Mildew identification method for medicinal and edible traditional Chinese medicinal materials

Abstract

The invention relates to the technical field of quality detection and mildew identification of traditional Chinese medicinal materials, and discloses a method for identifying mildew of traditional Chinese medicinal materials with homology of medicine and food, which comprises the following steps: selecting samples from traditional Chinese medicinal materials with homology of medicine and food, judging the mildew risk grade of the samples, collecting appearance, spectrum and smell characteristic information of the samples, processing the characteristic information by utilizing a characteristic selection algorithm, and identifying the mildew of the traditional Chinese medicinal materials with homology of medicine and food. The method comprises the following steps: detecting mildewing microorganisms in a sample to obtain existing mildewing types, detecting the types and quantity of the mildewing microorganisms in the sample through a real-time fluorescent quantitative PCR technology, comparing the existing mildewing types with mildewing microorganism detection data to obtain a mildewing recognition result of the sample, and combining an appearance image, spectral information, odor information and mildewing microorganism detection data. According to the method, the mildewing of the medicinal and edible traditional Chinese medicinal materials can be quickly and accurately identified, and an effective technical means is provided for quality control and safety detection of the traditional Chinese medicinal materials.

Description

Technical Field

[0001] The present invention relates to the technical field of quality detection and mildew identification of traditional Chinese medicines, and in particular to a method for identifying mildew of traditional Chinese medicines with medicinal and edible properties. Background Art

[0002] Traditional Chinese medicines that are both edible and medicinal are easily contaminated by mold during storage and transportation, leading to mold. Traditional methods for detecting mold mainly rely on sensory evaluation and culture of moldy microorganisms. The former relies on the experience of the inspectors, is highly subjective and difficult to quantify, and is prone to misjudgment. Although the latter can detect the presence of mold, it is time-consuming and usually takes several days to obtain results, which cannot meet the needs of rapid detection.

[0003] In recent years, some mold identification methods based on single technologies, such as spectral analysis or odor detection, have emerged. However, these methods have certain limitations. Although spectral analysis can provide chemical information of the sample, it is difficult to distinguish the type and degree of mold. Although odor detection can detect volatile organic compounds, it is easily interfered by environmental factors and cannot be used alone for accurate diagnosis of mold. In addition, the accuracy and reliability of these single technical methods when detecting complex samples need to be improved.

[0004] Identifying mold in Chinese medicinal materials that are both medicinal and edible requires not only determining whether mold has occurred, but also determining the state, degree, and type of mold. Existing detection technologies are difficult to meet these requirements simultaneously, especially in terms of rapid, accurate, and comprehensive identification. For samples with similar appearances but different degrees of mold, a single detection technology is often unable to effectively distinguish them, resulting in inaccurate test results.

[0005] In summary, existing technologies for detecting mold in edible and medicinal Chinese herbs suffer from slow detection speed, strong subjectivity, insufficient accuracy, and an inability to fully identify the state and type of mold. Therefore, a method that can quickly, accurately, and comprehensively identify mold in edible and medicinal Chinese herbs is urgently needed to meet the needs of practical applications. Summary of the Invention

[0006] The purpose of the present invention is to solve the problems in the prior art and to propose a method for identifying moldy Chinese medicinal materials that are both medicinal and edible.

[0007] In order to achieve the above object, the present invention adopts the following technical solution: a method for identifying moldy Chinese medicinal materials that are both medicinal and edible, comprising the following steps:

[0008] Step S1, selecting samples from Chinese medicinal materials that are both medicinal and edible, and determining the mold risk level of the samples;

[0009] Step S2, using a high-resolution camera to photograph the sample from multiple angles to obtain an appearance image;

[0010] Step S3, using a near-infrared spectrometer to perform spectral scanning on the sample to obtain spectral information;

[0011] Step S4, using an electronic nose device to detect volatile organic compounds released by the sample to obtain odor information;

[0012] Step S5, extracting and fusing the collected appearance image, spectral information, and odor information to obtain a comprehensive feature vector;

[0013] Step S6, using a feature selection algorithm to process the comprehensive feature vector and list the types of mold present;

[0014] Step S7, detecting the types and quantities of moldy microorganisms in the sample by real-time fluorescence quantitative PCR technology to obtain moldy microorganism detection data;

[0015] Step S8: Compare the existing mold types with the mold microorganism detection data to obtain a mold identification result of the sample.

