Fabric multi-attribute label intelligent generation method and system

Abstract

The application discloses a kind of fabric multi-attribute label intelligent generation method and system, it is related to fabric label technical field, and its technical solution key points include the following steps: obtaining the image data and text data of fabric, image data and text data are input to database to obtain the basic attribute data of fabric, the attribute condition of fabric is obtained according to basic attribute data;Fabric attribute item is set according to attribute category and the extraction parameter and extraction weight value corresponding to fabric attribute item are set;Attribute factor is generated according to fabric attribute item;The attribute correlation of fabric attribute item is obtained, and the correlation path and correlation influence coefficient between attribute factor are generated according to attribute correlation;Attribute correlation model is constructed according to attribute factor and correlation path;Effect is that multiple attribute labels can be directly presented fabric attribute condition according to abnormality and to-be-optimized factor, help to quickly judge whether fabric meets needs.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fabric label, more particularly, it relates to a fabric multi-attribute label intelligent generation method and system. BACKGROUND

[0002] Fabric label is a key carrier for transmitting core information of fabric. The traditional fabric attribute label generation method relies on manual detection and input, which is low in efficiency and prone to errors. For example, when recording the fiber composition of fabric manually, subjective judgment or operation errors may affect subsequent decision-making. At the same time, the correlation between attributes is not mined, only single attribute data is simply listed, and complex logic cannot be reflected, so it is difficult to predict the performance of fabric after obtaining the label, for example, the label marks "good moisture absorption" but does not associate the fabric structure, thus affecting product quality. SUMMARY

[0003] In view of the deficiencies in the prior art, the purpose of the present application is to provide a fabric multi-attribute label intelligent generation method and system.

[0004] To achieve the above purpose, the present application provides the following technical scheme:

[0005] A fabric multi-attribute label intelligent generation method, the method comprising the following steps:

[0006] Obtain image data and text data of the fabric, input the image data and text data into a database to obtain basic attribute data of the fabric, and obtain attribute status of the fabric according to the basic attribute data; set fabric attribute items according to attribute categories, and set extraction parameters and extraction weights corresponding to the fabric attribute items;

[0007] Generate attribute factors according to the fabric attribute items; obtain attribute correlation of the fabric attribute items, generate correlation paths and correlation influence coefficients between the attribute factors according to the attribute correlation; and construct an attribute correlation model according to the attribute factors and the correlation paths;

[0008] Collect real-time detection data of the fabric attribute items, obtain attribute matching coefficients of the fabric attribute items according to the real-time detection data, the extraction parameters and the extraction weights; and compare and analyze the attribute matching coefficients of the fabric attribute items with a preset matching coefficient threshold to obtain a to-be-optimized sub-item;

[0009] Obtain abnormal attribute factors corresponding to the to-be-optimized sub-item; obtain correlation attribute factors of the abnormal attribute factors according to the attribute correlation model, and obtain optimization influence coefficients of the correlation attribute factors on the abnormal attribute factors according to the attribute correlation model;

[0010] Obtain a comprehensive matching coefficient according to the optimization influence coefficients and the attribute matching coefficients corresponding to the correlation attribute factors; and compare and analyze the comprehensive matching coefficient with the preset matching coefficient threshold to obtain a to-be-optimized factor;

[0011] The fabric multi-attribute label is generated according to the abnormal attribute factor and the to-be-optimized factor.

[0012] Preferably, real-time detection data of the fabric attribute item is collected, and an attribute matching coefficient of the fabric attribute item is obtained according to the real-time detection data, the extraction parameter and the extraction weight value, and specifically includes the following steps:

[0013] Real-time detection data of the fabric attribute item is obtained based on the extraction parameter corresponding to the fabric attribute item;

[0014] A parameter deviation value corresponding to the fabric attribute item is obtained according to the real-time detection data and the extraction parameter;

[0015] The parameter deviation value corresponding to the fabric attribute item is weighted and calculated according to the extraction weight value to obtain the attribute matching coefficient of the fabric attribute item.

[0016] Preferably, the attribute matching coefficient of the fabric attribute item is compared and analyzed with a preset matching coefficient threshold to obtain a to-be-optimized sub-item, and specifically includes the following steps:

[0017] If the attribute matching coefficient of the fabric attribute item is greater than or equal to the preset matching coefficient threshold, it is judged that the attribute extraction of the fabric attribute item is normal;

[0018] If the attribute matching coefficient of the fabric attribute item is less than the preset matching coefficient threshold, it is judged that the attribute extraction of the fabric attribute item is abnormal, and the fabric attribute item is recorded as a to-be-optimized sub-item.

[0019] Preferably, an associated path and an associated influence coefficient between the attribute factors are generated according to the attribute association relationship, and specifically includes the following steps:

[0020] Associated sub-items of the fabric attribute item are obtained according to the attribute association relationship, and an attribute dependency relationship and an attribute logical structure of the fabric attribute item and the associated sub-items are obtained;

[0021] A factor association mark of the attribute factor and the associated attribute factor is obtained according to the attribute dependency relationship, and an associated path between the attribute factor and the associated attribute factor is generated according to the factor association mark;

[0022] An associated influence coefficient corresponding to the associated path is obtained according to the attribute logical structure between the attribute factor and the associated attribute factor.

