A method for adaptively controlling trouser waist pressing parameters

Abstract

The present invention discloses a method for adaptively controlling trouser waist pressing parameters, which relates to the field of garment processing automation. The method comprises: collecting target pressing forming images, extracting feature points, and obtaining pressing control points; performing data mining to generate a set of trouser waist pressing defect records; performing defect feature recognition, matching with pressing control points, and generating target pressing feature points; generating P pressing processes; performing pressing, and collecting process forming images; performing feature recognition, determining P parameter adjustment direction sets and P parameter adjustment step sets; performing adaptive adjustment, and generating P fine-tuning pressing process parameter sets; and performing parameter control on a pressing machine. The method solves the technical problems of the lack of real-time feedback and adaptive adjustment mechanism in the existing trouser waist pressing parameter control, resulting in low production efficiency and unstable pressing quality, and achieves the technical effect of improving production efficiency and pressing quality by automatically adjusting and optimizing trouser waist pressing parameters.

Description

Technical Field

[0001] The present application relates to the field of garment processing automation, and in particular to a method for adaptively controlling trouser waist pressing parameters. Background Art

[0002] In the garment manufacturing industry, the pressing process of trouser waists is a key step to ensure that the trousers are crisp and comfortable to wear. Traditionally, the trouser waist pressing process relies on manual experience to adjust the parameters of the pressing machine, such as temperature, pressure, time, etc., to achieve the ideal pressing effect. However, the traditional manual control method is time-consuming and highly dependent on the operator's experience and skill level, making it difficult to quickly respond to the pressing needs of trouser waists of different styles and materials on the production line.

[0003] In the current relevant technologies, there is a lack of real-time feedback and adaptive adjustment mechanism in the control of trouser waist pressing parameters, which leads to technical problems such as low production efficiency and unstable pressing quality. Summary of the invention

[0004] The present application provides a method for adaptively controlling trouser waist pressing parameters, which extracts feature points from a target pressing forming image to obtain pressing control points, obtains a set of trouser waist pressing defect records through data mining, matches the defect feature recognition results with the pressing control points, generates target pressing feature points, decomposes the pressing process, generates multiple pressing processes, presses the target trouser waist, collects process forming images, performs feature recognition on the process forming image set according to the target pressing feature points, determines a parameter adjustment direction set and a parameter adjustment pace set, adaptively adjusts the pressing process parameter set, performs parameter control on the pressing machine, and other technical means, thereby achieving the technical effect of automatically adjusting and optimizing the trouser waist pressing parameters to improve production efficiency and pressing quality.

[0005] The present application provides a method for adaptively controlling trouser waist pressing parameters, comprising:

[0006] A target pressing and forming image of a target waistband is collected, feature points are extracted from the target pressing and forming image to obtain K pressing control points, wherein the K pressing control points include K pressing control feature values; data mining is performed based on the waistband type and pressing process of the target waistband to generate a set of waistband pressing defect records; defect feature recognition is performed on the set of waistband pressing defect records, and the defect feature recognition results are matched with the K pressing control points to generate Q target pressing feature points; pressing process is decomposed according to the pressing process to generate P pressing processes, wherein the P pressing processes include P pressing process parameter sets; according to the P pressing processes, The target trouser waist is pressed according to a set of Q target pressing parameters, and after each pressing process is completed, process forming images are collected to obtain P process forming image sets; based on the Q target pressing feature points and the corresponding Q target pressing feature values, feature recognition is performed on the P process forming image sets in turn to determine P parameter adjustment direction sets and P parameter adjustment step sets; the P pressing process parameter sets are adaptively adjusted according to the P parameter adjustment direction sets and the P parameter adjustment step sets to generate P fine-tuning pressing process parameter sets; the pressing machine is parameter-controlled using the P fine-tuning pressing process parameter sets to perform pressing of the target trouser waist.

[0007] In a possible implementation, the following processing is performed:

[0008] Obtain a preset set of pressing control indicators; extract the waistline control area based on the preset set of pressing control indicators according to the pressing control indicator-waistline area mapping relationship, and generate a waistline mapping control area set; extract regional feature points based on the waistline mapping control area set, and generate the K pressing control points.

[0009] In a possible implementation, the following processing is performed:

[0010] Based on the waistband type and pressing process of the target waistband, a search is performed in the pressing database to obtain an initial waistband pressing defect record set; data difference identification is performed on the initial waistband pressing defect record set to generate an initial difference; when the initial difference does not meet a preset difference threshold, a cloud data mining instruction is obtained, and a cloud waistband pressing defect record set is obtained based on the cloud data mining instruction; the waistband pressing defect record set is generated according to the initial waistband pressing defect record set and the cloud waistband pressing defect record set.

[0011] In a possible implementation, the following processing is performed:

[0012] Using the Q target pressing feature points as indexes, feature recognition is performed on the P process forming image sets to generate P process pressing feature value sets; feature deviation recognition is performed on the P process pressing feature value sets based on the Q target pressing feature values ​​to obtain P feature value deviation sets and P feature value deviation direction sets; same-parameter clustering deviation analysis is performed on the P feature value deviation sets and the P feature value deviation directions to determine the P parameter adjustment direction sets and the P parameter adjustment step sets.

