Method for detecting uniformity of tobacco sheets

Abstract

The invention discloses a method for detecting the uniformity of a tobacco sheet, which is characterized by comprising the following steps of: image acquisition: acquiring a gray image of a tobacco sheet sample; determining target points: determining a required gray scale range based on the gray scale difference of pixel points in the gray scale image, and finding out all points in the required gray scale range in the gray scale image as the target points; and calculation and judgment: obtaining the area of the target point, calculating the dispersion degree of the target point according to the area of the target point, and judging the uniformity of the tobacco sheet sample according to the dispersion degree. Through the method, the uniformity of the tobacco sheet can be quantitatively detected, and the uniformity of a certain component of the tobacco sheet can be specifically detected, so that the production process can be helped to be optimized, and the production quality of the tobacco sheet is further improved.

Description

Technical Field

[0001] The present invention relates to the field of detection of processing raw and auxiliary materials, and particularly to a method for detecting the uniformity of tobacco sheet. Background Art

[0002] Tobacco sheet, also known as reconstituted tobacco leaf, is made by mixing tobacco leaves, tobacco dust, tobacco stems, etc. with adhesives and additives (such as flavorings, atomizing agents, etc.) and manufacturing through methods such as papermaking method and roll pressing method. It is an important raw material for tobacco products. After being cut into shreds, tobacco sheet with a specific formula can even constitute all of the smoking material of heated tobacco products.

[0003] Since the raw materials of tobacco sheet are mixed with various different components, there are large differences between different components, and the stress conditions in different directions during the pressing process may be different. All these factors will affect the uniformity of tobacco sheet, and the quality of tobacco sheet and its sensory experience, etc. are directly affected by the uniformity. Therefore, during the production process, it is necessary to detect the uniformity of tobacco sheet to evaluate its quality. Further, through the evaluation of uniformity, problems that may exist in the process can also be traced back. However, at present, there is a lack of quantitative evaluation methods for the evaluation of uniformity in the production of tobacco sheet, which also restricts the optimization of the production process. In addition, in view of the process characteristics of tobacco sheet, how to detect the uniformity of different components and different directions is also an urgent problem to be solved. Summary of the Invention

[0004] The present invention provides a method for detecting the uniformity of tobacco sheet to solve the above problems.

[0005] An embodiment of the present invention discloses a method for detecting the uniformity of tobacco sheet, including the following steps:

[0006] Image acquisition: acquiring a grayscale image of a tobacco sheet sample;

[0007] Determining target points: based on the grayscale differences of pixel points in the grayscale image, determining a required grayscale range, and finding all points within the required grayscale range in the grayscale image as target points;

[0008] Calculation and judgment: obtaining the area of the target points, calculating the degree of dispersion of the target points according to the area of the target points, and judging the uniformity of the tobacco sheet sample according to the degree of dispersion.

[0009] By adopting the above technical solution, quantitative uniformity detection can be carried out for different tobacco sheets. It is not necessary to preprocess the tobacco sheets before detection. The tobacco sheets can be detected as soon as they are produced in the workshop, which can timely discover possible problems in the production process and conduct investigation and improvement, facilitating the detection and improvement of the sheet quality. Moreover, by adopting the above technical solution, the steps of processing the image are simple, and it is easy to obtain image parameters and perform calculations, and the uniformity result can be obtained without consuming a large amount of manpower and resources.

[0010] Optionally, the image acquisition includes: obtaining a color image of a tobacco sheet sample by means of photographing or scanning, and then converting the color image into the grayscale image.

[0011] Optionally, the grayscale image is an 8-bit grayscale image.

[0012] Optionally, each target point is composed of one or more continuous and uninterrupted pixel points, and the area of the target point is the area in the grayscale image, and is obtained by the following method: obtaining the area of each pixel point, adding up the areas of all pixel points that make up the target point, and obtaining the area of the target point.

[0013] Optionally, the target point is a spot in the grayscale image that is different from the tobacco sheet substrate.

[0014] Optionally, the spot is a white point or a black point relative to the tobacco sheet substrate.

[0015] Optionally, the substance in the target point corresponds to a known additive, and the selection of the required gray level range satisfies the following condition: the proportion of the number of all pixel points in the target points selected within the required gray level range in the total number of pixel points does not exceed the volume proportion of the known additive in the entire tobacco sheet.

[0016] Optionally, determine the gray level value of the known additive in the grayscale image, and set the required gray level range to be within the required gray level range of the gray level value.

[0017] Optionally, the proportion of the number of all pixel points in the target points selected within the required gray level range in the total number of pixel points does not exceed 5%.

[0018] Optionally, the proportion of the number of all pixel points in the target point in the total number of pixel points is 1%-2%

[0019] Optionally, the calculating and judging step includes: evenly dividing the grayscale image into a plurality of regional units along a first direction, calculating the areas of all target points in each regional unit, and calculating a first degree of dispersion of the target points in each regional unit according to the areas of all target points in each regional unit; evenly dividing the grayscale image into a plurality of regional units along a second direction, calculating the areas of all target points in each regional unit, and calculating a second degree of dispersion of the target points in each regional unit according to the areas of all target points in each regional unit; judging the uniformity of the tobacco sheet sample according to the first degree of dispersion and the second degree of dispersion, wherein the first direction and the second direction intersect.

[0020] Optionally, the first direction is perpendicular to the second direction.

