AI traceability code identification system and transcoding packaging method

Abstract

The invention discloses an AI traceability code recognition system and a transcoding packaging method, and the system comprises an obtaining module which is used for obtaining a to-be-recognized image of a traceability code carved in a packaging bottle; the preprocessing module is used for carrying out image preprocessing on the to-be-recognized image to obtain a target image; and the identification module is used for identifying the target image based on the AI traceability code identification model to obtain an identification result. The traceability code can be conveniently identified, whether the traceability code is qualified or not can be accurately determined, the two-dimensional code on the preset object can be associated according to the determined traceability code, and traceability tracking of the product can be accurately realized.

Description

Technical Field

[0001] The present invention relates to the field of identification technology, and particularly relates to an AI traceability code identification system and a transcoding encapsulation method. Background Art

[0002] The traceability code on the bottle label is easily damaged, and it is impossible to trace the supply channel of the sold goods. Therefore, the traceability code is engraved inside the encapsulated bottle. However, when engraved inside the encapsulated bottle, it is not easy to identify the traceability code, and it is impossible to accurately determine whether the traceability code is qualified. Furthermore, it is not easy to associate the two-dimensional code on the bottle cap according to the determined traceability code, resulting in the inability to accurately achieve traceability tracking of the product. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems in the above technologies to some extent. To this end, the object of the present invention is to provide an AI traceability code identification system and a transcoding encapsulation method, which are convenient for identifying the traceability code, accurately determining whether the traceability code is qualified, and then facilitating the association of the two-dimensional code on the preset object according to the determined traceability code, so as to accurately achieve traceability tracking of the product.

[0004] To achieve the above object, a first aspect embodiment of the present invention provides an AI traceability code identification system, including:

[0005] An acquisition module, configured to acquire a to-be-identified image of the traceability code engraved inside the encapsulated bottle;

[0006] A preprocessing module, configured to perform image preprocessing on the to-be-identified image to obtain a target image;

[0007] An identification module, configured to identify the target image based on the AI traceability code identification model to obtain an identification result.

[0008] According to some embodiments of the present invention, the preprocessing module includes:

[0009] An evaluation module, configured to evaluate the brightness feature of the to-be-identified image to obtain a brightness evaluation value;

[0010] A brightness enhancement module, configured to compare the brightness evaluation value with a preset brightness threshold, and perform brightness enhancement processing when it is determined that the brightness evaluation value is less than the preset brightness threshold.

[0011] According to some embodiments of the present invention, the evaluation module includes:

[0012] A first calculation module, configured to:

[0013] Acquire the brightness value of each pixel point in the to-be-identified image, and determine the maximum brightness value and the minimum brightness value;

[0014] Calculate the average brightness value Z of the to-be-identified image;

[0015]

[0016] Among them, M is the length of the image to be recognized; N is the width of the image to be recognized; x is the maximum value of the abscissa of the pixel points in the image to be recognized; y is the maximum value of the ordinate of the pixel points in the image to be recognized; f(i, j) is the brightness value of the pixel point (i, j) in the image to be recognized;

[0017] Based on the average brightness value Z, the maximum brightness value, and the minimum brightness value of the image to be recognized as brightness features, determine the brightness evaluation value S of the image to be recognized;

[0018]

[0019] Among them, f max is the maximum brightness value of the pixel points in the image to be recognized; f min is the minimum brightness value of the pixel points in the image to be recognized.

[0020] According to some embodiments of the present invention, the recognition module includes:

[0021] A segmentation module, configured to perform character segmentation on the traceability code in the target image based on the AI traceability code recognition model, determine the upper and lower boundaries and left and right boundaries of each character, determine the size of each character according to the upper and lower boundaries and left and right boundaries, determine whether the size is consistent with the preset size, and determine the characters with inconsistent size and perform size adjustment;

[0022] A character content recognition module, configured to perform character content recognition on the adjusted traceability code based on the AI traceability code recognition model to obtain a recognition result.