[0023] Preferably, an associated influence coefficient corresponding to the associated path is obtained according to the attribute logical structure between the attribute factor and the associated attribute factor, and specifically includes the following steps:

[0024] A structure feature set is set, and the structure feature set includes a preset structure type and a structure influence coefficient corresponding to the preset structure type;

[0025] The attribute logic structure type of the fabric attribute item and the associated sub-item is acquired, the attribute logic structure type is compared with a preset structure type in a structure feature set to obtain a preset structure type corresponding to the attribute logic structure type, and a structure influence coefficient corresponding to the preset structure type is recorded as an influence association coefficient;

[0026] An influence association weight is set, and an association influence coefficient corresponding to the association path is obtained according to the influence association weight and the influence association coefficient.

[0027] Preferably, an associated attribute factor of an abnormal attribute factor is obtained according to the attribute association model, and an optimization influence coefficient of the associated attribute factor of the abnormal attribute factor is obtained according to the attribute association model, specifically including the following steps:

[0028] A factor association mark of the abnormal attribute factor is obtained according to the attribute association model, a target association path of the abnormal attribute factor is obtained according to the factor association mark, and the associated attribute factor corresponding to the abnormal attribute factor is obtained according to the target association path.

[0029] An association path between the abnormal attribute factor and the associated attribute factor is acquired, and an association influence coefficient corresponding to the association path is marked as a target association coefficient;

[0030] An optimization influence coefficient of the associated attribute factor of the abnormal attribute factor is obtained according to the target association coefficient and an attribute matching coefficient corresponding to the abnormal attribute factor.

[0031] Preferably, a comprehensive matching coefficient is obtained according to the optimization influence coefficient and the attribute matching coefficient corresponding to the associated attribute factor, and the comprehensive matching coefficient is compared with a preset matching coefficient threshold to obtain a factor to be optimized, specifically including the following steps:

[0032] The optimization influence coefficient and the attribute matching coefficient corresponding to the associated attribute factor are subjected to difference calculation to obtain a comprehensive matching coefficient of the associated attribute factor.

[0033] If the comprehensive matching coefficient of the associated attribute factor is less than the preset matching coefficient threshold, it is judged that the attribute extraction of the associated attribute factor is abnormal, and the associated attribute factor is marked as the factor to be optimized.

[0034] Preferably, a fabric multi-attribute label is generated according to the abnormal attribute factor and the factor to be optimized, specifically including the following steps:

[0035] A first abnormal proportion corresponding to the abnormal attribute factor is acquired;

[0036] A second abnormal proportion corresponding to the factor to be optimized is acquired;

[0037] A first optimization weight and a second optimization weight are set, and a label comprehensive coefficient of the fabric attribute is obtained according to the first optimization weight and the first abnormal proportion, the second optimization weight and the second abnormal proportion.

[0038] The fabric multi-attribute label is generated according to the label comprehensive coefficient.

[0039] A fabric multi-attribute label intelligent generation system comprises:

[0040] An acquisition module: acquiring basic attribute data of the fabric, and obtaining attribute conditions of the fabric according to the basic attribute data; setting fabric attribute items according to attribute categories, and setting extraction parameters and extraction weights corresponding to the fabric attribute items;

[0041] A construction module: generating attribute factors according to the fabric attribute items; obtaining attribute correlation relationships of the fabric attribute items, generating correlation paths and correlation influence coefficients between the attribute factors according to the attribute correlation relationships; and constructing an attribute correlation model according to the attribute factors and the correlation paths;

[0042] A comparison module: collecting real-time detection data of the fabric attribute items, obtaining attribute matching coefficients of the fabric attribute items according to the real-time detection data, the extraction parameters and the extraction weights; and comparing and analyzing the attribute matching coefficients of the fabric attribute items with a preset matching coefficient threshold to obtain to-be-optimized sub-items;

[0043] A processing module: obtaining abnormal attribute factors corresponding to the to-be-optimized sub-items; obtaining correlation attribute factors of the abnormal attribute factors according to the attribute correlation model, and obtaining optimization influence coefficients of the correlation attribute factors of the abnormal attribute factors according to the attribute correlation model;

[0044] An analysis module: obtaining comprehensive matching coefficients according to the optimization influence coefficients and attribute matching coefficients corresponding to the correlation attribute factors; and comparing and analyzing the comprehensive matching coefficients with the preset matching coefficient threshold to obtain to-be-optimized factors;

[0045] A generation module: generating the fabric multi-attribute label according to the abnormal attribute factors and the to-be-optimized factors.

[0046] A fabric multi-attribute label intelligent generation system comprises a memory, a processor and a computer program stored on the memory and executable on the processor, and the processor implements the fabric multi-attribute label intelligent generation method when executing the program.

[0047] Compared with the prior art, the present application has the following beneficial effects:

[0048] The application can focus on describing fabric attributes by acquiring basic attribute data and setting attribute items, extracting parameters and weight values, can mine hidden correlations between fabric attributes by generating attribute factors and constructing a correlation model, can clearly present the influence path and degree, and can help optimize the process in production. Real-time detection data is collected in the abnormal detection and optimization analysis stage to calculate the attribute matching coefficient, so that attribute extraction abnormalities can be found in time. By analyzing the correlation attribute factors and optimizing the influence coefficient with the aid of the correlation model, the abnormal influence range and degree can be comprehensively evaluated, and the multi-attribute label can be generated according to the abnormal and optimized factors to intuitively present the fabric attribute status, so as to help quickly judge whether the fabric meets the needs. BRIEF DESCRIPTION OF DRAWINGS

[0049] Fig. 1 A step schematic diagram of a fabric multi-attribute label intelligent generation method is provided for the application.

[0050] Fig. 2 A module schematic diagram of a fabric multi-attribute label intelligent generation system is provided for the application.

