Manufacturing data traceability analysis method for digital transfer printing paper

Abstract

The present invention relates to the field of commodity data tracing technology, and specifically to a manufacturing data tracing and analysis method for serving digital transfer printing paper. The method includes: obtaining manufacturing data and printing images of digital transfer printing paper; determining the complexity of printing tasks in combination with the color distribution of pixels in the printing image, and dividing all digital transfer printing papers; obtaining the printing process influence rate of each category in each dimension according to the difference in printing task complexity of different digital transfer printing papers in each category and the difference in manufacturing data in the same dimension; analyzing the distribution of manufacturing data in each dimension of all digital transfer printing papers in combination with the printing task complexity, and obtaining the printing-related uniqueness value of each dimension; integrating the printing process influence rate and the printing-related uniqueness value to perform abnormality detection on the manufacturing data and delete the abnormal manufacturing data, and use the remaining manufacturing data for tracing. The present invention improves the accuracy and credibility of the tracing results.

Description

Technical Field

[0001] The present invention relates to the technical field of commodity data tracing, and in particular to a manufacturing data tracing and analysis method for digital transfer printing paper. Background Art

[0002] Digital transfer printing paper is a special carrier used for digital printing of textiles, and it plays a key role in the digital transfer printing process. This paper is specially treated to carry and fix dyes or pigments until these pigment patterns are transferred to the fabric. The use of digital transfer printing paper has significantly improved the flexibility of pattern design and the fineness of printing, especially for small batches, personalized and complex pattern printing needs.

[0003] In order to ensure product quality, improve product technology in a timely manner, and enhance corporate image and competitiveness, it is necessary to trace and manage the manufacturing data of printed paper. The authenticity of data is an important attribute of the traceability system. Therefore, before uploading the manufacturing data of digital transfer printing paper to the traceability system, the data needs to be cleaned. However, existing data cleaning methods often tend to ignore the differences between manufacturing data under different printing processes, which may result in false detection or missed detection, thereby affecting the accuracy and credibility of subsequent traceability results. Summary of the invention

[0004] In order to solve the problem that the existing methods have false detection or missed detection when cleaning the manufacturing data of digital transfer printing paper, which affects the low accuracy and credibility of the traceability results of digital transfer printing products, the purpose of the present invention is to provide a manufacturing data traceability analysis method for digital transfer printing paper. The technical scheme adopted is as follows:

[0005] The present invention provides a manufacturing data tracing and analysis method for digital transfer printing paper, the method comprising the following steps:

[0006] Obtain manufacturing data of different digital transfer printing papers in different dimensions and corresponding printing images;

[0007] The complexity of the printing task is determined in combination with the color distribution of the pixel points in the printing image; all digital transfer printing papers are divided into multiple categories using the complexity of the printing task; the printing process influence rate of each category in each dimension is obtained according to the difference in the complexity of the printing task of different digital transfer printing papers in each category and the difference in the manufacturing data in the same dimension; the distribution of the manufacturing data in each dimension of all digital transfer printing papers is analyzed in combination with the complexity of the printing task to obtain the printing-related uniqueness value of each dimension;

[0008] The printing process influence rate and the printing-related uniqueness value are integrated to determine the interference anomaly index corresponding to each category in each dimension; the interference anomaly index is used to perform anomaly detection on all manufacturing data and delete the abnormal manufacturing data, and the remaining manufacturing data is used for tracing.

[0009] Preferably, the step of determining the complexity of the printing task in combination with the color distribution of pixels in the printing image includes:

[0010] Calculate the information entropy corresponding to each digital transfer printing paper according to the pixel value of the pixel point in the printing image of each digital transfer printing paper;

[0011] The product of the sum of the pixel values ​​of all the pixels in the printing image of each digital transfer printing paper and the information entropy is used as the color complexity of the corresponding digital transfer printing paper;

[0012] The normalized result of the color complexity is used as the printing task complexity of the corresponding digital transfer printing paper.

[0013] Preferably, all digital transfer printing papers are divided into multiple categories using the complexity of the printing task, including:

[0014] Based on the Euclidean distance between the printing task complexities of different digital transfer printing papers, the K-means clustering algorithm is used to cluster all the printing task complexities, and all the digital transfer printing papers corresponding to all the printing task complexities in each cluster are classified as one category.

[0015] Among them, the k value when using the K-means clustering algorithm for clustering is obtained by processing using the elbow method.