[0023] According to some embodiments of the present invention, the character content recognition module is configured to match the adjusted traceability code with the data table stored in the AI traceability code recognition model to obtain a recognition result.

[0024] According to some embodiments of the present invention, the recognition module includes:

[0025] A determination module, configured to:

[0026] Extract features from the target image based on the AI traceability code recognition model to determine the feature value of each pixel point in the target image;

[0027]

[0028] Among them, D i is the feature value of the i-th pixel point; R i is the R channel value of the i-th pixel point; G i is the G channel value of the i-th pixel point; B iis the B-channel value of the i-th pixel point;

[0029] Based on the feature values of each pixel point, a feature matrix D is formed. The feature matrix D has L rows and M columns;

[0030] A conversion module for converting the feature matrix D into a vector to obtain a target feature vector;

[0031] |D*D T -λE| = 0

[0032] where D T is the transpose of the feature matrix D; E is an L-order identity matrix, and the intermediate values solved from λ are L values. The L values are sorted from largest to smallest to form a target feature vector;

[0033] A matching module for matching the target feature vector with the preset feature vectors in the preset traceability code database, and determining the recognition result according to the matching result.

[0034] According to some embodiments of the present invention, it further includes: an adjustment module for segmenting the target image into N local images before the matching module matches the target feature vector with the preset feature vectors in the preset traceability code database; calculating the adjustment coefficients for each local image and performing adjustments.

[0035] According to some embodiments of the present invention, the adjustment module includes:

[0036] A second calculation module for calculating the energy function of each local image;

[0037]

[0038] where fCS N is the energy function of the N-th local image; W N is the N-th local image; CS N is the pixel matrix of the N-th local image; ‖W N -CS N ‖ F is the F-norm of W N -CS N ; is the weighted norm,

[0039] where j ∈ L, the value of j is L, L is the number of values contained in the target feature vector WV N d is the number of local images with an Euclidean distance less than 0.1, K is the weighted coefficient corresponding to the number of local images with an Euclidean distance less than 0.1; |SV N | j is the intermediate value;

[0040]

[0041] Among them, |WV N | j is the j-th value of the target feature vector WV N ; σ N is the variance of the N-th local image W N .

[0042] The third calculation module is used to calculate the adjustment coefficient LB based on the energy function of each local image N ;

[0043]

[0044] Adjust the corresponding local image based on the adjustment coefficient.

[0045] To achieve the above object, an embodiment of the second aspect of the present invention proposes a transcoding and encapsulation method, which applies the AI traceability code recognition system described above, including:

[0046] Recognize the traceability code engraved in the encapsulated bottle based on the AI traceability code recognition system to obtain the recognition result;

[0047] Convert it into a two-dimensional code according to the recognition result; the two-dimensional code includes the production source information, production date information, and production model information of the product;

[0048] Record the two-dimensional code on a preset object based on laser technology.

[0049] The present invention proposes an AI traceability code recognition system and a transcoding and encapsulation method, which are convenient for recognizing the traceability code, accurately determining whether the traceability code is qualified, and then facilitating the association of the two-dimensional code on the preset object according to the determined traceability code, and accurately realizing the traceability and tracking of the product.

[0050] Other features and advantages of the present invention will be described in the subsequent specification, and part of them will become obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in the written specification and the drawings.

[0051] The technical solutions of the present invention will be further described in detail below through the drawings and embodiments. Description of the Drawings

[0052] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0053] Figure 1It is a block diagram of an AI traceability code recognition system according to an embodiment of the present invention;

[0054] Figure 2 It is a schematic diagram of an image to be recognized of a traceability code according to an embodiment of the present invention;

[0055] Figure 3 It is a schematic diagram of recording a two-dimensional code on a preset object according to an embodiment of the present invention;

[0056] Figure 4 It is a flowchart of a transcoding and encapsulation method according to an embodiment of the present invention. Specific embodiments

[0057] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not intended to limit the present invention.

