Method and system for generating pattern file of hosiery knitter and medium

Abstract

The invention relates to the technical field of pattern file generation of hosiery knitters, in particular to a pattern file generation method and system of a hosiery knitter and a medium. Comprising the following steps: firstly, learning a pattern image by adopting a pattern parameter learning model; wherein the pattern image in the pattern parameter learning model is divided into a plurality of image layers; feature extraction is carried out on different layers, pattern related information is obtained from different dimensions, and the obtained feature information is converted into pattern parameters; full learning of patterns is realized; secondly, carrying out pattern parameter optimization by utilizing the pattern parameter optimization model; the pattern parameter optimization model analyzes matching of different types of socks and patterns, so that the purpose of optimizing pattern parameters is achieved; the pattern can be adjusted according to actual sock production; and finally, monitoring the socks and patterns in the sock production process through a feedback mechanism, and dynamically adjusting parameters in real time through a method of calculating the error amount.

Description

Technical Field

[0001] The invention relates to the technical field of pattern file generation for a hosiery knitting machine, and in particular to a pattern file generation method, system and medium for a hosiery knitting machine. Background Art

[0002] With the rapid development of science and technology and economy, people have higher and higher requirements for the quality of life, which is reflected in many aspects, including the comfort, functionality and fashion of clothing.

[0003] The development of socks reflects people's increasing demand for personalization. Whether in terms of materials, functions or designs, modern socks are constantly innovating to meet the needs of different consumers. In the production and manufacturing of socks, designers use various patterns and patterns to meet the aesthetic needs of different consumers; Nowadays, in order to meet people's needs, sock manufacturers will print various patterns on socks. The patterns on socks are generally produced by pattern files, and then the pattern files corresponding to the sock patterns to be produced are input into the sock knitting machine, and the sock knitting machine produces products according to various parameters of the pattern files.

[0004] The method of generating pattern files is mainly to store different types of patterns in a specific format. When socks need to be produced, the corresponding pattern parameters need to be set; if the sock type changes, the pattern parameters need to be reset. This method is not conducive to the efficient production of socks.

[0005] To this end, the present invention provides a method, system and medium for generating a pattern file of a hosiery machine. Summary of the invention

[0006] The purpose of the present invention is to provide a pattern file generation method, system and medium for a hosiery knitting machine, which are used for the automatic control of sock patterns; the method comprises: firstly, the present invention proposes a pattern parameter learning model for learning pattern feature parameters; the model comprises: a pattern receiving layer, a pattern stratification processing layer, an information marking layer, a feature extraction layer and a pattern parameter output layer; wherein, the pattern stratification processing layer is used to stratify the designed pattern patterns so as to have a more detailed understanding of the pattern; the feature extraction layer is used to perform feature learning on different pattern layer images so as to obtain color features, texture features, shape features, pattern features, detail features and depth features related to the pattern; secondly, a pattern parameter optimization model is proposed, which is used to combine the specific production The socks pattern is optimized; it includes: information acquisition layer, information processing layer, parameter analysis layer, parameter optimization layer and parameter output layer; among them, various information of the socks to be knitted is obtained through the information acquisition layer; the matching degree between the pattern and the socks to be knitted is analyzed by the parameter analysis layer to obtain the parameters that need to be optimized; and the parameter optimization layer is used to perform optimization operations to obtain the pattern parameters that are suitable for the socks to be knitted; finally, in order to meet the standard production requirements, a feedback mechanism is proposed in the production process; the mechanism can be dynamically adjusted according to the real-time monitoring of the pattern during the socks production process; the pattern after the production task is completed and the corresponding sock type are associated with parameters to generate a pattern file for reference of subsequent production materials.

[0007] To achieve the above object, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for generating a pattern file of a socks knitting machine, comprising: Get the sketch of the flower pattern design; Further, scanning the flower pattern design sketch to obtain a flower pattern design image; Furthermore, the flower design image is preprocessed to obtain a standard flower design image; wherein the preprocessing includes: enhancing the clarity of the flower design image to obtain a first flower design image; removing the background of the first flower design image to obtain a second flower design image; and standardizing the second flower design image to obtain the standard flower design image.

[0008] Further, the standard flower pattern design image is input into a flower pattern parameter learning model to obtain a flower pattern parameter set; Wherein, the pattern parameter learning model includes: A pattern receiving layer, used for receiving the standard pattern design image; The pattern layering processing layer is used to layer the standard pattern design image to obtain a pattern layering image set; wherein the pattern layering image set includes: a color layer image, a texture layer image, a shape layer image, a pattern layer image, a detail layer image and a depth layer image; wherein. The pattern layering processing layer includes: Performing color space conversion on the standard flower pattern design image to obtain a first standard flower pattern design image; analyzing various colors in the first standard flower pattern design image and setting a color threshold; segmenting various regions of the first standard flower pattern design image according to the color threshold; and composing the various regions into the color layer image; Extracting texture features of the standard pattern design image by using a filter; mapping the texture features to obtain a texture feature map; separating the texture feature map according to texture type to obtain a plurality of texture separation features; and composing the plurality of texture separation features into the texture layer image; Performing edge detection on the standard flower pattern design image to obtain a second standard flower pattern design image; extracting multiple shapes from the second standard flower pattern design image; classifying the multiple shapes and forming the shape layer image; Identify the specific pattern of the standard pattern design image; segment the specific pattern to obtain a main pattern and an auxiliary pattern; classify the main pattern and the auxiliary pattern to form the pattern layer image; Extracting detail features of the standard pattern design image; performing feature enhancement on the detail features to obtain detail enhancement features; classifying the detail enhancement features and forming the detail layer image; Extracting the depth information of each element in the standard pattern design image; dividing each element into foreground, middle ground and background according to the depth information, and forming the depth layer image.