[0016] Preferably, the printing process influence rate of each category in each dimension is obtained according to the difference in the complexity of printing tasks of different digital transfer printing papers in each category and the difference in manufacturing data in the same dimension, including:

[0017] The difference between the manufacturing data of each digital transfer printing paper in the candidate category and other digital transfer printing papers in each dimension is recorded as the first difference between the two digital transfer printing papers in each dimension; the difference in the complexity of the printing task between each digital transfer printing paper in the candidate category and other digital transfer printing papers is recorded as the second difference between the two digital transfer printing papers; the Pearson correlation coefficient between the manufacturing data of all digital transfer printing papers in the candidate category in each dimension and the complexity of the printing task of all digital transfer printing papers in the candidate category is calculated;

[0018] The overall difference between the first difference and the corresponding second difference of each digital transfer printing paper in each dimension and the Pearson correlation coefficient of other digital transfer printing papers in the candidate category are used to obtain the printing process influence rate of the candidate category in each dimension, wherein the overall difference is negatively correlated with the printing process influence rate, and the Pearson correlation coefficient is positively correlated with the printing process influence rate;

[0019] The candidate category is any category.

[0020] Preferably, the distribution of manufacturing data of all digital transfer printing papers in each dimension is analyzed in combination with the complexity of the printing task to obtain the printing-related uniqueness value of each dimension, including:

[0021] The manufacturing data of all digital transfer printing papers under the dimensions to be analyzed in the candidate category are sorted in order of the complexity of the printing tasks from small to large, and a curve fitting is performed on the sorted data to obtain a fitting curve;

[0022] Using points on the fitting curve where the derivative value is 0 as dividing points to divide the manufacturing data into monotonic intervals;

[0023] Obtain the data range of each monotonic interval, and obtain the printing-related uniqueness value of the dimension to be analyzed according to the mean of the data range of all monotonic intervals and the discreteness of the derivative values ​​corresponding to the manufacturing data of all digital transfer printing papers under the dimension to be analyzed in the candidate category on the fitting curve, wherein the discreteness is positively correlated with the printing-related uniqueness value, and the mean of the data range is negatively correlated with the printing-related uniqueness value;

[0024] The dimension to be analyzed is any dimension.

[0025] Preferably, the degree of discreteness of the derivative values ​​corresponding to the manufacturing data of all digital transfer printing papers in the candidate category under the dimension to be analyzed on the fitting curve includes: calculating the variance of the derivative values ​​corresponding to the manufacturing data of all digital transfer printing papers in the candidate category under the dimension to be analyzed on the fitting curve as the degree of discreteness.

[0026] Preferably, the fusing of the printing process influence rate and the printing related uniqueness value to determine the interference anomaly index corresponding to each category in each dimension includes:

[0027] According to the negative correlation normalization result of the printing-related uniqueness value of the dimension to be analyzed, the normalized value of the number of monotonic intervals divided by the fitting curve and the corresponding printing process influence rate, the interference anomaly index corresponding to the candidate category in the dimension to be analyzed is obtained.

[0028] Preferably, the interference anomaly index corresponding to the candidate category in the dimension to be analyzed is obtained based on the negative correlation normalization result of the printing-related uniqueness value of the dimension to be analyzed, the normalized value of the number of monotonic intervals divided by the fitting curve, and the corresponding printing process influence rate, including:

[0029] Calculate the difference between the constant 1 and the printing process influence rate of the candidate category under the dimension to be analyzed, and record the product of the difference and the negative correlation normalization result as the first eigenvalue;

[0030] The product of the normalized value of the number of monotonic intervals divided by the fitting curve and the printing process influence rate of the candidate category under the dimension to be analyzed is recorded as the second eigenvalue;

[0031] The sum of the first eigenvalue and the second eigenvalue is determined as the interference anomaly indicator corresponding to the candidate category in the dimension to be analyzed.

[0032] Preferably, performing anomaly detection on all manufacturing data using the interference anomaly index includes: performing anomaly detection on all manufacturing data using a LOF algorithm based on the interference anomaly index.

[0033] Preferably, when the LOF algorithm is used for detection, the distance measurement method between data is defined as weighted Euclidean distance, and the normalized result of the corresponding interference anomaly index in each dimension is used as the weight to measure the distance in different dimensions.

[0034] The present invention has at least the following beneficial effects:

[0035] The present invention analyzes the complexity of the printing task of digital transfer printing paper by combining the color distribution of pixel points in the printing image of digital transfer printing paper, divides all digital transfer printing papers into multiple categories by using the complexity of printing task, and then analyzes each category separately, and evaluates the printing process influence rate according to the difference in the complexity of printing task of different digital transfer printing papers in each category and the difference in manufacturing data under the same dimension, and further analyzes the distribution of manufacturing data under each dimension of all digital transfer printing papers in combination with the complexity of printing task, obtains the printing-related uniqueness value of each dimension, and finally performs abnormal detection on manufacturing data in combination with the printing process influence rate and the printing-related uniqueness value, and completes the cleaning of the manufacturing data of digital transfer printing paper to be uploaded to the traceability system. Compared with the prior art, the interference caused by different printing processes is reduced in the data cleaning process, and the abnormal probability of detection data of different printing processes is judged in combination with the change law of printing pattern complexity, the abnormal data detection accuracy is improved, the error rate of the tracing analysis process is reduced, the construction efficiency of the traceability system is improved, and the accuracy and credibility of the traceability results of digital transfer printing paper are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A flowchart of a manufacturing data tracing and analysis method for digital transfer printing paper provided by an embodiment of the present invention;

[0038] Figure 2 A flowchart of manufacturing data uploading provided by an embodiment of the present invention;

[0039] Figure 3 This is a structural block diagram of a manufacturing data tracing and analysis system for digital transfer printing paper provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the manufacturing data tracing and analysis method for digital transfer printing paper proposed by the present invention is described in detail below in combination with the accompanying drawings and preferred embodiments.

