A quality rating method and system for textile yarn

Abstract

The present invention provides a quality rating method and system for textile yarns, belonging to the technical field of quality evaluation, and the method includes: detecting the textile yarns according to the detection method to obtain the corresponding detection matrix; performing numerical cluster analysis on each detection matrix to determine the cluster feature sequence; performing index mapping on the quality results corresponding to the cluster feature sequence and the quality list matched by the yarn type of the textile yarn; determining the single coefficient of each quality indicator according to the index mapping result, and updating the initial weight of the corresponding quality indicator based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn; determining the textile optimization direction of the textile yarn according to all the single coefficients and the total rating coefficient and outputting a reminder. By performing intelligent analysis and index mapping on each detection method to determine the single coefficient of the indicator and the update of the weight, intelligent rating can be performed to improve the rating accuracy, thereby improving the reliability of yarn use.

Description

Technical Field

[0001] The present invention relates to the technical field of quality evaluation, and in particular to a quality rating method and system for textile yarns. Background Art

[0002] Yarn is a kind of textile, which is processed into products of a certain fineness from various textile fibers. It is used for weaving, rope making, thread making, knitting and embroidery, etc. It is divided into short fiber yarn, continuous filament, etc.

[0003] The quality of yarn depends on the quality of the finished products subsequently made from the yarn. Generally, the yarn generation parameters are set in advance for generation and production. After production is completed, manual inspection and quality determination are carried out. Since there are certain personal factors in manual determination, there will be certain errors in the inspection and determination process, which will lead to inaccurate quality rating results and reduce the reliability of yarn use.

[0004] Therefore, the present invention provides a method and system for grading the quality of textile yarns. Summary of the invention

[0005] The present invention provides a quality rating method and system for textile yarns, which are used to determine the detection method based on the purpose to ensure the rationality of the detection basis, and subsequently determine the single coefficient of the indicator and the update of the weight by performing intelligent analysis and indicator mapping on each detection method, so as to improve the accuracy of the rating through intelligent rating, thereby improving the reliability of yarn use.

[0006] The present invention provides a method for grading the quality of textile yarns, comprising:

[0007] Step 1: Determine the yarn usage of the textile yarn, and obtain the detection method for the corresponding quality index of the textile yarn from the usage-detection comparison table;

[0008] Step 2: Detecting the textile yarn according to the detection method to obtain a corresponding detection matrix;

[0009] Step 3: Perform numerical cluster analysis on each detection matrix to determine the cluster feature sequence;

[0010] Step 4: mapping the quality result corresponding to the cluster feature sequence with the quality list matching the yarn type of the textile yarn;

[0011] Step 5: determining a single coefficient of each quality indicator according to the indicator mapping result, and updating the initial weight of the corresponding quality indicator based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn;

[0012] Step 6: According to all the single coefficients and the total rating coefficient, determine the textile optimization direction of the textile yarn and output a reminder.

[0013] Preferably, the textile yarn is detected according to the detection method to obtain a corresponding detection matrix, including:

[0014] Randomly selecting a target detection area of ​​the textile yarn, wherein there are a plurality of target detection points in the target detection area, and the target detection points are distributed in an array;

[0015] Detect each target detection point according to the detection method, obtain the detection value of the corresponding target detection point, and arrange them according to the array distribution structure to obtain an initial matrix;

[0016] When the quality indicator corresponding to the detection method is a general indicator, the corresponding initial matrix is ​​regarded as a detection matrix;

[0017] Otherwise, the target detection area is illuminated as a whole to obtain a global detection signal, and at the same time, each target detection point in the target detection area is locally illuminated to obtain a local detection signal;

[0018] Expanding a unit range around each target detection point, and extracting a first detection signal in the expanded range from the global detection signal;

[0019] Determine a signal deviation according to a first detection signal and a local detection signal of each target detection point;

[0020] When the signal deviation is less than or equal to the preset deviation, extract the signal frequencies in the first detection signal and the local detection signal, determine the frequency overlapping range and the frequency non-overlapping range, and adjust the detection value of the target detection point in combination with the standard frequency range and the signal deviation of the corresponding quality index under the yarn use to obtain a detection matrix;

