Anti-counterfeiting identification method based on random fine texture pattern

Abstract

The invention discloses an anti-counterfeiting identification method based on a random fine texture pattern. The anti-counterfeiting identification method comprises the following steps: S1) generating an original identification image of a coding identification or an image-text identification; s2) generating an image of a random fine texture pattern for ink mark diffusion; s3) combining the random fine texture image with the original identification image to form an identification image; s4) printing an identification image containing a random fine texture pattern on the printing stock; s5) carrying out image acquisition on the fine texture ink mark pattern after the ink mark diffusion effect in the printing process, and binding the fine texture ink mark pattern with code identification or image-text identification information to serve as an anti-counterfeiting filing image; and S6) comparing the similarity between the image of the fine texture ink mark pattern on the printing stock and the anti-counterfeiting filing image to realize authenticity discrimination. According to the anti-counterfeiting identification method, the printing process is integrally formed, the printing process is simple, and the identification accuracy is high; the method is suitable for printing equipment with various resolutions in the market.

Description

Technical Field

[0001] The invention relates to an anti-counterfeiting identification method, in particular to an anti-counterfeiting identification method based on random fine texture patterns. Background Art

[0002] Anti-counterfeiting technology is a technical means used to identify authenticity and prevent counterfeiting, alteration, and cloning. There are many types of anti-counterfeiting technologies, such as embossed anti-counterfeiting technology, micro-anti-counterfeiting technology, paper anti-counterfeiting, watermark paper, anti-counterfeiting additives, laser anti-counterfeiting, barcode anti-counterfeiting technology, and electronic code anti-counterfeiting labels and telephone identification systems. However, these anti-counterfeiting technologies are generally expensive to use and have the risk of being easily counterfeited and copied.

[0003] The common QR codes in daily life do not have anti-counterfeiting functions. Currently, common anti-counterfeiting QR codes usually need the assistance of APP applications and corporate product information databases to achieve a certain anti-counterfeiting effect. The principle is to assign a unique QR code to each product and use difficult-to-copy materials and printing technology to ensure that the QR code is difficult to copy. When purchasing goods, consumers can use mobile phone APP to query the QR code and check the product information, manufacturer, etc. to distinguish the authenticity.

[0004] In order to increase the difficulty of copying QR codes, companies usually need to introduce advanced printing anti-counterfeiting technologies, such as engraving, gravure printing, rainbow printing, laser anti-counterfeiting and invisible images, etc., which often means the need to purchase expensive printing equipment and raw materials, and introduce complex printing processes and techniques, resulting in companies greatly increasing anti-counterfeiting costs while improving their anti-counterfeiting capabilities.

[0005] When printing images, the ink diffusion marks on paper are random curved micro-curves, which are random, unique and unreplicable. If the size of the fine texture in the image to be printed is close to or less than the highest printing precision of the printing equipment, the ink diffusion process becomes more irregular and unpredictable. Therefore, if this principle can be used to design a fine texture pattern that is more likely to cause random ink diffusion and has anti-counterfeiting recognition capabilities, it will be a good solution to the problem that printed patterns are easily counterfeited.

[0006] However, when the diffusion effect of ink in fine texture patterns during the printing process is actually used for anti-counterfeiting applications, there may be the following disadvantages: 1. Fine texture patterns and QR codes usually need to be printed on different materials using different printing parameters. If the fine texture pattern can be directly combined with the QR code so that it can be printed in one piece, the printing process can be greatly simplified and the printing difficulty and cost can be reduced; 2. The clarity of the texture pattern after ink diffusion may be affected by the printing resolution of the printing equipment, and the printing equipment has high requirements. If the size of the fine texture can be adjusted according to the printing resolution, the difficulty of promotion can be reduced; 3. The positioning accuracy of the fine texture pattern is directly related to the reliability of the recognition result. If the positioning accuracy of the fine texture pattern can be improved and the positioning error can be reduced, and if a certain offset tolerance can be introduced into the positioning result, the accuracy of the recognition result can be effectively improved. Summary of the invention

[0007] The technical problem to be solved by the present invention is to provide an anti-counterfeiting identification method based on random fine texture patterns. The printing process is integrated, the printing process is simple, and it has high identification accuracy. It is suitable for printing equipment of various resolutions on the market and is easy to implement and promote.