[0041] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0042] The specific scheme of the manufacturing data tracing and analysis method for digital transfer printing paper provided by the present invention is described in detail below with reference to the accompanying drawings.

[0043] Implementation example of manufacturing data traceability analysis method for digital transfer printing paper:

[0044] The specific scenario targeted by this embodiment is: in order to ensure the transparency and controllability of the digital transfer printing paper production process, and to locate problems in a timely manner and determine the direction of improvement, it is necessary to establish a traceability system for the printing paper production process. After uploading the printing paper production data in the traceability system, the traceability system can be accessed through the QR code or label on the product, and then the product production related information data can be obtained to realize product traceability.

[0045] This embodiment proposes a manufacturing data traceability analysis method for digital transfer printing paper, such as Figure 1 As shown, the manufacturing data traceability analysis method for digital transfer printing paper of this embodiment includes the following steps:

[0046] Step S1, obtaining manufacturing data of different digital transfer printing papers in different dimensions and corresponding printing images.

[0047] In the production process of digital transfer printing paper, uploading key data to the traceability system is crucial to ensure product quality and traceability. In order to ensure product quality and traceability, it is necessary to screen and upload manufacturing data in the production process. These manufacturing data specifically include: raw material data, production process data, quality inspection data, post-processing information and other data, among which raw material data includes supplier information, raw material batch specifications, quality inspection reports, etc., production process data includes production batches, production time, process parameters, etc., quality inspection data includes appearance, physical properties, chemical properties, etc., and post-processing information includes post-processing process information, packaging information, etc. In specific applications, implementers choose according to specific circumstances. It should be noted that: each type of data is manufacturing data under one dimension, and the above manufacturing data can be divided into digital data and text information data. For the convenience of calculation and analysis, for text information data, it is necessary to convert it into the form of digital data: extract keywords in each text information through the TF-IDF algorithm, and then use the word2vec algorithm to convert the extracted keywords into vector form. Then the manufacturing data of different digital transfer printing papers are formed into a multidimensional array, in which data of different dimensions represent different types of manufacturing data, and the data of each dimension is standardized.

[0048] A camera is used to capture a surface image of the digital transfer printing paper, the surface image being an RGB image, and the obtained surface image is recorded as a printing image.

[0049] At this point, the manufacturing data of each dimension of each digital transfer printing paper and the printing image of each digital transfer printing paper are obtained.

[0050] Step S2, determining the complexity of the printing task in combination with the color distribution of the pixel points in the printing image; using the printing task complexity to divide all digital transfer printing papers into multiple categories; obtaining the printing process influence rate of each category in each dimension according to the difference in printing task complexity of different digital transfer printing papers in each category and the difference in manufacturing data in the same dimension; analyzing the distribution of manufacturing data in each dimension of all digital transfer printing papers in combination with the printing task complexity to obtain the printing-related uniqueness value of each dimension.

[0051] For the manufacturing data of digital transfer printing paper, it is uploaded to the traceability system in batches according to the production time. The manufacturing data of the uploaded digital transfer printing paper and the printing image of each digital transfer printing paper are obtained, and these data are sent to the data processing module.

[0052] In a manufacturing data upload cycle, there are usually multiple printing tasks. Different printing tasks have different production processes and the difficulty of their production processes also varies. For the difficulty of printing with different production processes, the abnormal information that can be reflected in the manufacturing data also varies greatly. Therefore, it is necessary to evaluate the complexity of the printing tasks of the printing images of different digital transfer printing papers.

[0053] First, the pixel value of each pixel in the print image of each digital transfer print paper is collected, and then the information entropy corresponding to each digital transfer print paper is calculated according to the pixel values ​​of all pixels in the print image of each digital transfer print paper; the method for calculating the information entropy is a prior art and will not be described in detail here. The larger the information entropy, the more chaotic the distribution of pixel values ​​of the pixels in the print image, and the greater the printing difficulty when printing the digital transfer print paper. In addition, the larger the digital transfer print paper, the greater the printing complexity when printing. Based on this feature, the product of the sum of the pixel values ​​of all pixels in the print image of each digital transfer print paper and the information entropy is used as the color complexity of the corresponding digital transfer print paper. The normalized result of the color complexity is used as the printing task complexity of the corresponding digital transfer print paper.