[0009] An information marking layer, used for marking the stratification result of the pattern stratification processing layer and the pattern at corresponding positions; A feature extraction layer, used for extracting features from the image in the flower pattern layered image to obtain a flower pattern feature set; A pattern parameter generation layer, used for parameterizing the features in the pattern feature set to obtain pattern parameters; Wherein, the flower pattern parameter generation layer comprises: normalizing the features in the flower pattern feature set to obtain a normalized flower pattern feature set; Further, feature screening is performed on the normalized flower pattern feature set to obtain a screened flower pattern feature set; Further, feature encoding is performed on the screening flower pattern feature set to obtain a coded flower pattern feature set; Further, parameters of the features in the encoded pattern feature set are converted to obtain a plurality of pattern parameters; The pattern parameter output layer is used to output the pattern parameters and integrate them into the pattern parameter set.

[0010] Further, initial hosiery machine pattern parameters are generated through the pattern parameter set; Furthermore, the initial hosiery machine pattern parameters are optimized using a pattern optimization model to obtain optimized hosiery machine pattern parameters; wherein the pattern parameter optimization model includes: An information acquisition layer, used to acquire a sock code of the socks to be knitted, and acquire sock information according to the sock code; wherein the sock information includes: sock type, material information, size specification, shape information and design requirements; An information processing layer is used to process the socks information to obtain standardized socks data; The parameter analysis layer is used for analyzing the matching degree between the standardized sock data and the initial sock knitting machine pattern parameters, including: setting pattern evaluation criteria: pattern completeness, pattern position conformity, and pattern deformation; inputting the standardized sock data and the initial sock knitting machine pattern parameters into a parameter matching model; and performing evaluation according to the matching results of the parameter matching model; A parameter optimization layer, used for optimizing the initial hosiery machine pattern parameters according to the evaluation results; The parameter optimization layer includes: obtaining pattern requirements from the design requirements of the socks to be knitted; wherein the pattern requirements include: pattern type, pattern size and pattern position; Further, calculating the pattern error according to the matching result; Further, analyzing the cause of the error according to the pattern error; Further, the cause of the error is quantified and a quantitative index is established; Furthermore, the initial hosiery machine pattern parameters are optimized according to the quantitative index to obtain the optimized hosiery machine pattern parameters.

[0011] A parameter output layer, used for outputting the optimized hosiery machine pattern parameters; Further, a sock knitting task is performed according to the optimized pattern parameters of the sock knitting machine, and the optimized pattern parameters of the sock knitting machine are adjusted in real time; wherein, a feedback mechanism is used to adjust the optimized pattern parameters of the sock knitting machine in real time; the feedback mechanism includes: Get the current operating parameters of the sock knitting machine and the current fabric image of the sock knitting task; Further, the current sock knitting operation parameter is quantified to generate a current operation parameter vector; Further, the current fabric image is processed and segmented to obtain a current sock image and a current pattern image; Further, the current sock image and the current pattern image are extracted respectively to obtain current sock features and current pattern features; Further, the current sock feature and the current pattern feature are combined, and differences are identified according to the design requirements; Further, generating a difference vector according to the result of the difference identification; Further, a correlation analysis is performed on the difference vector and the current operation parameter vector to obtain a weaving correlation matrix; Further, the current operating parameter vector is subjected to parameter correction according to the weaving correlation matrix to obtain a corrected parameter vector; Furthermore, the correction parameter vector is applied to the hosiery machine as the adjusted pattern parameter of the hosiery machine and is continuously monitored.

[0012] Furthermore, after the sock knitting task is completed, the adjusted parameters are generated into a pattern parameter file and saved.

[0013] In a second aspect, the present invention provides a pattern file generation system for a hosiery machine, the system comprising: An input unit, for receiving a pattern design sketch; A preprocessing unit, used for processing the pattern design sketch to obtain a standard pattern design image; The pattern parameter learning unit is used to learn the characteristic parameters of the input pattern; the pattern parameter learning unit adopts a pattern parameter learning model, including: A pattern receiving layer, used for receiving a standard pattern design image; A pattern layering processing layer is used to layer the standard pattern design image to obtain a pattern layering image set; wherein the pattern layering image set includes: a color layer image, a texture layer image, a shape layer image, a pattern layer image, a detail layer image and a depth layer image; An information marking layer, used for marking the stratification result of the pattern stratification processing layer and the pattern at corresponding positions; A feature extraction layer, used for extracting features from the image in the flower pattern layered image to obtain a flower pattern feature set; The flower pattern parameter generation layer is used to parameterize the features in the flower pattern feature set to obtain flower pattern parameters; wherein the flower pattern parameter generation layer includes: normalizing the features in the flower pattern feature set to obtain a normalized flower pattern feature set; performing feature screening on the normalized flower pattern feature set to obtain a screened flower pattern feature set; performing feature encoding on the screened flower pattern feature set to obtain an encoded flower pattern feature set; performing parameter conversion on the features in the encoded flower pattern feature set to obtain multiple flower pattern parameters; The pattern parameter output layer is used to output the pattern parameters and integrate them into a pattern parameter set.

[0014] The pattern parameter optimization unit is used to optimize the initial pattern parameters of the sock knitting machine for the socks to be knitted; wherein the pattern parameter optimization unit uses the pattern parameter optimization model for optimization, including: An information acquisition layer, used to acquire a sock code of the socks to be knitted, and acquire sock information according to the sock code; wherein the sock information includes: sock type, material information, size specification, shape information and design requirements; An information processing layer is used to process the socks information to obtain standardized socks data; The parameter analysis layer is used for analyzing the matching degree between the standardized sock data and the initial hosiery machine pattern parameters, including: setting pattern evaluation criteria, including: pattern completeness, pattern position conformity, and pattern deformation; inputting the standardized sock data and the initial hosiery machine pattern parameters into a parameter matching model; and performing evaluation according to the matching results of the parameter matching model; A parameter optimization layer, used for optimizing the initial hosiery machine pattern parameters according to the evaluation results; The parameter output layer is used to output the optimized pattern parameters of the hosiery machine.