[0008] The technical solution adopted by the present invention to solve the above-mentioned technical problems is to provide an anti-counterfeiting identification method based on random fine texture patterns, comprising the following steps: S1) generating an original identification image of a coded identification or a graphic identification; S2) generating an image of a random fine texture pattern for ink diffusion; S3) combining the random fine texture image with the original identification image by means of, for example, adding, overlaying or nesting, to form an identification image containing a random fine texture pattern for printing; S4) printing the identification image containing the random fine texture pattern on a substrate; S5) capturing an image of the fine texture ink pattern after ink diffusion during the printing process, and binding it with the coded identification or graphic identification information as an anti-counterfeiting filing image; S6) achieving authenticity discrimination by comparing the similarity between the image of the fine texture ink pattern on the substrate and the anti-counterfeiting filing image.

[0009] Furthermore, the step S2 generates a random fine texture pattern for ink diffusion in the following manner: S21) randomly generating an original image with randomly distributed black and white pixels; S22) using a plurality of different preset mask images to perform XOR processing on the original image; S23) performing unfavorable factor penalty statistics on all mask-processed images, and taking the mask-processed image with the least total penalty as the best result image; S24) adding a full enclosing box or a half enclosing corner box for pattern positioning to the inner edge of the best result image to obtain a random fine texture image for ink diffusion.

[0010] Further, the step S22 includes 9 preset mask images, whose sizes are all set to be consistent with the original image size, and the pixel distribution characteristics of the 9 mask images are as follows: 1) all pixels are black; 2) pixels whose sum of row and column indexes is a multiple of 2 are black, and the others are white; 3) pixels whose row index is a multiple of 2 are black, and the others are white; 4) pixels whose column index is a multiple of 3 are black, and the others are white; 5) pixels whose sum of row and column indexes is a multiple of 3 are black, and the others are white; 6) pixels whose sum of the row index divided by 2 and the column index divided by 3 is a multiple of 2 are black, and the others are white;

[0011] 7) Pixels whose row-column index product is a multiple of 6 are black, and the others are white; 8) Pixels whose row-column index product is a multiple of 2 are black, and the others are white; 9) Pixels whose row-column index product is a multiple of 2 are black, and the others are white.

[0012] Furthermore, the unfavorable factor penalty standards for all masked images in step S23 are as follows: 1) horizontal penalty, whenever n horizontally continuous pixels of the same color appear in the image, and n>4, n-2 points are counted; 2) vertical penalty, whenever n vertically continuous pixels appear in the image, and n>4, n-2 points are counted; 3) 2*2 grid penalty, whenever 4 pixels in a 2*2 pixel grid in the image are of the same color, 3 points are counted; 4) color uneven penalty, the absolute value of the difference between the proportion of black blocks in the image and 50% is calculated, and 10 points are counted for every 5%.

[0013] Furthermore, the method of adding a full enclosing frame or a half enclosing corner frame for pattern positioning to the inner edge of the best result image in step S24 is as follows: 1) Full enclosing frame, setting the pixels within the outermost 5% image width range around the image to black; 2) Half enclosing corner frame, first setting the 8 rectangular areas near the four corners of the image to black, with their long edges close to the edge, the length of which is 25% of the image width, and the width of which is 5% of the image width, and then setting the pixels within a certain number of layers outward from the 4 black angled areas composed of these 8 rectangular areas to white to highlight the 4 black angled areas.

[0014] Furthermore, the random fine texture image in step S2 is a square, the coding identifier in step S3 is a QR code, and after scaling the original QR code image to a 3*3 area size that is consistent with the size of the random fine texture image, the random fine texture image is used to replace the 3*3 black square area inside one or more locators in the original QR code image to form a QR code image containing a random fine texture pattern for printing.

[0015] Furthermore, in step S3, the random fine texture image is combined with the original logo image by adding, overlaying or nesting; in step S5, the image of the fine texture ink pattern after the ink diffusion in the printing process is captured, and the image of the area where the coding logo or graphic logo is located is captured at the same time to obtain its coding information or graphic information; the anti-counterfeiting registration image is bound to the coding information or graphic information and retained to facilitate the search for the corresponding anti-counterfeiting registration image during subsequent verification.

[0016] Furthermore, in step S6, the process of authenticity identification of the subtle texture ink pattern on the substrate is as follows: S61) using a mobile phone to take an image of the area where the coded mark or graphic mark containing the random subtle texture pattern is located on the substrate; S62) identifying the coded information or graphic information in the image uploaded by the mobile phone, and searching for the corresponding anti-counterfeiting registration image based on the information; S63) accurately locating the subtle texture ink pattern area in the image uploaded by the mobile phone, and obtaining the subtle texture ink pattern image after cropping and preprocessing; S64) fine-tuning the subtle texture ink pattern image in three dimensions of translation, rotation and scaling to obtain the processing result image under various fine-tuning parameters; S65) using multiple similarity evaluation algorithms to evaluate the similarity between all processing result images and the anti-counterfeiting registration image, and retaining the evaluation value with the highest similarity for each algorithm, and finally comprehensively judging the authenticity of the subtle texture ink pattern in the uploaded image based on the similarity evaluation values ​​of various algorithms.

