. The signal-to-noise ratio (S / N) in the 1.2-1.5ppm range of each FID signal is calculated by the signal analysis module. When the S / N reaches the maximum value (≥30:1), the current magnetic field strength is recorded as the optimal value. The results show that the optimal static magnetic field strength is 1.1T for S10 (S / N=35:1), 1.0T for S11 (S / N=34:1), and 0.9T for S12 (S / N=32:1), all falling within the range of 0.5-1.5T.

[0212] (3) Radio frequency electromagnetic wave frequency matching: Based on the locked optimal static magnetic field strength, the precise matching between the radio frequency and the Zeeman energy level difference of alcohol atomic nuclei is achieved through "frequency sweeping-resonance signal capture". The steps are as follows: Calculate the theoretical resonance frequency: According to the formula ν=γ·B0 / 2π (where γ is the gyromagnetic ratio of hydrogen nuclei in linalools, taken as 42.58MHz / T; B0 is the optimal static magnetic field strength), we get: S10: ν=42.58×1.1≈46.84MHz; S11: ν=42.58×1.0≈42.58MHz; S12: ν=42.58×0.9≈38.32MHz; Frequency sweep: Sweep the frequency within the theoretical frequency ±0.5MHz range with a sweep step size of 0.01MHz; for each sweep, transmit a 90° radio frequency pulse (pulse width 10μs), collect the FID signal within 20s, and record the signal peak value; Optimal frequency lock: Plot the "RF frequency - signal peak" curve. The frequency corresponding to the highest point of the curve is the optimal RF frequency: S10 locks at 46.83MHz, S11 locks at 42.57MHz, and S12 locks at 38.31MHz, with frequency deviations all <0.02MHz. Specificity verification: Signals were collected from "pure linalool standard (concentration 10%)", "corresponding agarwood sample", and "pure lignin sample" at the locked frequency. Signal peak shape matching degree between agarwood samples and pure linalool standard: S10=97.2%, S11=97.5%, S12=96.8%; The proportions of lignin sample signal intensity to agarwood sample signal intensity were: S10=2.8%, S11=2.5%, and S12=3.0%, all <5%, confirming the absence of lignin and cellulose interference and successful frequency locking.

[0213] 3. Magnetic resonance signal acquisition: (1) Specific signal acquisition: Under the locked static magnetic field strength and radio frequency parameters, specific signal acquisition was performed on three samples for 20 minutes each: The system automatically records one set of NMR signal data every 1 second and generates a "time-signal intensity" NMR signal attenuation curve; Real-time monitoring of signal stability: Within 20 minutes, the coefficients of variation (CV) of signal strength of S10, S11 and S12 were 1.02%, 0.85% and 1.15% respectively, all ≤2%, which meets the signal stability requirements.

[0214] (2) Interference signal separation and acquisition: By combining spin echo sequences with CPMG sequences, the signals of water, oil, and alcohols were separated and acquired. Initial signal acquisition (S_total = moisture + oil + alcohols): Transmit a 90° radio frequency pulse, delay t1 = 30ms (t1 = 10-50ms, take the middle value to ensure that the moisture signal does not attenuate significantly), acquire the FID signal, and record the S_total intensity; Secondary signal acquisition (Soil alcohol = oil + alcohol): Apply a 180° radio frequency pulse, delay t2 = 60ms (t2 = 2t1, at which point the moisture signal has been completely attenuated due to the short T2), acquire the spin echo signal, and record the intensity of Soil alcohol. CPMG sequence acquisition: Apply the CPMG sequence (echo interval 10ms, total number of echoes 30) to the S oil alcohol signal and acquire 30 echo signals (E1-E30) for subsequent separation of alcohol and oil signals.

[0215] III. Signal Acquisition and Processing Methods: 1. Raw signal preprocessing: (1) Blank correction: Blank signal acquisition: Place an empty sample dish (with the same specifications as the sample dish used for sample detection) into the detection chamber, and acquire the background signal for 20 minutes under the same parameters. Take the average value after 3 consecutive acquisitions to obtain the blank signal intensity S_empty = 0.07mV. Sample signal correction: The original signals of the three samples are corrected using the formula Scorrection = Soriginal - Semption. After correction, the initial signal strengths (at t=0) of S10, S11, and S12 are 11.35mV, 11.08mV, and 8.37mV, respectively.