[0058] As Figures 1 - 2 shown, an embodiment of the first aspect of the present invention provides an AI traceability code recognition system, including:

[0059] An acquisition module, configured to acquire an image to be recognized of a traceability code engraved in a packaging bottle;

[0060] A preprocessing module, configured to perform image preprocessing on the image to be recognized to obtain a target image;

[0061] A recognition module, configured to recognize the target image based on an AI traceability code recognition model to obtain a recognition result.

[0062] The working principle of the above technical solution: The acquisition module uses a high-resolution camera or image acquisition device to ensure that the details of the traceability code can be clearly captured, and is mainly responsible for capturing the image of the traceability code engraved in the packaging bottle. Based on the AI traceability code recognition model, the preprocessed target image is recognized and the recognition result is output. This model is trained based on deep learning technology and can accurately recognize information such as characters, patterns or barcodes in the traceability code.

[0063] The beneficial effects of the above technical solution: It is convenient to recognize the traceability code, accurately determine whether the traceability code is qualified, and then it is convenient to associate the two-dimensional code on the bottle cap according to the determined traceability code, and accurately realize the traceability and tracking of the product. The AI traceability code recognition system is convenient for improving the recognition efficiency, reducing the cost of manual intervention, and enhancing the accuracy of data tracking and verification. By accurately recognizing the traceability code in the packaging bottle, enterprises can more effectively track the production, circulation and sales processes of products, ensuring product quality and safety. At the same time, consumers can also verify the authenticity and origin of products by scanning the traceability code, enhancing their trust in the brand.

[0064] In one embodiment, when the recognition module recognizes the target image based on the AI traceability code recognition model, it includes recognizing the positioning mark, and quickly positioning the traceability code in the target image based on the positioning mark, which is convenient for improving the recognition rate.

[0065] According to some embodiments of the present invention, the preprocessing module includes:

[0066] An evaluation module for evaluating the brightness feature of the image to be recognized to obtain a brightness evaluation value;

[0067] A brightness enhancement module for comparing the brightness evaluation value with a preset brightness threshold, and performing brightness enhancement processing when it is determined that the brightness evaluation value is less than the preset brightness threshold.

[0068] The working principle of the above technical solution: The preset brightness threshold is set according to the actual application scenario and image acquisition conditions. The preset brightness threshold is used to trigger brightness enhancement processing when the image brightness is insufficient, but at the same time, unnecessary processing is avoided when the image brightness is already sufficient.

[0069] The beneficial effect of the above technical solution: Through the collaborative work of the evaluation module and the brightness enhancement module, the preprocessing module can ensure that the image to be recognized has an appropriate brightness level, thereby improving the accuracy and reliability of traceability code recognition.

[0070] According to some embodiments of the present invention, the evaluation module includes:

[0071] A first calculation module for:

[0072] Obtaining the brightness value of each pixel point in the image to be recognized, and determining the maximum brightness value and the minimum brightness value;

[0073] Calculating the average brightness value Z of the image to be recognized;

[0074]

[0075] Where M is the length of the image to be recognized; N is the width of the image to be recognized; x is the maximum value of the abscissa of the pixel point in the image to be recognized; y is the maximum value of the ordinate of the pixel point in the image to be recognized; f(i, j) is the brightness value of the pixel point (i, j) in the image to be recognized;

[0076] Based on the average brightness value Z, the maximum brightness value and the minimum brightness value of the image to be recognized, as the brightness feature, determining the brightness evaluation value S of the image to be recognized;

[0077]

[0078] Where f max is the maximum brightness value of the pixel points in the image to be recognized; f minis the minimum brightness value of the pixel points in the image to be recognized.