[0017] Furthermore, in step S64, the offset tolerance range and offset step size of the three dimensions of translation, rotation and scaling are set respectively, and the offset is stepped with the set step size within the offset tolerance range of the three dimensions respectively. After processing the same initial fine texture ink pattern image, the processing result image under all step parameters is obtained.

[0018] Furthermore, the step S6 performs i, j, and k steps respectively within the offset tolerance range in the three dimensions, and obtains a total of i*j*k result images; the step S65 uses a cosine similarity evaluation algorithm, a mean square error similarity evaluation algorithm, a peak signal-to-noise ratio similarity evaluation algorithm, a gradient amplitude similarity deviation evaluation algorithm or a structural similarity evaluation algorithm to perform similarity evaluation on the i*j*k result images and the anti-counterfeiting filing image, and uses the evaluation results as a basis for authenticity judgment.

[0019] Compared with the prior art, the present invention has the following beneficial effects: the anti-counterfeiting identification method based on random fine texture patterns provided by the present invention embeds the random fine texture patterns used for anti-counterfeiting verification into the coding mark or graphic mark image used for coding marking before printing, so that only one printing is required and the printing process is integrated, which greatly simplifies the printing process and reduces the printing difficulty; the present invention utilizes the principle that ink will randomly diffuse irregularly in the fine texture pattern during the printing process to achieve resistance to counterfeiting and imitation printing, and the principle is simple and easy to promote; the present invention can freely adjust the texture size of the random fine texture pattern according to the printing resolution of the printing equipment, is suitable for printing equipment of various resolutions on the market, and is easy to implement and promote; after accurately positioning the position of the fine texture pattern, the present invention further improves the positioning accuracy by introducing an offset tolerance in three dimensions of translation, rotation and scaling. In addition, a variety of image quality evaluation algorithms and similarity evaluation algorithms are introduced to comprehensively realize the discrimination of authenticity and have a high recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the anti-counterfeiting identification process based on random fine texture patterns of the present invention;

[0021] Figure 2 Schematic diagram of nine kinds of mask images used when generating an image of a random fine texture pattern for ink diffusion based on the present invention;

[0022] Figure 3a A schematic diagram of an image of a random fine texture pattern for ink diffusion generated based on the present invention;

[0023] Figure 3b For use Figure 3a Schematic diagram of the QR code image after the black area inside the locator in the upper right corner of the QR code is replaced by a random fine texture pattern;

[0024] Figure 4a A schematic diagram of an anti-counterfeiting record collection image of a printed fine texture pattern based on the present invention;

[0025] Figure 4b This is a schematic diagram of the image capture of the printed QR code.

[0026] The markings in the figure are:

[0027] 1. Random fine texture pattern; 2. Black corner frame; 3. White space inside the corner frame; 4. Random fine texture pattern that replaces the inner area of ​​the locator in the upper right corner of the QR code. DETAILED DESCRIPTION

[0028] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0029] Figure 1 It is a schematic diagram of the anti-counterfeiting identification process based on random fine texture patterns of the present invention.

[0030] See also Figure 1 The anti-counterfeiting identification method based on random fine texture patterns provided by the present invention comprises the following steps:

[0031] S1) generating an original logo image of a coded logo or a graphic logo;

[0032] S2) generating an image of a random fine texture pattern for ink diffusion;

[0033] S3) combining the random fine texture image with the original logo image by adding, overlaying or nesting the random fine texture image to form a logo image containing the random fine texture pattern for printing;

[0034] S4) printing a logo image including a random fine texture pattern on a substrate;

[0035] S5) capturing an image of the fine texture ink pattern after the ink diffusion effect during the printing process, and binding it with the coding mark or graphic mark information as an anti-counterfeiting record image;

[0036] S6) The authenticity is determined by comparing the similarity between the image of the fine texture ink pattern on the substrate and the anti-counterfeiting filing image.