[0051] Fig. 3 Another electronic structure schematic diagram provided by the embodiment of the application.

[0052] 610, processor; 620, communication interface; 630, memory; 640, communication bus. DETAILED DESCRIPTION

[0053] In order to make the above-mentioned purposes, features and advantages of the application more obvious and easy to understand, the specific embodiments of the application will be described in detail below with reference to the accompanying drawings.

[0054] In the following description, many specific details are set forth in order to provide a thorough understanding of the application, but the application can also be implemented in other ways different from the description, and those skilled in the art can make similar extensions without departing from the concept of the application, therefore the application is not limited to the specific embodiments disclosed below.

[0055] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor does it mean an embodiment that is separate or selectively excluded from other embodiments.

[0056] Referring to Figs. 1-3 As shown in the figure.

[0057] The embodiments further illustrate the fabric multi-attribute label intelligent generation method and system provided by the application.

[0058] The application discloses an intelligent generation method of a fabric multi-attribute label.

[0059] Image data and text data of the fabric are acquired, and the image data and the text data are input into a database to obtain basic attribute data of the fabric.

[0060] First, image data is collected by unfolding the fabric, which can be acquired by professional fabric imaging equipment such as a camera. A piece of cotton shirt fabric is photographed to obtain visual information images containing fabric texture, color distribution and surface flaws. Text data mainly includes a commodity title and commodity attributes of an e-commerce platform, production records and ingredient instructions of the fabric. The database internally pre-sets rules for fabric attribute recognition and extraction, can analyze visual features of the input image data, and identify whether the texture of the fabric image is plain or diagonal, the color number and the color gamut range. Meanwhile, after semantic judgment on the text data, cotton fiber proportion values and production manufacturer information are extracted. Therefore, the database outputs basic attribute data of the fabric, such as the material composition, physical appearance characteristics and production-related information of the cotton fabric.

[0061] Attribute conditions of the fabric are obtained according to the basic attribute data; fabric attribute items are set according to attribute categories, and extraction parameters and extraction weights corresponding to the fabric attribute items are set;

[0062] Attribute factors are generated according to the fabric attribute items; attribute correlation relationships of the fabric attribute items are acquired, and correlation paths and correlation influence coefficients between the attribute factors are generated according to the attribute correlation relationships; and an attribute correlation model is constructed according to the attribute factors and the correlation paths;

[0063] Real-time detection data of the fabric attribute items are collected, attribute matching coefficients of the fabric attribute items are obtained according to the real-time detection data, the extraction parameters and the extraction weights; and the attribute matching coefficients of the fabric attribute items are compared and analyzed with a preset matching coefficient threshold to obtain to-be-optimized sub-items;

[0064] Corresponding abnormal attribute factors of the to-be-optimized sub-items are acquired; correlation attribute factors of the abnormal attribute factors are obtained according to the attribute correlation model, and optimization influence coefficients of the correlation attribute factors of the abnormal attribute factors are obtained according to the attribute correlation model;

[0065] Comprehensive matching coefficients are obtained according to the optimization influence coefficients and attribute matching coefficients corresponding to the correlation attribute factors; and the comprehensive matching coefficients are compared and analyzed with the preset matching coefficient threshold to obtain to-be-optimized factors;

[0066] The abnormal attribute factors and the to-be-optimized factors are used to generate a fabric multi-attribute label.

[0067] The basic attribute data processing includes the material composition and fiber length of the fabric. Through these basic data, the material composition and physical properties of the fabric can be clearly understood. The fabric attributes are divided into material attributes, physical performance attributes and chemical performance attributes. Specific fabric attribute items are set for each category, such as the fiber type and blending ratio attribute items in the material attribute category. At the same time, each attribute item is configured with corresponding extraction parameters and extraction weights. The extraction parameters specify the rules of the attribute item data; the extraction weights reflect the importance of different attribute items. For example, the fiber type has a great influence on the judgment of fabric properties, so the weight is high.

[0068] According to the set fabric attribute items, corresponding attribute factors are generated, such as polyester fiber as an attribute factor for fiber type. The correlation between fabric attribute items is mined, and based on this, the correlation path between attribute factors is determined, that is, the transmission route of the interaction and influence between attribute factors. For example, the fiber type affects the fabric wear resistance, that is, there is a correlation path between them. The correlation influence coefficient is used to measure the strength of the influence, such as the correlation influence coefficient of the fiber type of the polyester fiber material fabric on the wear resistance is high, because the wear resistance of the polyester fiber itself is obvious. Through the attribute factors and the correlation path, an attribute correlation model reflecting the relationship between the fabric attributes is constructed.

[0069] Real-time detection data of fabric attribute items are collected, and attribute matching coefficients are calculated by combining extraction parameters and extraction weights. Effective information is filtered from real-time detection data using extraction parameters. Parameter deviation values are calculated by comparing real-time data and extraction parameters. Then, attribute matching coefficients are obtained by weighting parameter deviation values using extraction weights. The deviation values of attribute items with high weights have a greater impact on matching coefficients. The attribute matching coefficient is compared with the preset matching coefficient threshold. If the attribute matching coefficient is less than the preset matching coefficient threshold, the corresponding fabric attribute item is a to-be-optimized sub-item. In this way, the attribute part that may have problems can be found, such as a fabric with a tensile strength attribute matching coefficient lower than the threshold, which is marked as a to-be-optimized sub-item.

