Mahalanobis distance method-based product quality stability evaluation method

Abstract

The invention provides a product quality stability evaluation method based on mahalanobis distance, belongs to the technical field of near-infrared nondestructive testing, and solves the problems that in the prior art, cigarette quality stability evaluation takes physical indexes and sensory evaluation as evaluation means, and a stability evaluation method based on chemical indexes is less in research and low in evaluation efficiency. And objectivity, comprehensiveness and accuracy of evaluation are not facilitated. According to the method, a near-infrared spectrometer is used for scanning tobacco shred spectrums of cigarette products which are of the same brand, are produced in different places and are produced at different times, and a near-infrared big data system is used for predicting chemical components; similarity analysis is carried out on key chemical components of tobacco shreds through the Mahalanobis distance method, the batches with more abnormal times are screened out through Mahalanobis distance similarity analysis to judge that fluctuation exists, the method is high in recognition accuracy and sensitivity, low in cost and easy to operate, and technical support is provided for product quality stability analysis of production enterprises.

Description

Technical Field

[0001] The invention relates to the technical field of near-infrared nondestructive testing, and in particular to a product quality stability evaluation method based on a Mahalanobis distance method. Background Art

[0002] Product quality is the cornerstone of an enterprise's survival and development, especially in the tobacco industry, where the improvement of product quality is directly related to the enterprise's market competitiveness and consumer satisfaction. The core goal of product research and development is to improve product quality and performance to meet market demand and consumer expectations. To this end, enterprises must always put product quality first during product research and development, and promote the improvement of product quality through continuous technological innovation and process improvement.

[0003] In terms of the evaluation of cigarette quality stability, physical indicators and sensory evaluation are currently mainly used as evaluation methods. Physical indicators include the appearance quality, weight, length, etc. of cigarettes. These indicators can be accurately measured by instruments and equipment to provide an objective quantitative basis for product quality. Sensory evaluation is a subjective evaluation method. Through professional evaluation personnel, the aroma, taste, aftertaste, etc. of cigarettes are comprehensively evaluated, which can intuitively reflect the sensory quality of the product. However, there are relatively few studies on stability evaluation methods based on chemical indicators. Chemical indicators involve the content of chemical components in cigarettes, such as tar, nicotine, etc. These indicators are of great significance for assessing the health risks and product quality of cigarettes. This is not conducive to breakthroughs in product quality control and quality and efficiency improvement, and is therefore not conducive to the objectivity, comprehensiveness and accuracy of the evaluation.

[0004] In view of this, a product quality stability evaluation method based on the Mahalanobis distance method is needed. Summary of the invention

[0005] In view of the problem that the existing technology of cigarette quality stability evaluation uses physical indicators and sensory evaluation as evaluation methods, and there are few studies on stability evaluation methods based on chemical indicators, which is not conducive to the objectivity, comprehensiveness and accuracy of the evaluation, the present invention provides a product quality stability evaluation method based on the Mahalanobis distance method. The product quality stability evaluation method can be established based on the Mahalanobis distance method. According to the discreteness of the data itself, the data is standardized through the covariance matrix, the distance between the data point and a distribution center point is calculated, and the eight indicator parameters of the key chemical components of tobacco, such as total sugar, total nitrogen, and total alkaloids, are assigned by the Mahalanobis distance method. The similarity and difference results between batches are ranked to provide a scientific basis for the quality product stability analysis. The specific technical scheme is as follows:

[0006] A product quality stability evaluation method based on Mahalanobis distance method comprises the following steps:

[0007] Scanning tobacco of cigarette products by near infrared spectrometer;

[0008] Use tobacco near-infrared big data system to predict chemical composition;

[0009] Taking the near-infrared data indicators of each sample as the object, the similarity analysis of the key chemical components of tobacco was carried out by Mahalanobis distance method, and the similarity difference similarity matrix between batches was obtained;

[0010] The similarity and difference results between batches were sorted, and batches with more abnormalities were screened out to determine whether they had quality fluctuations, and verification was performed using absorption evaluation.