[0016] Furthermore, in step S1, the following sub-steps are also included:

[0017] S1-1, using random sampling method, select Chinese herbal medicines from different batches, different origins, and different storage conditions;

[0018] S1-2, record the collection date, collection batch, and origin of the Chinese medicinal materials;

[0019] S1-3, record the storage conditions of Chinese medicinal materials, including temperature, humidity, and ventilation;

[0020] S1-4, establish a risk scoring system:

[0021] A time risk score is assigned based on the length of time between the collection date and the testing date, an origin risk score is assigned based on the climatic conditions of the origin, and a storage condition score is assigned based on the storage conditions;

[0022] The time risk score, origin risk score, and storage condition risk score are added together to obtain a comprehensive risk score. Based on the comprehensive risk score, the mold risk of Chinese medicinal materials is divided into three levels: low, medium, and high, and the mold risk level of the Chinese medicinal materials is recorded;

[0023] S1-5, select equal amounts of samples from Chinese medicinal materials of each mold risk level for mold testing.

[0024] Furthermore, in step S2, the following sub-steps are also included:

[0025] S2-1, place the sample on a shooting platform with uniform light and a flat surface;

[0026] S2-2, use a ring light device to supplement the light of the sample;

[0027] S2-3, using a 12-megapixel high-definition camera, photograph the sample from multiple angles, including the front, side, back, and oblique surfaces of the sample, to obtain an appearance image of the sample.

[0028] Furthermore, in step S3, the following sub-steps are also included:

[0029] S3-1, use a near-infrared spectrometer, set the near-infrared spectrometer wavelength range to 700-2500nm, and the resolution to 4cm -1 ;

[0030] S3-2, place the sample on the detection platform of the near-infrared spectrometer and fix it using a sample cell or a glass slide;

[0031] S3-3, starting a near-infrared spectrometer to detect the sample and obtain spectral information of the sample.

[0032] Furthermore, in step S4, the following sub-steps are also included:

[0033] S4-1, using an electronic nose device based on a metal oxide semiconductor (MOS) sensor, setting the detection range of the electronic nose device to 0-100 ppm and a resolution of 0.1 ppm;

[0034] S4-2, cut the sample into small pieces, fix them using a sample cell or glass slide, and place them in a sealed container for 30 minutes;

[0035] S4-3, starting the electronic nose device to detect the sample and obtain odor information of the sample. The odor information is a sensor response curve formed by the change of the response value of the metal oxide semiconductor (MOS) sensor in the electronic nose device over time.

[0036] Furthermore, in step S5, the following sub-steps are also included:

[0037] S5-1, extracting shape features of the appearance image to obtain a shape feature vector, wherein the shape features include the outline, area, and perimeter of the sample; calculating the average color value of the appearance image using the RGB channels to obtain a color feature vector; and extracting texture features of the appearance image using the gray-level co-occurrence matrix (GLCM) to obtain a texture feature vector;

[0038] S5-2, analyzing the spectral information of the sample to obtain a spectral curve feature vector, wherein the spectral information includes the position and intensity characteristics of the spectral absorption peak, and the slope and curvature characteristics of the spectral curve;

[0039] S5-3, extracting odor features from the odor information, wherein the odor features include the maximum value, minimum value, average value, and standard deviation of the sensor response curve, and performing dimensionality reduction processing on the odor features using principal component analysis (PCA) technology to obtain an odor feature vector;

[0040] S5-4, concatenating the extracted shape feature vector, color feature vector, texture feature vector, spectrum feature vector, and odor feature vector to obtain a comprehensive feature vector, which is:

[0041] Comprehensive feature vector = shape feature vector, color feature vector, texture feature vector, spectrum feature vector, smell feature vector.

[0042] Furthermore, in step S6, the following sub-steps are also included:

[0043] S6-1, using the recursive feature elimination algorithm RFE to filter the comprehensive feature vector and obtain the feature subset with the highest discrimination for mold identification;

[0044] S6-2, establishing a support vector machine (SVM) model, and training the SVM model using labeled samples, wherein the labeled samples include moldy samples and non-moldy samples;

[0045] S6-3, inputting the feature subset into the SVM model, determining whether the sample is moldy, listing the types of mold that exist, and obtaining a list of mold types;

[0046] S6-4: If the sample is moldy, the size and distribution of the colonies in the moldy area are obtained by analyzing the appearance image and spectral information.

[0047] Furthermore, in step S7, the following sub-steps are also included:

[0048] S7-1, extracting moldy microbial DNA from the sample using a moldy microbial DNA extraction kit to obtain a moldy microbial DNA template;

[0049] S7-2: Determine the target moldy microorganisms based on the mold species list, search for the specific gene sequence of the target moldy microorganisms, and design primers and probes based on the specific gene sequence;

[0050] S7-3: Prepare the reaction mixture in a sterile PCR tube according to the following ratio:

[0051] Moldy microbial DNA template: 2 μL;

[0052] Primer: unit concentration is 10 μM, input volume is 1 μL each for upstream and downstream;

[0053] Probe: unit concentration is 10 μM, input volume is 1 μL;