[0021] When the signal deviation is greater than a preset deviation, determining the number of expansion units corresponding to the target detection point, expanding the number of units around the corresponding target detection point, and extracting a second detection signal from the global detection signal;

[0022] ;

[0023] Among them, Nd is the corresponding number of expansion units; Indicates the rounding down symbol; xp indicates the signal deviation; Indicates the preset deviation; ln indicates the logarithmic function sign;

[0024] The second detection signal, the first detection signal and the local detection signal are subjected to variance calculation, and the detection values ​​of the corresponding target detection points are adjusted to obtain a detection matrix.

[0025] Preferably, performing numerical cluster analysis on each detection matrix to determine a cluster feature sequence includes:

[0026] Determine a first value that does not participate in clustering after the corresponding detection matrix is ​​analyzed based on the numerical clustering and a first position of the first value;

[0027] Obtaining the parcel position of each first position respectively to obtain the area to be analyzed corresponding to the first position;

[0028] Determine a first number of clustering elements and a second number of non-clustering elements in the region to be analyzed, and determine whether it is necessary to increase the step size of the corresponding first position in combination with the clustering clusters involved in the value of the first number of clustering elements;

[0029] If necessary, determine the unit amount of the step size increase in the four directions of up, down, left and right of the corresponding first position, and obtain the adjustment amount of the first value corresponding to the first position, and perform distance judgment with each cluster based on the adjusted first value, and classify it into the clustering result of the corresponding cluster;

[0030] If not, determining a first difference of first values ​​of the nearest neighbor participating in the cluster corresponding to the first value of the first position, and a second difference of first values ​​of the second nearest neighbor participating in the cluster;

[0031] Classifying the first value corresponding to the first position into the clustering result of the cluster corresponding to the smaller difference selected from the first difference and the second difference;

[0032] According to the latest clustering results, the updated clustering clusters are obtained, and the cluster feature sequence is constructed.

[0033] Preferably, the quality result corresponding to the cluster feature sequence is index-mapped to a quality list matching the yarn type of the textile yarn, comprising:

[0034] Obtaining a quality list matching the yarn type from a type-list comparison table;

[0035] The cluster feature sequence is input into the quality analysis model to obtain the detection quality of each cluster, and is mapped with the quality list respectively.

[0036] Preferably, a single coefficient of each quality indicator is determined according to the indicator mapping result, including:

[0037] ;

[0038] in, represents the single coefficient of the i1th quality indicator; represents the number of clusters existing under the i1th quality index; It indicates the number of detection values ​​in the latest clustering result corresponding to the j1th cluster under the i1th quality index; Represents the detection quality and corresponding quality list of the j1th cluster under the i1th quality index The mapping quality function of is in the range of (0,1); represents the variance of the detection quality of all clusters under the i1th quality index; Indicates the corresponding quality list Best quality in.

[0039] Preferably, the initial weights of the corresponding quality indicators are updated based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn, including:

[0040] Calculate the update weights of the corresponding quality indicators;

[0041] ;

[0042] in, represents the updated weight of the i1th quality indicator; represents the initial weight of the i1th quality indicator; y01 and y02 represent the set thresholds respectively; min represents the minimum value symbol; m1 represents the total number of quality indicators;

[0043] Calculating a total rating coefficient based on the updated weights and the single coefficient;

[0044] ;

[0045] Among them, Z represents the total rating coefficient.

[0046] Preferably, determining the textile optimization direction of the textile yarn according to all the single coefficients and the total rating coefficient comprises:

[0047] Obtain an optimized analysis model that matches the yarn usage from the usage-analysis comparison table;

[0048] The single coefficient and the total rating coefficient are input into the optimization analysis model in turn to obtain the textile optimization direction.