[0054] By adopting this method, the printing task complexity of each digital transfer printing paper can be obtained. Each digital transfer printing paper corresponds to a printing task complexity. The printing task complexity reflects the difficulty of the digital transfer printing paper process from the perspective of the printing image. For digital transfer printing papers with similar difficulty, the manufacturing data to be uploaded may have similar process parameters: the manufacturing data of products with similar printing task complexity are analyzed, and the influence of manufacturing data with different printing task complexity is evaluated.

[0055] Next, the different multi-dimensional data of each product are preliminarily classified based on the complexity of the printing task, and the printing process impact rate of different dimensional data within each classification is evaluated.

[0056] Based on the Euclidean distance between the printing task complexities of different digital transfer printing papers, the K-means clustering algorithm is used to cluster all printing task complexities, and all digital transfer printing papers corresponding to all printing task complexities in each cluster are taken as one category, that is, multiple categories are obtained; among them, the k value when clustering using the K-means clustering algorithm is obtained by processing the printing task complexity of all digital transfer printing papers using the elbow method. For digital transfer printing papers with large differences in printing task complexity, the process parameters of their manufacturing process may be quite different. Therefore, after classifying digital transfer printing papers by printing task complexity, it is more meaningful to analyze the manufacturing data of digital transfer printing papers of the same category.

[0057] Within different categories, the manufacturing data of different dimensions of some digital transfer printing papers change with the complexity of the printing task. That is, under the influence of the complexity of the printing task, some process parameters in the printing manufacturing process will also change accordingly. To reduce the degree of interference of anomaly detection by such changes, it is first necessary to evaluate the relationship between different printing process parameters and the complexity of the printing task.

[0058] Next, this embodiment is described by taking one category as an example, and other categories can be processed by the method provided in this embodiment.

[0059] Specifically, any category is recorded as a candidate category; the difference between the manufacturing data of each digital transfer printing paper in the candidate category and other digital transfer printing papers in each dimension is recorded as the first difference between the corresponding two digital transfer printing papers in each dimension; the difference in printing task complexity between each digital transfer printing paper in the candidate category and other digital transfer printing papers is recorded as the second difference between the corresponding two digital transfer printing papers; the Pearson correlation coefficient between the manufacturing data of all digital transfer printing papers in the candidate category in each dimension and the printing task complexity of all digital transfer printing papers in the candidate category is calculated; the method for calculating the Pearson correlation coefficient is a prior art and will not be elaborated here.

[0060] The overall difference between the first difference and the corresponding second difference of each digital transfer printing paper in the candidate category and other digital transfer printing papers in each dimension and the Pearson correlation coefficient are used to obtain the printing process influence rate of the candidate category in each dimension.

[0061] In this embodiment, a specific calculation formula for the printing process influence rate is given, and the printing process influence rate of the candidate category in the jth dimension can be expressed as:

[0062]

[0063] in, represents the printing process influence rate of the candidate category under the jth dimension, n represents the number of digital transfer printing papers in the candidate category, represents the Pearson correlation coefficient between the manufacturing data of all digital transfer printing papers in the candidate category under the jth dimension and the printing task complexity of all digital transfer printing papers in the candidate category, represents the difference between the manufacturing data of the i-th digital transfer printing paper in the candidate category under the j-th dimension and the manufacturing data of the b-th digital transfer printing paper in the candidate category except the i-th digital transfer printing paper in the candidate category under the j-th dimension, that is, the first difference between the i-th digital transfer printing paper and the b-th digital transfer printing paper in the candidate category in the j-th dimension; represents the difference in printing task complexity between the i-th digital transfer printing paper in the candidate category and the b-th digital transfer printing paper in the candidate category except the i-th digital transfer printing paper, that is, the second difference between the i-th digital transfer printing paper and the b-th digital transfer printing paper in the candidate category in the j-th dimension; represents the preset first adjustment parameter, tanh( ) represents the hyperbolic tangent function, and average( ) represents the averaging function.

[0064] The process of obtaining the difference between manufacturing data is: calculate the absolute value of the difference between two manufacturing data, and use the absolute value as the difference between the two manufacturing data; similarly, the process of obtaining the difference in printing task complexity is: calculate the absolute value of the difference between the two printing task complexities, and use the absolute value as the difference between the two printing task complexities.

[0065] In this embodiment, the preset first adjustment parameter is introduced in the calculation process of the printing process influence rate in order to prevent the denominator from being 0. The preset first adjustment parameter in this embodiment is 0.01. In specific applications, the implementer can set it according to the specific situation.