[0013] In a possible implementation, the following processing is performed:

[0014] The P eigenvalue deviation sets are clustered with the same parameters to generate P eigenvalue deviation clustering clusters; the P eigenvalue deviation clustering clusters and the corresponding P eigenvalue deviation direction clustering clusters are analyzed by using a mean shift algorithm to generate P parameter adjustment direction sets and the P parameter adjustment step sets.

[0015] In a possible implementation, the following processing is performed:

[0016] The P sets of pressing process parameters are randomly adjusted according to the P sets of parameter adjustment directions and the P sets of parameter adjustment steps to obtain P sets of stage pressing process parameters; fitness analysis is performed on the P sets of stage pressing process parameters to generate P stage fitnesses; the P sets of pressing process parameters are iteratively updated according to the P sets of stage pressing process parameters until a preset number of iterations is met to obtain P sets of fine-tuning pressing process parameters.

[0017] In a possible implementation, the following processing is performed:

[0018] Determine whether the difference between the P stage fitness of the P stage pressing process parameter sets and the P fitness of the P pressing process parameter sets meets the preset fitness difference; if not, update the P stage pressing process parameter sets as adjustment objects, and perform adjustment iterations according to the P parameter adjustment direction sets and the P parameter adjustment step sets until the preset number of iterations is met, thereby obtaining the P fine-tuning pressing process parameter sets.

[0019] In a possible implementation, the following processing is performed:

[0020] The preset pressing control index set includes pleat straightness, pleat depth, fabric flatness, dimensional stability, and symmetry of trouser waist and lining.

[0021] The present application proposes a method for adaptively controlling trouser waist pressing parameters. First, a target pressing forming image of a target trouser waist is collected, feature points are extracted from the target pressing forming image, and K pressing control points are obtained, wherein the K pressing control points include K pressing control feature values. Then, data mining is performed based on the trouser waist type and pressing process of the target trouser waist to generate a set of trouser waist pressing defect records. Then, defect feature recognition is performed on the set of trouser waist pressing defect records, and the defect feature recognition results are matched with the K pressing control points to generate Q target pressing feature points. Then, the pressing process is disassembled according to the pressing process to generate P pressing processes, wherein the P pressing processes include P pressing process parameter sets. Then, the pressing process parameters are sequentially matched according to the P pressing process parameter sets. The target trouser waist is pressed and ironed, and after each pressing process is completed, process forming images are collected to obtain P process forming image sets, and then feature recognition is performed on the P process forming image sets in turn based on Q target pressing feature points and corresponding Q target pressing feature values ​​to determine P parameter adjustment direction sets and P parameter adjustment step sets, and then the P pressing process parameter sets are adaptively adjusted according to the P parameter adjustment direction sets and the P parameter adjustment step sets to generate P fine-tuning pressing process parameter sets, and finally the P fine-tuning pressing process parameter sets are used to control the parameters of the pressing machine to execute the pressing of the target trouser waist, thereby achieving the technical effect of improving production efficiency and pressing quality by automatically adjusting and optimizing the trouser waist pressing parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the method according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0023] Figure 1 A schematic flow chart of a method for adaptively controlling trouser waist pressing parameters provided in an embodiment of the present application.

[0024] Figure 2 A schematic diagram of the process of generating a set of trouser waist pressing defect records in a trouser waist pressing parameter adaptive control method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0026] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0027] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0028] The present application embodiment provides a method for adaptively controlling trouser waist pressing parameters, such as Figure 1 As shown, the method includes:

[0029] Step S100, collect the target pressing and forming image of the target waistband, extract the feature points of the target pressing and forming image, and obtain K pressing and forming control points, wherein the K pressing and forming control points include K pressing and forming control feature values. Specifically, use a high-definition camera or image sensor to capture the pressing and forming image of the target waistband (the waistband portion that needs to be pressed and formed is the object controlled by this method), ensure that the image is clear, unobstructed, and covers the entire waistband area, and the target pressing and forming image is an image formed after the target waistband is processed by the pressing and forming process, reflecting the pressing and forming effect. De-noising is performed on the collected image to eliminate random noise and interference in the image, and image enhancement is performed, such as contrast adjustment, sharpening, etc., to improve the image quality. Image processing technology (such as edge detection, corner point detection, etc.) is used to extract key pressing and forming control points in the image, and these control points reflect the key features of the pressing and forming effect of the waistband, such as wrinkles, curves, etc. For each extracted pressing control point, its corresponding eigenvalue is calculated. The eigenvalue is a numerical value or vector that describes the characteristics of the feature point, including position coordinates (x, y), direction angle, scale information, shape parameters, descriptors, etc.