[0049] The present invention provides a quality rating system for textile yarns, comprising:

[0050] A method determination module, used to determine the yarn use of the textile yarn and obtain a detection method for a corresponding quality index of the textile yarn from a use-detection comparison table;

[0051] A matrix building module, used for detecting the textile yarn according to the detection method to obtain a corresponding detection matrix;

[0052] A sequence acquisition module is used to perform numerical cluster analysis on each detection matrix to determine the cluster feature sequence;

[0053] An index mapping module, used for performing index mapping between the quality result corresponding to the cluster feature sequence and the quality list matching the yarn type of the textile yarn;

[0054] A coefficient determination module, used for determining a single coefficient of each quality indicator according to the indicator mapping result, and updating the initial weight of the corresponding quality indicator based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn;

[0055] The optimization module is used to determine the textile optimization direction of the textile yarn according to all the single coefficients and the total rating coefficient and output a reminder.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] The detection method is determined based on the purpose to ensure the rationality of the detection basis. Subsequently, the single coefficient of the indicator and the weight update are determined through intelligent analysis and indicator mapping of each detection method. Intelligent rating can be used to improve the accuracy of rating and thus improve the reliability of yarn use.

[0058] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0059] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0061] Figure 1 It is a flow chart of a method for quality rating of textile yarn in an embodiment of the present invention;

[0062] Figure 2 4 is a structural diagram of a quality rating system for textile yarns in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0064] The present invention provides a method for grading the quality of textile yarns, such as Figure 1 As shown, including:

[0065] Step 1: Determine the yarn usage of the textile yarn, and obtain the detection method for the corresponding quality index of the textile yarn from the usage-detection comparison table;

[0066] Step 2: Detecting the textile yarn according to the detection method to obtain a corresponding detection matrix;

[0067] Step 3: Perform numerical cluster analysis on each detection matrix to determine the cluster feature sequence;

[0068] Step 4: mapping the quality result corresponding to the cluster feature sequence with the quality list matching the yarn type of the textile yarn;

[0069] Step 5: determining a single coefficient of each quality indicator according to the indicator mapping result, and updating the initial weight of the corresponding quality indicator based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn;

[0070] Step 6: According to all the single coefficients and the total rating coefficient, determine the textile optimization direction of the textile yarn and output a reminder.

[0071] In this embodiment, the use of yarn refers to the samples that can be made after using the yarn, for example, it can be used in weaving, rope making, thread making, knitting and embroidery, and the detection methods corresponding to different applications are different.

[0072] In this embodiment, the purpose-detection comparison table is pre-set, and the corresponding detection methods are different for different purposes. For example, for weaving purposes, there are quality index 1, quality index 2, and quality index 3, and quality index 1 matches detection method 1, quality index 2 matches detection method 2, and quality index 3 matches detection method 3. It should be noted that the quality indicators involved in different purposes may be different, and the quality indicators involve twist, line density, color fastness, weight deviation, coefficient of variation, single yarn breaking strength, yarn uniformity, cotton knots, impurities, weight per 100 meters, strength, washability, friction resistance, roughness, etc. The detection methods involved in each quality indicator are existing.

[0073] In this embodiment, the detection matrix includes several detection points of the textile yarn, and the detection points under different quality indicators are different. For example, based on the detection of several detection points specified in the square area on the sample made of textile yarn, a detection matrix for the square area can be constructed. For example, there are 3 rows and 3 columns in the positive direction area, so at this time, there are 9 evenly distributed detection points.

[0074] .

[0075] In this embodiment, the values ​​of the K-means algorithm matrix are clustered and classified, and a number of clusters can be automatically determined, which belongs to the prior art.

[0076] In this embodiment, cluster feature sequence = {detection quality of several clusters under the corresponding detection matrix};

[0077] In this embodiment, the quality list is predetermined, mainly as a reference for the quality of the textile yarn actually produced, and there is a corresponding quality list under each indicator, mainly to achieve mapping comparison analysis.

[0078] In this embodiment, there is a corresponding coefficient under each quality indicator, and the comprehensive result of all quality indicators is the total rating coefficient.

[0079] In this embodiment, after the purpose is determined, the weights of the quality indicators involved are directly determined, which serve as the basis for subsequent weight updates.

[0080] In this embodiment, the textile optimization direction may be the optimization of production parameters under a certain quality index. For example, if it is determined that there are burrs on the textile yarn, the optimization direction is to use as small a coarse yarn twist as possible without increasing the number of broken ends.