[0015] Adjustment unit, used for real-time adjustment of pattern parameters during production; A file generation unit is used to generate and save corresponding files after the production task is completed; Task unit, used to publish production tasks.

[0016] In a third aspect, the present invention proposes a pattern file generation medium for a hosiery machine, wherein the medium may be a computer program executed by an electronic device, and when the program runs on the electronic device, the electronic device executes the pattern file generation for the hosiery machine implementing any one of the first aspect or the second aspect.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention proposes a pattern parameter learning model for feature learning of various patterns; the model includes: a pattern receiving layer, a pattern receiving layer, a pattern stratification processing layer, a pattern stratification processing layer, an information marking layer, a feature extraction layer, a pattern parameter generation layer and a pattern parameter output layer; wherein, in the pattern stratification processing layer, the pattern is divided into multiple layers, including: a color layer image, a texture layer image, a shape layer image, a pattern layer image, a detail layer image and a depth layer image; in this way, the pattern can be understood more carefully from multiple dimensions; the features of different layers are obtained by using the feature extraction layer; and the obtained multiple features are learned through the pattern parameter generation layer to obtain the initial pattern parameters; the above use of the pattern parameter learning model not only helps to capture the complexity and subtle differences of the pattern, improve the understanding of the pattern, but also further improve the production efficiency of socks.

[0018] 2. A pattern parameter optimization model of the present invention is used to adjust pattern parameters for different types of socks; the model includes: an information acquisition layer, an information processing layer, a parameter analysis layer, a parameter optimization layer and a parameter output layer; wherein, in the parameter analysis layer, the parameters of the pattern are matched and calculated using the sock information, and the quality of the pattern is evaluated by the completeness of the pattern, the conformity of the pattern position and the degree of deformation of the pattern, so as to analyze the relationship of pattern parameter optimization; in the parameter optimization layer, the size and position of the pattern are optimized using the pattern requirements of the socks; through the pattern parameter optimization model, the pattern parameters can be set according to different types of socks in the actual production process, which is beneficial to the generation of pattern files of the sock knitting machine and improves the production efficiency of the socks.

[0019] 3. The present invention proposes a feedback mechanism for real-time dynamic adjustment of pattern parameters during the operation of a hosiery machine; the feedback mechanism obtains the current working parameters of the hosiery machine and the current fabric image, including a sock image and a pattern image; by extracting the features of the current fabric image of the hosiery machine, the difference from the actual design requirements is calculated, and the relationship with the current working parameters of the hosiery machine is analyzed to perform parameter correction; the corrected parameters are applied to the production process of the hosiery machine; the above method significantly improves the quality of pattern file generation of the hosiery machine, further improving the production efficiency of socks. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flow chart of a method for generating a pattern file for a hosiery machine provided by an embodiment of the present invention; Figure 2 A structural diagram of a pattern file generating system for a hosiery machine provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] Socks are an indispensable clothing in people's daily life. They include many types, such as sports socks, boat socks and stockings. As people's daily life improves, the requirements for socks also increase, especially in terms of the aesthetics of socks.

[0023] Nowadays, in order to meet people's needs, sock manufacturers will print various patterns on socks. For the patterns on socks, the manufacturer usually produces a pattern file, and then inputs the pattern file corresponding to the sock pattern to be produced into the sock knitting machine, and the sock knitting machine produces products according to the various parameters of the pattern file. Pattern files improve the efficiency and flexibility of sock production.

[0024] The pattern file generation method mainly stores different types of patterns in a specific format, and then sets the corresponding pattern parameters when socks need to be produced; however, this method is not conducive to the efficient production of socks. Therefore, the present invention proposes a pattern file generation method, system and medium for a sock knitting machine for generating different types of socks and pattern storage files. In order to explain the implementation process of the present invention in detail, two embodiments will be described below.

[0025] Embodiment 1: In the embodiment of the present application, the pattern file generation method and system of the hosiery machine proposed by the present invention are used to realize the generation of the pattern file. In order to illustrate the utility of the present invention in a practical scenario, the present invention is described in the embodiment of the present application for the production task of the hosiery production line 1 of manufacturer A; wherein, it is known that the production task of the hosiery production line 1 is to knit a batch of stockings with cartoon dog patterns on the stockings; Figure 1 and Figure 2 Describe the content in detail; Figure 1 The specific steps of the method proposed in the present invention include: S10. obtaining a pattern design sketch; S20. scanning the pattern design sketch to obtain a pattern design image; S30. preprocessing the pattern design image; S40. obtaining pattern parameters; S50. optimizing pattern parameters; S60. executing production tasks and adjusting pattern parameters in real time; S70. generating a pattern file; Figure 2 The structure of the pattern file generation system of a hosiery machine is given in the paper, including: an input unit, a preprocessing unit, a pattern parameter optimization unit, an adjustment unit, a file generation unit and a task unit.

[0026] According to the above content, the implementation process of the present invention includes: According to step S10, a pattern design sketch is obtained; Furthermore, the input unit of the system is used to receive a flower design sketch to obtain a flower design image, corresponding to step S20; wherein the input unit scans the flower design sketch through a scanning device connected to an external interface and saves it in a computer.

[0027] Furthermore, the preprocessing unit of the system preprocesses the flower design image, corresponding to the above-mentioned step S30; wherein the preprocessing process of the image by the preprocessing unit includes: enhancing the clarity of the flower design image to obtain a first flower design image; removing the background of the first flower design image to obtain a second flower design image; and standardizing the second flower design image to obtain the standard flower design image.