[0030] In a possible implementation, step S100 further includes step S110, obtaining a preset pressing control index set. Specifically, according to the technical requirements and industry standards for trouser waist pressing, a series of key pressing control indexes are defined, including temperature, pressure, time, wrinkle depth, wrinkle width, indentation clarity, etc. These control indexes are organized into a set, namely, a preset pressing control index set, which contains a series of key indicators for evaluating and optimizing the trouser waist pressing process. For each control index, a reasonable threshold range is set to determine whether the pressing effect reaches the expected standard. Step S120, according to the pressing control index-waist region mapping relationship, the trouser waist control region is extracted based on the preset pressing control index set, and a trouser waist mapping control region set is generated. Specifically, the mapping relationship between each pressing control index and a specific region on the trouser waist is analyzed. For example, the wrinkle depth is mainly related to the central region of the front piece of the trouser waist, and the indentation clarity involves the entire trouser waist edge. Based on the mapping relationship, the trouser waist is divided into multiple control regions, each of which is associated with one or more pressing control indicators. The divided control areas are organized into a set, namely, a waistband mapping control area set. Step S130, regional feature point extraction is performed based on the waistband mapping control area set to generate the K pressing control points. Specifically, the waistband image in each waistband mapping control area is carefully analyzed to identify key features directly related to the pressing effect. Representative feature points are extracted in each waistband mapping control area, and these feature points are used to accurately reflect the key information of the pressing effect in the area, such as the starting point of the wrinkle, the edge size point of the waistband, etc. The feature points in all waistband mapping control areas are integrated to form a set of K pressing control points, each of which contains its position coordinates, corresponding control indicators and other attribute information. This implementation method ensures that the pressing process has clear goals and standards by presetting the set of pressing control indicators, and realizes the accurate division of the waistband area by establishing the mapping relationship between the pressing control indicators and the waistband area, and dividing the control area accordingly, thereby improving the pertinence of feature point extraction and achieving the technical effect of obtaining a comprehensive and representative set of pressing control points.

[0031] In a possible implementation, step S110 further includes step S111, wherein the preset pressing control index set includes wrinkle straightness, pleat depth, fabric flatness, dimensional stability, and waistband lining symmetry. Specifically, wrinkle straightness is a standard for measuring whether the pleats of the waistband are straight and neat after pressing, and reflects the control ability of the pressing process on the wrinkle morphology. Pleat depth refers to the degree of depression of the pleats after pressing, which determines the three-dimensional sense and aesthetics of the pleats. Fabric flatness refers to the overall smoothness of the waistband fabric after pressing, which is related to the wearing comfort and overall appearance of the waistband. Dimensional stability refers to whether the size of the waistband remains unchanged or changes within an acceptable range after pressing. Waistband lining symmetry refers to the degree of symmetry between the inner and outer sides of the waistband after pressing, which is related to the overall aesthetics and wearing effect of the waistband. This implementation ensures that there are clear control targets and evaluation standards in the waistband pressing process by clearly presetting the pressing control index set, and achieves the technical effect of providing clear guidance for the division of the waistband control area and the extraction of feature points.

[0032] Step S200, data mining is performed based on the waistband type and pressing process of the target waistband to generate a set of waistband pressing defect records. Specifically, a database or data warehouse is established to store historical waistband pressing data. These data are collected from various channels such as production records, quality inspection reports, user feedback, etc., including pressing results under different waistband types (such as elastic waistbands, belt buckle waistbands, etc.) and various pressing processes (processes that use heat and pressure to make the waistband reach a predetermined shape and texture, such as temperature, pressure, time and other parameter combinations). The collected data is preprocessed, including removing duplicate data, processing missing values, correcting erroneous data, etc., to ensure the accuracy and completeness of the data. The data is classified according to the waistband type and pressing process, and each record is annotated using manual annotation or machine learning methods to indicate whether there is a defect and its type (such as uneven wrinkles, too deep indentations, scorching, etc.). Use data mining algorithms (association rule mining, cluster analysis, classification and prediction, etc.) to conduct in-depth analysis on the cleaned and labeled data to discover the association rules or patterns between waistband type, pressing process and defects, such as analyzing which waistband types are prone to which types of defects under which pressing processes. Based on the data mining results, a set of waistband pressing defect records is generated, which contains key information such as waistband type, pressing process, defect type that occurs during waistband pressing and its probability of occurrence.