[0081] The beneficial effects of the above technical solution are: determining the detection method based on the purpose to ensure the rationality of the detection basis, and subsequently determining the single coefficient of the indicator and the update of the weight by intelligent analysis and indicator mapping of each detection method, which can improve the accuracy of the rating through intelligent rating and thus improve the reliability of yarn use.

[0082] The present invention provides a method for grading the quality of textile yarns, which detects the textile yarns according to the detection method to obtain a corresponding detection matrix, including:

[0083] Randomly selecting a target detection area of ​​the textile yarn, wherein there are a plurality of target detection points in the target detection area, and the target detection points are distributed in an array;

[0084] Detect each target detection point according to the detection method, obtain the detection value of the corresponding target detection point, and arrange them according to the array distribution structure to obtain an initial matrix;

[0085] When the quality indicator corresponding to the detection method is a general indicator, the corresponding initial matrix is ​​regarded as a detection matrix;

[0086] Otherwise, the target detection area is illuminated as a whole to obtain a global detection signal, and at the same time, each target detection point in the target detection area is locally illuminated to obtain a local detection signal;

[0087] Expanding a unit range around each target detection point, and extracting a first detection signal in the expanded range from the global detection signal;

[0088] Determine a signal deviation according to a first detection signal and a local detection signal of each target detection point;

[0089] When the signal deviation is less than or equal to the preset deviation, extract the signal frequencies in the first detection signal and the local detection signal, determine the frequency overlapping range and the frequency non-overlapping range, and adjust the detection value of the target detection point in combination with the standard frequency range and the signal deviation of the corresponding quality index under the yarn use to obtain a detection matrix;

[0090] When the signal deviation is greater than a preset deviation, determining the number of expansion units corresponding to the target detection point, expanding the number of units around the corresponding target detection point, and extracting a second detection signal from the global detection signal;

[0091] ;

[0092] Among them, Nd is the corresponding number of expansion units; Indicates the rounding down symbol; xp indicates the signal deviation; Indicates the preset deviation; ln indicates the logarithmic function sign;

[0093] The second detection signal, the first detection signal and the local detection signal are subjected to variance calculation, and the detection values ​​of the corresponding target detection points are adjusted to obtain a detection matrix.

[0094] In this embodiment, the target detection area is a regular detection area. Generally, for the convenience of detection, a textile yarn surface made of several textile yarns is obtained, and a regular area, such as a square or a rectangle, is locked from the surface.

[0095] In this embodiment, the purpose of the array distribution is to evenly lock the detection points to facilitate a more comprehensive analysis of the target detection area.

[0096] In this embodiment, for example, the strength detection is to measure the tension at both ends of each detection point to determine the strength value of the point, that is, the detection values ​​are arranged in position to obtain an initial matrix.

[0097] In this embodiment, general indicators refer to twist, linear density, color fastness, weight deviation, coefficient of variation, single yarn breaking strength, yarn uniformity, neps, impurities, weight per 100 meters, strength, washability, and friction resistance, and roughness indicators are regarded as non-general indicators.

[0098] In this embodiment, during the detection based on the roughness index, detection light is emitted to the target detection area and the reflection signal based on this area is received as the global detection signal.

[0099] In this embodiment, the local illumination is obtained by detecting with a diameter of 1 cm centered on the detection point to obtain the local detection signal of this point.

[0100] In this embodiment, the unit range expansion is performed in units of the two nearest target detection points.

[0101] In this embodiment, the global detection signal includes the signals of each position point in the target detection area. Therefore, when the positions match, the detection signals in the corresponding range can be directly extracted.

[0102] In this embodiment, the signal deviation = the absolute value of the difference between the average signal value of the first detection signal of the target detection point and the average signal value of the corresponding layout detection signal.

[0103] In this embodiment, the preset deviation is set in advance.

[0104] In this embodiment, the signal frequency is obtained from the reflection signal - frequency comparison table, which includes different degrees of reflection signals and the matching signal frequencies. That is, when determining the frequency, according to this comparison table, the signal frequency of each position point can be directly determined. By statistically analyzing the signal frequencies, the frequency range of the first detection signal is (f1, f2), and the frequency range of the local detection signal is (f3, f4), where f1 < f3 < f2 < f4. At this time, the frequency overlap range is (f3, f2), and the frequency non - overlap ranges are (f1, f3) and (f2, f4). The standard frequency range under roughness is set in advance as (f01, f02).