[0028] In the embodiment of the present application, the pattern design image scanned into the computer is preprocessed, including image enhancement, background removal and standardization. Through the above preprocessing steps, the system can provide higher quality input data for subsequent pattern learning and pattern file generation. This not only improves the efficiency and effect of model training, but also makes the generated pattern file more accurate and practical.

[0029] Further, the pattern parameter learning unit of the system is used to learn the standard pattern design image, corresponding to step S40; wherein the pattern parameter learning unit adopts a pattern parameter learning model to learn the pattern; the specific process includes: receiving the standard pattern design image by the pattern receiving layer; Further, the standard pattern design image is layered processed by the pattern layering processing layer to obtain a pattern layered image set; wherein the pattern layered image set includes: a color layer image, a texture layer image, a shape layer image, a pattern layer image, a detail layer image and a depth layer image; For the various layered images mentioned above, the main layering operations include: (1) Color layer image: performing color space conversion on the standard flower pattern design image to obtain a first standard flower pattern design image; wherein the standard flower pattern design image is converted from RGB to HSV color space; Further, various colors in the first standard pattern design image are analyzed, and color thresholds are set; Among them, the histogram analysis method is used for setting the color threshold; that is, the histograms of H (hue), S (saturation) and V (brightness) are calculated in the HSV color space; by observing the histogram of the H channel, the hue peak of the main color in the image can be identified. The histograms of the S and V channels are used to analyze the saturation and brightness of the color; according to the peak position of the histogram, the HSV threshold range is set for each main color. For example, red may correspond to an H value range of 0 to 10, and green may correspond to an H value range of 40 to 80; using the set HSV threshold, a mask image is generated to extract the corresponding color area; refer to Table 1, which gives the color threshold setting of the flower pattern in the application embodiment; Table 1 Color threshold setting description

[0030] In Table 1, black is used for the outline, eyes and nose of the dog; white is used for the bright parts of the dog's body or face; brown is used for the main body color of the dog; gray is used for shadows or color changes in specific parts; red, green, blue, yellow and pink can be used for decorative elements of the pattern; Further, each region of the first standard flower pattern design image is segmented according to the color threshold; wherein the first standard flower pattern design image is segmented using the generated mask to extract regions of each color; Further, each of the regions is combined into the color layer image; Table 2 shows the description of the color layer image of the cartoon dog in the embodiment of the present application; Table 2 Color layer image description

[0031] (2) Texture layer image: using a filter to extract texture features of the standard pattern design image; wherein the filter is a Gabor filter or a wavelet filter; Further, the texture features are mapped to obtain a texture feature map; wherein the obtained texture features are weighted averaged to highlight important features; Further, the texture feature map is separated according to the texture type to obtain a plurality of texture separation features; wherein, in the embodiment of the application, the corresponding texture features of the flower pattern include: hair texture, shadow texture, decorative texture and background texture, etc.; different flower patterns correspond to different texture types, so the corresponding texture type is set according to the flower pattern during the production process; Further, a plurality of the texture separation features are combined into the texture layer image; referring to Table 3, Table 3 gives a description of the texture layer image of a cartoon dog in the embodiment of the present application; Table 3 Pattern texture description

[0032] (3) Shape layer image: performing edge detection on the standard pattern design image to obtain a second standard pattern design image; wherein the edges in the image are identified using Canny edge detection or Sobel operator; Further, extracting a plurality of shapes from the second standard pattern design image; wherein, for the extraction of shapes, a contour extraction algorithm may be used to extract relevant shapes in the image; Further, the plurality of shapes are classified and formed into the shape layer image; see Table 4, which gives a description of the shape layer image; Table 4 Description of flower shape

[0033] (4) Pattern layer image: identifying a specific pattern of the standard flower design image; wherein the specific pattern in the image is obtained by using a template matching algorithm; the process of template matching is as follows: creating one or more template patterns according to the standard flower design; preprocessing the input standard flower design image, including denoising, graying and normalization; calculating the correlation between the template and the local area of ​​the image; by setting a threshold, filtering out areas with low matching degrees to ensure that only high-confidence matches are extracted; wherein the matching threshold is set to 0.5; the identified specific pattern and its position in the original image; Further, the specific pattern is segmented to obtain a main pattern and an auxiliary pattern; Further, the main pattern and the auxiliary pattern are classified and form the pattern layer image; refer to Table 5, which shows the pattern layering of the flower pattern in the embodiment of the present application; Table 5 Flower pattern layering description

[0034] (5) Detail layer image Extracting detail features of the standard pattern design image; wherein the relevant detail features in the extracted image are obtained by using filtering technology; Further, the detail feature is enhanced to obtain a detail enhanced feature; Further, the detail enhancement features are classified and formed into the detail layer image; refer to Table 6, which shows the detail layering of the pattern in the embodiment of the present application; Table 6 Flower pattern details layering description

[0035] (6) Depth layer image: extracting the depth information of each element in the standard pattern design image; Further, each of the elements is divided into foreground, middle ground and background according to the depth information, and the depth layer image is formed; refer to Table 7, which shows the depth layering of the flower pattern in the embodiment of the present application; Table 7 Description of pattern depth stratification

[0036] In the embodiment of the present application, a pattern parameter learning model is used for learning the pattern; in order to fully understand the pattern, a pattern hierarchical processing layer is set in the model to decompose the pattern in multiple dimensions, including: color, texture, shape, pattern, detail and depth; the pattern hierarchical processing layer provides a systematic and detailed analysis framework for pattern learning, improves the depth and breadth of learning, and thus effectively improves the quality of pattern file generation.