[0070] The corresponding abnormal attribute factor is located for the found to-be-optimized sub-item, such as the to-be-optimized sub-item being insufficient tensile strength, and the corresponding abnormal attribute factor being low fiber interweaving density. With the help of the constructed attribute correlation model, the associated attribute factors of the abnormal attribute factor are found, that is, other attribute factors affected by the abnormal factor, such as the low fiber interweaving density associated with the fabric hand feeling and fabric breathability attribute factors. At the same time, the optimization influence coefficient of the abnormal attribute factor on the associated attribute factor is determined, which measures the degree of the effect of optimizing the abnormal factor on the associated factor. The comprehensive matching coefficient is calculated by combining the optimization influence coefficient and the attribute matching coefficient of the associated attribute factor itself, and the to-be-optimized factor is determined, further clarifying the range of attribute factors that need to be optimized, such as the fabric hand feeling comprehensive matching coefficient being lower than the threshold, which is marked as a to-be-optimized factor.

[0071] The abnormal attribute factors and the factors to be optimized are integrated to generate a fabric multi-attribute label. The label clearly presents the problems of the fabric in which attribute factors need to be optimized, such as the abnormal attribute factor being low fiber interlacing density, and the factors to be optimized being fabric hand feeling and fabric air permeability content, thereby providing intuitive and targeted guidance for fabric quality evaluation and subsequent improvement direction.

[0072] Real-time detection data of the fabric attribute item is collected, and an attribute matching coefficient of the fabric attribute item is obtained according to the real-time detection data, the extraction parameter and the extraction weight value, and specifically includes the following steps:

[0073] Real-time detection data of the fabric attribute item is collected based on the extraction parameter corresponding to the fabric attribute item;

[0074] A parameter deviation value corresponding to the fabric attribute item is obtained according to the real-time detection data and the extraction parameter;

[0075] The parameter deviation value corresponding to the fabric attribute item is weighted and calculated according to the extraction weight value to obtain the attribute matching coefficient of the fabric attribute item.

[0076] The extraction parameter of each attribute item of the fabric is pre-set in the present application, and the extraction parameter clearly detects the compliance conditions of the attribute, such as detecting the fabric thickness attribute item, and the extraction parameter specifies that the measurement is performed in a constant temperature environment of 25°C. Real-time detection data of the fabric attribute item is collected based on these extraction parameters.

[0077] The real-time detection data is compared with the extraction parameter, if the extraction parameter requires the fabric thickness to be 0.5mm, the actual real-time detection data is 0.6mm, and the parameter deviation value is calculated to be 0.1mm through the difference between the two.

[0078] Different fabric attribute items have different importance in evaluating the overall characteristics of the fabric, such as the waterproofness attribute in the fabric of outdoor jackets has a large impact on product performance, and the extraction weight value is set to 0.8; the color degree has a relatively small impact, and the weight value is set to 0.2. The parameter deviation value and the extraction weight value of each attribute item are weighted and operated, such as the waterproofness parameter deviation value of a certain fabric is 0.1, the weight value is 0.8, the color degree parameter deviation value is 0.2, and the weight value is 0.2, and the attribute matching coefficient is calculated to be 0.12. The coefficient comprehensively reflects the degree of fit between the actual performance of the fabric attribute item and the standard requirement, and the smaller the value, the greater the standard deviation, thereby providing a quantitative basis for subsequent judgment of whether the fabric attribute needs to be optimized.

[0079] The attribute matching coefficient of the fabric attribute item is compared and analyzed with the preset matching coefficient threshold value to obtain the sub-item to be optimized, specifically including the following steps:

[0080] If the attribute matching coefficient of the fabric attribute item is greater than or equal to the preset matching coefficient threshold, it is determined that the attribute extraction of the fabric attribute item is normal.

[0081] If the attribute matching coefficient of the fabric attribute item is less than the preset matching coefficient threshold, it is determined that the attribute extraction of the fabric attribute item is abnormal, and the fabric attribute item is marked as a to-be-optimized sub-item.

[0082] The application pre-sets a matching coefficient threshold, which is determined based on the quality standard and production requirement factors of the fabric, and represents a critical value that the extraction result of the fabric attribute item meets the expectation. For example, for pure cotton fabric used to make high-end shirts, the matching coefficient threshold of the fiber neatness attribute item is set to 0.8.

[0083] After obtaining the attribute matching coefficient of the fabric attribute item, classification judgment is performed. If the attribute matching coefficient is greater than or equal to the preset matching coefficient threshold, it means that the actual extraction of the fabric attribute item is well matched with the preset standard, and it is determined that the attribute extraction is normal. For example, the attribute matching coefficient of fiber neatness is 0.85, which is greater than the matching coefficient threshold, indicating that the attribute extraction of the fabric in terms of fiber neatness meets the requirements.

[0084] If the attribute matching coefficient is less than the preset matching coefficient threshold, it indicates that the actual extraction result of the fabric attribute item deviates greatly from the preset standard, and it is determined that the attribute extraction is abnormal. For example, the attribute matching coefficient of fiber neatness is 0.7, which is less than the matching coefficient threshold, indicating that the extraction of the fiber neatness attribute of the fabric does not meet the expectation. At this time, the fiber neatness fabric attribute item is marked as a to-be-optimized sub-item, and subsequent improvement and optimization can be performed on this to-be-optimized sub-item, such as adjusting the production process parameters, to improve the accuracy and standardization of fabric attribute extraction.