[0033] like Figure 2As shown, in a possible implementation, step S200 further includes step S210, searching in the pressing database based on the waistband type and pressing process of the target waistband to obtain an initial waistband pressing defect record set. Specifically, identify the specific type of the target waistband (such as high waist, mid waist, low waist, etc.) and the pressing process used (such as hot pressing, steam pressing, etc.). Using the identified waistband type and pressing process as query conditions, search in the pressing database (such as the pressing record data repository within the workshop), filter out all waistband pressing defect records that match the target waistband type and pressing process from the pressing database, and form an initial waistband pressing defect record set. Step S220, identify the data difference of the initial waistband pressing defect record set to generate an initial difference. Specifically, the difference refers to the diversity and richness of the characteristics such as type, quantity, and distribution recorded in the initial waistband pressing defect record set. An algorithm or statistical method (such as entropy method, variance analysis, etc.) is used to process the data in the initial waistband pressing defect record set to calculate the initial difference. The larger the initial difference, the richer the data and the higher the reference value for subsequent analysis. Step S230: When the initial difference does not meet the preset difference threshold, a cloud data mining instruction is obtained, and a cloud waistband pressing defect record set is obtained based on the cloud data mining instruction. Specifically, it is determined whether the initial difference is less than the preset difference threshold. If so, it means that the initial waistband pressing defect record set is insufficient to support subsequent analysis and additional data is required. At this time, a data mining instruction is sent to the cloud to request the cloud to provide additional data support. According to the cloud data mining instruction, a defect record set related to the target waistband type and pressing process is obtained from the cloud waistband pressing defect record database. Step S240: The waistband pressing defect record set is generated according to the initial waistband pressing defect record set and the cloud waistband pressing defect record set. Specifically, the initial trouser waist pressing defect record set and the cloud trouser waist pressing defect record set are merged to form a more complete and richer defect record set. The duplicate records in the merged data set are removed to ensure the uniqueness of each record. The data set after merging and deduplication processing is the final trouser waist pressing defect record set. This implementation method quickly obtains the initial trouser waist pressing defect record set related to the target trouser waist type and pressing process by retrieving the pressing database within the workshop. Since the data within the workshop is limited and has limitations (such as insufficient data volume, single type, etc.), cloud data mining is used to supplement and expand the data set. The cloud database contains a wider and more comprehensive data resource, and provides a richer defect record set. By merging the initial data set and the cloud data set, a more comprehensive and representative trouser waist pressing defect record set is generated, which achieves the technical effect of ensuring the richness and diversity of the data and providing more reliable data support for subsequent analysis and optimization work.

[0034] Step S300, performing defect feature recognition on the set of trouser waist pressing defect records, and matching the defect feature recognition results with the K pressing control points to generate Q target pressing feature points. Specifically, using image processing, machine learning or deep learning and other technologies, key feature information, such as the shape, size, position, color, etc. of the defects, is extracted from the set of trouser waist pressing defect records, and the extracted feature information is encoded into a computer-recognizable format, such as converting the features into numerical or vector form. According to the type, nature and other characteristics of the defects, the defect records are classified, such as into multiple types such as wrinkles, deformations, and burns, and a specific set of features is defined for each type of defect for subsequent matching with the pressing control points.

[0035] Perform correlation analysis on K pressing control points and defect features, and use matching algorithms (such as nearest neighbor search, similarity calculation, etc.) to match defect features with K pressing control points to find out which pressing control points have direct or indirect connections with defect features. According to the matching results, screen out pressing control points that are highly correlated with defect features as target pressing feature points, that is, target pressing feature points are pressing control points that have a significant impact on defects after defect feature identification and pressing control point matching. Determine the specific eigenvalues ​​of each target pressing feature point for parameter adjustment in subsequent processes. Finally, generate Q target pressing feature points and their corresponding eigenvalues, where Q is less than or equal to K.

[0036] Step S400, according to the pressing process, the pressing process is disassembled to generate P pressing processes, wherein the P pressing processes include P pressing process parameter sets. Specifically, the pressing process adopted by the target trouser waist is analyzed, including its overall process, key technical points, required equipment parameters, etc., and the variable factors existing in the pressing process are analyzed, such as temperature, pressure, time, pressing plate shape, fabric material, etc., which will affect the pressing effect. According to the complexity of the pressing process and actual needs, the entire pressing process is divided into several sub-processes (i.e., P pressing processes), wherein each pressing process is relatively independent, has clear operating steps and expected goals, and is closely related to the overall pressing effect. Based on past experience, process requirements, equipment performance and other factors, a set of initial parameter sets are set for each pressing process, and these parameter sets include specific numerical ranges or set values ​​of key parameters such as temperature, pressure, and time. Construct the logical relationship and execution sequence between each pressing process to ensure the continuity and effectiveness of the entire pressing process, determine the starting and ending conditions of each pressing process, and the connection method between processes (such as whether cooling, transfer, etc. are required).

[0037] Step S500, press the target waistband according to the P pressing process parameter sets in turn, and collect the process forming image after each pressing process is completed to obtain P process forming image sets. Specifically, confirm that the target waistband has been placed in the correct position of the pressing machine, and the pressing machine is ready according to the initial settings, and check whether all the pressing process parameter sets (P) have been correctly input into the control system of the pressing machine. According to the first set of the P pressing process parameter sets, start the pressing machine to perform the pressing operation of the first process. During the pressing process, monitor the system status in real time to ensure that the pressing machine operates stably according to the preset parameters. When the first pressing process is completed, use the image acquisition device to shoot or scan the pressed waistband to obtain the forming image of the first process, save the collected forming image to a specified location, and mark it as the forming image of the first process. For the remaining P-1 pressing processes, repeat the above steps of pressing and process forming image collection, and collect and save the forming image after each process is completed until all P processes are completed. After the above steps, P process forming images are obtained, and these images are arranged in the order of the processes to form a process forming image set.