[0021] Optionally, the degree of dispersion is characterized by the coefficient of variation of the target point areas in a plurality of regional units in the first direction and the second direction:

[0022]

[0023]

[0024] In formula (1), n represents the number of a plurality of regional units, n≥2, i = 1, 2, 3......n, S xi represents the area of the target points in the i-th regional unit in the first direction, represents the mean value of the target point areas in n regional units, CV xnS represents the coefficient of variation of the target point areas in n regional units in the first direction;

[0025] In formula (2), n represents the number of a plurality of regional units, n≥2, i = 1, 2, 3......n, S yi represents the area of the target points in the i-th regional unit in the second direction, represents the mean value of the target point areas in n regional units, CV ynS represents the coefficient of variation of the target point areas in n regional units in the second direction.

[0026] Optionally, the degree of dispersion is also characterized by the overall distribution coefficient of variation C of the target points in the first direction and the second direction ns characterized by:

[0027]

[0028] In the formula, n represents the number of a plurality of regional units, n≥2, CV xnS represents the coefficient of variation of the target point areas in n regional units in the first direction, CV ynSRepresents the coefficient of variation of the area of the target points in n regional units in the second direction.

[0029] Optionally, the detection method further includes performing statistical calculations on the selected target points to determine the centroid and obtain the centroid coordinates, where the centroid is determined by the following method: calculating the sum of the products of the distances between each pixel point in the image and the target points and the areas of the target points through the following formula to obtain G1, G2, G3... Gq, where q represents the number of pixel points, comparing the calculation results of G1, G2, G3... Gq, and determining the pixel point with the smallest calculation result as the centroid.

[0030]

[0031] In the formula, m represents the number of target points, d p Represents the distance between each target point and the pixel point, S p Represents the area of each target point, p = 1, 2, 3...... m.

[0032] Optionally, the detection method further includes calculating the dispersion of the target points, including the following steps:

[0033] Calculating the area S of each target point p The ratio A to the total area S of the target points p %;

[0034] Calculating the overall eccentricity distance d of the target points c :

[0035]

[0036] In the formula, m represents the total number of target points, d p Represents the distance between the target point and the centroid, p = 1, 2, 3...... m.

[0037] Optionally, the calculation of the dispersion of the target points further includes dimensionless processing of the overall eccentricity distance d c :

[0038]

[0039] In the formula, D represents the overall eccentricity, and a and b are respectively the length and width of the acquired image

[0040] Optionally, it further includes binarizing the grayscale image so that the target points present one of white or black, and the other parts outside the target points present the other of white or black. Description of the Drawings

[0041] Figures 1(a), 1(b), and 1(c) show the surface color images of Sample 1, Sample 2, and Sample 3;

[0042] Figures 2(a), 2(b), and 2(c) show the 8-bit grayscale images of the surfaces of Sample 1, Sample 2, and Sample 3;

[0043] Figures 3(a) and 3(b) show schematic diagrams of selecting target points in the grayscale image of Sample 1 according to the required grayscale range;

[0044] Figures 4(a) and 4(b) show the binarized images of the target points corresponding to the black spots and white spots of Sample 1 after screening;

[0045] Figures 4(c) and 4(d) show the binarized images of the target points corresponding to the black spots and white spots of Sample 2 after screening;

[0046] Figures 4(e) and 4(f) show the binarized images of the target points corresponding to the black spots and white spots of Sample 3 after screening;

[0047] Figures 5(a) and 5(b) show schematic diagrams of dividing the image into n regional units along the first direction and the second direction;

[0048] Figures 6(a), 6(b), and 6(c) show the color images of the surfaces of Sample 4, Sample 5, and Sample 6;

[0049] Figures 7(a), 7(b), and 7(c) show the 8-bit grayscale images of the surfaces of Sample 4, Sample 5, and Sample 6;

[0050] Figures 8(a) and 8(b) show schematic diagrams of selecting target points in the grayscale image of Sample 4 according to the required grayscale range;

[0051] Figures 9(a) and 9(b) show the binarized images of the target points corresponding to the black spots and white spots of Sample 4 after screening;

[0052] Figures 9(c) and 9(d) show the binarized images of the target points corresponding to the black spots and white spots of Sample 5 after screening;

[0053] Figures 9(e) and 9(f) show the binarized images of the target points corresponding to the black spots and white spots of Sample 6 after screening; Detailed implementation manners

[0054] The following specific embodiments illustrate the implementation manners of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Although the description of the present invention will be introduced in conjunction with preferred embodiments, this does not mean that the features of this invention are limited to this implementation manner. On the contrary, the purpose of introducing the invention in conjunction with the implementation manner is to cover other alternatives or modifications that may be extended based on the claims of the present invention. In order to provide a deep understanding of the present invention, many specific details will be included in the following description. The present invention can also be implemented without using these details. In addition, in order to avoid confusing or obscuring the key points of the present invention, some specific details will be omitted in the description. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0055] It should be noted that in this specification, similar reference numerals and letters indicate similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0056] The terms "first", "second", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0057] To make the purpose, technical solutions and advantages of the present invention clearer, the implementation manners of the invention will be further described in detail below with reference to the drawings.

[0058] In recent years, reconstituted tobacco has been increasingly widely used as a raw material for traditional cigarettes and new tobacco products. On the one hand, it can make the most of tobacco raw materials and save costs; on the other hand, it can also reduce the release of harmful substances. The preparation process of reconstituted tobacco has been relatively mature. Taking the commonly used roll pressing method as an example, it includes steps such as raw material dispersion, raw material mixing, and pressing into shape. Among them, the components for mixing include tobacco raw materials, various additives, adhesives, etc. Uniformity is an important aspect for evaluating the quality of reconstituted tobacco, which affects various performances of reconstituted tobacco, but there is yet no method capable of quantitatively analyzing the uniformity of reconstituted tobacco. In fact, due to the special production method of reconstituted tobacco, the surface of reconstituted tobacco mainly presents a yellowish-brown color. However, there are often black, white or other colored spots on it, and these spots are usually the raw materials mixed during production. For example, most of the tobacco powder in the raw materials is yellowish-brown, but there are also a small number of black particles, and the additives in the raw materials are often white powder. These components are dispersed throughout the reconstituted tobacco through the processing procedures of reconstituted tobacco and are also distributed on the surface of reconstituted tobacco. Moreover, since the raw materials of each group of reconstituted tobacco are limited and known, and the color differences between them are relatively large, there is a corresponding relationship between the spots on the surface of reconstituted tobacco and the raw materials.