[0085] According to the attribute association relationship, an association path and an association influence coefficient between attribute factors are generated, which specifically includes the following steps:

[0086] According to the attribute association relationship, the associated sub-item of the fabric attribute item is obtained, and the attribute dependency relationship and the attribute logical structure of the fabric attribute item and the associated sub-item are obtained;

[0087] According to the attribute dependency relationship, a factor association mark of the attribute factor and the associated attribute factor is obtained, and an association path between the attribute factor and the associated attribute factor is generated according to the factor association mark;

[0088] According to the attribute logical structure between the attribute factor and the associated attribute factor, an association influence coefficient corresponding to the association path is obtained.

[0089] The application determines the associated sub-items of the fabric attribute item according to the fabric attribute association relationship, and determines the attribute dependency relationship and the attribute logical structure between the fabric attribute item and the associated sub-items. For example, the associated sub-items of the moisture absorption property of the cotton fabric are fiber structure and environmental humidity adaptability. The attribute dependency relationship is that the fiber structure affects the moisture absorption, such as the porous structure of the cotton fiber determines the basic moisture absorption; the attribute logical structure refers to the logical association of the attribute items and the associated sub-items.

[0090] According to the attribute dependency relationship, the factor association mark of the attribute factor and the associated attribute factor is determined, and then the association path therebetween is generated. Taking the cotton fabric as an example, the moisture absorption corresponds to the attribute factor, and the fiber structure is the associated attribute factor. The association path from the fiber structure attribute factor to the moisture absorption attribute factor is generated according to the fiber structure dependency relationship affecting the moisture absorption, and the conduction direction of the influence between the attribute factors is clearly presented.

[0091] The associated influence coefficient corresponding to the association path is obtained according to the attribute logical structure between the attribute factor and the associated attribute factor. For example, in the logical structure of the fiber structure on the moisture absorption, if the fiber structure is tight, the promotion effect on the moisture absorption is strong, and the corresponding associated influence coefficient is high; if the fiber structure is loose, the influence is weak, and the coefficient is low. In this way, the degree of the influence between the attribute factors is quantified, which provides data support for subsequent analysis of the fabric attribute influence based on the association relationship.

[0092] The associated influence coefficient corresponding to the association path is obtained according to the attribute logical structure between the attribute factor and the associated attribute factor, which specifically includes the following steps:

[0093] Setting a structure feature set, the structure feature set including a preset structure type and a structure influence coefficient corresponding to the preset structure type;

[0094] Obtaining the attribute logical structure type of the fabric attribute item and the associated sub-item, comparing the attribute logical structure type with the preset structure type in the structure feature set to obtain the preset structure type corresponding to the attribute logical structure type, and recording the structure influence coefficient corresponding to the preset structure type as the influence association coefficient;

[0095] Setting an influence association weight; obtaining the associated influence coefficient corresponding to the association path according to the influence association weight and the influence association coefficient.

[0096] The structure feature set includes the preset structure type and the structure influence coefficient corresponding to the preset structure type. For example, for the fabric attribute logical structure, the preset linear conduction type and the mesh interaction type structure type are set, the linear conduction type structure influence coefficient is set to 0.6, which means that the attribute influence degree in this logical structure is medium; the mesh interaction type structure influence coefficient is set to 0.8, which means that the mutual influence degree between the attributes is higher.

[0097] The attribute logic structure type of the fabric attribute item and the associated sub-item is obtained, which is compared with the preset structure type in the structure feature set to determine the corresponding preset structure type, and then the influence correlation coefficient is obtained. For example, the dyeing fastness of a certain fabric is an attribute item, and the dye stability and fabric fiber are structure associated sub-items. The attribute logic structure of the attribute item and the structure associated sub-item is a mesh interaction type, which is matched with the preset mesh interaction type in the structure feature set. The structure influence coefficient 0.8 corresponding to the preset structure type is used as the influence correlation coefficient, which is used to measure the basic influence degree of the attribute correlation under this logic structure.

[0098] For example, if the production focus is on the attribute, different influence correlation weights are set for different logic structures. The influence correlation weight and the influence correlation coefficient are weighted to obtain the correlation influence coefficient corresponding to the correlation path. Assuming that the influence correlation weight of the mesh interaction type logic structure is 1.2, then the correlation influence coefficient is 0.8x1.2=0.96, which comprehensively reflects the degree of influence of the correlation path on the attribute under the attribute logic structure factor.

[0099] According to the attribute correlation model, the correlation attribute factor of the abnormal attribute factor is obtained, and the optimization influence coefficient of the abnormal attribute factor on the correlation attribute factor is obtained according to the attribute correlation model. The specific steps include the following steps:

[0100] According to the attribute correlation model, the factor correlation mark of the abnormal attribute factor is obtained, the target correlation path of the abnormal attribute factor is obtained according to the factor correlation mark, and the correlation attribute factor corresponding to the abnormal attribute factor is obtained according to the target correlation path;

[0101] The correlation path between the abnormal attribute factor and the correlation attribute factor is obtained, and the correlation influence coefficient corresponding to the correlation path is marked as the target correlation coefficient;

[0102] According to the target correlation coefficient and the attribute matching coefficient corresponding to the abnormal attribute factor, the optimization influence coefficient of the abnormal attribute factor on the correlation attribute factor is obtained.

[0103] With the help of the pre-constructed attribute correlation model, the factor correlation mark of the abnormal attribute factor is obtained, and such mark is an identification of the correlation characteristics of the abnormal attribute factor and other attribute factors. Based on this mark, the target correlation path of the abnormal attribute factor is combed, that is, the specific transmission route of the abnormal attribute factor affecting other attribute factors. According to the target correlation path, the correlation attribute factor corresponding to the abnormal attribute factor is locked, and the object affected by the abnormal factor is determined. For example, in wool fabric, fiber embrittlement is an abnormal attribute factor. Through the attribute correlation model, the factor correlation mark points to the correlation between fiber toughness and fabric wear resistance, the target correlation path is fiber embrittlement, fiber toughness reduction and fabric wear resistance reduction, and the corresponding correlation attribute factors are fiber toughness and fabric wear resistance.