[0011] Preferably, the process of obtaining the similarity difference similarity matrix between batches is as follows:

[0012] Taking the eight core indicators of finished tobacco near-infrared data of each sample as the object, the Mahalanobis distance between all batches of finished tobacco was calculated;

[0013] By calculating the Mahalanobis distance between all batches in the product specification data, the similarity matrix D of all batches is constructed, and the similarity difference similarity matrix between batches is obtained as follows:

[0014]

[0015] Among them, d i,j =d j,i , both represent the Mahalanobis distance between the i-th batch and the j-th batch, n is the number of data points, and D is a symmetric matrix with all diagonal elements being 0.

[0016] Preferably, the calculation process of the Mahalanobis distance is as follows:

[0017] Calculate the mean vector u:

[0018]

[0019] Where n is the number of data points, x i is the ith data point.

[0020] Calculate the covariance matrix S:

[0021]

[0022] Calculate the inverse matrix S of the covariance matrix ―1 .

[0023] Calculate the Mahalanobis distance: For each data point x, calculate its Mahalanobis distance with the mean vector μ:

[0024]

[0025] Preferably, the key chemical component indicators of the model are total alkaloids, reducing sugars, total sugars, total nitrogen, K, chlorine, pH, and starch.

[0026] Preferably, a near infrared big data system is used, and the moisture content of the sample is modulated in a range of 6-8% before near infrared scanning.

[0027] Preferably, the step of crushing the sample into 60-80 mesh before scanning is included, and the sample is shredded tobacco, tobacco stems and / or tobacco dust.

[0028] A computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the product quality stability evaluation method based on the Mahalanobis distance method as described above.

[0029] A processor is used to run a program, wherein the program executes the product quality stability evaluation method based on the Mahalanobis distance method as described above when running.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention scans the tobacco of cigarette products by a near-infrared spectrometer, and uses a tobacco near-infrared big data system to predict the chemical composition. Then, taking the near-infrared data indicators of each sample as the object, the key chemical components of the tobacco are analyzed for similarity by the Mahalanobis distance method, and the similarity difference similarity matrix between batches is obtained. Finally, the similarity difference results between batches are sorted, and batches with a large number of abnormalities are screened out to determine that they have quality fluctuations, and verification is performed by smoking evaluation. In short, the present invention uses the Mahalanobis distance method to quantitatively analyze the near-infrared spectrum and establish a product quality stability analysis model. The method of the present invention is simple, fast, non-destructive to samples, and non-polluting to the environment. In addition, the method of the present invention also has the advantages of low cost and high efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0033] Figure 1 is a flow chart of the method of the present invention;

[0034] Figure 2 is the Mahalanobis distance similarity matrix between batches of brand “A”;

[0035] Figure 3 is the Mahalanobis distance similarity matrix between batches of “B” brand;

[0036] Figure 4 This is the original near-infrared scanning spectrum of cigarette products. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0039] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an" and "the" are intended to include plural forms unless the context clearly indicates otherwise.

[0040] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0041] Near infrared spectrum is an electromagnetic radiation wave between visible light (Vis) and mid-infrared (MIR), and is the first non-visible light region discovered in the absorption spectrum. In organic molecules, the atoms that make up chemical bonds or functional groups are in a state of constant vibration, and their vibration frequency is equivalent to the vibration frequency of infrared light. Near infrared spectrum is the frequency-doubled and combined absorption spectrum of organic molecules. It can not only obtain the structure, composition, and state information of molecules, but also obtain physical state information such as density, particle size, degree of polymerization of polymer compounds, and diameter of fibers from the near infrared reflection spectrum. It has the advantages of being fast, efficient, low cost, non-destructive to samples, and non-consumption of chemical reagents.

[0042] The present invention establishes a product quality stability evaluation method based on the Mahalanobis distance method, sorts the similarity difference results between batches, counts several batches with a large number of occurrences, judges their fluctuations according to the number of abnormalities, and realizes rapid evaluation of the quality stability of cigarette products.