[0054] 2×qPCR Mix solution: The input volume is 10 μL. The 2×qPCR Mix solution is a pre-prepared reaction mixture containing Taq enzyme, dNTPs, Mg 2+ ;

[0055] Use sterile water to make up the reaction system to 20 μL;

[0056] S7-4: Set the temperature cycle parameters of the real-time fluorescence quantitative PCR instrument according to the design of primers and probes:

[0057] Initial denaturation: 95°C for 10 min;

[0058] Cyclic denaturation: 95°C for 15 seconds;

[0059] Annealing and extension: 60°C for 1 min;

[0060] Number of cycles: 40-45 times;

[0061] Select the fluorescence detection channel according to the probe type;

[0062] S7-5, set the baseline and fluorescence threshold using the real-time fluorescence quantitative PCR instrument software;

[0063] S7-6: Add the prepared reaction system to the reaction wells of the real-time fluorescence quantitative PCR instrument, start the real-time fluorescence quantitative PCR instrument, run the preset temperature cycle program, and detect the fluorescence signal after each cycle;

[0064] S7-7, the real-time fluorescence quantitative PCR instrument calculates the Ct value of each sample through the fluorescence signal, and determines whether the target moldy microorganisms are present in the sample by comparing the Ct value and the fluorescence threshold. If the target moldy microorganisms are present, the number of target moldy microorganisms is calculated by the standard curve method to obtain the moldy microorganism detection data.

[0065] Furthermore, in step S8, the following sub-steps are also included:

[0066] S8-1, comparing the type of mold with the mold microorganism detection data to determine whether the sample is moldy and obtain the mold status. If the sample is moldy, confirm the type of mold;

[0067] S8-2, assessing the degree of mold based on the number of moldy microorganisms, colony size in the moldy area, and distribution range. The degree of mold can be classified as mild, moderate, or severe.

[0068] S8-3, determining the type of mildew based on the degree and type of mildew, wherein the mildew type includes single-species mildew and multi-species mildew;

[0069] S8-4, recording the mold identification result in a report, wherein the mold identification result includes the mold state, mold type, mold degree, and mold type;

[0070] S8-5, determining whether the mold identification result meets the mold risk level described in step S1, and recording the determination result in a report.

[0071] The beneficial effects brought about by the technical solution provided by the present invention include at least:

[0072] The present invention combines multiple technical means such as high-resolution cameras to capture appearance images, near-infrared spectrometers to scan spectral information, and electronic noses to detect odor information. It can obtain multi-dimensional information of samples in a short period of time, quickly generate comprehensive feature vectors through feature extraction and fusion, and list the types of mold through feature selection algorithms. Compared with traditional moldy microbial culture methods, the present invention greatly shortens detection time and improves detection efficiency.

[0073] The present invention more comprehensively reflects the moldy characteristics of samples through the integration of multi-dimensional information, and further verifies the types and quantities of moldy microorganisms through real-time fluorescence quantitative PCR technology, making the detection results more accurate and reliable. This multi-technology integration method can effectively avoid the limitations of single technology detection and improve the accuracy of mold identification.

[0074] The present invention can further determine the state, degree and type of mold by extracting and fusion-generated comprehensive feature vectors, combining them with mold microbial detection data to comprehensively identify mold characteristics, and cross-validating the mold identification results with the initial risk level. This method can meet the comprehensive requirements for mold detection in traditional Chinese medicines that are both medicinal and edible in practical applications, while effectively improving the reliability and traceability of mold identification results, providing strong support for the quality control and safety assessment of traditional Chinese medicines.

[0075] The various detection technologies used in the present invention are all mature technologies, with relatively low equipment costs and simple operation. Through algorithm processing and automated analysis, the complexity and errors of manual operation are reduced. Compared with traditional moldy microorganism culture methods, the present invention does not require a long culture process, reduces detection costs and operation difficulty, and is suitable for large-scale promotion and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0077] Figure 1A flowchart of a method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0078] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a method for identifying moldy Chinese medicinal materials of medicinal and edible origin proposed by the present invention, its specific implementation, structure, features and effects. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable form.

[0079] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0080] The following examples are for illustrative purposes only and are not intended to limit the scope of the present invention.

[0081] The specific scheme of the method for identifying moldy Chinese medicinal materials with both medicinal and edible properties provided by the present invention is described in detail below with reference to the accompanying drawings.