[0104] After the correlation path between the abnormal attribute factor and the correlation attribute factor is obtained, the correlation influence coefficient corresponding to the path is extracted and marked as a target correlation coefficient. The target correlation coefficient is used to measure the basic degree of influence of the correlation path on the attribute. Taking wool fabric as an example, if the correlation influence coefficient of the fiber brittleness to fiber toughness correlation path is 0.7, it represents the basic strength of the influence of fiber brittleness on fiber toughness.

[0105] The target correlation coefficient is combined with the attribute matching coefficient of the abnormal attribute factor itself to obtain an optimized influence coefficient. The attribute matching coefficient reflects the degree of fit between the actual performance of the abnormal attribute factor and the standard. If the attribute matching coefficient of the fiber brittleness abnormal attribute factor is 0.3 and the target correlation coefficient is 0.7, then the optimized influence coefficient is obtained by operating the target correlation coefficient and the attribute matching coefficient, that is, 0.7 x (1-0.3) = 0.49. It is used to quantify the optimization influence degree of the abnormal factor on the correlation attribute factor.

[0106] According to the optimized influence coefficient and the attribute matching coefficient corresponding to the correlation attribute factor, a comprehensive matching coefficient is obtained; the comprehensive matching coefficient is compared with a preset matching coefficient threshold to obtain a factor to be optimized, which specifically includes the following steps:

[0107] The optimized influence coefficient and the attribute matching coefficient corresponding to the correlation attribute factor are difference calculated to obtain a comprehensive matching coefficient of the correlation attribute factor;

[0108] If the comprehensive matching coefficient of the correlation attribute factor is less than the preset matching coefficient threshold, it is judged that the attribute extraction of the correlation attribute factor is abnormal, and the correlation attribute factor is marked as a factor to be optimized.

[0109] The optimized influence coefficient reflects the optimization effect degree of the abnormal attribute factor on the correlation attribute factor, and the attribute matching coefficient reflects the fit between the correlation attribute factor itself and the standard. Taking cotton fabric as an example, assuming that high fiber impurity content is an abnormal attribute factor, its optimized influence coefficient on the fabric cleanliness correlation attribute factor is 0.3, and the original attribute matching coefficient of the fabric cleanliness is 0.6. Difference calculation of the two obtains the comprehensive matching coefficient of the fabric cleanliness correlation attribute factor is 0.3, which comprehensively considers the optimization influence of the abnormal factor and the matching situation of itself, and presents the actual matching state of the correlation attribute factor after being affected by the abnormality.

[0110] The comprehensive matching coefficient of the associated attribute factor is compared with a preset matching coefficient threshold. If the comprehensive matching coefficient is less than the threshold, it indicates that the actual performance of the associated attribute factor after being affected by the abnormal attribute factor has a large standard deviation, and it is judged that the attribute extraction is abnormal. The associated attribute factor is marked as a factor to be optimized. Taking cotton fabric as an example, the preset fabric cleanliness matching coefficient threshold is 0.5, and the comprehensive matching coefficient is 0.3. Therefore, the comprehensive matching coefficient is less than the preset fabric cleanliness matching coefficient threshold, and the fabric cleanliness associated attribute factor is marked as a factor to be optimized. Subsequent improvement measures are taken for these factors to be optimized, such as optimizing the fabric impurity removal process to improve the matching degree of the associated attribute factor, thereby improving the overall attribute performance of the fabric.

[0111] A fabric multi-attribute label is generated according to the abnormal attribute factor and the factor to be optimized, specifically including the following steps:

[0112] A first abnormal proportion corresponding to the abnormal attribute factor is obtained.

[0113] A second abnormal proportion corresponding to the factor to be optimized is obtained.

[0114] A first optimization weight and a second optimization weight are set, and a label comprehensive coefficient of the fabric attribute is obtained according to the first optimization weight and the first abnormal proportion, the second optimization weight and the second abnormal proportion.

[0115] A fabric multi-attribute label is generated according to the label comprehensive coefficient.

[0116] The first abnormal proportion corresponding to the abnormal attribute factor and the second abnormal proportion corresponding to the factor to be optimized are obtained. For example, in cotton fabric, the abnormal attribute factor is insufficient fiber strength, and the first abnormal proportion of this abnormality in fabric attribute judgment is 30% after statistics; the factor to be optimized is fabric color fastness that does not meet the standard, and the second abnormal proportion in the judgment is 20%. These proportions reflect the influence degree of different abnormal factors in fabric attribute problems.

[0117] The first optimization weight and the second optimization weight are set according to actual needs. If more attention is paid to the abnormality of the basic attribute, such as fiber strength, the first optimization weight is set to 0.6; the second optimization weight of the factor to be optimized (such as color fastness) is set to 0.4. Then, the label comprehensive coefficient of the fabric attribute is calculated according to the weight and the abnormal proportion, that is, the first optimization weight x the first abnormal proportion + the second optimization weight x the second abnormal proportion, so that the label comprehensive coefficient of the fabric attribute is 0.26. This coefficient comprehensively reflects the overall situation of the fabric attribute abnormality.