[0105] ;

[0106] Among them, represents the range ratio determination function, with a value range of (0, 1); represents the fine - tuning factor based on the detection signal.

[0107] In this embodiment, the number of expansion units refers to expanding by a multiple of Nd centered on the detection point. For example, the original unit range expansion is from point 1 to point 2. At this time, the distance between point 1 and point 2 is the unit length. At this time, this unit length is expanded by Nd times, and the final range is the range after expansion according to the number of units.

[0108] In this embodiment, when the signal deviation is greater than the preset deviation: (1+(first variance between the first detection signal and the local detection signal+variance between the second detection signal and the local detection signal) / 2)+initial detection value=adjusted detection value.

[0109] The beneficial effect of the above technical solution is: by testing general indicators according to the corresponding testing methods, it is convenient to effectively obtain the corresponding testing indicators, and for non-general indicators, a global and local combination of methods is used to adjust the deviation of the initial detection value to improve the detection quality, providing a reliable basis for subsequent analysis of yarn quality.

[0110] The present invention provides a quality rating method for textile yarns, which performs numerical cluster analysis on each detection matrix to determine a cluster feature sequence, including:

[0111] Determine a first value that does not participate in clustering after the corresponding detection matrix is ​​analyzed based on the numerical clustering and a first position of the first value;

[0112] Obtaining the parcel position of each first position respectively to obtain the area to be analyzed corresponding to the first position;

[0113] Determine a first number of clustering elements and a second number of non-clustering elements in the region to be analyzed, and determine whether it is necessary to increase the step size of the corresponding first position in combination with the clustering clusters involved in the value of the first number of clustering elements;

[0114] If necessary, determine the unit amount of the step size increase in the four directions of up, down, left and right of the corresponding first position, and obtain the adjustment amount of the first value corresponding to the first position, and perform distance judgment with each cluster based on the adjusted first value, and classify it into the clustering result of the corresponding cluster;

[0115] If not, determining a first difference of first values ​​of the nearest neighbor participating in the cluster corresponding to the first value of the first position, and a second difference of first values ​​of the second nearest neighbor participating in the cluster;

[0116] Classifying the first value corresponding to the first position into the clustering result of the cluster corresponding to the smaller difference selected from the first difference and the second difference;

[0117] According to the latest clustering results, the updated clustering clusters are obtained, and the cluster feature sequence is constructed.

[0118] In this embodiment, the first value not involved in clustering refers to a detection value that is not in any clustering result, and the first position refers to the position of the target detection point corresponding to the first value based on the detection matrix.

[0119] In this embodiment, the package position refers to the upper neighbor, lower neighbor, left neighbor, right neighbor, upper left neighbor, lower left neighbor, upper right neighbor, and lower right neighbor of the first position. If a value exists in the corresponding neighboring position, it is retained. If not, the position is ignored, thereby forming an area to be analyzed.

[0120] In this embodiment, determining whether the step length of the corresponding first position needs to be increased includes:

[0121] ;

[0122] in, is the judgment coefficient;

[0123] When Py is greater than or equal to P0, it is determined that the step size needs to be increased, otherwise, it is determined that the step size does not need to be increased, wherein P0 is a preset coefficient.

[0124] ;

[0125] In this embodiment, the adjustment amount=the first value+the average value of the sum of the unit amounts in the four directions.

[0126] In this embodiment, the distance judgment refers to the difference between two values.

[0127] In this embodiment, the nearest neighbor and the second neighbor refer to the neighbors of the positions in the matrix.

[0128] In this embodiment, the cluster refers to a cluster that is re-determined based on the updated clustering result, generally based on the sum of the average value of all updated detection values ​​and the variance of all updated detection values ​​in the clustering result.