[0043] The present invention grinds a series of tobacco shreds of the same brand but from different batches at different times into tobacco powder for near-infrared scanning by using a near-infrared spectrometer, and uses a tobacco near-infrared big data system to predict chemical components; and uses the Mahalanobis distance method to perform product similarity analysis on key chemical components of tobacco shreds.

[0044] The specific modeling steps are as follows:

[0045] (1) Taking the eight core indicators of the finished tobacco near-infrared data of each sample as the object, the Mahalanobis distance between all batches of finished tobacco is calculated:

[0046] 1) Calculate the mean vector u:

[0047]

[0048] Where n is the number of data points, x i is the ith data point.

[0049] 2) Calculate the covariance matrix S:

[0050]

[0051] 3) Calculate the inverse matrix S of the covariance matrix ―1 .

[0052] 4) Calculate the Mahalanobis distance: For each data point x, calculate its Mahalanobis distance with the mean vector μ:

[0053]

[0054] (2) By calculating the Mahalanobis distance between all batches in the product specification data, the similarity matrix D of all batches is constructed, and the similarity difference similarity matrix between batches is obtained as follows:

[0055]

[0056] Among them, d i,j =d j,i , both represent the Mahalanobis distance between the i-th batch and the j-th batch, so D is a symmetric matrix with all diagonal elements being 0.

[0057] (3) Sort the similarity difference results between batches: the greater the difference between batch pair data, the more significant the difference between the batch and other batches.

[0058] (4) Count the batches with the highest number of occurrences and make a preliminary judgment on whether these batches may have quality fluctuations compared with other batches, which can be further verified through sensory evaluation.

[0059] The specific implementation modes of the present invention are further described in detail below with reference to the accompanying drawings.

[0060] Example 1

[0061] Experimental instruments

[0062] MPA Fourier near-infrared spectrometer, ZM200 precision pulverizer, and oven produced by BRUKER (Germany).

[0063] Sample collection:

[0064] In order to make the established classification model have wide applicability, a total of 30 samples of "A" brand products of the same brand produced by Guangxi Zhongyan Nanning Factory in different time periods were selected for modeling.

[0065] Sample preparation:

[0066] The cigarette shreds were dried in an oven at 40°C to keep the moisture content of the samples basically consistent, and then fully crushed with a ZM200 precision grinder, passed through a 60-mesh sieve, and adjusted to a moisture content of 6-8%.

[0067] Spectral scanning and data processing:

[0068] The spectrum of tobacco leaf samples was scanned using an MPA Fourier near-infrared spectrometer (with a near-infrared quantitative analysis diffuse reflectance gold-plated large integrating sphere and sample rotator sampling accessories) produced by BRUKER (Germany), and the spectrum was processed using the qualitative analysis software QUANT7.0 in Bruker OPUS. The specific operation is as follows: put tobacco powder into the sample cup, the height in the cup is about 3 cm, press the weight on the sample for 10 seconds and then take it out, wipe the quartz glass at the bottom of the cup with gauze, and then place the sample cup on a rotating platform for NIR scanning. The operating parameters are: spectral scanning range 12 000~4 000cm -1 , spectral resolution 8cm -1 , scanning times 64 times (about 30S). Spectral data were collected in a transmission mode and processed into the first-order differential of the absorption spectrum. The original scan of the cigarette tobacco is shown in Figure 4 In the modeling process, in order to eliminate the influence of noise and baseline, the standard normal variable transformation method was used to eliminate the differences caused by the particle size, surface scattering and optical path change of solid samples, the spectrum was smoothed and filtered by wavelet transform, and the second-order derivative was used for preprocessing. After the sample was scanned, the spectral data was processed using statistical software.