[0037] Furthermore, the stratification result of the pattern is marked by using an information marking layer; wherein the information marking layer marks the image information obtained in the pattern stratification processing layer according to the relative positions of each area of ​​the pattern image; Furthermore, each labeled layered image is input into a feature extraction layer for feature extraction to obtain a flower pattern feature set; wherein the feature extraction layer adopts a CNN convolutional neural network; Further, the features in the pattern feature set are input into the pattern parameter generation layer to generate pattern parameters; The flower pattern parameter generation layer includes: normalizing the features in the flower pattern feature set to obtain a normalized flower pattern feature set; performing feature screening on the normalized flower pattern feature set to obtain a screened flower pattern feature set; performing feature encoding on the screened flower pattern feature set to obtain a coded flower pattern feature set; performing parameter conversion on the features in the coded flower pattern feature set to obtain a plurality of flower pattern parameters; Furthermore, the flower pattern parameter output layer is used to integrate multiple flower patterns into a flower pattern parameter set and output it; Furthermore, the output pattern parameter set is used as initial hosiery machine pattern parameters; wherein the initial hosiery machine pattern parameters involve pattern design information and manufacturing process parameters.

[0038] In the embodiment of the present application, a pattern parameter learning model is used to learn the pattern manufacturing parameters; the model performs multi-level learning on the received standard pattern images to achieve a more accurate understanding of the pattern; and the acquired multiple features are learned through the pattern parameter generation layer to obtain the initial pattern parameters; the above use of the pattern parameter learning model not only helps to capture the complexity and subtle differences of the pattern, but also improves the understanding of the pattern.

[0039] Further, the pattern parameter optimization unit is used to optimize the initial hosiery machine pattern parameters, corresponding to the above step S50; wherein the optimization of parameters includes: pattern size, pattern knitting position and manufacturing process; for the above-mentioned contents, according to the type of socks in the actual production task, the pattern parameter optimization model is used for optimization; the specific process is as follows: The sock code in the production task of the embodiment of the present application is obtained through information; wherein, according to the production schedule of the sock production line 1, it can be known that the sock code is S010; Further, sock information is obtained according to the sock code, including: sock type, material information, size specification, shape information and design requirements; refer to Table 8, which shows the sock information of the embodiment of the present application; Table 8 S010 Socks Information Description

[0040] Further, the acquired socks information is processed by an information processing unit to obtain standard socks data; wherein the socks information processing includes: The socks information is divided into basic information, material characteristics, size characteristics, shape characteristics and design characteristics; Extract basic information features and convert "stockings" into category features; Material feature extraction, decomposing "cotton 70% + polyester 30%" into: Material 1: cotton, proportion: 70%; Material 2: polyester, proportion: 30%; Create binary features: Use Cotton: Yes; Use Polyester: Yes; Record the proportion of each material as a numerical feature: proportion of cotton: 70; proportion of polyester: 30; Size feature extraction, extracting the middle value from the size range, length: the calculated middle value is 26cm (25cm~27cm); foot circumference: the calculated middle value is 23cm (22cm~24cm); ankle circumference: the calculated middle value is 19cm (18cm~20cm); sock width: the calculated middle value is 10cm (9cm~11cm); Each dimension is recorded as a numerical feature.

[0041] Shape feature extraction, extract features from "round head, straight tube design": round head: yes; straight tube design: yes; flower pattern: extracted as category feature (flower pattern: cartoon dog); flower pattern size: decomposed into: flower pattern width: 10cm; flower pattern height: 15cm; Pattern position: extracted as categorical features (position: above the ankle); All extracted features are integrated into a structured data table. See Table 9, which shows the standard data after processing the socks information; Table 9 Socks information processing instructions

[0042] Furthermore, the standardized sock data and the initial hosiery machine pattern parameters are input into the parameter analysis layer for matching analysis; the specific process is: setting pattern evaluation criteria: pattern completeness, pattern position conformity and pattern deformation; wherein the pattern completeness refers to whether the pattern is fully presented on the socks provided by the implementation of the present application; the pattern position conformity refers to the degree of deviation of the pattern position from the design requirements of the socks due to the shape of the pattern; the pattern deformation refers to the degree of deformation of the pattern; Further, the standardized socks data and the initial socks knitting machine pattern parameters are input into a parameter matching model; wherein the parameter matching model comprises: inputting the standardized socks data and the initial socks knitting machine pattern parameters from an input layer; Further, the input data is input into a matching layer; wherein the matching layer extracts corresponding features of the standardized sock data and the initial sock knitting machine pattern parameters respectively, and calculates the matching degree of the corresponding features through a similarity algorithm; Further, the matching result of the parameter matching model is evaluated; wherein the matching result is evaluated according to the flower pattern evaluation standard, including: (1) Calculate the completeness percentage: ; (2) Calculate the deviation between the actual position and the designed position: ;in, and Represented as pattern element coordinate information; (3) Calculate the deformation degree of the pattern: ;in, and They are respectively represented as the distances of element feature points before and after weaving; Expressed as the number of element feature points; Further, the parameter optimization layer performs parameter optimization according to the evaluation result of the parameter analysis layer; wherein the parameter optimization layer comprises: obtaining pattern requirements from the design requirements of the socks to be knitted; wherein the pattern requirements comprise: pattern type, pattern size and pattern position; Further, the pattern error is calculated according to the matching result; wherein the pattern error includes: integrity error, position error and deformation error; Further, the cause of the error is analyzed according to the pattern error; wherein a multiple regression analysis model is used for the analysis of the cause of the error; the initial hosiery machine pattern parameters and standardized socks data are used as variables, and the pattern error is used as an adjustment parameter; Furthermore, the causes of the errors are quantified and quantitative indicators are established; wherein the quantitative indicators include the absolute value and relative value of the errors and the proportion of each error type; Further, the initial hosiery machine pattern parameters are optimized according to the quantitative index to obtain the optimized hosiery machine pattern parameters. Referring to Table 10, Table 10 shows the optimization results of the pattern parameters of the embodiment of the present application.