[0038] Step S600, based on the Q target pressing feature points and the corresponding Q target pressing feature values, the P process forming image sets are sequentially feature recognized, and P parameter adjustment direction sets and P parameter adjustment step sets are determined. Specifically, the parameters and thresholds required by the image recognition algorithm are set to ensure that the algorithm can accurately recognize the target pressing feature points, and the Q target pressing feature points and their corresponding feature values ​​are loaded as reference standards in the recognition process. For each image in the P process forming image set (in process order), the following steps are performed:

[0039] The molding image of the current process is preprocessed, including denoising, contrast enhancement, edge detection and other operations to improve the accuracy of feature recognition. Use the image recognition algorithm to search for feature points that match the Q target pressing feature points in the molding image of the current process. For the successfully identified feature points, their feature values ​​are compared with the corresponding target pressing feature values. Analyze the difference in feature values ​​to determine in which aspects the current process does not achieve the expected effect (such as insufficient pressing temperature, excessive pressure, etc.). According to the comparison results of the feature values, determine the direction of the parameters that need to be adjusted. For example, if the feature value indicates that the pressing temperature is insufficient, the adjustment direction should be to increase the temperature. For each process, generate a parameter adjustment direction set, which contains all the parameters that need to be adjusted in the process and their directions (such as increasing temperature, reducing pressure, etc.). According to the degree of difference in feature values ​​and the requirements of the pressing process, determine the pace of each parameter adjustment (i.e., the amplitude or increment of adjustment), and for each process, generate a parameter adjustment step set corresponding to the parameter adjustment direction set. Summarize the parameter adjustment direction set and parameter adjustment step set of all processes to form a complete adjustment plan.

[0040] In a possible implementation, step S600 further includes step S610, using the Q target pressing feature points as indexes, performing feature recognition on the P process forming image sets, and generating P process pressing feature value sets. Specifically, each image in the P process forming image sets is loaded in turn, and each image is preprocessed, such as denoising, contrast enhancement, etc., so as to better extract features. Using the known Q target pressing feature points as indexes, the corresponding positions of these feature points are located in the preprocessed image through template matching, feature descriptor matching, etc. Methods. Around each located feature point, feature values ​​describing the feature point are extracted, and these feature values ​​can be quantitative representations of attributes such as color, texture, and shape. The extracted feature values ​​are organized according to the process sequence and feature point index to form P process pressing feature value sets. Step S620, feature deviation recognition is performed on the P process pressing feature value sets based on the Q target pressing feature values, and P feature value deviation sets and P feature value deviation direction sets are obtained. Specifically, each characteristic value in the P process pressing characteristic value set is compared with the corresponding Q target pressing characteristic values, and the difference or deviation degree between the two is calculated. Whether the characteristic value deviates from the target value is determined according to the degree of difference, and the value and direction of the deviation are recorded. Among them, the deviation direction can be a positive deviation (greater than the target value) or a negative deviation (less than the target value). The characteristic value deviation values ​​and directions of all processes are organized into P characteristic value deviation sets and P characteristic value deviation direction sets respectively. Step S630, the P characteristic value deviation sets and the P characteristic value deviation directions are subjected to the same parameter clustering deviation analysis to determine the P parameter adjustment direction sets and the P parameter adjustment step sets. Specifically, the characteristic value deviations involving the same parameter in the P processes are clustered, that is, all characteristic value deviations affecting the same pressing parameter are analyzed together. The degree and direction of the characteristic value deviation in each cluster are analyzed to determine which parameters need to be adjusted and the direction of the adjustment. According to the degree of deviation and process requirements, a suitable adjustment pace is determined for each parameter that needs to be adjusted, and the determination of the pace is used to balance the adjustment effect and production efficiency. The parameter adjustment direction and pace obtained by analysis are organized into P parameter adjustment direction sets and P parameter adjustment pace sets. This implementation method achieves the technical effect of accurately determining the parameters that need to be adjusted during the pressing process and their adjustment direction and pace through meticulous feature recognition, deviation recognition and same-parameter clustering deviation analysis.