[0079] The working principle of the above technical solution: The first calculation module first traverses each pixel point in the image to be recognized to accurately obtain its brightness value. Subsequently, among all the obtained brightness values, the maximum brightness value and the minimum brightness value are determined. Calculate the average brightness value of the image to be recognized. Based on the average brightness value Z of the image to be recognized, the maximum brightness value, and the minimum brightness value, as brightness features, determine the brightness evaluation value of the image to be recognized. It comprehensively considers the deviation between the pixel brightness and the average brightness (i.e., the degree of dispersion of the brightness distribution) and the difference between the maximum and minimum brightness values, thereby providing a comprehensive evaluation of the brightness features. The brightness evaluation value S not only reflects the overall distribution of the image brightness but also reflects the dynamic range of the brightness values (i.e., the difference between the maximum brightness and the minimum brightness). When the image brightness distribution is relatively uniform, the deviation between the pixel brightness and the average brightness will be small, resulting in a relatively low S value. On the contrary, if there are significant brightness changes or contrast enhancements in the image, the deviation between the pixel brightness and the average brightness will increase, thereby increasing the S value.

[0080] The beneficial effects of the above technical solution: The first calculation module provides a solid foundation for subsequent brightness enhancement processing by accurately calculating the brightness features of the image to be recognized. This not only helps improve the accuracy of traceability code recognition but also ensures the efficient operation of the entire AI traceability code recognition system.

[0081] According to some embodiments of the present invention, the recognition module includes:

[0082] A segmentation module for segmenting the traceability code in the target image based on the AI traceability code recognition model, determining the upper and lower boundaries and the left and right boundaries of each character, determining the size of each character according to the upper and lower boundaries and the left and right boundaries, judging whether the size is consistent with the preset size, and determining the characters with inconsistent size and performing size adjustment;

[0083] A character content recognition module for recognizing the character content of the adjusted traceability code based on the AI traceability code recognition model to obtain the recognition result.

[0084] Working principle of the above technical solution: The segmentation module first uses the AI traceability code recognition model to accurately segment the traceability code in the target image. This process aims to clearly separate each character in the traceability code, laying a foundation for subsequent character content recognition. Image segmentation algorithms in deep learning technology are adopted, such as semantic segmentation, instance segmentation, or segmentation methods based on edge detection. After completing the character segmentation, the module further determines the upper and lower boundaries and left and right boundaries of each character. Calculate the size of each character according to the boundary information and compare it with the preset size. The preset size is usually based on the standard design specifications of the traceability code to ensure the accuracy of recognition. If it is found that the character size is inconsistent with the preset size, the module will adjust the size of these characters. The adjustment methods may include scaling, stretching, or interpolation, etc., to ensure that all characters have a consistent size in the subsequent recognition process. After the character segmentation and size adjustment are completed, the character content recognition module will accurately recognize the character content of the adjusted traceability code based on the AI traceability code recognition model. According to the optical character recognition (OCR) technology in deep learning, this technology can automatically recognize the character content in the image and convert it into an editable text format. The character content recognition module will output the recognition result. This result includes the complete character sequence in the traceability code and the confidence score of each character.

[0085] Beneficial effects of the above technical solution: The segmentation module and the character content recognition module in the recognition module cooperate together to achieve the accurate segmentation and character content recognition of the traceability code in the target image. This process not only improves the accuracy of recognition but also ensures the efficient operation of the entire AI traceability code recognition system.

[0086] According to some embodiments of the present invention, the character content recognition module is used to match the adjusted traceability code with the data table stored in the AI traceability code recognition model to obtain the recognition result.

[0087] Working principle of the above technical solution: During the training process of the AI traceability code recognition model, it will learn and store a large number of traceability code character samples and their corresponding labels (i.e., character content). These samples and labels are sorted and stored in a data table for subsequent character recognition. When the character content recognition module receives the adjusted traceability code image, it extracts the character features in the image and matches these features with the character samples stored in the data table. Once the most matching character sample is found, the character content recognition module will output the label corresponding to the sample as the recognition result. This process ensures the accuracy and reliability of the recognition result.

[0088] Beneficial effects of the above technical solution: The character content recognition module realizes the accurate recognition of the character content of the adjusted traceability code by matching with the data table stored in the AI traceability code recognition model.