[0118] The fabric multi-attribute label is generated according to the calculated label comprehensive coefficient. For example, the label corresponding to the label comprehensive coefficient 0.26 can be marked as the attribute abnormality that the fabric has insufficient fiber strength and substandard color fastness. Such a label can clearly present the key factors and comprehensive influence of the fabric attribute abnormality, and provide an intuitive basis for fabric quality evaluation and improvement, facilitating production personnel and quality inspection personnel to quickly understand the fabric attribute problems and then carry out targeted optimization work.

[0119] A fabric multi-attribute label intelligent generation system comprises:

[0120] An acquisition module: acquiring basic attribute data of a fabric, obtaining attribute conditions of the fabric according to the basic attribute data; setting fabric attribute items according to attribute categories and setting extraction parameters and extraction weights corresponding to the fabric attribute items;

[0121] A construction module: generating attribute factors according to the fabric attribute items; obtaining attribute association relationships of the fabric attribute items, generating association paths and association influence coefficients between the attribute factors according to the attribute association relationships; and constructing an attribute association model according to the attribute factors and the association paths;

[0122] A comparison module: collecting real-time detection data of the fabric attribute items, obtaining attribute matching coefficients of the fabric attribute items according to the real-time detection data, the extraction parameters and the extraction weights; and comparing and analyzing the attribute matching coefficients of the fabric attribute items with a preset matching coefficient threshold to obtain to-be-optimized sub-items;

[0123] A processing module: obtaining abnormal attribute factors corresponding to the to-be-optimized sub-items; obtaining associated attribute factors of the abnormal attribute factors according to the attribute association model, and obtaining optimization influence coefficients of the associated attribute factors of the abnormal attribute factors according to the attribute association model;

[0124] An analysis module: obtaining a comprehensive matching coefficient according to the optimization influence coefficients and attribute matching coefficients corresponding to the associated attribute factors; and comparing and analyzing the comprehensive matching coefficient with the preset matching coefficient threshold to obtain to-be-optimized factors;

[0125] A generation module: generating a fabric multi-attribute label according to the abnormal attribute factors and the to-be-optimized factors.

[0126] A fabric multi-attribute label intelligent generation system comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements a fabric multi-attribute label intelligent generation method when executing the program.

[0127] As Fig. 3As shown, the fabric multi-attribute label intelligent generation system includes a processor 610, a communication interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communication interface 620 and the memory 630 complete mutual communication through the communication bus 640. The processor 610 can call the logical instructions in the memory 630 to execute a fabric multi-attribute label intelligent generation method.

[0128] In addition, the logical instructions in the memory 630 described above can be realized in the form of a software functional unit and sold or used as an independent product, and can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and various media that can store program codes.

[0129] On the other hand, the present application also provides a computer program product, which includes a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program is executed by a processor, so that the computer can execute a fabric multi-attribute label intelligent generation method.

[0130] In yet another aspect, the present application also provides a non-transitory computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement a fabric multi-attribute label intelligent generation method.

[0131] The device embodiments described above are only schematic, wherein the units shown as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e. they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the present embodiment. Those skilled in the art can understand and implement it without creative labor.

[0132] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software with the necessary general hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0133] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for intelligent generation of multi-attribute labels of fabrics, characterized in that, The method comprises the following steps: Obtaining image data and text data of the fabric, inputting the image data and the text data into a database to obtain basic attribute data of the fabric, and obtaining attribute conditions of the fabric according to the basic attribute data; setting a fabric attribute item according to an attribute category, and setting extraction parameters and extraction weights corresponding to the fabric attribute item; Generating an attribute factor according to the fabric attribute item; obtaining an attribute correlation relationship of the fabric attribute item, and generating a correlation path and a correlation influence coefficient between attribute factors according to the attribute correlation relationship; and constructing an attribute correlation model according to the attribute factors and the correlation path; Collecting real-time detection data of the fabric attribute item, and obtaining an attribute matching coefficient of the fabric attribute item according to the real-time detection data, the extraction parameters and the extraction weights; and comparing and analyzing the attribute matching coefficient of the fabric attribute item with a preset matching coefficient threshold to obtain a to-be-optimized sub-item; Obtaining an abnormal attribute factor corresponding to the to-be-optimized sub-item; obtaining a correlation attribute factor of the abnormal attribute factor according to the attribute correlation model, and obtaining an optimization influence coefficient of the abnormal attribute factor on the correlation attribute factor according to the attribute correlation model; Obtaining a comprehensive matching coefficient according to the optimization influence coefficient and the attribute matching coefficient corresponding to the correlation attribute factor; and comparing and analyzing the comprehensive matching coefficient with the preset matching coefficient threshold to obtain a to-be-optimized factor; Generating a fabric multi-attribute label according to the abnormal attribute factor and the to-be-optimized factor.

2. The method of claim 1, wherein, Collecting real-time detection data of the fabric attribute item, and obtaining an attribute matching coefficient of the fabric attribute item according to the real-time detection data, the extraction parameters and the extraction weights, specifically comprising the following steps: Obtaining real-time detection data of the fabric attribute item based on the extraction parameters corresponding to the fabric attribute item; Obtaining a parameter deviation value corresponding to the fabric attribute item according to the real-time detection data and the extraction parameters; Weighting and calculating the parameter deviation value corresponding to the fabric attribute item according to the extraction weights to obtain the attribute matching coefficient of the fabric attribute item.

3. The method of claim 2, wherein, Comparing and analyzing the attribute matching coefficient of the fabric attribute item with a preset matching coefficient threshold to obtain a to-be-optimized sub-item, specifically comprising the following steps: If the attribute matching coefficient of the fabric attribute item is greater than or equal to the preset matching coefficient threshold, it is judged that the attribute extraction of the fabric attribute item is normal; If the attribute matching coefficient of the fabric attribute item is less than the preset matching coefficient threshold, it is judged that the attribute extraction of the fabric attribute item is abnormal, and the fabric attribute item is recorded as a to-be-optimized sub-item.