[0041] In a possible implementation, step S630 further includes step S631, performing same-parameter clustering on the P eigenvalue deviation sets to generate P eigenvalue deviation clustering clusters. Specifically, the same-parameter clustering is a process of clustering analysis based on eigenvalue deviations that are commonly affected by one or some process parameters. For each process, the relevant eigenvalue deviations (which may be deviations of multiple feature points) are combined into a feature vector, and these feature vectors are used as input data for clustering. Using a clustering algorithm (such as K-means, hierarchical clustering, etc.) and clustering parameters (such as the number of clusters, distance measurement, etc.), the P eigenvectors are clustered to generate several clusters. Each cluster represents a group of processes with similar eigenvalue deviations, which may be affected by the same or similar process parameters. The eigenvalue deviation patterns in each cluster are analyzed to find out their associations with specific process parameters. For example, if "too deep folds" and "easy to reproduce folds" appear in a cluster at the same time, this indicates that these phenomena may be related to uneven pressure. Step S632, using the mean shift algorithm to perform deviation concentration analysis on the P eigenvalue deviation clustering clusters and the corresponding P eigenvalue deviation direction clustering clusters, and generate P parameter adjustment direction sets and the P parameter adjustment pace sets. Specifically, the deviation direction information in each eigenvalue deviation clustering cluster is extracted to form the input data of the eigenvalue deviation direction clustering cluster. The mean shift algorithm is selected as an analysis tool because it can find the natural clustering structure in the data without presetting the number of clusters. The mean shift algorithm is applied to the eigenvalue deviation clustering cluster and the eigenvalue deviation direction clustering cluster to further refine the clustering results and identify the central trend of deviation. The clustering results generated by the mean shift algorithm are analyzed to identify the central trend and pattern of deviation in each clustering cluster, and the parameter adjustment direction and pace corresponding to each clustering cluster are determined based on these trends and patterns. The parameter adjustment direction and pace obtained by the analysis are organized into P parameter adjustment direction sets and P parameter adjustment pace sets. This implementation method first identifies process groups with similar characteristic value deviations through clustering with the same parameters. These groups may be affected by the same process parameters. Then, these clusters are further analyzed using the mean shift algorithm to discover the central trend and pattern of deviations, achieving the technical effect of improving the accuracy of determining the direction and pace of parameter adjustment.

[0042] Step S700, adaptively adjust the P pressing process parameter sets according to the P parameter adjustment direction sets and the P parameter adjustment step sets, and generate P fine-tuning pressing process parameter sets. Specifically, for each pressing process (P in total), initialize a fine-tuning pressing process parameter set, which can be empty or the same as the original pressing process parameter set at the beginning. Traverse each parameter adjustment direction set (P in total), each set corresponds to a pressing process. At the same time, traverse the parameter adjustment step set corresponding to the parameter adjustment direction set. For the pressing process currently traversed, find the specific process parameters (such as temperature, pressure, time, etc.) indicated by each direction in the corresponding parameter adjustment direction set. According to the corresponding step value (i.e., the amplitude or proportion of adjustment) in the parameter adjustment step set, adjust the process parameter up or down. Among them, the step value can be a specific numerical increment or a percentage. Add the adjusted process parameter to the corresponding fine-tuning pressing process parameter set. After each parameter adjustment, check whether the adjusted parameter value is within the allowable range of the equipment, such as whether the temperature is too high or too low, whether the pressure is within the safe range, etc. If the adjusted parameter value is unreasonable, it is corrected according to the preset rules, such as limiting the parameter value to the maximum or minimum allowable value of the equipment. After completing the parameter adjustment of all pressing processes, each pressing process corresponds to a fine-tuning pressing process parameter set containing the adjusted parameters. That is, the fine-tuning pressing process parameter set is a parameter set obtained after adaptive adjustment based on the original pressing process parameter set according to the parameter adjustment direction set and the parameter adjustment step set. These parameters are used in the actual pressing process to improve the pressing quality.

[0043] In a possible implementation, step S700 further includes step S710, randomly adjusting the P pressing process parameter sets according to the P parameter adjustment direction sets and the P parameter adjustment step sets to obtain P stage pressing process parameter sets. Specifically, a random adjustment factor is assigned to each pressing process parameter set, and the factor is ranged according to the step value in the parameter adjustment step set to ensure that the adjustment is within a reasonable range. For each pressing process parameter set, the parameters are randomly adjusted according to the direction in the corresponding parameter adjustment direction set (such as increase, decrease or remain unchanged) combined with the random adjustment factor. This adjustment can be based on the percentage change of the current parameter or a specific numerical increase or decrease. The randomly adjusted parameter sets are saved as stage pressing process parameter sets, which represent the pressing parameter configurations of different stages. Step S720, fitness analysis is performed on the P stage pressing process parameter sets to generate P stage fitness. Specifically, based on each stage pressing process parameter set, the pressing process is predicted or actually performed through simulation or small-scale experiments, and pressing effect data is collected. According to the pressing effect data (such as pressing uniformity, wrinkle conditions, material damage degree, etc.), the quality of the pressing process parameter set of each stage is evaluated. The evaluation result is converted into a fitness value. The higher the fitness value, the better the pressing effect of the pressing process parameter set of this stage. Step S730, iteratively update the P pressing process parameter sets according to the P stage pressing process parameter sets until the preset number of iterations is met, and P fine-tuning pressing process parameter sets are obtained. Specifically, in each iteration cycle, the optimal stage pressing process parameter set is selected according to the stage fitness value. Based on the optimal parameter set, combined with the favorable factors in other parameter sets (such as the configuration of certain specific parameters), the original pressing process parameter set is updated. Repeat steps S710 to S720, continuously generate new stage pressing process parameter sets and evaluate them, and update the original pressing process parameter set at the same time. This process continues until the preset number of iterations is reached or other stop conditions are met. When the iterative process ends, the final updated pressing process parameter set is output as the fine-tuning pressing process parameter set. This implementation method, by introducing a mechanism of random adjustment and iterative update, can gradually approach the optimal parameter configuration of the pressing process while maintaining a certain degree of flexibility, achieving a technical effect that helps explore potential better solutions in the parameter space.