[0129] The beneficial effects of the above technical solution are: by locking the values ​​and positions that do not participate in clustering to lock the large analysis area, and then concentrating on the step size judgment of the area to ensure the accuracy of clustering, and when the step size needs to be increased, the unit quantity is determined from four directions to achieve effective update of the logarithmic value, thereby ensuring the rationality of the sequence and providing a reliable basis for quality rating.

[0130] The present invention provides a method for quality rating of textile yarns, which performs index mapping between the quality results corresponding to the cluster feature sequence and the quality list matching the yarn type of the textile yarn, including:

[0131] Obtaining a quality list matching the yarn type from a type-list comparison table;

[0132] The cluster feature sequence is input into the quality analysis model to obtain the detection quality of each cluster, and is mapped with the quality list respectively.

[0133] In this embodiment, the type-list comparison table includes different yarn types and corresponding quality lists, and the quality list includes corresponding quality standards under different clusters.

[0134] In this embodiment, the quality analysis model is obtained by pre-training the neural network model using the sequence and the expert's quality evaluation results of the sequence as samples, so as to facilitate the detection quality of each cluster.

[0135] In this embodiment, the mapping is to determine the quality result of the corresponding detection quality.

[0136] The beneficial effect of the above technical solution is: by matching the list from the comparison table and analyzing the sequence based on the model, effective mapping is achieved.

[0137] The present invention provides a quality rating method for textile yarns, which determines a single coefficient of each quality index according to an index mapping result, including:

[0138] ;

[0139] in, represents the single coefficient of the i1th quality indicator; represents the number of clusters existing under the i1th quality index; It indicates the number of detection values ​​in the latest clustering result corresponding to the j1th cluster under the i1th quality index; Represents the detection quality and corresponding quality list of the j1th cluster under the i1th quality index The mapping quality function of is in the range of (0,1); represents the variance of the detection quality of all clusters under the i1th quality index; Indicates the corresponding quality list Best quality in.

[0140] The beneficial effect of the above technical solution is: a single coefficient is calculated by the number of parameter clusters, the number of detection values, the mapping quality function, etc., which provides a basis for the subsequent calculation of the total rating coefficient.

[0141] The present invention provides a quality rating method for textile yarns, which updates the initial weights of corresponding quality indicators based on cluster feature sequences to obtain a total rating coefficient of the textile yarns, including:

[0142] Calculate the update weights of the corresponding quality indicators;

[0143] ;

[0144] in, represents the updated weight of the i1th quality indicator; represents the initial weight of the i1th quality indicator; y01 and y02 represent the set thresholds respectively; min represents the minimum value symbol; m1 represents the total number of quality indicators;

[0145] Calculating a total rating coefficient based on the updated weights and the single coefficient;

[0146] ;

[0147] Among them, Z represents the total rating coefficient.

[0148] The beneficial effect of the above technical solution is: by updating the weights, each indicator is effectively kept in the latest state, and then the total rating coefficient is calculated by combining the weights and coefficients.

[0149] The present invention provides a quality rating method for textile yarns, which determines the textile optimization direction of the textile yarns according to all single coefficients and the total rating coefficient, comprising:

[0150] Obtain an optimized analysis model that matches the yarn usage from the usage-analysis comparison table;

[0151] The single coefficient and the total rating coefficient are input into the optimization analysis model in turn to obtain the textile optimization direction.

[0152] In this embodiment, the usage-analysis comparison table includes different yarn usages and a pre-trained optimization analysis model, and the model is obtained by training the neural network model with single coefficients under different indicators, total rating coefficients and optimization directions as samples, so the optimization direction can be directly obtained.

[0153] The beneficial effect of the above technical solution is: through further analysis of the single coefficient and the total rating coefficient, the optimization direction is determined to ensure the quality of the textile yarn.