[0059] Taking advantage of the above characteristics of the reconstituted tobacco, a method for detecting the uniformity of the reconstituted tobacco is specifically provided in this application, including the following steps:

[0060] Image acquisition: Obtain a grayscale image of the reconstituted tobacco sample;

[0061] Determine the target points: Based on the gray-scale differences of the pixel points in the grayscale image, determine the required gray-scale range, and find all the points within the required gray-scale range in the grayscale image as the target points;

[0062] Calculation and judgment: Obtain the area of the target points, calculate the degree of dispersion of the target points according to the area of the target points, and judge the uniformity of the reconstituted tobacco sample according to the degree of dispersion.

[0063] By using the spots of different colors existing on the surface of the reconstituted tobacco, one or several colors of the spots can be selected for the detection of uniformity, and thus the uniformity of the corresponding raw material components can be specifically studied. When observing the reconstituted tobacco with the naked eye, the spots on the thin sheet, whether white, black or other colors, have differences from the color of the thin sheet substrate and can be recognized. And this kind of difference is manifested as the difference in gray-scale values in the grayscale image. By using the gray-scale value difference, the gray-scale value range corresponding to one or several colors of spots can be specifically selected, that is, the required gray-scale value range, so as to screen out the corresponding points of such spots in the grayscale image, and the screened points are called "target points". Through the parameter calculation and dispersion degree analysis of the target points, the uniformity of the reconstituted tobacco can be quantitatively detected. Calculate multiple groups of reconstituted tobacco, compare the numerical values corresponding to the dispersion degree, the greater the dispersion degree, the worse the uniformity. The smaller the dispersion degree, the better the uniformity. And since it is possible to know which type of raw material components bring the spots of a specific color, the uniformity of a certain type of components can be specifically evaluated. For example, the calculation of the target points corresponding to the black spots can reflect the distribution uniformity of the tobacco powder, while the calculation results of the target points corresponding to the white spots reflect the distribution of the additives. When the uniformity of a certain type of components is known, when its uniformity is poor, corresponding adjustments and optimizations can be made for the problems that may exist in the steps such as dispersion and mixing of this type of components.

[0064] In a specific embodiment of the present invention, image acquisition further includes obtaining a color image of the surface of the reconstituted tobacco sample by taking a photo or scanning. When taking a photo, a mobile phone or a camera can be used; when scanning, a printer can be used. As long as a color image reflecting the surface topography of the reconstituted tobacco can be formed, there is no limitation on the specific acquisition method. When using a mobile phone or a camera to take a photo, there is no special requirement for the light source conditions, which can be natural light or under a fluorescent lamp. There is also no requirement on where to obtain the image on the surface of the reconstituted tobacco sample. In a specific embodiment, a relatively flat part of the sample surface can be selected. Before acquiring the image, no treatment is required for the reconstituted tobacco sample. This method is simple and immediate. Even without taking the reconstituted tobacco out of the workshop, only obtaining a photo is sufficient for detection.

[0065] Image acquisition further includes converting the color image into a grayscale image. The color of each pixel point in the color image is composed of components in three dimensions: R (red), G (green), and B (blue). If directly screening the color image, that is, performing numerical screening in the RGB three dimensions, it is difficult to determine what the corresponding values in each dimension are and perform screening of target points. Moreover, it is very difficult to implement screening of the numerical ranges in the three dimensions on general software. Converting the color image into a grayscale image can make there be only one-dimensional representation in the image, which is more convenient for processing and screening.

[0066] Furthermore, in a specific embodiment, the grayscale image is an 8-bit grayscale image. Compared with 2-bit and 4-bit images, 8-bit images have higher precision and 256 gradations. While 2-bit and 4-bit images only have 4 and 16 gradations respectively, with small distinguishability, which is not conducive to screening and statistics. Although 16-bit images have higher precision, for reconstituted tobacco samples, such high precision is not required, and the processing of 16-bit images also has higher requirements for computer configurations.

[0067] In a specific embodiment of the present invention, the method for screening target points is as follows: When selecting black spots on the tobacco sheet for uniformity detection, since the gray value corresponding to pure black is 0, the required gray range is determined starting from 0. Additionally, the raw material corresponding to the black spots is generally tobacco powder. In some cases, the tobacco powder obtained by crushing the same tobacco leaf may be both yellowish-brown and black. In other cases, the black tobacco powder belongs to another type of tobacco powder different from the main body in the yellowish-brown tobacco sheet. Since the addition amount of tobacco powder in the tobacco sheet is known, the proportion of the number of all pixel points in the target points screened within the required range in the total number of pixel points should not exceed the volume proportion of the known tobacco powder in the entire tobacco sheet. The required gray range is, for example, specifically 0 - 80, 0 - 65. In this way, other possible interference factors can be prevented from being excluded from the screened target points. Moreover, since the raw materials of the tobacco sheet are limited and all known, and the color differences between the components are relatively large, it is difficult for the screening of target points to be affected by other components. At the same time, the selection of the required gray range should also ensure that the proportion of the number of all pixel points in the target points in the total number of pixel points does not exceed 5%. If the proportion is too large, the bounding of the target points will not be representative and lose statistical value. A proportion of 5% can not only ensure the representativeness of the target point statistics but also not exceed the dosage of the corresponding raw material component. Preferably, the proportion of the number of all pixel points in the target points in the total number of pixel points is 1% - 2%. If the proportion is above 1%, it can prevent the number of target points from being too small to perform accurate statistics. If the proportion is below 2%, it can further distinguish the target points corresponding to a specific component from the points corresponding to other components, ensure the accuracy of screening the target points, and at the same time reflect the uniformity of the specific component.