[0069] Using the "Near Infrared Big Data Application System" to predict the chemical composition of tobacco:

[0070] The predicted results of the main chemical compositions of 27 batches of "A" brand products produced by Guangxi China Tobacco Co., Ltd. in 2023 are shown in Table 1:

[0071] Table 1 Prediction results of “Near Infrared Big Data Application System” (excerpt of main components)

[0072]

[0073]

[0074] Establishment of Mahalanobis distance model

[0075] The steps to build the model are as follows:

[0076] (1) Taking the eight core indicators of the finished tobacco near-infrared data of each sample as the object, the Mahalanobis distance between all batches of finished tobacco is calculated:

[0077] Calculate the mean vector u:

[0078]

[0079] Where n is the number of data points, x i is the ith data point.

[0080] Calculate the covariance matrix S:

[0081]

[0082] Calculate the inverse matrix S of the covariance matrix ―1 .

[0083] Calculate the Mahalanobis distance: For each data point x, calculate its Mahalanobis distance with the mean vector μ:

[0084]

[0085] By calculating the Mahalanobis distance between all batches in the product specification data, the similarity matrix D of all batches is constructed, and the similarity difference similarity matrix between batches is obtained as follows:

[0086]

[0087] Among them, d i,j =d j,i , both represent the Mahalanobis distance between the i-th batch and the j-th batch, so D is a symmetric matrix with all diagonal elements being 0.

[0088] The similarity difference results between batches are sorted, and the first 27 pairs (the number of pairs can be set manually) with the largest distance difference are selected. The larger the distance difference between batch pairs, the more significant the difference between the batch and other batches.

[0089] Among the first 27 pairs of batch pairs with the farthest distances, abnormal batches are analyzed and identified, and several batches with the largest number of occurrences are counted (the specific selected batches can be set according to the accuracy requirements of stability control). It is believed that these batches may have quality deviations compared with other batches.

[0090] According to the data in Table 1, the Mahalanobis distance method is used to perform similarity analysis on the near-infrared data. Figure 2 .

[0091] The top 25 pairs of data with the longest distance for brand “A”:

[0092] Batch_6 and Batch_8: 6.32

[0093] Batch_1 and Batch_7: 6.18

[0094] Batch_8 and Batch_20: 5.55

[0095] Batch_1 and Batch_19: 5.54

[0096] Batch_1 and Batch_6: 5.42

[0097] Batch_7 and Batch_16: 5.40

[0098] Batch_6 and Batch_10: 5.38

[0099] Batch_1 and Batch_17: 5.37

[0100] Batch_20 and Batch_22: 5.35

[0101] Batch_1 and Batch_10: 5.33

[0102] Batch_1 and Batch_2: 5.29

[0103] Batch_4 and Batch_8: 5.27

[0104] Batch_1 and Batch_20: 5.26

[0105] Batch_4 and Batch_26: 5.26

[0106] Batch_3 and Batch_20: 5.21

[0107] Batch_2 and Batch_9: 5.20

[0108] Batch_11 and Batch_20: 5.19

[0109] Batch_4 and Batch_15: 5.17

[0110] Batch_3 and Batch_19: 5.16

[0111] Batch_6 and Batch_13: 5.13

[0112] Batch_1 and Batch_8: 5.09

[0113] Batch_11 and Batch_26: 5.09

[0114] Batch_4 and Batch_6: 5.09

[0115] Batch_2 and Batch_10: 5.09

[0116] Batch_6 and Batch_17: 5.02

[0117] Batch_1 and Batch_3: 5.01

[0118] Batch_22 and Batch_27: 5.00

[0119] Batch_1 and Batch_12: 4.98

[0120] Batch_1 and Batch_15: 4.96

[0121] Batch_2 and Batch_20: 4.96

[0122] The top 10 abnormal batches with the largest number of "A" brands are:

[0123] Batch_1: 7 times

[0124] Batch_4: 3 times

[0125] Batch_6: 2 times

[0126] Batch_8: 1 time

[0127] Batch_7: 1 time

[0128] Batch_20: 1 time

[0129] Batch_3: 1 time

[0130] Batch_2: 1 time

[0131] Batch_11: 1 time

[0132] From the data analysis, we can see that batch 1 of "A" brand cigarettes appeared more often, 7 times, and batch 4 appeared 3 times, indicating that there are certain quality fluctuations in these two batches of products and they need to be paid special attention.