[0216] (2) Outlier removal: The Grubbs test was used (α=0.05, sample size n=1200, critical value G). 0.05 =3.3) Remove outliers (such as signal mutation points and noise interference points) from the corrected signal: S10: Remove two outliers (occurring at t=8min and t=15min respectively, with signal deviation > 3 times the standard deviation); S11: Remove 1 outlier (occurring at t=12min, signal deviation > 3 times standard deviation); S12: Remove two outliers (occurring at t=5min and t=18min respectively, with signal deviation > 3 times the standard deviation); The outlier removal rate was ≤0.2%, which did not affect the overall data validity.

[0217] (3) Time-domain filtering: The corrected signal was processed using a Hanning window function (window width 5ms) to suppress signal tail noise (noise intensity ≤0.03mV after 15ms), while retaining the effective signal within the first 15ms (alcohol signal proportion ≥80%). After filtering, the signal smoothness of S10, S11 and S12 is improved by 30%, 32% and 28% respectively, and the signal baseline is stable without fluctuation.

[0218] 2. Interference signal separation and processing: (1) Moisture signal separation: The pure water signal is calculated based on the signal difference between two acquisitions of the spin echo sequence: Formula: Swater = Stotal - Soilol; Calculation results: S10=0.78mV, S11=0.82mV, S12=0.88mV; Verification: The moisture signal intensity was positively correlated with the sample moisture content (S10=7.2%, S11=7.5%, S12=8.0%) (correlation coefficient R=0.96), confirming that the moisture signal separation was accurate.

[0219] (2) Alcohol-oil signal separation: Based on the difference in transverse relaxation time (T2) between alcohols and oils, 30 echo signals acquired from CPMG sequences were analyzed: Relaxation characteristics analysis: Through experimental determination, the T2 value of alcoholic substances in white agarwood in this embodiment is 85-110ms, and the T2 value of oil is 210-280ms. The difference between the two is >40%, which provides a basis for separation. Attenuation curve fitting: The double exponential fitting function of Origin 2023 software was used to fit the alcohol attenuation curves to the CPMG echo signals (E1-E30) of S10, S11, and S12 respectively. With oil content decay curve (Where A0 is the initial signal intensity of alcohols, O0 is the initial signal intensity of oils, T2A is the T2 value of alcohols, and T2O is the T2 value of oils.) The fitting results are shown in the table below:

[0220] Goodness of fit of all samples (R) 2 (≥0.995), proving that the attenuation curve model is reliable and can accurately separate alcohol and oil signals.

[0221] (3) Frequency domain purification of alcohol signals: To eliminate the interference of trace amounts of bound water (T2=50-80ms) on alcohol signals, frequency domain purification was performed: Signal conversion: The initial alcohol signal A0 obtained by fitting was converted into a frequency domain spectrum by Fourier transform. The characteristic peaks of alcohol were concentrated in the range of 1.2-1.5 ppm, and the interference peak of bound water was located near 4.7 ppm. Digital filtering: Design a bandpass filter (center frequency 1.35ppm, bandwidth 0.5ppm) to filter the frequency domain spectrum, achieving a combined water signal attenuation of ≥99% near 4.7ppm; Purity verification: The purity of the filtered alcohol signal (alcohol signal intensity / total signal intensity × 100%) was calculated. S10 = 99.6%, S11 = 99.5%, and S12 = 99.4%, all of which are ≥ 99.0%, meeting the requirements for subsequent quantitative analysis.

[0222] 3. Calculation of alcohol content: Based on a pre-established standard curve of alcohol concentration-signal intensity (fitting equation: S_avg=25.8×c, (R 2 =0.999), where S_avg is the alcohol signal intensity and c is the alcohol content), substituting the pure alcohol signal intensity A0 of the three samples, calculate the alcohol content: S10 (Nha Trang, Vietnam insect leak): c10 = A0 / S_{slope} = 10.85 / 25.8 ≈ 42.1%; S11 (inducible by Hainan fungi): c11 = 10.58 / 25.8 ≈ 41.0%; S12 (physical induction in Dianbai, Guangdong): c12 = 8.15 / 25.8 ≈ 31.6%.