[0068] Similarly, when selecting white spots on the tobacco sheet for uniformity detection, since the gray value corresponding to pure white is 255, the required gray range is determined starting from 255. Additionally, since the raw material corresponding to the white spots is additive powder and the addition amount of the additive powder is known, the proportion of the number of all pixel points in the target points selected within the required range in the total number of pixel points should not exceed the volume proportion of the known additive powder in the entire tobacco sheet. The required gray range is, for example, specifically 110 - 255, 155 - 255. This can prevent other possible interference factors from being excluded from the selected target points. Moreover, since the raw materials of the tobacco sheet are limited and all known, and the color differences between the components are relatively large, it is difficult for the selection of target points to be affected by other components. At the same time, the selection of the required gray range should also ensure that the proportion of the number of all pixel points in the target points in the total number of pixel points does not exceed 5%. A too large proportion will make the bounding of the target points unrepresentative and lose statistical value. A 5% proportion can both ensure the representativeness of the target point statistics and not exceed the dosage of the corresponding raw material component. Preferably, the proportion of the number of all pixel points in the target points in the total number of pixel points is 1% - 2%. A proportion above 1% can prevent the number of target points from being too small to enable accurate statistics, and a proportion below 2% can further distinguish the target points corresponding to a specific component from those corresponding to other components, ensuring the accuracy of screening the target points and reflecting the uniformity of the specific component.

[0069] In addition, if it is known that other raw material components can exhibit spots with distinct contrast relative to the substrate and the volume ratio of the incorporated raw material component is known, the required gray range can be determined based on the gray value corresponding to the color of the raw material component, and the proportion of the target points selected by the required gray range can be made to correspond to the volume ratio of the raw material component.

[0070] In a specific embodiment of the present invention, the target points are determined among the spots in the image of the tobacco sheet, and the spots are white or black points relative to the substrate of the tobacco sheet. Since the main body of the tobacco sheet is tobacco powder, which is yellowish-brown and appears gray in the gray-scale image, while the additives or components different from the main tobacco powder have different colors from the yellowish-brown color and will appear darker or lighter relative to the gray background main body in the gray-scale image. The darker color in the gray-scale image will be relatively black, and the lighter color will be white relative to the background. Therefore, selecting white or black points as target points to study their distribution uniformity reflects the distribution uniformity of a certain type of substance in the tobacco sheet. Specifically, the white or black points on the gray-scale image of the tobacco sheet are the white or black points that can be recognized from the image.

[0071] In a specific embodiment of the present invention, the parameter calculation for the target point is specifically the area of the target point. In some existing uniformity detection methods, uniformity is characterized by calculating the number of particles. However, specifically in tobacco sheet, since different components may adhere to each other, it is difficult to count the number of particles, and it is even more impossible to evaluate the uniformity of a specific component. Moreover, even the same component may have different degrees of adhesion, and component patches of different sizes may appear at different positions. Taking the same tobacco sheet sample as an example, there may be a certain component with the same number of particles within different area ranges on it, but this does not mean that this component is uniform in the tobacco sheet because the sizes of the particles are not necessarily the same. That is to say, the statistical calculation of the number is not applicable to tobacco sheet.

[0072] In addition, there are also some existing methods that characterize uniformity by the gray value and the degree of dispersion of the combined gray value. First of all, the statistical calculation of the gray value should be for each pixel point, rather than the combined result within a certain area, because adding the gray values or performing other similar calculations has no specific meaning. In addition, only by statistically calculating the gray value cannot reflect the distribution of a certain type of component, because the difference in gray value can only reflect the difference in color but cannot reflect the distribution difference of such color spots, let alone detect in the complex situation where there may be adhesion between components in tobacco sheet.

[0073] In a specific embodiment of the present invention, in the gray-scale image, each target point is composed of one or more continuous and non-interrupted pixel points, and the area of the target point is the area in the gray-scale image and is obtained by the following method: obtaining the area of each pixel point, summing up the areas of all pixel points that make up the target point, and obtaining the area of the target point. The area of each pixel point can be obtained from the image side length of each pixel point.

[0074] In a specific embodiment of the present invention, the area of the target point can be the area in the image, and the degree of dispersion is calculated using the area of the target point in the image. Additionally, it can also be the actual area in the tobacco sheet used to calculate the degree of dispersion. At this time, tools such as a ruler can be used to measure the side length of the tobacco sheet sample, and the actual side length is corresponded to the image side length to obtain the scale, so as to obtain the actual area size corresponding to each pixel point in the image. Since the image may only be a part of the tobacco sheet sample, by calculating the actual area size of the target point, it can be compared with the actual area of the tobacco sheet sample, which can reflect the actual distribution of the target point on the tobacco sheet sample, and it can also indicate that the sizes of a certain number of target points are within the visible range of the human eye and can be observed in the actual sample.

[0075] Further, the calculation and judgment steps further include: evenly dividing the grayscale image into multiple regional units along a first direction, calculating the areas of all target points in each regional unit, and calculating a first degree of dispersion of the target points in each regional unit according to the areas of all target points in each regional unit; evenly dividing the grayscale image into multiple regional units along a second direction, calculating the areas of all target points in each regional unit, and calculating a second degree of dispersion of the target points in each regional unit according to the areas of all target points in each regional unit; judging the uniformity of the tobacco sheet sample according to the first degree of dispersion and the second degree of dispersion, wherein the first direction and the second direction intersect.