[0082] See also Figure 1 , which shows a method flow chart of a method for identifying moldy Chinese medicinal materials with medicinal and edible properties provided by one embodiment of the present invention, the method comprising the following steps:

[0083] Step S1, selecting samples from Chinese medicinal materials that are both medicinal and edible, and determining the mold risk level of the samples;

[0084] Wherein, step S1 further includes the following sub-steps:

[0085] S1-1, using random sampling method, select Chinese herbal medicines from different batches, different origins, and different storage conditions;

[0086] S1-2, record the collection date, collection batch, and origin of the Chinese medicinal materials;

[0087] S1-3, record the storage conditions of Chinese medicinal materials, including temperature, humidity, and ventilation;

[0088] S1-4, establish a risk scoring system:

[0089] A time risk score is assigned based on the length of time between the collection date and the testing date, an origin risk score is assigned based on the climatic conditions of the origin, and a storage condition score is assigned based on the storage conditions;

[0090] The time risk score, origin risk score, and storage condition risk score are added together to obtain a comprehensive risk score. Based on the comprehensive risk score, the mold risk of Chinese medicinal materials is divided into three levels: low, medium, and high, and the mold risk level of the Chinese medicinal materials is recorded;

[0091] S1-5, select equal amounts of samples from Chinese medicinal materials of each mold risk level for mold testing.

[0092] It should be noted that:

[0093] Random sampling is a commonly used sample selection technique in statistics, which aims to ensure that the sample is representative of the population, thereby improving the reliability and universality of the research results.

[0094] The collection date is an important time dimension factor in assessing the risk of mold. The climatic conditions in different periods have a significant impact on the mold risk of Chinese medicinal materials. The high temperature and high humidity in summer (temperature 20℃-30℃, relative humidity above 70%) are ideal environments for mold growth. The risk of mold in Chinese medicinal materials collected in summer is higher. The low temperature and dryness in winter (temperature below 10℃, relative humidity below 60%) are generally not conducive to mold growth. The risk of mold in Chinese medicinal materials collected in winter is lower.

[0095] Origin is an important spatial dimension factor in assessing mold risk. The climatic conditions in different origins vary significantly. The southern region usually has a humid climate, which is suitable for mold growth. Therefore, Chinese medicinal materials produced in the south have a higher risk of mold; the northern region is relatively dry, and Chinese medicinal materials produced in the north have a lower risk of mold. Secondly, the soil type in different origins affects the quality of Chinese medicinal materials. Some soils may contain more mold spores, and samples from this origin have a higher risk of mold.

[0096] Storage conditions are key environmental factors in assessing mold risk, and their impact is mainly reflected in the following aspects:

[0097] The suitable temperature range for mold growth is usually 20℃-30℃. Chinese medicinal materials stored at room temperature have a high risk of mold, while those stored at low temperatures have a low risk of mold.

[0098] The suitable relative humidity for mold growth is usually above 70%. Chinese medicinal materials stored in a high humidity environment have a high risk of mold, while those stored in a low humidity environment have a low risk of mold.

[0099] Good ventilation can reduce the humidity and temperature of the storage environment and reduce the risk of mildew.

[0100] Step S2, using a high-resolution camera to photograph the sample from multiple angles to obtain an appearance image;

[0101] Wherein, in step S2, the following sub-steps are also included:

[0102] S2-1, place the sample on a shooting platform with uniform light and a flat surface;

[0103] S2-2, use a ring light device to supplement the light of the sample;

[0104] S2-3 uses a 12-megapixel high-definition camera to shoot the sample from multiple angles, including the front, side, back and oblique surface of the sample, to obtain the appearance image of the sample.

[0105] It should be noted that:

[0106] Ring lights are specialized lighting devices widely used in industrial inspection, scientific research, medical imaging, and photography. Providing a uniform, shadowless, ring-shaped light source, they effectively enhance detail and contrast in observed objects, making them particularly suitable for applications requiring high-precision imaging.

[0107] The 12-megapixel high-definition camera can clearly capture tiny details on the surface of Chinese medicinal materials, such as spots in the early stages of mold, texture changes and color differences. The high-pixel image remains clear after magnification, making it easier to observe the microstructure of Chinese medicinal materials, such as mold hyphae and spores, thereby improving the accuracy of mold identification. In addition, the rich pixel information provides sufficient data support for subsequent image processing and feature extraction.

[0108] Step S3, using a near-infrared spectrometer to perform spectral scanning on the sample to obtain spectral information;

[0109] Wherein, in step S3, the following sub-steps are also included:

[0110] S3-1, use a near-infrared spectrometer, set the near-infrared spectrometer wavelength range to 700-2500nm, and the resolution to 4cm -1 ;

[0111] S3-2, place the sample on the detection platform of the near-infrared spectrometer and fix it using a sample cell or a glass slide;

[0112] S3-3, starting a near-infrared spectrometer to detect the sample and obtain spectral information of the sample.

[0113] It should be noted that the near-infrared spectrometer is an analytical instrument based on the principle of near-infrared light absorption. It utilizes the vibrational harmonics and summed frequency absorption characteristics of chemical bonds in organic matter to obtain the chemical information of the sample by measuring the sample's absorption spectrum of near-infrared light. Its advantages include being fast, non-destructive, and requiring no complex sample pre-treatment. It is suitable for the detection of samples in various forms (such as solids, liquids, powders, etc.), can be used for component analysis, quality control, and origin identification, and is widely used in food, medicine, agriculture, petrochemicals and other fields.