[0043] Table 10 Optimization of initial knitting machine pattern parameters of S010 socks

[0044] Table 10 shows some optimization results of the initial hosiery machine pattern parameters; for example, the pattern size is optimized from 10cm×15cm to 10cm×12cm, the pattern width is reduced to avoid pattern missing; the pattern height is reduced to inhibit pattern deformation; the knitting position is changed from "above the ankle" to "middle of the ankle", and the pattern proportion is suitable for the "stockings" in the embodiment of the present application.

[0045] In the embodiment of the present application, the initial hosiery machine pattern parameters are optimized through the parameter optimization layer; this layer calculates the pattern error according to the matching result, and analyzes the cause of the error using a multivariate regression analysis model to obtain the accurate cause; the analyzed cause is quantified and the original parameters are corrected; not only the precise parameter setting of the pattern of different types of socks is achieved, but also the productivity is further improved.

[0046] Furthermore, the parameter output layer outputs the optimized parameters; Further, the adjustment unit of the system performs parameter adjustment, corresponding to the above step S70; wherein the adjustment unit is used for the task unit of the system to perform real-time dynamic parameter adjustment using a feedback mechanism during the production process; the specific process includes: obtaining the current operating parameters of the hosiery knitting machine and the current fabric image of the hosiery knitting task; Further, the current sock knitting operation parameter is quantified to generate a current operation parameter vector; Further, the current fabric image is processed and segmented to obtain a current sock image and a current pattern image; Furthermore, the current sock image and the current pattern image are extracted respectively to obtain the current sock features and the current pattern features; wherein the current sock features include shape features: using edge detection (such as the Canny algorithm) to extract the contour information of the socks; texture features: applying the gray-level co-occurrence matrix (GLCM) to extract texture information, such as contrast, correlation, etc.; color features: extracting the color histogram of the sock image, analyzing the main color tone and its distribution; integrating the extracted shape, texture and color features to form the current sock feature vector. The current pattern features include pattern features: identifying the basic shape, boundary and symmetry of the pattern; color features: extracting the color information of the pattern, analyzing the color combination and its distribution; frequency domain features: Fourier transform can be used to extract the frequency domain features of the pattern and analyze the frequency characteristics of the pattern; integrating the extracted pattern, color and frequency domain features to form the current pattern feature vector.

[0047] Further, the current sock feature and the current pattern feature are combined, and differences are identified according to the design requirements; Further, a difference vector is generated according to the result of the difference identification; wherein the process of obtaining the difference vector includes: comparing the combined feature vector with the preset design requirements to identify the difference in features; calculating the difference measure between the comprehensive feature vector and the design requirements using Euclidean distance, Manhattan distance, etc.; and quantifying the identified difference into a difference vector according to the comparison result; Furthermore, the difference vector and the current operation parameter vector are subjected to correlation analysis to obtain a knitting correlation matrix; wherein the process of obtaining the knitting correlation matrix includes: standardizing the current operation parameter vector and the difference vector; using a correlation analysis method (such as the Pearson correlation coefficient or the Spearman rank correlation coefficient) to calculate the correlation between the difference vector and the current operation parameter vector. The correlation coefficients between all operation parameters and difference indicators are organized into a matrix; refer to Table 11, which shows the correlation analysis of the difference vector and the current operation parameter vector; Table 11 Correlation analysis

[0048] Further, the current operating parameter vector is subjected to parameter correction according to the weaving correlation matrix to obtain a corrected parameter vector; Furthermore, the correction parameter vector is applied to the hosiery machine as the adjusted pattern parameter of the hosiery machine and is continuously monitored.

[0049] In the embodiment of the present application, a feedback mechanism is proposed for real-time dynamic adjustment of pattern parameters during the operation of a hosiery machine. The feedback mechanism obtains the current working parameters of the hosiery machine and the current fabric image, including a sock image and a pattern image; calculates the difference from the actual design requirements by extracting the features of the current fabric image of the hosiery machine; performs parameter correction by analyzing the correlation between the difference and the current working parameters of the hosiery machine; and feeds back the corrected parameters to the hosiery machine. The above method significantly improves the quality of pattern file generation of the hosiery machine, and further improves the production efficiency of socks.

[0050] Furthermore, the system file generation unit saves the pattern, sock type, design requirements and loom working parameters in the embodiment of the present application to generate a pattern file for the sock knitting machine, corresponding to the above-mentioned step S70.

[0051] In the application embodiment, the production task of the socks production line 1 of manufacturer A is taken as the implementation object, and the pattern file generation of the socks type is long stockings and the pattern is a cartoon dog is realized through the method and system of the present invention. The embodiment of the present application mainly starts from the following points. First, a pattern parameter learning model is used to learn the feature parameters of the pattern image; in this model, the image is layered from different dimensions through the pattern layering processing layer, the purpose is to fully understand the pattern; the feature extraction layer is used to obtain the features of different layers, and the features are generated into corresponding pattern parameters; this is conducive to the control of pattern manufacturing in actual production; secondly, a pattern parameter optimization model is proposed to optimize the parameters; the model analyzes the applicability of the pattern according to the information of different types of socks, the purpose is to achieve flexible parameter adjustment when knitting socks; finally, a feedback mechanism is used to adjust the socks and patterns in the production process in real time and dynamically; the feedback mechanism is quantified according to the error of the knitting process, and the correlation between the error and the working parameters is analyzed and corrected; this is conducive to the correctness of the pattern in the socks manufacturing process. The above methods not only realize the generation of files related to patterns and socks, but also provide sufficient data support for future production tasks based on the generated files.