[0076] In view of the characteristics of the tobacco sheet preparation process, after steps such as dispersion, mixing, and forming, the final product may exhibit anisotropy, that is, the uniformity may not be the same in different directions. Therefore, it may not be sufficient to only detect the uniformity of the tobacco sheet in one direction. Even if a certain component has good uniformity in one direction, it does not mean that it can also exhibit the same uniformity in other directions. Therefore, considering this characteristic of the tobacco sheet, statistics and calculations can be carried out separately from two directions to respectively confirm the uniformity conditions of the two directions, that is, the first direction x and the second direction y. It can be understood that the first direction x and the second direction y should at least intersect with each other, so as to have statistical value. Preferably, the first direction x and the second direction y are perpendicular to each other. For example, the first direction x is the length direction of the image, and the second direction y is the width direction of the image. In this way, the uniformity can be detected respectively in the two directions with the largest difference, and the uniformity of the tobacco sheet in the length and width directions can be determined correspondingly.

[0077] In addition, it further includes evenly dividing the image into multiple regional units along the first direction x and the second direction y respectively. For example, n regional units, it can be understood that n is a positive integer greater than or equal to 2. On this basis, calculate the total area of the target points in each regional unit, and calculate the average value of the target point areas in the n regional units. Subsequently, calculate the first degree of dispersion in the first direction x and the second degree of dispersion in the second direction y. The first degree of dispersion is the coefficient of variation CV of the target point areas in the n regional units in the first direction x xnS The second degree of dispersion is the coefficient of variation CV of the target point areas in the n regional units in the second direction y ynS The calculation method is:

[0078]

[0079]

[0080] In formula (1), n represents the number of multiple regional units, n≥2, i = 1, 2, 3......n, S xiDenotes the area of all target points within the \(i\)-th regional unit in the first direction. Denotes the mean value of the target point areas within \(n\) regional units, that is CV xnS Denotes the coefficient of variation of the target point areas within \(n\) regional units in the first direction;

[0081] In Equation (2), \(n\) represents the number of multiple regional units, \(n\geq2\), \(i = 1, 2, 3......n\), \(S\) yi Denotes the area of all target points within the \(i\)-th regional unit in the second direction, Denotes the mean value of the target point areas within \(n\) regional units, that is CV ynS Denotes the coefficient of variation of the target point areas within \(n\) regional units in the second direction.

[0082] Thus, according to Equation (1) and Equation (2), the first degree of dispersion in the first direction \(x\) and the second degree of dispersion in the second direction \(y\) can be obtained, and the uniformity in the first direction \(x\) and the second direction \(y\) can be judged respectively according to the first degree of dispersion and the second degree of dispersion. The coefficient of variation can also eliminate the influence of dimensions and can accurately reflect the degree of uniformity of substances in the tobacco sheet.

[0083] In another embodiment of the present invention, the degree of dispersion is also characterized by the overall distribution coefficient of variation \(C\) of the target points in the first direction \(x\) and the second direction \(y\) ns Characterize:

[0084]

[0085] In the formula, \(n\) represents the number of multiple regional units, \(n\geq2\), \(CV\) xnS Denotes the coefficient of variation of the target point areas within \(n\) regional units in the first direction \(x\), \(CV\) ynS Denotes the coefficient of variation of the target point areas within \(n\) regional units in the second direction \(y\). This overall distribution coefficient of variation of the target points can comprehensively evaluate the uniformity of the first direction \(x\) and the second direction \(y\). The larger this value is, the greater the difference in the target point areas between regions, and the worse the uniformity of the distribution of the target points in the tobacco sheet sample. In particular, due to the introduction of exponential calculation, when the coefficients of variation in the two directions differ greatly, compared with the case where the difference is small, the growth of the value is more obvious. Therefore, when the value of \(CV\) xnS is relatively large, it can also play a role in reminding to pay attention to the difference in uniformity between the two directions.

[0086] In another specific embodiment of the present invention, the detection method further includes performing statistical calculations on the selected target points to determine the center of gravity and obtain the center of gravity coordinates, where the center of gravity is determined in the following manner: Calculate the sum of the products of the distances between each pixel point in the image and the target points and the areas of the target points through the following formula to obtain G1, G2, G3... Gq, where q represents the number of pixel points, compare the calculation results of G1, G2, G3... Gq, and determine the pixel point with the smallest calculation result as the center of gravity.

[0087]

[0088] In the formula, m represents the number of target points, d p represents the distance between each target point and the pixel point, and S p represents the area of each target point, and p = 1, 2, 3...... m.

[0089] Specifically, the distance between the pixel point and the target point is calculated through coordinates, and the coordinates of the target point are represented as the center of the pixel points it occupies. The specific calculation is as follows: Add up the center coordinates of each pixel point in the target point and take the arithmetic average to obtain the coordinates of the target point.

[0090] On this basis, the dispersion of the target points can be calculated. The specific steps are as follows:

[0091] Calculate the area S p of each target point and the ratio A p % of the total area S of the target points;

[0092] Calculate the overall eccentricity d c :

[0093]

[0094] In the formula, m represents the total number of target points, d p represents the distance between the target point and the center of gravity, and p = 1, 2, 3...... m.

[0095] In fact, the overall eccentricity d c is also affected by the image size. For example, when it is necessary to compare and evaluate multiple groups of tobacco sheet samples, since the sizes of the acquired images may be different, it is necessary to eliminate the size influence to realize the comparative study on the dispersion between the components of different tobacco sheet samples. Therefore, it is also necessary to c dimensionalize the overall eccentricity d

[0096]

[0097] In the formula, D represents the overall eccentricity, and a and b are respectively the length and width of the acquired image.