[0133] Example 2

[0134] In this embodiment, the model is applied to the homogenization analysis of the same brand of cigarettes "B" produced in different places of origin.

[0135] 1. Experimental instruments

[0136] MPA Fourier near-infrared spectrometer, ZM200 precision pulverizer, and oven produced by BRUKER (Germany).

[0137] 2. Sample collection

[0138] In order to make the established classification model have wide applicability, a total of 64 samples of "B" brand products of the same brand produced by Guangxi Tobacco Nanning and Liuzhou factories in different time periods and different production lines were selected for modeling.

[0139] 3. Sample preparation

[0140] The cigarette shreds were dried in an oven at 40°C to keep the moisture content of the samples basically consistent, and then fully crushed with a ZM200 precision grinder, passed through a 60-mesh sieve, and adjusted to a moisture content of 6-8%.

[0141] 4. Spectral scanning and data processing

[0142] Spectral scanning and data processing are the same as in Example 1.

[0143] 5. Using the “Near Infrared Big Data Application System” to predict the chemical composition of tobacco

[0144] The predicted results of the main chemical composition of 29 batches of "B" brand products are shown in Table 2

[0145] Table 2 Prediction results of “Tobacco Near Infrared Big Data Application System” (excerpt of main components)

[0146]

[0147]

[0148] 6. Establishment of Mahalanobis distance analysis model

[0149] The steps of establishing the model are the same as those in Example 1 above.

[0150] According to the data in Table 2, the Mahalanobis distance method is used to perform similarity analysis on the near-infrared data. Figure 3 The 29 pairs of data with the longest distance for the “B” brand are:

[0151] Batch_11 and Batch_22: 6.49

[0152] Batch_15 and Batch_22: 6.31

[0153] Batch_1 and Batch_13: 6.27

[0154] Batch_11 and Batch_27:6.19

[0155] Batch_1 and Batch_11: 6.07

[0156] Batch_11 and Batch_26: 6.01

[0157] Batch_15 and Batch_20: 5.95

[0158] Batch_11 and Batch_13: 5.94

[0159] Batch_15 and Batch_23: 5.87

[0160] Batch_11 and Batch_17: 5.84

[0161] Batch_11 and Batch_29: 5.82

[0162] Batch_11 and Batch_15: 5.74

[0163] Batch_13 and Batch_25: 5.60

[0164] Batch_5 and Batch_11: 5.59

[0165] Batch_2 and Batch_11: 5.53

[0166] Batch_4 and Batch_11: 5.50

[0167] Batch_11 and Batch_20: 5.48

[0168] Batch_17 and Batch_22: 5.45

[0169] Batch_15 and Batch_18: 5.44

[0170] Batch_13 and Batch_20: 5.42

[0171] Batch_13 and Batch_29: 5.33

[0172] Batch_22 and Batch_27: 5.30

[0173] Batch_11 and Batch_21: 5.28

[0174] Batch_13 and Batch_16: 5.23

[0175] Batch_11 and Batch_18: 5.22

[0176] Batch_13 and Batch_22: 5.21

[0177] Batch_13 and Batch_23: 5.18

[0178] Batch_5 and Batch_29: 5.18

[0179] Batch_19 and Batch_23: 5.17

[0180] The top 10 abnormal batches with the largest number of "A" brands are:

[0181] Batch_11: 8 times

[0182] Batch_15: 4 times

[0183] Batch_13: 3 times

[0184] Batch_1: 2 times

[0185] Batch_5: 1 time

[0186] Batch_2: 1 time

[0187] Batch_4: 1 time

[0188] Batch_17: 1 time

[0189] From the data analysis, we can see that batch 11 of "B" brand cigarettes appeared the most times, 8 times, while batch 15 and batch 13 appeared 4 times and 3 times respectively, indicating that there are certain quality fluctuations in these three batches of products and they need to be paid special attention.