[0066] It is used to reflect the overall difference between the manufacturing data of the i-th digital transfer printing paper in the candidate category in the j-dimension and the manufacturing data of other digital transfer printing papers in the j-dimension and the difference in the complexity of the printing task between the i-th digital transfer printing paper and other digital transfer printing papers. The larger the value, the greater the overall difference. When the overall difference between the manufacturing data of the i-th digital transfer printing paper in the candidate category and other digital transfer printing papers in the j-dimension and the difference in the complexity of the printing task between the i-th digital transfer printing paper and other digital transfer printing papers is smaller, and the Pearson correlation coefficient between the manufacturing data of all digital transfer printing papers in the candidate category in the j-dimension and the complexity of the printing task of all digital transfer printing papers in the candidate category is larger, it means that the process parameters represented by the manufacturing data of the j-dimension in the current category are more affected by the complexity of the printing pattern, that is, the printing process influence rate of the candidate category in the j-dimension is greater.

[0067] The calculation process of the relationship between process parameters of different dimensions and printing complexity is optimized through individual differences in data, reducing the interference of abnormal data on the calculation of process parameters of different dimensions affected by printing complexity.

[0068] The printing process influence rate corresponding to different dimensional data can reflect the influence of printing complexity on the manufacturing data of each product in the difficulty of similar printing pattern tasks. To further improve the accuracy of the abnormal data detection process, the abnormal probability of manufacturing data of different dimensions can be analyzed in combination with the influence of printing complexity on process parameters.

[0069] Specifically, any dimension is recorded as the dimension to be analyzed, and the manufacturing data of all digital transfer printing papers under the dimension to be analyzed in the candidate category are sorted in order from small to large according to the complexity of the printing task, and the least squares method is used to perform curve fitting on the sorted data to obtain a fitting curve. After monotonically sorting by the complexity of the printing task, under ideal conditions, the dimensional data with a higher printing process influence rate will change monotonically with the sorting order, but under the interference of abnormal data, the monotonicity of the dimensional data will fluctuate in the sorting order. In addition, for data with a lower printing process influence rate, the changes in process parameters reflected by these dimensional data are weakly related to the changes in the complexity of the printing pattern. Therefore, on the fitting curve, the discrete degree of the change rate of the data at each point is relatively high; in addition, since the complexity of the printing tasks in the same category is relatively similar, there is usually not much difference in the change amplitude between the process parameters of each dimension. The derivative value of each point on the fitting curve is obtained, and the points on the fitting curve where the derivative value is 0 are used as dividing points to divide the manufacturing data into multiple monotonic intervals; the data range of each monotonic interval is obtained, and the printing-related uniqueness value of the dimension to be analyzed is obtained according to the mean of the data range of all monotonic intervals and the discreteness of the derivative values ​​corresponding to the manufacturing data of all digital transfer printing papers in the candidate category under the dimension to be analyzed on the fitting curve, the discreteness is positively correlated with the printing-related uniqueness value, and the mean of the data range is negatively correlated with the printing-related uniqueness value.

[0070] In this embodiment, the process of obtaining the discrete degree of the derivative value is: calculating the variance of the derivative value corresponding to the manufacturing data of all digital transfer printing papers in the candidate category under the dimension to be analyzed on the fitting curve, as the discrete degree of the derivative value corresponding to the manufacturing data of all digital transfer printing papers in the candidate category under the dimension to be analyzed on the fitting curve.

[0071] Among them, a positive correlation means that the dependent variable will increase as the independent variable increases, and the dependent variable will decrease as the independent variable decreases. It can be an additive relationship, a multiplicative relationship, etc., which is determined by actual application; a negative correlation means that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases. It can be a subtractive relationship, a division relationship, etc., which is determined by actual application.

[0072] In this embodiment, the ratio between the discreteness of the derivative values ​​corresponding to the manufacturing data of all digital transfer printing papers in the candidate category under the dimension to be analyzed on the fitting curve and the mean of the data range of all monotonic intervals is taken. The larger the ratio is, the more ideal the degree of independence between the process parameters represented by the manufacturing data under the current dimension and the complexity of the printing pattern, and the less affected by abnormal data, that is, the greater its printing-related uniqueness is. Therefore, the ratio is used as the printing-related uniqueness value of the dimension to be analyzed.

[0073] By using the above method, the print-related uniqueness value of each dimension can be obtained.

[0074] Step S3, integrating the printing process influence rate and the printing-related uniqueness value to determine the interference anomaly index corresponding to each category in each dimension; using the interference anomaly index to perform anomaly detection on all manufacturing data and delete the abnormal manufacturing data, and using the remaining manufacturing data for tracing.

[0075] For dimensions where the data changes are similar to the changes in the complexity of the printing pattern, the data change amplitude on the fitting curve should be relatively gentle, and the number of monotonic intervals should be small; while for dimensions where the data changes are different from the changes in the complexity of the printing pattern, the printing-related uniqueness should be relatively large.