[0044] In a possible implementation, step S730 further includes step S731, judging whether the difference between the P stage fitness of the P stage pressing process parameter set and the P fitness of the P pressing process parameter set meets the preset fitness difference, if not, updating the P stage pressing process parameter set as the adjustment object, and adjusting and iterating according to the P parameter adjustment direction set and the P parameter adjustment step set until the preset number of iterations is met, and obtaining the P fine-tuning pressing process parameter set. Specifically, the difference between the stage fitness of each stage pressing process parameter set and the fitness of the corresponding original pressing process parameter set is calculated, that is, the advantages and disadvantages of the two sets of parameter sets in the pressing effect are compared. It is judged whether the absolute values ​​of these differences all meet the preset fitness difference threshold. The preset fitness difference is a standard set according to actual needs, which is used to measure whether the parameter adjustment brings significant performance improvement. If there is at least one stage fitness difference with the original fitness that does not meet the preset fitness difference, it means that there is room for further optimization of the current stage pressing process parameter set. Therefore, the set of parameters of the stage pressing process that do not meet the conditions are updated as the objects of a new round of adjustment, in preparation for the next round of adjustment iterations.

[0045] For the stage pressing process parameter set selected as the adjustment object, the adjustment is performed according to the P parameter adjustment direction sets and the P parameter adjustment step sets, and steps S710 (random adjustment), S720 (fitness analysis) and the fitness difference evaluation (S731) and the update adjustment object in this step are repeatedly performed until the fitness difference between the fitness of all stage pressing process parameter sets and the original fitness meets the preset fitness difference, or reaches the preset number of iterations. When any of the above stop conditions is met, the iterative process ends. At this time, the final stage pressing process parameter set is output as the fine-tuning pressing process parameter set. These parameter sets are optimized through multiple rounds of iterations, and it is expected that better results can be achieved in the pressing process. This implementation method ensures the effectiveness of parameter adjustment by introducing a judgment mechanism for fitness difference. Only when the performance improvement brought by the adjustment reaches a certain level (that is, the fitness difference meets the preset threshold value), the iterative process will continue or stop, achieving the technical effect of avoiding invalid adjustment iterations and saving computing resources and time.

[0046] Step S800, the parameters of the pressing machine are adjusted by using the P fine-tuning pressing process parameter sets to perform pressing of the target trouser waist. Specifically, the generated P fine-tuning pressing process parameter sets are loaded into the control system of the pressing machine, that is, the parameter set of each process (such as temperature, pressure, time, etc.) is transmitted to the corresponding control module of the pressing machine. Before starting pressing, the pressing machine is initialized, including checking the equipment status, preheating the equipment, calibrating the sensor, etc., to ensure that the pressing machine is in the best working state. According to the loaded fine-tuning pressing process parameter set, the various control parameters of the pressing machine are set, including adjusting the temperature controller, pressure regulating valve, time timer, etc., to match the specific requirements of each pressing process. According to the fine-tuned pressing process parameters, the pressing process of the target trouser waist is performed in sequence. After each process is completed, the pressing machine automatically or according to preset conditions switches to the parameter setting of the next process. During the pressing process, the operating status and pressing effect of the pressing machine are continuously monitored. If any abnormality or deviation from the expected situation is found, fine-tuning is performed automatically or according to the instructions of the operator to ensure the pressing quality. When all the pressing processes are completed, the pressing machine stops working and outputs the pressed target waistband. At this time, a final quality check is performed to ensure that the pressing effect meets the requirements. The embodiment of the present application uses feature point extraction of the target pressing forming image to obtain the pressing control point, obtains the set of waistband pressing defect records through data mining, matches the defect feature recognition results with the pressing control point, generates the target pressing feature point, performs pressing process disassembly, generates multiple pressing processes, presses the target waistband, and collects the process forming image, performs feature recognition on the process forming image set according to the target pressing feature point, determines the parameter adjustment direction set and the parameter adjustment pace set, adaptively adjusts the pressing process parameter set, and performs parameter control on the pressing machine. The technical effect of automatically adjusting and optimizing the waistband pressing parameters to improve production efficiency and pressing quality is achieved.