[0089] According to some embodiments of the present invention, the recognition module includes:

[0090] A determination module, configured to:

[0091] Extract features from the target image based on the AI traceability code recognition model to determine the feature values of each pixel point in the target image;

[0092]

[0093] where D i is the feature value of the i-th pixel point; R i is the R-channel value of the i-th pixel point; G i is the G-channel value of the i-th pixel point; B i is the B-channel value of the i-th pixel point;

[0094] Based on the feature values of each pixel point, a feature matrix D is formed. The feature matrix D is L rows and M columns;

[0095] A conversion module, configured to perform vector conversion on the feature matrix D to obtain a target feature vector;

[0096] |D*D T -λE| = 0

[0097] where D T is the transpose of the feature matrix D; E is an identity matrix of order L, and the intermediate values solved from λ are L values. The L values are sorted from largest to smallest to form a target feature vector;

[0098] A matching module, configured to match the target feature vector with a preset feature vector in a preset traceability code database, and determine a recognition result according to the matching result.

[0099] Working principle of the above technical solution: The determination module first uses the AI traceability code recognition model to extract features from the target image, facilitating the extraction of key information that can characterize the traceability code characters from the image. For each pixel point in the target image, the determination module calculates a feature value based on the values of its three channels: R (red), G (green), and B (blue). The contributions of the RGB three channels to the feature value are comprehensively considered, but with different weights, which are set based on the distribution characteristics of the traceability code characters in the color space. After calculating the feature values of all pixel points, the determination module organizes these feature values into a feature matrix D. The dimension of the feature matrix D is L rows and M columns, where L and M represent the number of rows and columns of the target image, respectively. The conversion module is used to perform vector conversion on the feature matrix D to obtain the target feature vector; the matching module matches the target feature vector with the preset feature vectors in the preset traceability code database. The preset traceability code database contains a large number of feature vectors of known traceability codes and their corresponding labels (i.e., traceability code content). The matching process uses measurement methods such as cosine similarity and Euclidean distance to calculate the similarity between the target feature vector and the preset feature vector. According to the matching result, the matching module finds the preset feature vector that is most similar to the target feature vector and outputs its corresponding label as the recognition result.

[0100] Beneficial effects of the above technical solution: Through the collaborative work of the determination module, the conversion module, and the matching module, the recognition module achieves accurate recognition of the traceability code in the target image.

[0101] According to some embodiments of the present invention, it further includes: an adjustment module, which is used to divide the target image into N local images before the matching module matches the target feature vector with the preset feature vectors in the preset traceability code database; calculate the adjustment coefficient for each local image and perform adjustment.

[0102] Working principle of the above technical solution: The adjustment module first divides the target image into N local images. The purpose of this step is to divide the larger image into smaller and more easily processed regions for separate feature extraction and adjustment of each region. For each segmented local image, the adjustment module calculates an adjustment coefficient. After calculating the adjustment coefficient, the adjustment module adjusts each local image. The adjustment operation is noise reduction, which is convenient for improving the image quality and enhancing the recognizability of the traceability code characters.

[0103] Beneficial effects of the above technical solution: The introduction of the adjustment module further improves the recognition process, enhancing the accuracy and efficiency of traceability code recognition. Through the segmentation and adjustment of local images, it ensures that each region can enter the subsequent feature extraction and matching process in the best state, thereby improving the performance of the entire recognition system.

[0104] According to some embodiments of the present invention, the adjustment module includes:

[0105] A second calculation module for calculating the energy function of each local image;

[0106]

[0107] where fCS N is the energy function of the Nth local image; W N is the Nth local image; CS N is the pixel matrix of the Nth local image; ‖W N -CS N ‖ F is the F-norm of W N -CS N , measuring the difference between the two; is the weighted norm,

[0108] where j ∈ L, the value of j is L, L is the number of values contained in the target feature vector WV N , d is the number of local images with an Euclidean distance less than 0.1, K is the weighting coefficient corresponding to the number of local images with an Euclidean distance less than 0.1; |SV N | j is the intermediate value;

[0109]

[0110] where |WV N | j is the jth value of the target feature vector WV N ; σ N is the variance of the Nth local image W N ;

[0111] A third calculation module for calculating the adjustment coefficient LB N ;

[0112]

[0113] Adjusting the corresponding local image based on the adjustment coefficient.