[0076] Next, the dimension to be analyzed is still used as an example for explanation. The method provided in this embodiment can be used to process other dimensions. According to the negative correlation normalized result of the printing-related uniqueness value of the dimension to be analyzed, the normalized value of the number of monotonic intervals divided by the fitting curve and the corresponding printing process influence rate, the interference anomaly index corresponding to the candidate category in the dimension to be analyzed is obtained.

[0077] Specifically, the negative correlation normalized result of the printing-related uniqueness value of the dimension to be analyzed is calculated, the difference between the constant 1 and the printing process influence rate of the candidate category in the dimension to be analyzed is calculated, and the product of the difference and the negative correlation normalized result is recorded as the first eigenvalue; the product of the normalized value of the number of monotonic intervals divided by the fitting curve and the printing process influence rate of the candidate category in the dimension to be analyzed is recorded as the second eigenvalue; the sum of the first eigenvalue and the second eigenvalue is determined as the interference anomaly index corresponding to the candidate category in the dimension to be analyzed.

[0078] The interference anomaly index corresponding to the candidate category in the jth dimension can be expressed as:

[0079] in, It represents the interference anomaly index corresponding to the candidate category in the jth dimension, represents the printing process influence rate of the candidate category in the jth dimension, A normalized value representing the number of monotonic intervals divided by the fitted curve, represents the print-related uniqueness value of the j-th dimension, and exp( ) is an exponential function with a natural constant as the base.

[0080] represents the negative correlation normalization result of the print-related uniqueness value of the j-th dimension, represents the first eigenvalue, The number of monotonic intervals is normalized by using the maximum and minimum values ​​of the number of monotonic intervals divided by the fitting curves corresponding to all dimensions. The maximum and minimum value normalization method is an existing technology and will not be described in detail here.

[0081] In summary, after adjusting the number of monotonic intervals and the range of data change rate by the printing process influence rate, the analysis of the influence of different printing complexities is increased, the probability of data anomaly is judged in combination with the change of printing complexity, and the data of different printing complexities are analyzed in a personalized way to improve the accuracy of subsequent detection. The larger the interference anomaly index, the higher the probability that the manufacturing data of the corresponding dimension still has anomalies after being affected by the printing process.

[0082] By using the above method, the interference anomaly index corresponding to each category in each dimension can be obtained. Next, the LOF algorithm is used to perform anomaly detection on all manufacturing data based on the interference anomaly index.

[0083] Specifically, the LOF algorithm is used to detect anomalies in manufacturing data. In order to further reduce the impact of different printing processes on the detection process, the K value of the LOF algorithm is set to the mean of the amount of data contained in all categories; the distance measurement method between data is defined as weighted Euclidean distance, and the distance measurement of different dimensions uses the normalized value of the interference anomaly index as the weight to amplify the data in the dimension with a higher probability of anomaly, and then filter the abnormal data. The LOF algorithm is a prior art and will not be described in detail here.

[0084] The abnormal data screened out include noise and irrelevant information such as duplication and errors. After screening the abnormal data, delete the corresponding data points, re-query the same batch and re-test the production data of the corresponding goods, and then re-acquire the data to realize the data cleaning process.

[0085] This embodiment implements the traceability system of digital transfer printing paper through blockchain technology. After screening the manufacturing data, the manufacturing data in the production process of digital transfer printing paper is uploaded to the blockchain. In the actual upload process, in addition to the manufacturing data, it also includes raw material procurement and quality inspection, production process records, product quality inspection and sales traceability marking information, and the entry process of this information is the same. Taking product quality inspection data as an example, the processed test data information is packaged and entered into the quality inspection node and database at the same time. The SHA3 algorithm is used to generate an information summary for the data entered by the node and sent to the blockchain. After the participating nodes reach a consensus, the information summary is written to the block. At this time, the block will return a hash value stored in the database, which can be used for data indexing.

[0086] During the process of tracing back the source code in the database, if you are skeptical about the information you have queried, you can compare the information summary in the blockchain with the information summary generated by the same SHA3 algorithm for the query information in the database to determine whether the data has been changed. If the information summary is consistent, the data is authentic and valid. Figure 2 As shown in the figure, this is a flowchart for uploading manufacturing data. At this point, the digital transfer printing paper traceability system has been established.

[0087] This embodiment analyzes the complexity of the printing task of digital transfer printing paper by combining the color distribution of pixel points in the printing image of digital transfer printing paper, and divides all digital transfer printing papers into multiple categories using the complexity of printing tasks. Each category is analyzed separately, and the printing process influence rate is evaluated according to the difference in the complexity of printing tasks of different digital transfer printing papers in each category and the difference in manufacturing data in the same dimension. The distribution of manufacturing data in each dimension of all digital transfer printing papers is further analyzed in combination with the complexity of printing tasks, and the printing-related uniqueness value of each dimension is obtained. Finally, the manufacturing data is detected for abnormality in combination with the printing process influence rate and the printing-related uniqueness value, and the cleaning of the manufacturing data of the digital transfer printing paper to be uploaded to the traceability system is completed. Compared with the prior art, the interference caused by different printing processes is reduced in the data cleaning process, and the abnormal probability of the detection data of different printing processes is judged in combination with the change law of the complexity of the printing pattern, thereby improving the accuracy of abnormal data detection, reducing the error rate of the tracing analysis process, improving the construction efficiency of the traceability system, and improving the accuracy and credibility of the traceability results of digital transfer printing paper.