[0154] The present invention provides a quality rating system for textile yarns, such as Figure 2 As shown, including:

[0155] A method determination module, used to determine the yarn use of the textile yarn and obtain a detection method for each quality indicator of the textile yarn from a use-detection comparison table;

[0156] A matrix building module, used for detecting the textile yarn according to the detection method to obtain a corresponding detection matrix;

[0157] A sequence acquisition module is used to perform numerical cluster analysis on each detection matrix to determine the cluster feature sequence;

[0158] An index mapping module, used for performing index mapping between the quality result corresponding to the cluster feature sequence and the quality list matching the yarn type of the textile yarn;

[0159] A coefficient determination module, used for determining a single coefficient of each quality indicator according to the indicator mapping result, and updating the initial weight of the corresponding quality indicator based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn;

[0160] The optimization module is used to determine the textile optimization direction of the textile yarn according to all the single coefficients and the total rating coefficient and output a reminder.

[0161] The beneficial effects of the above technical solution are: determining the detection method based on the purpose to ensure the rationality of the detection basis, and subsequently determining the single coefficient of the indicator and the update of the weight by intelligent analysis and indicator mapping of each detection method, which can improve the accuracy of the rating through intelligent rating and thus improve the reliability of yarn use.

[0162] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for grading the quality of textile yarns, characterized in that: include: Step 1: Determine the yarn usage of the textile yarn, and obtain the detection method for the corresponding quality index of the textile yarn from the usage-detection comparison table; Step 2: Detecting the textile yarn according to the detection method to obtain a corresponding detection matrix; Step 3: Perform numerical cluster analysis on each detection matrix to determine the cluster feature sequence; Step 4: mapping the quality result corresponding to the cluster feature sequence with the quality list matching the yarn type of the textile yarn; Step 5: determining a single coefficient of each quality indicator according to the indicator mapping result, and updating the initial weight of the corresponding quality indicator based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn; Step 6: Determine the textile optimization direction of the textile yarn according to all the single coefficients and the total rating coefficient and output a reminder; Among them, a single coefficient of each quality indicator is determined according to the indicator mapping result, including: ; in, represents the single coefficient of the i1th quality indicator; represents the number of clusters existing under the i1th quality index; It indicates the number of detection values ​​in the latest clustering result corresponding to the j1th cluster under the i1th quality index; Represents the detection quality and corresponding quality list of the j1th cluster under the i1th quality index The mapping quality function of is in the range of (0,1); represents the variance of the detection quality of all clusters under the i1th quality index; Indicates the corresponding quality list The best quality in Wherein, the initial weights of the corresponding quality indicators are updated based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn, including: Calculate the update weights of the corresponding quality indicators; ; in, represents the updated weight of the i1th quality indicator; represents the initial weight of the i1th quality indicator; y01 and y02 represent the set thresholds respectively; min represents the minimum value symbol; m1 represents the total number of quality indicators; Calculating a total rating coefficient based on the updated weights and the single coefficient; ; Among them, Z represents the total rating coefficient.

2. The method for grading the quality of textile yarn according to claim 1, characterized in that: The textile yarn is tested according to the testing method to obtain a corresponding testing matrix, including: Randomly selecting a target detection area of ​​the textile yarn, wherein there are a plurality of target detection points in the target detection area, and the target detection points are distributed in an array; Detect each target detection point according to the detection method, obtain the detection value of the corresponding target detection point, and arrange them according to the array distribution structure to obtain an initial matrix; When the quality index corresponding to the detection method is a general index, the corresponding initial matrix is ​​regarded as a detection matrix; Otherwise, the target detection area is illuminated as a whole to obtain a global detection signal, and at the same time, each target detection point in the target detection area is locally illuminated to obtain a local detection signal; Expanding a unit range around each target detection point, and extracting a first detection signal in the expanded range from the global detection signal; Determine a signal deviation according to a first detection signal and a local detection signal of each target detection point; When the signal deviation is less than or equal to the preset deviation, extract the signal frequencies in the first detection signal and the local detection signal, determine the frequency overlapping range and the frequency non-overlapping range, and adjust the detection value of the target detection point in combination with the standard frequency range and the signal deviation of the corresponding quality index under the yarn use to obtain a detection matrix; When the signal deviation is greater than a preset deviation, determining the number of expansion units corresponding to the target detection point, expanding the number of units around the corresponding target detection point, and extracting a second detection signal from the global detection signal; ; Among them, Nd is the corresponding number of expansion units; Indicates the rounding down symbol; xp indicates the signal deviation; Yp indicates the preset deviation; ln indicates the logarithmic function symbol; The second detection signal, the first detection signal and the local detection signal are subjected to variance calculation, and the detection values ​​of the corresponding target detection points are adjusted to obtain a detection matrix.