[0052] Embodiment 2: In the first embodiment, the production task of the socks production line 1 of manufacturer A is taken as the implementation object, and the pattern file generation of the socks type is long stockings and the pattern is a cartoon dog is realized through the method and system of the present invention; in order to illustrate that the pattern file generation method proposed by the present invention in the production of different types of socks has high applicability; in the embodiment of the present application, the production task of the socks production line 2 of manufacturer A is taken as the implementation object; wherein, the production task of the socks production line 2 is to produce a batch of socks with a cartoon dog pattern; The following is the implementation process: obtain the flower design sketch; Further, scanning the flower pattern design sketch to obtain a flower pattern design image; Further, the flower pattern design image is preprocessed to obtain a standard flower pattern design image; Further, the standard flower pattern design image is input into the flower pattern parameter learning model to obtain a flower pattern parameter set; wherein, the flower pattern design sketch in the embodiment of the present application is the same as the flower pattern design sketch in the first embodiment, so the process of obtaining the flower pattern parameter set is also the same, which will not be described in detail here, and reference will be made to the corresponding content of the first embodiment; Further, the pattern optimization model is used to optimize the initial hosiery machine pattern parameters to obtain optimized hosiery machine pattern parameters; wherein the pattern parameter optimization model includes: an information acquisition layer, used to obtain the sock code of the socks to be knitted, and obtain sock information according to the sock code; wherein the sock information includes: sock type, material information, size specifications, shape information and design requirements; refer to Table 12, which is the sock information of the embodiment of the present application; Table 12 S020 socks information description

[0053] An information processing layer is used to process the socks information to obtain standardized socks data; The parameter analysis layer is used for analyzing the matching degree between the standardized sock data and the initial sock knitting machine pattern parameters, including: setting pattern evaluation criteria: pattern completeness, pattern position conformity, and pattern deformation; inputting the standardized sock data and the initial sock knitting machine pattern parameters into a parameter matching model; and performing evaluation according to the matching results of the parameter matching model; A parameter optimization layer, used for optimizing the initial hosiery machine pattern parameters according to the evaluation results; The parameter output layer is used to output the optimized hosiery machine pattern parameters; refer to Table 13, which shows the parameter optimization of the embodiment of the present application.

[0054] Table 13 Optimization of initial knitting machine pattern parameters of S020 socks

[0055] Further, a sock knitting task is performed according to the optimized sock knitting machine pattern parameters, and the optimized sock knitting machine pattern parameters are adjusted in real time; Furthermore, a pattern file of a sock machine with a sock pattern of short socks and a pattern of a cartoon dog is generated according to the adjusted parameters.

[0056] Embodiment three: A pattern file generation medium for a hosiery machine, wherein the medium can be a computer program executed by an electronic device, and when the program is run on the electronic device, the electronic device can execute all the contents of embodiment one or embodiment two.

[0057] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for generating a pattern file for a hosiery machine, characterized in that: include: Get the sketch of the flower pattern design; Scanning the pattern design sketch to obtain a pattern design image; Preprocessing the flower pattern design image to obtain a standard flower pattern design image; Inputting the standard flower pattern design image into a flower pattern parameter learning model to obtain a flower pattern parameter set; Generate initial hosiery machine pattern parameters through the pattern parameter set; The initial hosiery machine pattern parameters are optimized by using a pattern parameter optimization model to obtain optimized hosiery machine pattern parameters; wherein the pattern parameter optimization model includes: An information acquisition layer, used to acquire a sock code of the socks to be knitted, and acquire sock information according to the sock code; An information processing layer is used to process the socks information to obtain standardized socks data; The parameter analysis layer is used for the matching degree analysis between the standardized socks data and the initial hosiery machine pattern parameters, including: setting pattern evaluation criteria: pattern completeness, pattern position conformity and pattern deformation; inputting the standardized socks data and the initial hosiery machine pattern parameters into a parameter matching model; and evaluating the matching results of the parameter matching model; A parameter optimization layer, used for optimizing the initial hosiery machine pattern parameters according to the evaluation results; A parameter output layer, used for outputting the optimized hosiery machine pattern parameters; Performing a sock knitting task according to the optimized sock knitting machine pattern parameters, and adjusting the optimized sock knitting machine pattern parameters in real time; After the sock knitting task is completed, the adjusted parameters are used to generate a pattern parameter file and the file is saved.

2. The method for generating a pattern file for a hosiery machine according to claim 1, characterized in that: The preprocessing includes: enhancing the clarity of the flower pattern design image to obtain a first flower pattern design image; removing the background of the first flower pattern design image to obtain a second flower pattern design image; and standardizing the second flower pattern design image to obtain the standard flower pattern design image.

3. The method for generating a pattern file for a hosiery machine according to claim 1, characterized in that: The pattern parameter learning model includes: A pattern receiving layer, used for receiving the standard pattern design image; A pattern layering processing layer is used to layer the standard pattern design image to obtain a pattern layering image set; wherein the pattern layering image set includes: a color layer image, a texture layer image, a shape layer image, a pattern layer image, a detail layer image and a depth layer image; An information marking layer, used for marking the stratification result of the pattern stratification processing layer and the pattern at corresponding positions; A feature extraction layer, used for extracting features from the image in the flower pattern layered image to obtain a flower pattern feature set; The flower pattern parameter generation layer is used to parameterize the features in the flower pattern feature set to obtain flower pattern parameters; wherein the flower pattern parameter generation layer includes: normalizing the features in the flower pattern feature set to obtain a normalized flower pattern feature set; performing feature screening on the normalized flower pattern feature set to obtain a screened flower pattern feature set; performing feature encoding on the screened flower pattern feature set to obtain an encoded flower pattern feature set; performing parameter conversion on the features in the encoded flower pattern feature set to obtain multiple flower pattern parameters; The pattern parameter output layer is used to output the pattern parameters and integrate them into the pattern parameter set.