[0223] 4. Moisture distribution uniformity analysis: The uniformity of moisture distribution in agarwood samples is calculated by utilizing the slope change of the moisture signal attenuation curve (formula: U=1-|slope_{max}-slope_{min}| / slope_{avg}, the closer U is to 1, the more uniform the moisture distribution): Slope calculation: Divide the moisture signal decay curve (0-20min) into 5-minute segments, for a total of 4 segments, and calculate the slope DeltaS / Deltat for each segment; Uniformity results: S10: Slope_{max}=0.041mV / min, Slope_{min}=0.037mV / min, Slope_{avg}=0.039mV / min, U_{10}=1-|0.041-0.037| / 0.039=0.897≈0.90; S11: Slope_{max}=0.044mV / min, slope_{min}=0.042mV / min, slope_{avg}=0.043mV / min, U_{11}=1-|0.044-0.042| / 0.043=0.953≈0.95; S12: Slope_{max}=0.049mV / min, slope_{min}=0.044mV / min, slope_{avg}=0.046mV / min, U_{12}=1-|0.049-0.044| / 0.046=0.891≈0.89.

[0224] IV. Data Analysis and Result Verification: 1. Comparative Analysis: (1) Analysis of differences in alcohol content among different resin formation methods: Content ranking: Natural insect-induced resin (S10, 42.1%) ≈ Fungal-induced resin (S11, 41.0%) > Physically induced resin (S12, 31.6%). Reasons for the difference: S10 (Natural Insect-Bug): Insect infestation causes a stress response in agarwood trees, resulting in vigorous resin secretion and a long natural aging time (≥10 years). Alcohols (such as agaric spirol and benzyl acetone) accumulate sufficiently, and the "ventilation channels" formed by the insect holes promote the oxidation and transformation of resin, further increasing the alcohol content. S11 (Fungal Induction): Artificially inoculated white agarwood-specific Paecilomyces fungi can secrete specific enzymes (such as lignin-degrading enzymes) to accelerate the conversion of xylem into resin, and the resin formation cycle is controllable (3-5 years). The enzymatic reaction is highly uniform, and the alcohol content is close to that of natural insect-damaged material. S12 (physically induced): Resin secretion is stimulated only by mechanical damage, lacking the conversion effect of microbial enzymes, resulting in low resin synthesis efficiency (only 75% of that induced by fungi). Furthermore, the "local stress" caused by physical damage leads to uneven distribution of alcohols, with the overall content being significantly lower than the former two.

[0225] (2) Correlation analysis between signal stability and sample characteristics: Signal CV values ​​ranked as follows: fungal induction (S11, 0.85%) < natural insect leakage (S10, 1.02%) < physical induction (S12, 1.15%). Correlation logic: S11 is a regular cubic block with a uniform sheet-like distribution of oil (75% coverage), and the signal response is highly consistent under the influence of a magnetic field; S10 has uneven density in some areas due to insect holes (density in insect-hole areas is 1.02 g / cm³, and in non-insect-hole areas it is 1.08 g / cm³), and the signal fluctuation is slightly larger; S12 is a cuboid with linear oil distribution along the texture (density in textured areas is 1.01 g / cm³, and in non-textured areas it is 1.05 g / cm³), and the density difference is more obvious, resulting in slightly poorer signal stability. However, the CV values ​​of all three are ≤2%, which meets the detection requirements.

[0226] (3) Correlation analysis between moisture distribution uniformity and resin formation method: Evenness ranking: fungal induction (S11, 0.95) > natural insect leakage (S10, 0.90) ≈ physical induction (S12, 0.89). Related logic: During fungal-induced agarwood formation, the fungal hyphae grow uniformly inside the wood (hyphae distribution density variation coefficient < 5%), which can promote the uniform transport of water and nutrients and avoid local water accumulation; natural insect-damaged agarwood is affected by the randomness of insect borer paths, and water tends to accumulate near the borer holes (moisture content 7.8% in borer hole areas and 6.8% in non-borer areas); the moisture distribution of physically induced agarwood is limited by the damaged texture (moisture content 8.5% in textured grooves and 7.5% in non-textured areas), and the uniformity is low. This result is completely consistent with the microstructural characteristics of the sample (verified by scanning electron microscopy).