[0114] Step S4, using an electronic nose device to detect volatile organic compounds released by the sample to obtain odor information;

[0115] Wherein, in step S4, the following sub-steps are also included:

[0116] S4-1, using an electronic nose device based on a metal oxide semiconductor (MOS) sensor, setting the detection range of the electronic nose device to 0-100 ppm and a resolution of 0.1 ppm;

[0117] S4-2, cut the sample into small pieces, fix them using a sample cell or glass slide, and place them in a sealed container for 30 minutes;

[0118] S4-3, starting the electronic nose device to detect the sample and obtain odor information of the sample. The odor information is a sensor response curve formed by the change of the response value of the metal oxide semiconductor (MOS) sensor in the electronic nose device over time.

[0119] It should be noted that the electronic nose device is an intelligent detection instrument that simulates human sense of smell. It is mainly composed of a gas sensor array, a signal processing unit and a pattern recognition system. It realizes rapid detection and analysis of odors by simulating the sensing technology and pattern recognition algorithm of biological olfaction. It has the advantages of high sensitivity, fast response speed and strong repeatability. It can detect extremely low concentrations of odor molecules and is widely used in food quality testing, environmental monitoring, medical diagnosis and industrial production.

[0120] Step S5: extracting and fusing the collected appearance image, spectral information, and odor information to obtain a comprehensive feature vector;

[0121] Wherein, in step S5, the following sub-steps are also included:

[0122] S5-1, extract the shape features of the appearance image to obtain a shape feature vector. The shape features include the outline, area, and perimeter of the sample. Calculate the average color value of the appearance image through the RGB channel to obtain a color feature vector. Extract the texture features of the appearance image through the gray-level co-occurrence matrix (GLCM) to obtain a texture feature vector.

[0123] S5-2, analyzing the spectral information of the sample to obtain a spectral curve feature vector, the spectral information including the position and intensity characteristics of the spectral absorption peak, and the slope and curvature characteristics of the spectral curve;

[0124] S5-3, extracting odor features from the odor information. The odor features include the maximum value, minimum value, average value, and standard deviation of the sensor response curve. The odor features are reduced in dimension using the principal component analysis (PCA) technique to obtain an odor feature vector.

[0125] S5-4, the extracted shape feature vector, color feature vector, texture feature vector, spectrum feature vector and odor feature vector are spliced ​​to obtain a comprehensive feature vector, which is:

[0126] Comprehensive feature vector = [shape feature vector, color feature vector, texture feature vector, spectrum feature vector, smell feature vector].

[0127] It should be noted that the Gray-Level Co-occurrence Matrix (GLCM) is a method used for texture analysis. Based on the gray value distribution of pixel pairs in an image, it calculates the frequency of occurrence of pixel pairs in a specific direction and distance, and describes texture features by counting the symbiotic relationship of pixel gray values ​​in the image. GLCM is widely used in image processing, computer vision, and medical image analysis.

[0128] Principal component analysis (PCA) is a classic statistical analysis and dimensionality reduction technique used to extract key information from high-dimensional data. It uses a few principal components to explain most of the variation in the data, reducing the data dimension while retaining the most important features. PCA is widely used in data preprocessing, feature extraction, image compression, genetic data analysis and other fields.

[0129] Step S6: Process the comprehensive feature vector using a feature selection algorithm to list the types of mold present;

[0130] Wherein, in step S6, the following sub-steps are also included:

[0131] S6-1, using the recursive feature elimination algorithm RFE to filter the comprehensive feature vector and obtain the feature subset with the highest discrimination for mold identification;

[0132] S6-2, establishing a support vector machine (SVM) model and training the SVM model using labeled samples, where the labeled samples include moldy samples and non-moldy samples;

[0133] S6-3, inputting the feature subset into the SVM model, determining whether the sample is moldy, listing possible mold types, and obtaining a list of mold types;

[0134] S6-4: If the sample is moldy, the size and distribution of the colonies in the moldy area are obtained by analyzing the appearance image and spectral information.

[0135] It should be noted that Recursive Feature Elimination (RFE) is a feature selection technique that aims to select the optimal feature subset by recursively removing features. RFE can be used for regression and classification problems, and is often used to reduce feature dimensionality and improve model performance.