3. The method for grading the quality of textile yarn according to claim 1, characterized in that: Numerical clustering analysis is performed on each detection matrix to determine the cluster feature sequence, including: Determine a first value that does not participate in clustering after the corresponding detection matrix is ​​analyzed based on the numerical clustering and a first position of the first value; Obtaining the parcel position of each first position respectively to obtain the area to be analyzed corresponding to the first position; Determine a first number of clustering elements and a second number of non-clustering elements in the region to be analyzed, and determine whether it is necessary to increase the step size of the corresponding first position in combination with the clustering clusters involved in the value of the first number of clustering elements; If necessary, determine the unit amount of the step size increase in the four directions of up, down, left and right of the corresponding first position, and obtain the adjustment amount of the first value corresponding to the first position, and perform distance judgment with each cluster based on the adjusted first value, and classify it into the clustering result of the corresponding cluster; If not, determining a first difference of first values ​​of the nearest neighbor participating in the cluster corresponding to the first value of the first position, and a second difference of first values ​​of the second nearest neighbor participating in the cluster; Classifying the first value corresponding to the first position into the clustering result of the cluster corresponding to the smaller difference selected from the first difference and the second difference; According to the latest clustering results, the updated clustering clusters are obtained, and the cluster feature sequence is constructed.

4. The method for grading the quality of textile yarn according to claim 1, characterized in that: The quality result corresponding to the cluster feature sequence is index-mapped with a quality list matching the yarn type of the textile yarn, including: Obtaining a quality list matching the yarn type from a type-list comparison table; The cluster feature sequence is input into the quality analysis model to obtain the detection quality of each cluster, and is mapped with the quality list respectively.

5. The method for grading the quality of textile yarn according to claim 1, characterized in that: According to all the single coefficients and the overall rating coefficient, the textile optimization direction of the textile yarn is determined, including: Obtain an optimized analysis model that matches the yarn usage from the usage-analysis comparison table; The single coefficient and the total rating coefficient are input into the optimization analysis model in turn to obtain the textile optimization direction.

6. A quality rating system for textile yarns, characterized in that: include: A method determination module, used to determine the yarn use of the textile yarn and obtain a detection method for a corresponding quality index of the textile yarn from a use-detection comparison table; A matrix building module, used for detecting the textile yarn according to the detection method to obtain a corresponding detection matrix; A sequence acquisition module is used to perform numerical cluster analysis on each detection matrix to determine the cluster feature sequence; An index mapping module, used for performing index mapping between the quality result corresponding to the cluster feature sequence and the quality list matching the yarn type of the textile yarn; A coefficient determination module, used for determining a single coefficient of each quality indicator according to the indicator mapping result, and updating the initial weight of the corresponding quality indicator based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn; An optimization module, for determining a textile optimization direction of the textile yarn according to all single coefficients and the total rating coefficient and outputting a reminder; Among them, a single coefficient of each quality indicator is determined according to the indicator mapping result, including: ; in, represents the single coefficient of the i1th quality indicator; represents the number of clusters existing under the i1th quality index; It indicates the number of detection values ​​in the latest clustering result corresponding to the j1th cluster under the i1th quality index; Represents the detection quality and corresponding quality list of the j1th cluster under the i1th quality index The mapping quality function of is in the range of (0,1); represents the variance of the detection quality of all clusters under the i1th quality index; Indicates the corresponding quality list The best quality in Wherein, the initial weights of the corresponding quality indicators are updated based on the cluster feature sequence to obtain the total rating coefficient of the textile yarn, including: Calculate the update weights of the corresponding quality indicators; ; in, represents the updated weight of the i1th quality indicator; represents the initial weight of the i1th quality indicator; y01 and y02 represent the set thresholds respectively; min represents the minimum value symbol; m1 represents the total number of quality indicators; Calculating a total rating coefficient based on the updated weights and the single coefficient; ; Among them, Z represents the total rating coefficient.

Images