[0227] 2. Result Verification: (1) Repeatability verification: Five parallel tests were performed on each sample (using the same equipment, parameters, and different operators), and the relative standard deviation (RSD) of the alcohol content was calculated. S10: The five test values ​​were 42.1%, 42.3%, 41.9%, 42.2%, and 42.0%, respectively, with an average value of 42.1% and an RSD of 0.38%. S11: The five test values ​​were 41.0%, 41.2%, 40.8%, 41.1%, and 40.9%, respectively, with an average value of 41.0% and an RSD of 0.36%. S12: The five test values ​​were 31.6%, 31.8%, 31.4%, 31.7%, and 31.5%, respectively, with an average value of 31.6% and an RSD of 0.45%. The RSDs are all <1.0%, which proves that the method has excellent repeatability and the operational error is controllable.

[0228] (2) Accuracy verification: The alcohol content of three samples was determined by high performance liquid chromatography (HPLC, referring to the national forestry recommended standard "Agarwood" LY / T2904-2017), and the results were compared with the baseline values ​​using this method. HPLC detection conditions: C18 column (250 mm × 4.6 mm, 5 μm), mobile phase methanol-water (85:15, v / v), flow rate 1.0 mL / min, detection wavelength 280 nm, column temperature 30 ℃, injection volume 10 μL; Comparison results:

[0229] The RE values ​​were all <1.0%, meeting the accuracy requirements of the detection method (allowable error ±2%), proving that the detection results of this method are highly consistent with the standard HPLC method, and the data are reliable.

[0230] In conjunction with the above embodiments, please refer to Figure 2 , Figure 2 This is a structural block diagram of an agarwood alcohol content detection device 200 provided in an embodiment of this application. The device 200 can be a module, program segment, or code on an electronic device. It should be understood that the device 200 corresponds to the above method embodiment and is capable of performing the various steps involved in the method embodiment. The specific functions of the device 200 can be found in the description above. To avoid repetition, detailed descriptions are appropriately omitted here.

[0231] Optionally, the device 200 includes: The scanning module 210 is used to perform nuclear magnetic resonance scanning on the agarwood sample to be tested, adjust the static magnetic field strength and radio frequency, and determine the optimal magnetic field strength and optimal radio frequency for exciting the hydrogen nuclei of alcohol substances. The signal acquisition module 220 is used to transmit a preset pulse sequence and acquire the target mixed signal while maintaining the optimal magnetic field strength and the optimal radio frequency. The feature acquisition module 230 is used to acquire the signal features of alcohols based on the target mixed signal; The concentration determination module 240 is used to determine the concentration of alcohols in the agarwood sample to be tested based on the signal characteristics.

[0232] Optionally, the signal acquisition module 220 is configured to transmit a first radio frequency pulse sequence, transmit a second radio frequency pulse sequence after a first delay time, and acquire a first mixed signal after a second delay time, wherein the first mixed signal includes signals of oil and alcohol, the first delay time is less than the transverse relaxation time of water, and the second delay time is greater than the transverse relaxation time of water; transmit a third radio frequency pulse sequence, and acquire multiple second mixed signals arranged in chronological order, wherein the second mixed signal includes signals of oil and alcohol; wherein the target mixed signal includes the first mixed signal and the second mixed signal.

[0233] Optionally, the feature acquisition module 230 is used to perform exponential decay curve fitting on multiple second mixed signals based on the difference in lateral relaxation time between alcohols and oils; and to acquire the initial signal intensity of alcohols based on the fitting result and the first mixed signal, wherein the signal feature includes the initial signal intensity.

[0234] Optionally, the concentration determination module 240 is used to determine the concentration of alcohol corresponding to the initial signal intensity based on a preset correlation between alcohol concentration and signal intensity.

[0235] Optionally, the fitting result includes the alcohol decay signal, and the feature acquisition module 230 is further configured to perform FFT transformation on the alcohol decay signal to obtain a frequency signal; perform bandpass filtering on the frequency signal to obtain a filtered alcohol decay signal, and determine the initial signal intensity of the alcohol based on the filtered alcohol decay signal.