[0136] In Python, you can use the RFE class in the sklearn library to implement recursive feature elimination. The following is an example code:

[0137] import pandas as pd

[0138] from sklearn.ensemble import RandomForestClassifier as RFC

[0139] from sklearn.feature_selection import RFE

[0140] from sklearn.model_selection import cross_val_score

[0141] #Read data

[0142] data=pd.read_csv('WFs1.csv')

[0143] X = data.iloc[:,1:]

[0144] Y = data.iloc[:,0]

[0145] #Build a random forest model

[0146] model=RFC()

[0147] #Recursive feature elimination method

[0148] selector=RFE(model,n_features_to_select=3,step=1)

[0149] selector = selector.fit(X,Y)

[0150] # Output results

[0151] print("Number of selected features:",selector.n_features_)

[0152] print("Feature ranking:",selector.ranking_)

[0153] X_selected=selector.transform(X)

[0154] score=cross_val_score(model,X_selected,Y,cv=9).mean()

[0155] print("Cross validation score:",score)

[0156] Support Vector Machine (SVM) is a powerful supervised learning algorithm mainly used for classification and regression tasks. Its core idea is to separate data points of different categories by finding the optimal hyperplane, while maximizing the interval between the hyperplane and the nearest data point (support vector). This interval maximization strategy gives SVM good generalization ability and can work effectively even in high-dimensional space. SVM is widely used in image recognition, text classification, bioinformatics and other fields.

[0157] Step S7: Detect the types and quantities of moldy microorganisms in the sample using real-time fluorescence quantitative PCR technology to obtain moldy microorganism detection data;

[0158] Wherein, in step S7, the following sub-steps are also included:

[0159] S7-1, extracting moldy microbial DNA from the sample using a moldy microbial DNA extraction kit to obtain a moldy microbial DNA template;

[0160] S7-2: Determine the target moldy microorganisms based on the mold species list, search for the specific gene sequence of the target moldy microorganisms, and design primers and probes based on the specific gene sequence;

[0161] S7-3: Prepare the reaction mixture in a sterile PCR tube according to the following ratio:

[0162] Moldy microbial DNA template: 2 μL;

[0163] Primer: unit concentration is 10 μM, input volume is 1 μL each for upstream and downstream;

[0164] Probe: unit concentration is 10 μM, input volume is 1 μL;

[0165] 2×qPCR Mix solution: The input volume is 10 μL. 2×qPCR Mix solution is a pre-prepared reaction mixture containing Taq enzyme, dNTPs, Mg 2+ ;

[0166] Use sterile water to make up the reaction system to 20 μL;

[0167] S7-4: Set the temperature cycle parameters of the real-time fluorescence quantitative PCR instrument according to the design of primers and probes:

[0168] Initial denaturation: 95°C for 10 min;

[0169] Cyclic denaturation: 95°C for 15 seconds;

[0170] Annealing and extension: 60°C for 1 min;

[0171] Number of cycles: 40-45 times;

[0172] Select the fluorescence detection channel according to the probe type;

[0173] S7-5, set the baseline and fluorescence threshold using the real-time fluorescence quantitative PCR instrument software;

[0174] S7-6: Add the prepared reaction system to the reaction wells of the real-time fluorescence quantitative PCR instrument, start the real-time fluorescence quantitative PCR instrument, run the preset temperature cycle program, and detect the fluorescence signal after each cycle;

[0175] S7-7, the real-time fluorescence quantitative PCR instrument calculates the Ct value of each sample through the fluorescence signal, and determines whether the target moldy microorganisms are present in the sample by comparing the Ct value and the fluorescence threshold. If the target moldy microorganisms are present, the number of target moldy microorganisms is calculated by the standard curve method to obtain the moldy microorganism detection data.

[0176] It should be noted that the moldy microbial DNA extraction kit is a tool used to extract high-quality genomic DNA from moldy microbial samples. It is widely used in the field of scientific research. It usually adopts the centrifugal adsorption column method or magnetic bead method, which can efficiently lyse moldy microbial cells and purify DNA. The extracted DNA can be directly used in downstream experiments such as PCR and sequencing.

[0177] A real-time fluorescence quantitative PCR instrument is a molecular biology instrument used to accurately detect and quantify DNA or RNA. By adding fluorescently labeled primers or probes to the PCR reaction system, the changes in the fluorescence signal are used to monitor the DNA amplification process in real time. As amplification proceeds, the fluorescence signal increases, and the instrument records the fluorescence intensity and generates an amplification curve to determine the initial concentration of the target nucleic acid. It has high sensitivity, high specificity, and high repeatability, can complete detection in a short time, and can detect multiple target genes simultaneously. Real-time fluorescence quantitative PCR instruments are widely used in medical diagnosis, biological research, food safety, and environmental monitoring.

[0178] Step S8: Compare the existing mold types with the mold microorganism detection data to obtain the mold identification result of the sample;

[0179] Wherein, in step S8, the following sub-steps are also included:

[0180] S8-1, comparing the possible types of mold with the mold microorganism detection data to determine whether the sample is moldy and obtain the mold status. If the sample is moldy, confirm the type of mold;

[0181] S8-2, evaluate the degree of mold based on the number of moldy microorganisms, colony size and distribution range in the moldy area. The degree of mold can be classified into mild, moderate and severe.