[0088] Implementation example of manufacturing data traceability analysis system for digital transfer printing paper:

[0089] See also Figure 3 , which shows a structural block diagram of a manufacturing data tracing and analysis system for digital transfer printing paper provided by an embodiment of the present invention. The system may include a data acquisition module, a data processing module and a data tracing module.

[0090] Among them, the data acquisition module is used to obtain the manufacturing data of different digital transfer printing papers in different dimensions and the corresponding printing images;

[0091] A data processing module is used to determine the complexity of the printing task in combination with the color distribution of the pixel points in the printing image; divide all digital transfer printing papers into multiple categories using the printing task complexity; obtain the printing process influence rate of each category in each dimension according to the difference in printing task complexity of different digital transfer printing papers in each category and the difference in manufacturing data in the same dimension; analyze the distribution of manufacturing data in each dimension of all digital transfer printing papers in combination with the printing task complexity to obtain the printing-related uniqueness value of each dimension;

[0092] The data tracing module is used to integrate the printing process influence rate and the printing-related uniqueness value to determine the interference anomaly index corresponding to each category in each dimension; use the interference anomaly index to perform anomaly detection on all manufacturing data and delete the abnormal manufacturing data, and use the remaining manufacturing data for tracing.

[0093] It should be understood that Figure 3 The structural block diagram of the manufacturing data traceability analysis system for digital transfer printing paper shown and its modules can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art can understand that the above-mentioned method and system can be implemented using computer executable instructions and / or included in a processor control code, for example, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system and its modules of this specification can not only be implemented by hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software (for example, firmware).

[0094] For more details about the above modules, please refer to other places in this manual and will not be repeated here.

[0095] In other embodiments, a manufacturing data tracing and analysis device for digital transfer printing paper is provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes the manufacturing data tracing and analysis method for digital transfer printing paper. The device can be a chip, a component or a module, and the chip can include a connected processor and a memory; wherein the memory is used to store instructions, and when the processor calls and executes the instructions, the chip can execute the manufacturing data tracing and analysis method for digital transfer printing paper provided in the above embodiment.

[0096] In other embodiments, a computer program product is also provided. When the computer program product is run on a computer, the computer executes the above-mentioned related steps to implement the manufacturing data tracing and analysis method for digital transfer printing paper provided in the above-mentioned embodiment.

[0097] In other embodiments, a computer-readable storage medium is also provided, in which a computer program code is stored. When the computer program code is run on a computer, the computer executes the above-mentioned related method steps to implement the manufacturing data traceability analysis method for digital transfer printing paper provided in the above-mentioned embodiment.