[0236] Optionally, the first radio frequency pulse sequence is a 90° spin echo sequence, the second radio frequency pulse sequence is a 180° spin echo sequence, and the third radio frequency pulse sequence is a CPMG sequence.

[0237] Optionally, the device 200 further includes: The uniformity analysis module is used to acquire a third mixed signal after the first delay time; obtain a moisture signal based on the first mixed signal and the third mixed signal; acquire the decay rate change at different time periods based on the moisture signal; and evaluate the uniformity of moisture distribution inside the agarwood sample based on the decay rate change.

[0238] Optionally, the scanning module 210 is used to adjust the magnetic field strength in increments of a set step size within a preset range, and to acquire the magnetic field signal at each magnetic field strength; to determine the magnetic field strength corresponding to the maximum signal-to-noise ratio of the alcohol characteristics in the magnetic field signal as the optimal magnetic field strength; to calculate the theoretical radio frequency based on the optimal magnetic field strength; to perform scanning within a set proximity range of the theoretical radio frequency and acquire the response signal at each frequency point; and to determine the frequency corresponding to the peak value of the response signal as the optimal radio frequency.

[0239] Optionally, the signal acquisition module 220 is used to transmit a single excitation pulse while maintaining the optimal magnetic field strength and the optimal radio frequency, and to acquire a baseline signal of a preset duration; if the intensity fluctuation of the baseline signal does not exceed a set threshold, then the step of transmitting the preset pulse is executed.

[0240] It should be noted that those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0241] Please refer to Figure 3 , Figure 3 This is a schematic diagram of an electronic device for performing a method for detecting the alcohol content of agarwood, provided in an embodiment of this application. The electronic device may include: at least one processor 310, such as a CPU; at least one communication interface 320; at least one memory 330; and at least one communication bus 340. The communication bus 340 is used to establish communication between these components. In this embodiment, the communication interface 320 is used for signaling or data communication with other node devices. The memory 330 may be a high-speed RAM or a non-volatile memory, such as at least one disk storage device. Optionally, the memory 330 may also be at least one storage device located remotely from the aforementioned processor. The memory 330 stores computer-readable instructions, which, when executed by the processor 310, cause the electronic device to perform the aforementioned method process.

[0242] Understandable. Figure 3 The structure shown is for illustrative purposes only; the electronic device may also include components that are more advanced than those shown. Figure 3 The more or fewer components shown, or having the same Figure 3 The different configurations shown. Figure 3 The components shown can be implemented using hardware, software, or a combination thereof.

[0243] This application provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it performs the method process executed by the electronic device in the above method embodiments.

[0244] This embodiment discloses a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can perform the methods provided in the above-described method embodiments, such as including: Nuclear magnetic resonance scanning was performed on the agarwood sample to be tested. The static magnetic field strength and radio frequency were adjusted to determine the optimal magnetic field strength and optimal radio frequency for exciting the hydrogen nuclei of alcohols. While maintaining the optimal magnetic field strength and the optimal radio frequency, a preset pulse sequence is emitted to acquire the target mixed signal; Based on the target mixed signal, the signal characteristics of alcohols are obtained; Based on the signal characteristics, the concentration of alcohols in the agarwood sample to be tested is determined.

[0245] In summary, the embodiments of this application provide a method, electronic device, storage medium, and program product for detecting alcohol content in agarwood. This method can achieve non-destructive testing of agarwood samples using nuclear magnetic resonance technology, without the need for complex pretreatment processes, which can significantly improve detection efficiency. By locking the optimal magnetic field strength and optimal radio frequency for exciting hydrogen nuclei of alcohol substances, combined with targeted acquisition and signal feature extraction of target mixed signals, the specificity and signal-to-noise ratio of alcohol signals are significantly improved. It effectively removes interference from non-target components such as moisture, oil, and lignin in agarwood samples, ensuring detection accuracy and repeatability.