[0182] S8-3, determine the type of mildew based on the degree and type of mildew, which includes single-species mildew and multi-species mildew;

[0183] S8-4, recording the mold identification results in a report, wherein the mold identification results include mold status, mold type, mold degree, and mold type;

[0184] S8-5: Determine whether the mold identification result meets the mold risk level and record the determination result in the report.

[0185] It should be noted that mildew of Chinese medicinal materials refers to the phenomenon that during the storage process of Chinese medicinal materials, mold grows due to moisture, high temperature, poor ventilation, etc., causing the medicinal materials to deteriorate. Common mildew characteristics include the appearance of mold spots and hair-like substances on the surface of the medicinal materials, and a musty smell. Mildew will not only destroy the appearance and smell of Chinese medicinal materials, but also cause the degradation of their effective ingredients, reduce their efficacy, and even produce toxins, causing harm to human health.

[0186] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for identifying moldy Chinese medicinal materials that are both medicinal and edible, characterized in that: The method includes: Step S1, selecting samples from Chinese medicinal materials that are both medicinal and edible, and determining the mold risk level of the samples; Step S2, using a high-resolution camera to photograph the sample from multiple angles to obtain an appearance image; Step S3, using a near-infrared spectrometer to perform spectral scanning on the sample to obtain spectral information; Step S4, using an electronic nose device to detect volatile organic compounds released by the sample to obtain odor information; Step S5, extracting and fusing the collected appearance image, spectral information, and odor information to obtain a comprehensive feature vector; Step S6, using a feature selection algorithm to process the comprehensive feature vector and list the types of mold present; Step S7, detecting the types and quantities of moldy microorganisms in the sample by real-time fluorescence quantitative PCR technology to obtain moldy microorganism detection data; Step S8: Compare the existing mold types with the mold microorganism detection data to obtain a mold identification result of the sample.

2. The method for identifying mildewed Chinese medicinal materials of edible and medicinal properties according to claim 1, wherein: Wherein, step S1 further includes the following sub-steps: S1-1, using random sampling method, select Chinese herbal medicines from different batches, different origins, and different storage conditions; S1-2, record the collection date, collection batch, and origin of the Chinese medicinal materials; S1-3, record the storage conditions of Chinese medicinal materials, including temperature, humidity, and ventilation; S1-4, establish a risk scoring system: A time risk score is assigned based on the length of time between the collection date and the testing date, an origin risk score is assigned based on the climatic conditions of the origin, and a storage condition score is assigned based on the storage conditions; The time risk score, origin risk score, and storage condition risk score are added together to obtain a comprehensive risk score. Based on the comprehensive risk score, the mold risk of Chinese medicinal materials is divided into three levels: low, medium, and high, and the mold risk level of the Chinese medicinal materials is recorded; S1-5, select equal amounts of samples from Chinese medicinal materials of each mold risk level for mold testing.

3. The method for identifying mildewed Chinese medicinal materials of edible and medicinal properties according to claim 1, wherein: Wherein, in step S2, the following sub-steps are also included: S2-1, place the sample on a shooting platform with uniform light and a flat surface; S2-2, use a ring light device to supplement the light of the sample; S2-3, using a 12-megapixel high-definition camera, photograph the sample from multiple angles, including the front, side, back, and oblique surfaces of the sample, to obtain an appearance image of the sample.

4. The method for identifying mildewed Chinese medicinal materials of edible and medicinal properties according to claim 1, wherein: Wherein, in step S3, the following sub-steps are also included: S3-1, use a near-infrared spectrometer, set the near-infrared spectrometer wavelength range to 700-2500nm, and the resolution to 4cm -1 ; S3-2, place the sample on the detection platform of the near-infrared spectrometer and fix it using a sample cell or a glass slide; S3-3, starting a near-infrared spectrometer to detect the sample and obtain spectral information of the sample.

5. The method for identifying mildewed Chinese medicinal materials of edible and medicinal properties according to claim 1, wherein: Wherein, in step S4, the following sub-steps are also included: S4-1, using an electronic nose device based on a metal oxide semiconductor (MOS) sensor, setting the detection range of the electronic nose device to 0-100 ppm and a resolution of 0.1 ppm; S4-2, cut the sample into small pieces, fix them using a sample cell or glass slide, and place them in a sealed container for 30 minutes; S4-3, starting the electronic nose device to detect the sample and obtain odor information of the sample. The odor information is a sensor response curve formed by the change of the response value of the metal oxide semiconductor (MOS) sensor in the electronic nose device over time.