[0098] The overall eccentricity reflects the degree of dispersion of the target points. If there is only one target point, it indicates that one or more pixel points are concentrated at this point, and the eccentricity is the smallest at this time. The smaller the value of D, the higher the concentration of the target points near the centroid and the smaller the degree of dispersion; conversely, it indicates that the target points are more dispersed compared to the centroid. When it is known that the uniformity of the target points of a certain sample is poor, calculating the overall eccentricity of the target points can assist in judging the reasons for the uniformity difference. For example, when the coefficient of variation of the tobacco sheet is large and the overall eccentricity is small, it indicates that the particles are relatively concentrated, and there may be problems in raw material dispersion or raw material mixing. The above a and b can be the length and width of the obtained color image of the tobacco sheet, or the length and width of the converted grayscale image.

[0099] In another specific embodiment of the present invention, it further includes performing a binarization process on the grayscale image to make the target points appear white and the other parts outside the target points appear black, making the observation clearer, more in line with human eye perception, and helpful for human-assisted judgment. In other embodiments, after the binarization process, the target points can also be made to appear black and the other parts outside the target points can be made to appear white.

[0100] The above technical solutions and technical effects of the invention will be further explained and illustrated below in combination with more specific embodiments.

[0101] Example 1:

[0102] The sample uses roll-pressed tobacco sheets, and its technological process is as follows: raw material dispersion, including the dispersion of tobacco powder and additives in their respective containers; raw material mixing, that is, various raw materials are put into one container and mixed evenly; pressing into shape. Three types of roll-pressed tobacco sheets are selected, namely Sample 1, Sample 2, and Sample 3, and the sample size is: length 12 - 18 cm, width 9 - 17 cm.

[0103] Use an HP Officejet Pro X567dw MFP PCL 6 printer to scan the three samples respectively; the scanned images are RGB images with a resolution of 600 PPI. The scanned images of Sample 1, Sample 2, and Sample 3 are as Figure 1(a) - Figure 1(c) shown.

[0104] Use ImageJ software to open the above RGB images and convert them into 8-bit grayscale images. The 8-bit grayscale images of Sample 1, Sample 2, and Sample 3 are as Figure 2(a) - Figure 2(c) shown.

[0105] The colors of various raw materials for producing tobacco sheets are different. For example, the main component, tobacco powder, is yellowish-brown in color, and there are also black particles in the tobacco powder, which can form black spots on the surface of the tobacco sheet. The appearance of the additive is a white powder, which can form white spots on the surface of the tobacco sheet.

[0106] Based on this, first, the black spots are selected as the basis for determining the required gray-scale range and screening the target points. Since the gray-scale value corresponding to black is 0, the required gray-scale range is determined starting from 0, and the selection that satisfies the required gray-scale range makes the proportion of all pixel points in the target points in the total number of pixel points 1%-2%. This proportion is less than the addition amount of tobacco powder, and it is beneficial to the subsequent statistical calculation of uniformity. Therefore, the required gray-scale range is determined to be 0-80, and the target points with gray-scale values of 0-80 are screened out. At this time, the target points within this gray-scale range that are framed appear red, as shown in Fig. 3(a) (taking sample 1 as an example). The gray-scale images of the three samples that have been screened for the target points corresponding to the black spots are binarized, as shown in Figs. 4(a), 4(c), and 4(e).

[0107] The white spots can also be selected as the basis for determining the required gray-scale range and screening the target points. Since the gray-scale value corresponding to white is 255, the required gray-scale range is determined starting from 255, and the selection that satisfies the required gray-scale range makes the proportion of all pixel points in the target points in the total number of pixel points 1%-2%. This proportion is less than the addition amount of the additive powder, and it is beneficial to the subsequent statistical calculation of uniformity. Therefore, the required gray-scale range is determined to be 110-255, and the target points with gray-scale values of 110-255 are screened out. At this time, the target points within this gray-scale range that are framed appear red, as shown in Fig. 3(b) (taking sample 1 as an example). The gray-scale images of the three samples that have been screened for the target points corresponding to the white spots are binarized, as shown in Figs. 4(b), 4(d), and 4(f).

[0108] Subsequently, as shown in Figs. 5(a) and 5(b), the length of the image is a and the width is b. The image is evenly divided into n regional units along the first direction x and the second direction y, such as S x1 、S x2 ……S xn ,S y1 、S y2 ……S yn shown. In this embodiment, n is taken as 10, that is, it is evenly divided into 10 regional units. Subsequently, the area and mean value of the target points in each regional unit in the two directions are obtained respectively, and the relevant calculations of the uniformity and dispersion of the foregoing target points are carried out to obtain the first dispersion degree CV x10S in the first direction x, the second dispersion degree CV y10S in the second direction y, and the overall distribution coefficient of variation C 10s of the target points in the first direction x and the second direction y. Further, the centroid of the screened target points is determined and the overall eccentricity D is calculated. The statistical results are shown in Table 1 below:

[0109] Table 1 Statistical results of tobacco sheet samples 1, 2, and 3

[0110]

[0111] C in the above table 10s As the judgment basis for the uniformity of the area distribution of the sample target points, the larger the value, the more uneven the distribution of the target points on the sample. Comparing the uniformity results of the black spot distributions, it can be seen that sample 3 is the best, sample 1 is the second best, and sample 2 is the worst. From the results of the overall eccentricity D, the dispersions of the black spots of the three are close, indicating that the uniformity difference is not caused by the excessive aggregation of the black spots. Therefore, the powder itself has good dispersion, and the uniformity difference may be affected by the mixing process.