[0114] The working principle of the above technical solution: The energy function is an index that measures the difference between a local image and a certain ideal state. ‖W N -CS N ‖ F is the F-norm of W N -CS N , measuring the difference between the two. Calculating the adjustment coefficient LB N, the difference between the energy function and a threshold based on the weighted norm is considered. If the difference is greater than 0, the adjustment coefficient takes this difference value; otherwise, it takes 0.

[0115] Advantages of the above technical solution: By calculating the energy function and the adjustment coefficient of each local image, the adjustment module can achieve fine adjustment of the local image, improving the accuracy of traceability code recognition. The adjustment module can make personalized adjustments according to the characteristics of different local images, enhancing the adaptability of the entire recognition system. By improving the image quality, the adjustment module helps to enhance the robustness of the recognition system against adverse factors such as light changes and noise interference. Through the cooperation of the second calculation module and the third calculation module, the adjustment module realizes fine adjustment of the local image, providing higher-quality image input for subsequent feature extraction and matching processes.

[0116] As Figures 3 - 4 shown, to achieve the above object, the second aspect embodiment of the present invention proposes a transcoding and encapsulation method, applying the AI traceability code recognition system as described above, including steps S1 - S3:

[0117] S1. Based on the AI traceability code recognition system, identify the traceability code engraved in the packaging bottle to obtain the recognition result;

[0118] S2. Convert it into a two-dimensional code according to the recognition result; the two-dimensional code includes the production source information, production date information, and production model information of the product;

[0119] S3. Record the two-dimensional code on a preset object based on laser technology.

[0120] Working principle and beneficial effects of the above technical solution: Utilize an advanced AI traceability code recognition system to accurately identify the traceability code carefully engraved inside the packaging bottle. Convert the key information (such as production source, production date, production model, etc.) in the traceability code into a QR code format. Adopt an efficient QR code generation algorithm to ensure that the generated QR code contains rich information while maintaining a small size and high readability. With the help of precise laser technology, accurately imprint the QR code generated in step S2 on a preset object. The preset object includes at least one of the packaging bottle cap, the bottom of the packaging bottle, the bottle wall, and the packaging body. The QR code is firmly recorded on the preset object, facilitating consumers to scan and query product information, while enhancing the anti-counterfeiting and traceability of the product. The automated recognition and transcoding process significantly shortens the time for product information processing and packaging. Combining AI recognition and laser marking technology provides a difficult-to-replicate anti-counterfeiting mark for the product. Consumers can easily obtain product details by simply scanning the QR code on the bottle cap, enhancing the user experience. A perfect traceability system helps enterprises quickly respond to quality problems and safeguard consumers' rights and interests. By integrating the AI traceability code recognition system and laser technology, the efficient conversion and packaging of product information from the traceability code to the QR code are achieved, not only improving production efficiency but also enhancing the anti-counterfeiting and traceability of the product, bringing a more convenient and secure shopping experience for consumers.

[0121] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these changes and modifications.

Claims

1. An AI traceability code recognition system, characterized in that, Including: An acquisition module, configured to acquire a to-be-recognized image of a traceability code engraved in a packaging bottle; A preprocessing module, configured to perform image preprocessing on the to-be-recognized image to obtain a target image; A recognition module, configured to recognize the target image based on an AI traceability code recognition model to obtain a recognition result.

2. The AI traceability code recognition system according to claim 1, characterized in that, The preprocessing module includes: An evaluation module, configured to evaluate the brightness feature of the to-be-recognized image to obtain a brightness evaluation value; A brightness enhancement module, configured to compare the brightness evaluation value with a preset brightness threshold, and perform brightness enhancement processing when it is determined that the brightness evaluation value is less than the preset brightness threshold.