6. The method for identifying mildewed Chinese medicinal materials of edible and medicinal properties according to claim 1, wherein: Wherein, in step S5, the following sub-steps are also included: S5-1, extracting shape features of the appearance image to obtain a shape feature vector, wherein the shape features include the outline, area, and perimeter of the sample; calculating the average color value of the appearance image using the RGB channels to obtain a color feature vector; and extracting texture features of the appearance image using the gray-level co-occurrence matrix (GLCM) to obtain a texture feature vector; S5-2, analyzing the spectral information of the sample to obtain a spectral curve feature vector, wherein the spectral information includes the position and intensity characteristics of the spectral absorption peak, and the slope and curvature characteristics of the spectral curve; S5-3, extracting odor features from the odor information, wherein the odor features include the maximum value, minimum value, average value, and standard deviation of the sensor response curve, and performing dimensionality reduction processing on the odor features using principal component analysis (PCA) technology to obtain an odor feature vector; S5-4, concatenating the extracted shape feature vector, color feature vector, texture feature vector, spectrum feature vector, and odor feature vector to obtain a comprehensive feature vector, which is: Comprehensive feature vector = shape feature vector, color feature vector, texture feature vector, spectrum feature vector, smell feature vector.

7. The method for identifying mildewed Chinese medicinal materials of edible and medicinal properties according to claim 1, wherein: Wherein, in step S6, the following sub-steps are also included: S6-1, using the recursive feature elimination algorithm RFE to filter the comprehensive feature vector and obtain the feature subset with the highest discrimination for mold identification; S6-2, establishing a support vector machine (SVM) model, and training the SVM model using labeled samples, wherein the labeled samples include moldy samples and non-moldy samples; S6-3, inputting the feature subset into the SVM model, determining whether the sample is moldy, listing the types of mold that exist, and obtaining a list of mold types; S6-4: If the sample is moldy, the size and distribution of the colonies in the moldy area are obtained by analyzing the appearance image and spectral information.

8. The method for identifying mildewed Chinese medicinal materials of edible and medicinal properties according to claim 1, wherein: Wherein, in step S7, the following sub-steps are also included: S7-1, extracting moldy microbial DNA from the sample using a moldy microbial DNA extraction kit to obtain a moldy microbial DNA template; S7-2: Determine the target moldy microorganisms based on the mold species list, search for the specific gene sequence of the target moldy microorganisms, and design primers and probes based on the specific gene sequence; S7-3: Prepare the reaction mixture in a sterile PCR tube according to the following ratio: Moldy microbial DNA template: 2 μL; Primer: unit concentration is 10 μM, input volume is 1 μL each for upstream and downstream; Probe: unit concentration is 10 μM, input volume is 1 μL; 2×qPCR Mix solution: The input volume is 10 μL. The 2×qPCR Mix solution is a pre-prepared reaction mixture containing Taq enzyme, dNTPs, Mg 2+ ; Use sterile water to make up the reaction system to 20 μL; S7-4: Set the temperature cycle parameters of the real-time fluorescence quantitative PCR instrument according to the design of primers and probes: Initial denaturation: 95°C for 10 min; Cyclic denaturation: 95°C for 15 seconds; Annealing and extension: 60°C for 1 min; Number of cycles: 40-45 times; Select the fluorescence detection channel according to the probe type; S7-5, set the baseline and fluorescence threshold using the real-time fluorescence quantitative PCR instrument software; S7-6: Add the prepared reaction system to the reaction wells of the real-time fluorescence quantitative PCR instrument, start the real-time fluorescence quantitative PCR instrument, run the preset temperature cycle program, and detect the fluorescence signal after each cycle; S7-7, the real-time fluorescence quantitative PCR instrument calculates the Ct value of each sample through the fluorescence signal, and determines whether the target moldy microorganisms are present in the sample by comparing the Ct value and the fluorescence threshold. If the target moldy microorganisms are present, the number of target moldy microorganisms is calculated by the standard curve method to obtain the moldy microorganism detection data.

9. The method for identifying mildewed Chinese medicinal materials of edible and medicinal properties according to claim 1, wherein: Wherein, in step S8, the following sub-steps are also included: S8-1, comparing the type of mold with the mold microorganism detection data to determine whether the sample is moldy and obtain the mold status. If the sample is moldy, confirm the type of mold; S8-2, assessing the degree of mold based on the number of moldy microorganisms, colony size in the moldy area, and distribution range. The degree of mold can be classified as mild, moderate, or severe. S8-3, determining the type of mildew based on the degree and type of mildew, wherein the mildew type includes single-species mildew and multi-species mildew; S8-4, recording the mold identification result in a report, wherein the mold identification result includes the mold state, mold type, mold degree, and mold type; S8-5, determining whether the mold identification result meets the mold risk level described in step S1, and recording the determination result in a report.

Images