[0112] The uniformity results of the white spot distributions from good to bad are: sample 2, sample 1, sample 3. Among them, the C of sample 3 10s value is as high as 0.933, significantly higher than that of samples 1 and 2. It can be seen that the white spot uniformity of sample 3 is poor, which may be affected by multiple processes such as raw material mixing and raw material dispersion. Combining the results of the eccentricity D, the white spot eccentricity of sample 3 is the smallest and the aggregation degree is the highest, and it may be greatly affected by the raw material dispersion process.

[0113] Further comparing the coefficient of variation of the target point area in the first direction x and the second direction y, it can be found that the CV x10S and CV y10S results corresponding to the white spots in samples 1 and 3 have large differences; reflected in the polarization images of FIGS. 4(b) and 4(f), it can be clearly seen that there are arrangements of white spots in the first direction x in samples 1 and 3, that is, the white lines in the first direction x. This phenomenon is particularly obvious in sample 3, and it also corresponds to the relatively larger differences in the CV x10S and CV y10S results. In addition, the differences in the CV x10S and CV y10s results may be due to uneven forces during the pressing process of the thin sheet, resulting in the white spots being arranged in a line along a certain direction.

[0114] Example 2:

[0115] The sample production process is the same as that of Example 1, which are samples 4, 5, and 6 respectively, and the sample size is 10 cm × 10 cm.

[0116] Under natural light conditions, use a mobile phone camera, parallel to the tobacco thin sheet, and take images at a position 20 cm away to obtain RGB images with a resolution of 72 PPI. The scanned images of samples 4, 5, and 6 are as Figure 6(a) - Figure 6(c) shown.

[0117] Use ImageJ software to open the above RGB images and convert them into 8-bit grayscale images. The 8-bit grayscale images of samples 4, 5, and 6 are as Figure 7(a) - Figure 7(c) shown..

[0118] First, select the black spots as the basis for determining the required gray-scale range and screening the target points. Since the gray-scale value corresponding to black is 0, start determining the required gray-scale range from 0. And the selection that satisfies the required gray-scale range makes the proportion of all pixel points in the target points in the total number of pixel points be 1%-2%. This proportion is less than the addition amount of tobacco powder, and it is beneficial to the subsequent statistical calculation of uniformity. Therefore, determine the required gray-scale range as 0-65, and screen out the target points with gray-scale values of 0-65. At this time, the target points within this gray-scale range that are framed appear red, as shown in Figure 8(a) (taking Sample 4 as an example). Binarize the gray-scale images of the three samples that have screened out the target points corresponding to the black spots, as shown in Figures 9(a), 9(c), and 9(e).

[0119] It is also possible to select the white spots as the basis for determining the required gray-scale range and screening the target points. Since the gray-scale value corresponding to white is 255, start determining the required gray-scale range from 255. And the selection that satisfies the required gray-scale range makes the proportion of all pixel points in the target points in the total number of pixel points be 1%-2%. This proportion is less than the addition amount of the additive powder, and it is beneficial to the subsequent statistical calculation of uniformity. Therefore, determine the required gray-scale range as 155-255, and screen out the target points with gray-scale values of 155-255. At this time, the target points within this gray-scale range that are framed appear red, as shown in Figure 8(b) (taking Sample 4 as an example). Binarize the gray-scale images of the three samples that have screened out the target points corresponding to the white spots, as shown in Figures 9(b), 9(d), and 9(f).

[0120] Subsequently, as shown in Figures 5(a) and 5(b), the length of the image is a and the width is b. Divide the image into n regional units along the first direction x and the second direction y respectively, such as S x1 、S x2 ……S xn ,S y1 、S y2 ……S yn shown. At this time, take n as 10, that is, divide it into 10 regional units respectively. Subsequently, obtain the area and mean value of the target points within each regional unit in the two directions respectively, and perform the relevant calculations of the uniformity and dispersion of the aforementioned target points to obtain the first degree of dispersion CV x10s in the first direction x, the second degree of dispersion CV y10S in the second direction y, and the overall distribution coefficient of variation C 10s of the target points in the first direction x and the second direction y. Further determine the centroid of the screened target points and calculate the overall eccentricity D. The statistical results are shown in Table 2 below:

[0121] Table 2 Statistical Results of Tobacco Sheet Samples 4, 5, and 6

[0122]

[0123] As can be seen from the above table, the order of the area uniformity of the black spots in Samples 4, 5, and 6 from good to poor is: Sample 4, Sample 5, Sample 6; the order of the dispersion degree of the black spots is: Sample 5, Sample 4, Sample 6. The order of the area uniformity of the white spots in the three samples from good to poor is: Sample 5, Sample 4, Sample 6; the order of the dispersion degree of the white spots is: Sample 6, Sample 4, Sample 5. From this, the reasons for the uniformity differences and the corresponding process optimizations can be further analyzed.

[0124] The above uniformity detection method of the present invention is based on the process characteristics and product characteristics of tobacco sheet, can be simply and quickly processed and calculated, quantitatively detect the uniformity of tobacco sheet, and moreover, can determine the uniformity characteristics of different components in different directions, and based on this, the relevant processes can be optimized targeted.

[0125] Although the present invention has been illustrated and described by referring to some preferred embodiments of the present invention, those of ordinary skill in the art should understand that the above content is a further detailed description of the present invention in combination with specific embodiments, and it cannot be determined that the specific implementation of the present invention is only limited to these descriptions. Those skilled in the art can make various changes in form and details, including making several simple deductions or substitutions, without departing from the spirit and scope of the present invention.