[0037] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns, wherein the coding mark or graphic mark in step S1 is preferably a QR code, and the random fine texture image generated in step S2 is preferably a square (see the schematic diagram for details). Figure 3a ); In step S3, a square random fine texture image is used to replace the 3*3 black square area inside the locator in the upper right corner of the original two-dimensional code image to form a two-dimensional code image containing a random fine texture pattern for printing (see the schematic diagram for details). Figure 3b ); In step S5, by collecting a two-dimensional code image (see the schematic diagram for details) Figure 4b ), decode the QR code content, and send the anti-counterfeiting filing image (see the schematic diagram for details) Figure 4a ) is bound to the QR code information and retained so as to facilitate the search for the corresponding anti-counterfeiting registration image during subsequent verification; in the step S6, the QR code content in the QR code image printed on the substrate to be verified is identified, and the corresponding anti-counterfeiting registration image is found according to its information to perform authenticity verification.

[0038] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. In step S2, the random fine texture pattern for ink diffusion is generated in the following manner (the image schematic diagram of the generated random fine texture pattern is shown in FIG. Figure 3a ):

[0039] S21) randomly generating an original image with randomly distributed black and white pixels;

[0040] S22) using a plurality of different preset mask images to perform XOR processing on the original image;

[0041] S23) performing penalty statistics on all masked images that are unfavorable for subsequent comparison, and taking the masked image with the least total penalty as the best result image;

[0042] S24) adding a full enclosing box or a half enclosing corner box for pattern positioning to the inner edge of the best result image to obtain a random subtle texture image for ink diffusion.

[0043] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. The step S22 includes 9 preset mask images, and the sizes of the 9 mask images are all set to be consistent with the original image size. Figure 2 , and its pixel distribution characteristics are as follows:

[0044] (1) All pixels are black;

[0045] (2) Pixels whose row and column indices sum to a multiple of 2 are black, and the others are white;

[0046] (3) Pixels whose row index is a multiple of 2 are black, and the others are white;

[0047] (4) Pixels whose column index is a multiple of 3 are black, and the others are white;

[0048] (5) Pixels whose row and column index sum is a multiple of 3 are black, and the others are white;

[0049] (6) The row index is divided by 2 and the column index is divided by 3. The pixels whose sum is a multiple of 2 are black, and the others are white.

[0050] (7) Pixels whose row and column index product is a multiple of 6 are black, and the others are white;

[0051] (8) The product of the row and column indices is divided by 3 and the remainder is added to the product of the row and column indices. Pixels that are multiples of 2 are black, and the others are white.

[0052] (9) The product of the row and column indices is divided by 3 and the modulo is added to the sum of the row and column indices. Pixels that are multiples of 2 are black, and the others are white.

[0053] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. In step S23, the unfavorable factors and penalty standards for statistically analyzing all masked images are as follows:

[0054] (1) Horizontal penalty: whenever there are n horizontally consecutive pixels of the same color in the image, and n>4, n-2 points are counted;

[0055] (2) Vertical penalty: whenever there are n vertically continuous pixels in the image and n>4, n-2 points are counted;

[0056] (3) 2*2 grid penalty: 3 points are scored whenever 4 pixels in a 2*2 pixel grid in the image are of the same color;

[0057] (4) Color unevenness penalty: calculate the proportion of black blocks in the image. The normal proportion of black blocks is 50%. Every 5% difference will be counted as 10 points.

[0058] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. The method of adding a full enclosing frame or a half enclosing corner frame for pattern positioning to the inner edge of the best result image in step S24 is as follows:

[0059] (1) Full bounding box: set the pixels within the outermost 5% of the image width around the image to black;

[0060] (2) Half-enclosed corner frame: first, set the eight rectangular areas near the four corners of the image to black, with the length of the long side being 25% of the image width and the width being 5% of the image width. Then, set the pixels within a certain number of layers outside the four black corner areas formed by these eight rectangular areas to white to highlight the four black corners. See the schematic diagram of the half-enclosed corner frame for details. Figure 3a .

[0061] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. In step S3, the random fine texture pattern used for anti-counterfeiting verification is embedded into a coding mark or a graphic mark image used for coding marking to obtain a mark image containing the random fine texture pattern for printing. The schematic diagram is shown in FIG. Figure 3b The image only needs to be printed once and is formed in one piece, which greatly simplifies the printing process and reduces the difficulty of printing.

[0062] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. In step S4, the ink will be randomly diffused irregularly in the fine texture pattern during the printing process (see the schematic diagram for details). Figure 4a ) to resist counterfeiting and imitation printing. The principle is simple and easy to promote.