4. The method of claim 3, wherein, Generating a correlation path and a correlation influence coefficient between attribute factors according to an attribute correlation relationship, specifically comprising the following steps: Obtaining a correlation sub-item of the fabric attribute item according to the attribute correlation relationship, and obtaining an attribute dependency relationship and an attribute logical structure of the fabric attribute item and the correlation sub-item; Obtaining a factor correlation mark between the attribute factor and the correlation attribute factor according to the attribute dependency relationship, and generating a correlation path between the attribute factor and the correlation attribute factor according to the factor correlation mark; Obtaining a correlation influence coefficient corresponding to the correlation path according to the attribute logical structure between the attribute factor and the correlation attribute factor.

5. The method of claim 4, wherein, Obtaining a correlation influence coefficient corresponding to the correlation path according to the attribute logical structure between the attribute factor and the correlation attribute factor, specifically comprising the following steps: Setting a structure feature set, wherein the structure feature set comprises a preset structure type and a structure influence coefficient corresponding to the preset structure type; An attribute logical structure type of the fabric attribute item and the associated sub-item is acquired, the attribute logical structure type is compared with a preset structure type in a structure feature set to obtain a preset structure type corresponding to the attribute logical structure type, and a structure influence coefficient corresponding to the preset structure type is recorded as an influence association coefficient; An influence association weight is set, and an association influence coefficient corresponding to the association path is obtained according to the influence association weight and the influence association coefficient.

6. The method of claim 5, wherein, An associated attribute factor of an abnormal attribute factor is obtained according to the attribute association model, and an optimization influence coefficient of the abnormal attribute factor on the associated attribute factor is obtained according to the attribute association model, specifically including the following steps: A factor association mark of the abnormal attribute factor is obtained according to the attribute association model, a target association path of the abnormal attribute factor is obtained according to the factor association mark, and the associated attribute factor corresponding to the abnormal attribute factor is obtained according to the target association path; An association path between the abnormal attribute factor and the associated attribute factor is acquired, and an association influence coefficient corresponding to the association path is marked as a target association coefficient; An optimization influence coefficient of the abnormal attribute factor on the associated attribute factor is obtained according to the target association coefficient and an attribute matching coefficient corresponding to the abnormal attribute factor.

7. The method of claim 6, wherein, A comprehensive matching coefficient is obtained according to the optimization influence coefficient and the attribute matching coefficient corresponding to the associated attribute factor; the comprehensive matching coefficient is compared with a preset matching coefficient threshold to obtain a factor to be optimized, specifically including the following steps: The optimization influence coefficient and the attribute matching coefficient corresponding to the associated attribute factor are subjected to difference calculation to obtain a comprehensive matching coefficient of the associated attribute factor; If the comprehensive matching coefficient of the associated attribute factor is less than the preset matching coefficient threshold, it is judged that the attribute extraction of the associated attribute factor is abnormal, and the associated attribute factor is marked as the factor to be optimized.

8. The method of claim 7, wherein, A fabric multi-attribute label is generated according to the abnormal attribute factor and the factor to be optimized, specifically including the following steps: A first abnormal proportion corresponding to the abnormal attribute factor is acquired; A second abnormal proportion corresponding to the factor to be optimized is acquired; A first optimization weight and a second optimization weight are set, and a label comprehensive coefficient of the fabric attribute is obtained according to the first optimization weight and the first abnormal proportion, the second optimization weight and the second abnormal proportion; The fabric multi-attribute label is generated according to the label comprehensive coefficient.

9. A fabric multi-attribute label intelligent generation system applied to the fabric multi-attribute label intelligent generation method of any one of claims 1 to 8, characterized in that, It includes: An acquisition module: acquiring basic attribute data of a fabric, and obtaining an attribute status of the fabric according to the basic attribute data; Fabric attribute items are set according to attribute categories, and extraction parameters and extraction weights corresponding to the fabric attribute items are set; A construction module: generating attribute factors according to fabric attribute items; acquiring attribute association relationships of the fabric attribute items, generating association paths and association influence coefficients between the attribute factors according to the attribute association relationships; and constructing an attribute association model according to the attribute factors and the association paths; A comparison module: collecting real-time detection data of the fabric attribute items, obtaining attribute matching coefficients of the fabric attribute items according to the real-time detection data, the extraction parameters and the extraction weights; and comparing the attribute matching coefficients of the fabric attribute items with a preset matching coefficient threshold to obtain factors to be optimized; A processing module: acquiring abnormal attribute factors corresponding to the factors to be optimized; According to the attribute correlation model, an abnormal attribute factor is obtained, and an optimization influence coefficient of the abnormal attribute factor on the correlation attribute factor is obtained according to the attribute correlation model; The analysis module: according to the optimization influence coefficient and the attribute matching coefficient corresponding to the correlation attribute factor, a comprehensive matching coefficient is obtained; the comprehensive matching coefficient is compared with a preset matching coefficient threshold to obtain a to-be-optimized factor; The generation module: according to the abnormal attribute factor and the to-be-optimized factor, a fabric multi-attribute label is generated. 10.The intelligent generation system of fabric multi-attribute label according to claim 9, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein, The processor implements the fabric multi-attribute label intelligent generation method according to any one of claims 1 to 8 when executing the program.

Images