[0246] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0247] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0248] Furthermore, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0249] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0250] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for detecting the alcohol content of agarwood, characterized by, The method comprises: nuclear magnetic resonance scanning is performed on the to-be-tested eaglewood sample, a static magnetic field strength and a radio frequency are adjusted, and optimal magnetic field strength and optimal radio frequency for exciting hydrogen nuclei of alcohol substances are determined; a preset pulse sequence is transmitted under the optimal magnetic field strength and the optimal radio frequency, and a target mixed signal is collected; signal characteristics of alcohol substances are obtained according to the target mixed signal; the alcohol substance concentration of the to-be-tested eaglewood sample is determined according to the signal characteristics.

2. The method of claim 1, wherein, The transmitting of the preset pulse sequence and the collecting of the target mixed signal comprise: a first radio frequency pulse sequence is transmitted, a second radio frequency pulse sequence is transmitted after a first delay time, and a first mixed signal is collected after a second delay time, wherein the first mixed signal comprises signals of oil and alcohol substances, the first delay time is less than the transverse relaxation time of water, and the second delay time is greater than the transverse relaxation time of water; a third radio frequency pulse sequence is transmitted, and a plurality of second mixed signals arranged in time sequence are collected, wherein the second mixed signals comprise signals of oil and alcohol substances; the target mixed signal comprises the first mixed signal and the second mixed signals.

3. The method of claim 2, wherein, The obtaining of the signal characteristics of alcohol substances according to the target mixed signal comprises: exponential decay curve fitting is performed on the plurality of second mixed signals based on the difference in transverse relaxation time between alcohol substances and oil; initial signal strength of alcohol substances is obtained according to the fitting result and the first mixed signal, and the signal characteristics comprise the initial signal strength.

4. The method of claim 3, wherein, The determining of the alcohol substance concentration of the to-be-tested eaglewood sample according to the signal characteristics comprises: the alcohol substance concentration corresponding to the initial signal strength is determined according to a preset correlation between alcohol substance concentration and signal strength.

5. The method of claim 3, wherein, The fitting result comprises alcohol substance decay signals, and after the initial signal strength of alcohol substances is obtained, the method further comprises: FFT transformation is performed on the alcohol substance decay signals to obtain frequency signals; band-pass filtering is performed on the frequency signals to obtain filtered alcohol substance decay signals, and the initial signal strength of alcohol substances is determined according to the filtered alcohol substance decay signals.

6. The method of claim 2, wherein, The first radio frequency pulse sequence is a 90° spin echo sequence, the second radio frequency pulse sequence is a 180° spin echo sequence, and the third radio frequency pulse sequence is a CPMG sequence.

7. The method of claim 2, wherein, The method further comprises: a third mixed signal is collected after the first delay time; a water signal is obtained according to the first mixed signal and the third mixed signal; a change in decay rate in different time periods is obtained according to the water signal; the distribution uniformity of internal water in the to-be-tested eaglewood sample is evaluated according to the change in decay rate.

8. The method of claim 1, wherein, The adjusting of the static magnetic field strength and the radio frequency to determine the optimal magnetic field strength and the optimal radio frequency for exciting hydrogen nuclei of alcohol substances comprises: the magnetic field strength is adjusted in a preset range with a set step length, and a magnetic field signal under each magnetic field strength is collected; the optimal magnetic field strength is determined as the magnetic field strength corresponding to the maximum signal-to-noise ratio of alcohol characteristics in the magnetic field signal; and calculating a theoretical radio frequency according to the optimal magnetic field strength; scanning in a set adjacent range of the theoretical radio frequency, and collecting response signals of each frequency point; determining a frequency corresponding to a peak of the response signal as an optimal radio frequency.

9. The method of claim 1, wherein, Before the step of transmitting the preset pulse sequence, the method further comprises: maintaining the optimal magnetic field strength and the optimal radio frequency, transmitting a single excitation pulse, and collecting a baseline signal of a preset time length; if the intensity fluctuation of the baseline signal does not exceed a set threshold, performing the step of transmitting the preset pulse.

10. An electronic device, comprising: The computer program is executed by the processor to run the method of any one of claims 1-9.

11. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to run the method of any one of claims 1-9.

12. A computer program product, characterised in that, The computer program is executed by the processor to run the method of any one of claims 1-9.

Images