[0063] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. In step S5, the fine texture ink pattern after the ink diffusion effect in the printing process is imaged and bound with the coding mark or graphic mark information as an anti-counterfeiting filing image (see the schematic diagram for details). Figure 4a ); at the same time, collect the image of the area where the coding mark or graphic mark is located (see the schematic diagram for details) Figure 4b ), and obtain its coding information or graphic information. The anti-counterfeiting filing image is bound to the coding information or graphic information and retained, so as to find the corresponding anti-counterfeiting filing image during subsequent verification.

[0064] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. In step S6, the process of distinguishing the authenticity of the fine texture ink pattern on the substrate is as follows:

[0065] S61) using a mobile phone to take an image of the area where the coded mark or graphic mark containing the random fine texture pattern is located on the printed object;

[0066] S62) identifying the coding information or graphic information in the image uploaded by the mobile phone, and searching for the corresponding anti-counterfeiting filing image according to the information;

[0067] S63) accurately locating the fine texture ink pattern area in the image uploaded by the mobile phone, and obtaining the fine texture ink pattern image after cropping and preprocessing;

[0068] S64) performing fine-tuning processing on the fine texture ink pattern image in three dimensions of translation, rotation and scaling to obtain processing result images under various fine-tuning parameters;

[0069] S65) Using a variety of similarity evaluation algorithms, the similarity between all the processed result images and the anti-counterfeiting filing images is evaluated, and the evaluation value with the highest similarity is retained for each algorithm. Finally, the similarity evaluation values ​​of various algorithms are used to comprehensively determine the authenticity of the fine texture ink pattern in the uploaded image.

[0070] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. In step S64, the offset tolerance range and offset step size of the three dimensions of translation, rotation and scaling are set respectively, and the offset step size is set within the offset tolerance range of the three dimensions. After processing the same initial fine texture ink pattern image, the processing result image under all step parameters is obtained. In this process, if i, j, and k steps are performed within the offset tolerance range of the three dimensions respectively, a total of i*j*k result images will be obtained, and each algorithm in step S65 will respectively evaluate the similarity between the i*j*k result images and the anti-counterfeiting filing image. This process further improves the positioning accuracy and the accuracy of the verification process.

[0071] The present invention provides an anti-counterfeiting identification method based on random fine texture patterns. In step S65, the following five similarity evaluation algorithms are introduced, and the authenticity is comprehensively judged based on the evaluation results, which has a high recognition accuracy:

[0072] (1) Cosine similarity;

[0073] (2) mean square error;

[0074] (3) Peak signal-to-noise ratio;

[0075] (4) gradient amplitude similarity deviation;

[0076] (5) Structural similarity.

[0077] Although the present invention has been disclosed as above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art may make some modifications and improvements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be based on the definition of the claims.

Claims

1. An anti-counterfeiting identification method based on random fine texture patterns, characterized in that: The steps include: S1) generating an original logo image of a coded logo or a graphic logo; S2) generating an image of a random fine texture pattern for ink diffusion; S3) combining the random fine texture image with the original logo image to form a logo image containing the random fine texture pattern for printing; S4) printing a logo image including a random fine texture pattern on a substrate; S5) capturing an image of the fine texture ink pattern after the ink diffusion effect during the printing process, and binding it with the coding mark or graphic mark information as an anti-counterfeiting record image; S6) Authenticity discrimination is achieved by comparing the similarity between the image of the fine texture ink pattern on the substrate and the anti-counterfeiting filing image.

2. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 1, characterized in that: The step S2 generates a random fine texture pattern for ink diffusion in the following manner: S21) randomly generating an original image with randomly distributed black and white pixels; S22) using a plurality of different preset mask images to perform XOR processing on the original image; S23) performing unfavorable factor penalty statistics on all masked images, and taking the masked image with the least total penalty as the best result image; S24) adding a full enclosing box or a half enclosing corner box for pattern positioning to the inner edge of the best result image to obtain a random subtle texture image for ink diffusion.

3. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 2, characterized in that: The step S22 includes 9 preset mask images, the sizes of which are all set to be consistent with the original image size. The pixel distribution characteristics of the 9 mask images are as follows: 1) All pixels are black; 2) Pixels whose row and column indices are multiples of 2 are black, and the others are white; 3) Pixels whose row index is a multiple of 2 are black, and the others are white; 4) Pixels with column indices that are multiples of 3 are black, and the others are white; 5) Pixels whose row and column index sum is a multiple of 3 are black, and the others are white; 6) The row index is divided by 2 and the column index is divided by 3. The pixels whose sum is a multiple of 2 are black, and the others are white. 7) Pixels whose row and column index product is a multiple of 6 are black, and the others are white; 8) The product of the row and column indices is divided by 3 and the remainder is added to the product of the row and column indices. Pixels that are multiples of 2 are black, and the others are white. 9) The product of the row and column indices is divided by 3 and the modulo is added to the sum of the row and column indices. Pixels that are multiples of 2 are black, and the others are white.

4. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 2, characterized in that: The unfavorable factor penalty standard for statistically analyzing all masked images in step S23 is as follows: 1) Horizontal penalty: whenever there are n horizontally continuous pixels of the same color in the image, and n>4, n-2 points will be counted; 2) Vertical penalty: whenever there are n vertically continuous pixels in the image, and n>4, n-2 points are counted; 3) 2*2 grid penalty: 3 points are scored every time 4 pixels in a 2*2 pixel grid in the image are of the same color; 4) Color unevenness penalty: calculate the absolute value of the difference between the proportion of black blocks in the image and 50%, and 10 points will be counted for every 5% reached.

5. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 2, characterized in that: The method of adding a full enclosing frame or a half enclosing corner frame for pattern positioning to the inner edge of the best result image in step S24 is as follows: 1) Full bounding box, set the pixels within the outermost 5% of the image width around the image to black; 2) Semi-enclosed corner frame: first set the 8 rectangular areas near the four corners of the image to black, with the length of their long edges being 25% of the image width and the width being 5% of the image width. Then set the pixels within a certain number of layers outward from the 4 black angled areas formed by these 8 rectangular areas to white to highlight the 4 black angled areas.

6. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 1, characterized in that: The random fine texture image in step S2 is a square, and the coding mark in step S3 is a two-dimensional code. After the original two-dimensional code image is scaled to a 3*3 area size that is consistent with the size of the random fine texture image, the random fine texture image is used to replace the 3*3 black square area inside one or more locators in the original two-dimensional code image to form a two-dimensional code image containing a random fine texture pattern for printing.

7. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 1, characterized in that: In step S3, the random fine texture image is combined with the original logo image by adding, overlaying or nesting. In step S5, the fine texture ink pattern after the ink diffusion effect in the printing process is imaged, and the image of the area where the coded logo or graphic logo is located is collected to obtain its coded information or graphic information. The anti-counterfeiting registration image is bound to the coding information or graphic information and retained to facilitate the search for the corresponding anti-counterfeiting registration image during subsequent verification.

8. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 1, characterized in that: In step S6, the process of determining the authenticity of the fine texture ink pattern on the substrate is as follows: S61) using a mobile phone to take an image of the area where the coded mark or graphic mark containing the random fine texture pattern is located on the printed object; S62) identifying the coding information or graphic information in the image uploaded by the mobile phone, and searching for the corresponding anti-counterfeiting filing image according to the information; S63) accurately locating the fine texture ink pattern area in the image uploaded by the mobile phone, and obtaining the fine texture ink pattern image after cropping and preprocessing; S64) performing fine-tuning processing on the fine texture ink pattern image in three dimensions of translation, rotation and scaling to obtain processing result images under various fine-tuning parameters; S65) Using a variety of similarity evaluation algorithms, the similarity between all the processed result images and the anti-counterfeiting filing images is evaluated, and the evaluation value with the highest similarity is retained for each algorithm. Finally, the similarity evaluation values ​​of various algorithms are used to comprehensively determine the authenticity of the fine texture ink pattern in the uploaded image.

9. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 8, characterized in that: In step S64, the offset tolerance ranges and offset step sizes of the three dimensions of translation, rotation and scaling are set respectively, and the offset is stepped with the set step size within the offset tolerance ranges of the three dimensions respectively. After processing the same initial fine texture ink pattern image, the processing result image under all step parameters is obtained.

10. The anti-counterfeiting identification method based on random fine texture patterns as claimed in claim 8, characterized in that: The step S6 performs i, j, and k steps respectively within the offset tolerance range in three dimensions, and obtains a total of i*j*k result images; the step S65 adopts a cosine similarity evaluation algorithm, a mean square error similarity evaluation algorithm, a peak signal-to-noise ratio similarity evaluation algorithm, a gradient amplitude similarity deviation evaluation algorithm or a structural similarity evaluation algorithm to perform similarity evaluation on the i*j*k result images and the anti-counterfeiting filing image, and uses the evaluation results as the basis for authenticity judgment.

Images