Claims

1. A method for detecting the uniformity of tobacco sheet, characterized in that, Including the following steps: Image acquisition: acquiring a grayscale image of a tobacco sheet sample; Determining target points: based on the grayscale differences of pixel points in the grayscale image, determining a required grayscale range, and finding all points within the required grayscale range in the grayscale image as target points; Calculation and judgment: obtaining the area of the target points, calculating the degree of dispersion of the target points according to the area of the target points, and judging the uniformity of the tobacco sheet sample according to the degree of dispersion.

2. The detection method for the uniformity of reconstituted tobacco according to claim 1, characterized in that The image acquisition includes: acquiring a color image of a tobacco sheet sample by means of photographing or scanning, and then converting the color image into the grayscale image.

3. The method for detecting the uniformity of reconstituted tobacco according to claim 1, characterized in that, The grayscale image is an 8-bit grayscale image.

4. The detection method for the uniformity of reconstituted tobacco according to claim 1, characterized in that, Each target point is composed of one or more continuous and uninterrupted pixel points. The area of the target point is the area in the grayscale image and is obtained by the following method: obtaining the area of each pixel point, summing up the areas of all pixel points that make up the target point, and obtaining the area of the target point.

5. The method for detecting the uniformity of reconstituted tobacco according to claim 1, wherein The target points are spots in the grayscale image that are different from the tobacco sheet substrate.

6. The detection method for the uniformity of reconstituted tobacco according to claim 5, characterized in that, The spots are white or black points relative to the tobacco sheet substrate.

7. The method for detecting the uniformity of tobacco sheet according to claim 5, characterized in that, The substances in the target points correspond to known additives. The selection of the required grayscale range satisfies the following condition: the proportion of the number of all pixel points in the target points selected within the required grayscale range in the total number of pixel points does not exceed the volume proportion of the known additives in the entire tobacco sheet.

8. The method for detecting the uniformity of reconstituted tobacco according to claim 7, wherein, Determining the grayscale value of the known additives in the grayscale image, and setting the required grayscale range to be within the grayscale value.

9. The method for detecting the uniformity of tobacco sheet according to claim 7, characterized in that, The proportion of the number of all pixel points in the target points selected within the required grayscale range in the total number of pixel points does not exceed 5%.

10. The method for detecting the uniformity of tobacco sheet according to claim 8, wherein The proportion of the number of all pixel points in the target points in the total number of pixel points is 1%-2%.

11. The detection method for the uniformity of tobacco sheet according to claim 1, characterized in that, The calculation and judgment step includes: evenly dividing the grayscale image into multiple regional units along a first direction, calculating the areas of all target points in each regional unit, and calculating the first degree of dispersion of the target points in each regional unit according to the areas of all target points in each regional unit; evenly dividing the grayscale image into multiple regional units along a second direction, calculating the areas of all target points in each regional unit, and calculating the second degree of dispersion of the target points in each regional unit according to the areas of all target points in each regional unit; judging the uniformity of the tobacco sheet sample according to the first degree of dispersion and the second degree of dispersion, wherein the first direction and the second direction intersect.

12. The method for detecting the uniformity of reconstituted tobacco according to claim 11, characterized in that, The first direction is perpendicular to the second direction.

13. The detection method for the uniformity of reconstituted tobacco according to claim 11, wherein, The degree of dispersion is characterized by the coefficient of variation of the areas of target points in multiple regional units in the first direction and the second direction: In formula (1), n represents the number of multiple regional units, n ≥ 2, i = 1, 2, 3......n, S xi represents the area of the target point within the ith regional unit in the first direction, represents the mean value of the areas of the target points within the n regional units, CV xnS represents the coefficient of variation of the areas of the target points within the n regional units in the first direction; In formula (2), n represents the number of multiple regional units, where n ≥ 2, and i = 1, 2, 3......n, S yi represents the area of the target point within the i-th regional unit in the second direction, represents the mean value of the areas of the target points within the n regional units, and CV ynS represents the coefficient of variation of the areas of the target points within the n regional units in the second direction.

14. The method for detecting the uniformity of tobacco sheet according to claim 13, characterized in that, The degree of dispersion is also characterized by the coefficient of variation C of the overall distribution of the target points in the first direction and the second direction ns characterized by: where n represents the number of multiple regional units, n≥2, CV xnS represents the coefficient of variation of the target point area in n regional units in the first direction, CV ynS represents the coefficient of variation of the target point area in n regional units in the second direction.

15. The method for detecting the uniformity of reconstituted tobacco according to claim 1, wherein, The detection method further includes statistically calculating the selected target points to determine the centroid and obtain the centroid coordinates, wherein the centroid is determined by the following method: calculating the sum of the products of the distances between each pixel point in the image and the target points and the target point area through the following formula to obtain G1, G2, G3... Gq, where q represents the number of pixel points, comparing the calculation results of G1, G2, G3... Gq, and determining the pixel point with the smallest calculation result as the centroid. Where m represents the number of target points, and d p represents the distance between each target point and the pixel point, and S p represents the area of each target point, where p = 1, 2, 3......m.

16. The method for detecting the uniformity of reconstituted tobacco according to claim 15, wherein, The detection method further includes the calculation of the dispersion of the target points, including the following steps: Calculate the area S of each of the target points p The ratio A to the total area S of the target points p %; Calculate the overall eccentricity distance d of the target point c : Wherein, m represents the total number of target points, and d p represents the distance between the target point and the center of gravity, and p = 1, 2, 3......m.

17. The method for detecting the uniformity of reconstituted tobacco according to claim 16, wherein, The calculation of the dispersion of the target points also includes the overall centrifugal distance d c Dimensionless quantity: In the formula, D represents the overall eccentricity, and a and b are respectively the length and width of the obtained image.

18. The method for detecting the uniformity of reconstituted tobacco according to claim 1, wherein It also includes binarizing the grayscale image so that the target points present one of white or black, and the other parts outside the target points present the other of white or black.

Images