3. The AI traceability code recognition system according to claim 2, characterized in that, The evaluation module includes: A first calculation module, configured to: Obtain the brightness value of each pixel point in the to-be-recognized image, and determine the maximum brightness value and the minimum brightness value; Calculate the average brightness value Z of the to-be-recognized image; Wherein, M is the length of the to-be-recognized image; N is the width of the to-be-recognized image; x is the maximum value of the abscissa of the pixel points in the to-be-recognized image; y is the maximum value of the ordinate of the pixel points in the to-be-recognized image; f(i, j) is the brightness value of the pixel point (i, j) in the to-be-recognized image; Determine the brightness evaluation value S of the to-be-recognized image according to the average brightness value Z, the maximum brightness value and the minimum brightness value of the to-be-recognized image as the brightness feature; Among them, f max is the maximum luminance value of the pixel points in the image to be recognized; f min is the minimum luminance value of the pixel points in the image to be recognized.

4. The AI traceability code recognition system according to claim 1, characterized in that, The recognition module includes: A segmentation module, configured to perform character segmentation on the traceability code in the target image based on the AI traceability code recognition model, determine the upper and lower boundaries and the left and right boundaries of each character, determine the size of each character according to the upper and lower boundaries and the left and right boundaries, judge whether the size is consistent with a preset size, and determine the characters with inconsistent size with the preset size and perform size adjustment; A character content recognition module, configured to perform character content recognition on the adjusted traceability code based on the AI traceability code recognition model to obtain a recognition result.

5. The AI traceability code recognition system according to claim 4, wherein, The character content recognition module is configured to match the adjusted traceability code with a data table stored in the AI traceability code recognition model to obtain a recognition result.

6. The AI traceability code recognition system according to claim 1, characterized in that, The recognition module includes: A determination module, configured to: Extract features from the target image based on the AI traceability code recognition model to determine the feature value of each pixel point in the target image; Among them, D i is the eigenvalue of the i-th pixel; R i is the R-channel value of the i-th pixel; G i is the G-channel value of the i-th pixel; B i is the B-channel value of the i-th pixel; Construct a feature matrix D based on the feature values of each pixel point, and the feature matrix D is L rows and M columns; A conversion module, configured to perform vector conversion on the feature matrix D to obtain a target feature vector; |D*D T -λE| = 0 Among them, D T is the transpose of the feature matrix D; E is the identity matrix of order L, and the intermediate values solved from λ are L values. The L values are sorted from large to small to form the target feature vector; A matching module, configured to match the target feature vector with a preset feature vector in a preset traceability code database, and determine the recognition result according to the matching result.

7. The AI traceability code recognition system according to claim 1, characterized in that, It further includes: An adjustment module, configured to segment the target image into N local images before the matching module matches the target feature vector with the preset feature vector in the preset traceability code database; calculate the adjustment coefficient for each local image and perform adjustment.

8. The AI traceability code recognition system according to claim 7, characterized in that, The adjustment module includes: A second calculation module, configured to calculate the energy function of each local image; Among them, fCS N is the energy function of the Nth local image; W N is the Nth local image; CS N is the pixel matrix of the Nth local image; ‖W N - CS N ‖ F is the F-norm of W N - CS N ; ‖CS N ‖ w* is the weighted norm, where j ∈ L is the value of j, L is the number of values contained in the target feature vector WV N the number of values, d is the number of local images with Euclidean distance less than 0.1, and K is the weighted coefficient corresponding to the number of local images with Euclidean distance less than 0.1; |SV N | j is the intermediate value; Among them, |WV N | j is the j-th value of the target feature vector WV N ; σ N is the variance of the N-th local image W N ; A third calculation module, configured to calculate an adjustment coefficient LB based on the energy function of each local image N ; Adjust the corresponding local image based on the adjustment coefficient.

9. A transcoding and encapsulation method, applied to the AI traceability code recognition system according to any one of claims 1-8, characterized in that Including: Recognize the traceability code engraved in the packaging bottle based on the AI traceability code recognition system to obtain a recognition result; Convert the recognition result into a two-dimensional code; the two-dimensional code includes the production source information, production date information, and production model information of the product; Record a QR code on a preset object based on laser technology.

Images