[0098] Among them, the provided system, electronic device, computer program product, and computer-readable storage medium are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0099] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A manufacturing data tracing and analysis method for digital transfer printing paper, characterized in that: The method comprises the following steps: Obtain manufacturing data of different digital transfer printing papers in different dimensions and corresponding printing images; The complexity of the printing task is determined in combination with the color distribution of the pixel points in the printing image; all digital transfer printing papers are divided into multiple categories using the complexity of the printing task; the printing process influence rate of each category in each dimension is obtained according to the difference in the complexity of the printing task of different digital transfer printing papers in each category and the difference in the manufacturing data in the same dimension; the distribution of the manufacturing data in each dimension of all digital transfer printing papers is analyzed in combination with the complexity of the printing task to obtain the printing-related uniqueness value of each dimension; The printing process influence rate and the printing-related uniqueness value are integrated to determine the interference anomaly index corresponding to each category in each dimension; the interference anomaly index is used to perform anomaly detection on all manufacturing data and delete the abnormal manufacturing data, and the remaining manufacturing data is used for tracing; According to the difference in the complexity of printing tasks of different digital transfer printing papers in each category and the difference in manufacturing data in the same dimension, the printing process influence rate of each category in each dimension is obtained, including: The difference between the manufacturing data of each digital transfer printing paper in the candidate category and other digital transfer printing papers in each dimension is recorded as the first difference between the two digital transfer printing papers in each dimension; the difference in the complexity of the printing task between each digital transfer printing paper in the candidate category and other digital transfer printing papers is recorded as the second difference between the two digital transfer printing papers; the Pearson correlation coefficient between the manufacturing data of all digital transfer printing papers in the candidate category in each dimension and the complexity of the printing task of all digital transfer printing papers in the candidate category is calculated; The overall difference between the first difference and the corresponding second difference of each digital transfer printing paper in each dimension and the Pearson correlation coefficient of other digital transfer printing papers in the candidate category are used to obtain the printing process influence rate of the candidate category in each dimension, wherein the overall difference is negatively correlated with the printing process influence rate, and the Pearson correlation coefficient is positively correlated with the printing process influence rate; The candidate category is any category; The distribution of manufacturing data in each dimension of all digital transfer printing papers is analyzed in combination with the complexity of the printing task to obtain the printing-related uniqueness value of each dimension, including: The manufacturing data of all digital transfer printing papers under the dimensions to be analyzed in the candidate category are sorted in order of the complexity of the printing tasks from small to large, and a curve fitting is performed on the sorted data to obtain a fitting curve; Using points on the fitting curve where the derivative value is 0 as dividing points to divide the manufacturing data into monotonic intervals; Obtain the data range of each monotonic interval, and obtain the printing-related uniqueness value of the dimension to be analyzed according to the mean of the data range of all monotonic intervals and the discreteness of the derivative values ​​corresponding to the manufacturing data of all digital transfer printing papers under the dimension to be analyzed in the candidate category on the fitting curve, wherein the discreteness is positively correlated with the printing-related uniqueness value, and the mean of the data range is negatively correlated with the printing-related uniqueness value; The dimension to be analyzed is any dimension; The fusing of the printing process influence rate and the printing related uniqueness value to determine the interference anomaly index corresponding to each category in each dimension includes: According to the negative correlation normalization result of the printing-related uniqueness value of the dimension to be analyzed, the normalized value of the number of monotonic intervals divided by the fitting curve and the corresponding printing process influence rate, the interference anomaly index corresponding to the candidate category in the dimension to be analyzed is obtained.

2. The manufacturing data tracing and analysis method for digital transfer printing paper according to claim 1, characterized in that: The step of determining the complexity of the printing task in combination with the color distribution of pixels in the printing image includes: Calculate the information entropy corresponding to each digital transfer printing paper according to the pixel value of the pixel point in the printing image of each digital transfer printing paper; The product of the sum of the pixel values ​​of all the pixels in the printing image of each digital transfer printing paper and the information entropy is used as the color complexity of the corresponding digital transfer printing paper; The normalized result of the color complexity is used as the printing task complexity of the corresponding digital transfer printing paper.

3. The manufacturing data tracing and analysis method for digital transfer printing paper according to claim 1, characterized in that: The complexity of the printing task is used to divide all digital transfer printing papers into multiple categories, including: Based on the Euclidean distance between the printing task complexities of different digital transfer printing papers, the K-means clustering algorithm is used to cluster all the printing task complexities, and all the digital transfer printing papers corresponding to all the printing task complexities in each cluster are classified as one category. Among them, the k value when using the K-means clustering algorithm for clustering is obtained by processing using the elbow method.

4. The manufacturing data tracing and analysis method for digital transfer printing paper according to claim 1, characterized in that: The degree of discreteness of the derivative values ​​corresponding to the manufacturing data of all digital transfer printing papers in the candidate category under the dimension to be analyzed on the fitting curve includes: calculating the variance of the derivative values ​​corresponding to the manufacturing data of all digital transfer printing papers in the candidate category under the dimension to be analyzed on the fitting curve as the degree of discreteness.

5. The manufacturing data tracing and analysis method for digital transfer printing paper according to claim 1, characterized in that: The negative correlation normalization result of the printing-related uniqueness value of the dimension to be analyzed, the normalized value of the number of monotonic intervals divided by the fitting curve, and the corresponding printing process influence rate are used to obtain the interference anomaly index corresponding to the candidate category under the dimension to be analyzed, including: Calculate the difference between the constant 1 and the printing process influence rate of the candidate category under the dimension to be analyzed, and record the product of the difference and the negative correlation normalization result as the first eigenvalue; The product of the normalized value of the number of monotonic intervals divided by the fitting curve and the printing process influence rate of the candidate category under the dimension to be analyzed is recorded as the second eigenvalue; The sum of the first eigenvalue and the second eigenvalue is determined as the interference anomaly indicator corresponding to the candidate category in the dimension to be analyzed.

6. The manufacturing data tracing and analysis method for digital transfer printing paper according to claim 1, characterized in that: Using the interference anomaly index to perform anomaly detection on all manufacturing data includes: using a LOF algorithm based on the interference anomaly index to perform anomaly detection on all manufacturing data.

7. The manufacturing data tracing and analysis method for digital transfer printing paper according to claim 6, characterized in that: When the LOF algorithm is used for detection, the distance measurement method between data is defined as weighted Euclidean distance, and the normalized result of the corresponding interference anomaly index in each dimension is used as the weight to measure the distance in different dimensions.

Images