[0190] 8. Model Validation

[0191] In order to better verify the recognition ability of the model, this experiment adopts the method of evaluation and verification, and conducts secret evaluation on the samples that are identified to have fluctuations by professional evaluation experts. The results are shown in Table 3:

[0192] Table 3 Evaluation and verification recognition results

[0193]

[0194]

[0195] The results showed that the 9 samples with quality fluctuations identified by Mahalanobis distance similarity analysis of different batches of samples were all verified to have sensory fluctuations through smoking evaluation, with an identification rate of 100%, indicating that the prediction accuracy of the established model is high and can be used for quality stability analysis of cigarette products.

[0196] It can be seen from the above embodiments that the present invention utilizes a product quality stability evaluation model based on the Mahalanobis distance method, which has a high recognition rate for cigarette quality fluctuations. It can be seen that the application of near-infrared analysis technology can be well used for cigarette quality stability analysis, which is an effective and feasible method with strong practicality and realistic significance.

[0197] Those of ordinary skill in the art will appreciate that the units of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0198] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0199] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0200] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-0nlyMemory), random access memory (RAM, RandomAccessMemory), mobile hard disk, magnetic disk or optical disk, etc., which can store program code.

[0201] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention 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 replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A product quality stability evaluation method based on Mahalanobis distance method, characterized in that: The following steps are involved: Scanning tobacco of cigarette products by near infrared spectrometer; Use tobacco near-infrared big data system to predict chemical composition; Taking the near-infrared data indicators of each sample as the object, the similarity analysis of the key chemical components of tobacco was carried out by Mahalanobis distance method, and the similarity difference similarity matrix between batches was obtained; The similarity and difference results between batches were sorted, and batches with more abnormalities were screened out to determine whether they had quality fluctuations, and verification was performed using absorption evaluation.

2. The method for evaluating product quality stability based on Mahalanobis distance method according to claim 1, characterized in that: The process of deriving the similarity difference similarity matrix between batches is as follows: Taking the eight core indicators of finished tobacco near-infrared data of each sample as the object, the Mahalanobis distance between all batches of finished tobacco was calculated; By calculating the Mahalanobis distance between all batches in the product specification data, the similarity matrix D of all batches is constructed, and the similarity difference similarity matrix between batches is obtained as follows: Among them, d i,j =d j,i , both represent the Mahalanobis distance between the i-th batch and the j-th batch, n is the number of data points, and D is a symmetric matrix with all diagonal elements being 0.

3. The method for evaluating product quality stability based on Mahalanobis distance method according to claim 2, characterized in that: The calculation process of Mahalanobis distance is as follows: Calculate the mean vector u: Where n is the number of data points, x i is the ith data point. Calculate the covariance matrix S: Calculate the inverse matrix S of the covariance matrix ―1 ; Calculate the Mahalanobis distance: For each data point x, calculate its Mahalanobis distance with the mean vector μ:

4. The method for evaluating product quality stability based on Mahalanobis distance method according to claim 1, characterized in that: The key chemical composition indicators of the model are total alkaloids, reducing sugar, total sugar, total nitrogen, K, chlorine, pH, and starch.

5. The method for evaluating product quality stability based on Mahalanobis distance method according to claim 1, characterized in that: Including the use of near-infrared big data system, the moisture content modulation range of the sample before near-infrared scanning is 6-8%.

6. The method for evaluating product quality stability based on Mahalanobis distance method according to claim 1, characterized in that: The method comprises the step of crushing the sample into 60-80 meshes before scanning, and the sample is shredded tobacco, tobacco stems and / or tobacco dust.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the product quality stability evaluation method based on the Mahalanobis distance method as described in any one of claims 1 to 6.

8. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the product quality stability evaluation method based on the Mahalanobis distance method as described in any one of claims 1 to 6.

Images