4. The method for generating a pattern file for a hosiery machine according to claim 3, characterized in that: The pattern layering treatment layer comprises: Performing color space conversion on the standard flower pattern design image to obtain a first standard flower pattern design image; analyzing various colors in the first standard flower pattern design image and setting a color threshold; segmenting various regions of the first standard flower pattern design image according to the color threshold; and composing the various regions into the color layer image; Extracting texture features of the standard pattern design image by using a filter; mapping the texture features to obtain a texture feature map; separating the texture feature map according to texture type to obtain a plurality of texture separation features; and composing the plurality of texture separation features into the texture layer image; Performing edge detection on the standard flower pattern design image to obtain a second standard flower pattern design image; extracting multiple shapes from the second standard flower pattern design image; classifying the multiple shapes and composing the shape layer image; Identify the specific pattern of the standard pattern design image; segment the specific pattern to obtain a main pattern and an auxiliary pattern; classify the main pattern and the auxiliary pattern to form the pattern layer image; Extracting detail features of the standard pattern design image; performing feature enhancement on the detail features to obtain detail enhancement features; classifying the detail enhancement features and forming the detail layer image; Extracting the depth information of each element in the standard pattern design image; dividing each element into foreground, middle ground and background according to the depth information, and forming the depth layer image.

5. The method for generating a pattern file for a hosiery machine according to claim 1, characterized in that: The parameter optimization layer includes: obtaining pattern requirements from the design requirements of the socks to be knitted; wherein the pattern requirements include: pattern type, pattern size and pattern position; calculating pattern errors according to the matching results; analyzing error causes according to the pattern errors; quantifying the error causes and establishing quantitative indicators; optimizing the initial hosiery machine pattern parameters according to the quantitative indicators to obtain the optimized hosiery machine pattern parameters.

6. The method for generating a pattern file for a hosiery machine according to claim 1, characterized in that: A feedback mechanism is used to adjust the pattern parameters of the hosiery knitting machine in real time; wherein the feedback mechanism includes: Get the current operating parameters of the sock knitting machine and the current fabric image of the sock knitting task; quantifying the current operating parameters of the hosiery knitting machine to generate a current operating parameter vector; Processing and segmenting the current fabric image to obtain a current sock image and a current pattern image; Extracting the current sock image and the current pattern image respectively to obtain current sock features and current pattern features; Combining the current sock feature with the current pattern feature, and performing difference identification according to the design requirements; generating a difference vector according to the result of the difference identification; Performing correlation analysis on the difference vector and the current operating parameter vector to obtain a weaving correlation matrix; Performing parameter correction on the current operating parameter vector according to the weaving correlation matrix to obtain a corrected parameter vector; The correction parameter vector is applied to the hosiery machine as the adjusted hosiery machine pattern parameter and is continuously monitored.

7. A pattern file generation system for a hosiery machine, characterized in that: The system comprises: an input unit for receiving a pattern design sketch; a preprocessing unit for processing the pattern design sketch to obtain a standard pattern design image; a pattern parameter learning unit for learning characteristic parameters of an input pattern; a pattern parameter optimization unit for optimizing initial hosiery machine pattern parameters for socks to be knitted; an adjustment unit for real-time adjustment of pattern parameters during production; a file generation unit for generating and saving corresponding files after a production task is completed; and a task unit for issuing a production task.

8. A pattern file generating system for a hosiery machine according to claim 7, characterized in that: The pattern parameter learning unit adopts a pattern parameter learning model, including: A pattern receiving layer, used for receiving a standard pattern design image; A pattern layering processing layer is used to layer the standard pattern design image to obtain a pattern layering image set; wherein the pattern layering image set includes: a color layer image, a texture layer image, a shape layer image, a pattern layer image, a detail layer image and a depth layer image; An information marking layer, used for marking the stratification result of the pattern stratification processing layer and the pattern at corresponding positions; A feature extraction layer, used for extracting features from the image in the flower pattern layered image to obtain a flower pattern feature set; The flower pattern parameter generation layer is used to parameterize the features in the flower pattern feature set to obtain flower pattern parameters; wherein the flower pattern parameter generation layer includes: normalizing the features in the flower pattern feature set to obtain a normalized flower pattern feature set; performing feature screening on the normalized flower pattern feature set to obtain a screened flower pattern feature set; performing feature encoding on the screened flower pattern feature set to obtain an encoded flower pattern feature set; performing parameter conversion on the features in the encoded flower pattern feature set to obtain multiple flower pattern parameters; The pattern parameter output layer is used to output the pattern parameters and integrate them into a pattern parameter set.

9. The pattern file generating system for a hosiery machine according to claim 7, characterized in that: The pattern parameter optimization unit uses the pattern parameter optimization model to perform optimization, including: An information acquisition layer, used to acquire a sock code of the socks to be knitted, and acquire sock information according to the sock code; wherein the sock information includes: sock type, material information, size specification, shape information and design requirements; An information processing layer is used to process the socks information to obtain standardized socks data; The parameter analysis layer is used for analyzing the matching degree between the standardized sock data and the initial sock knitting machine pattern parameters, including: setting pattern evaluation criteria: pattern completeness, pattern position conformity, and pattern deformation; inputting the standardized sock data and the initial sock knitting machine pattern parameters into a parameter matching model; and performing evaluation according to the matching results of the parameter matching model; A parameter optimization layer, used for optimizing the initial hosiery machine pattern parameters according to the evaluation results; The parameter output layer is used to output the optimized pattern parameters of the hosiery machine.

10. A pattern file generation medium for a hosiery machine, characterized in that: The medium is a computer program that can be executed by an electronic device. When the program is run on the electronic device, the electronic device executes the pattern file generation method for a hosiery machine as described in any one of claims 1 to 6.

Images