[0047] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some implementation specifications, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for adaptively controlling trouser waist pressing parameters, characterized in that: include: Collecting a target pressing and forming image of a target trouser waist, extracting feature points of the target pressing and forming image, and obtaining K pressing and forming control points; wherein the K pressing and forming control points correspond to K pressing and forming control feature values; Performing data mining based on the waistband type and pressing process of the target waistband to generate a waistband pressing defect record set; Perform defect feature recognition on the set of trouser waist pressing defect records, and match the defect feature recognition results with the K pressing control points, and select pressing control points related to the defect features according to the matching results as target pressing feature points, and finally generate Q target pressing feature points, and obtain Q target pressing feature values ​​corresponding to the Q target pressing feature points, where Q is less than or equal to K; Decomposing the pressing process according to the pressing process to generate P pressing processes; wherein a set of initial parameters is set for each pressing process to obtain P pressing process parameter sets; Pressing and ironing the target trouser waist according to the P pressing and ironing process parameter sets in sequence, and collecting process forming images after each pressing and ironing process is completed to obtain P pressing and ironing process forming image sets; Based on the Q target pressing feature points and the corresponding Q target pressing feature values, feature recognition is performed on the P pressing process forming image sets in turn to determine P parameter adjustment direction sets and P parameter adjustment step sets; Adaptively adjusting the P pressing process parameter sets according to the P parameter adjustment direction sets and the P parameter adjustment step sets to generate P fine-tuning pressing process parameter sets; Utilizing the P fine-tuning pressing process parameter sets to adjust parameters of the pressing machine, and performing pressing of the target trouser waist; The method of sequentially performing feature recognition on the P sets of pressing process forming images based on the Q target pressing feature points and the corresponding Q target pressing feature values, and determining P sets of parameter adjustment directions and P sets of parameter adjustment steps, comprises: Using the Q target pressing feature points as indexes, feature recognition is performed on the P pressing process forming image sets, the positions of the Q target pressing feature points are located in the pressing process forming images, and corresponding feature values ​​are extracted around each located feature point to obtain a pressing feature value set for each pressing process, thereby generating a pressing feature value set for P pressing processes; Based on the Q target pressing feature values, feature deviation identification is performed on the pressing feature value sets of the P pressing processes to obtain P feature value deviation sets and P feature value deviation direction sets; Performing same-parameter clustering deviation analysis on the P eigenvalue deviation sets and the P eigenvalue deviation directions; analyzing the differences in eigenvalues ​​to determine aspects of the current pressing process that fail to achieve the expected results; determining the parameter directions that need to be adjusted based on the eigenvalue comparison results, generating a parameter adjustment direction set for each pressing process, and obtaining P parameter adjustment direction sets; determining the pace of each parameter adjustment based on the degree of eigenvalue difference and the requirements of the pressing process, generating a parameter adjustment step set for each pressing process, and obtaining P parameter adjustment step sets, the parameter adjustment step set corresponding to the parameter adjustment direction set.

2. A method for adaptively controlling trouser waist pressing parameters as claimed in claim 1, characterized in that: The step of collecting a target pressed and formed image of a target trouser waist, extracting feature points from the target pressed and formed image, and obtaining K pressed and formed control points includes: Get a preset set of pressing control indicators; According to the mapping relationship between the pressing control index and the waist area, the waist area control area is extracted based on the preset pressing control index set to generate a waist area mapping control area set; Regional feature points are extracted based on the waistband mapping control area set to generate the K pressing control points.

3. The method for adaptively controlling trouser waist pressing parameters according to claim 1, characterized in that: The data mining is performed based on the waistband type and pressing process of the target waistband to generate a waistband pressing defect record set, including: Based on the waistband type and pressing process of the target waistband, a pressing database is searched to obtain an initial waistband pressing defect record set; Performing data difference identification on the initial trouser waist pressing defect record set to generate an initial difference; When the initial difference does not meet the preset difference threshold, obtaining a cloud data mining instruction, and obtaining a cloud trouser waist pressing defect record set based on the cloud data mining instruction; The trouser waist pressing defect record set is generated according to the initial trouser waist pressing defect record set and the cloud trouser waist pressing defect record set.

4. The method for adaptively controlling trouser waist pressing parameters according to claim 1, characterized in that: The process of determining P sets of parameter adjustment directions and P sets of parameter adjustment steps based on the same-parameter clustering deviation analysis includes: Performing same-parameter clustering on the P eigenvalue deviation sets to generate P eigenvalue deviation clusters; The mean shift algorithm is used to perform deviation concentration analysis on the P eigenvalue deviation clusters and the corresponding P eigenvalue deviation direction clusters to generate P parameter adjustment direction sets and the P parameter adjustment step sets.

5. The method for adaptively controlling trouser waist pressing parameters according to claim 1, characterized in that: The step of adaptively adjusting the P pressing process parameter sets according to the P parameter adjustment direction sets and the P parameter adjustment step sets to generate P fine-tuning pressing process parameter sets comprises: Randomly adjusting the P sets of pressing process parameters according to the P sets of parameter adjustment directions and the P sets of parameter adjustment steps to obtain P sets of pressing process parameters in stages; Performing fitness analysis on the P-stage pressing process parameter sets to generate P-stage fitness; In each iteration cycle, the optimal stage pressing process parameter set is selected according to the stage fitness to iteratively update the P pressing process parameter sets until the preset number of iterations is met to obtain P fine-tuning pressing process parameter sets.

6. A method for adaptively controlling trouser waist pressing parameters as claimed in claim 5, characterized in that: In each iteration cycle, the optimal stage pressing process parameter set is selected according to the stage fitness to iteratively update the P pressing process parameter sets until the preset number of iterations is met to obtain P fine-tuning pressing process parameter sets, including: Determine whether the difference between the P stage fitness of the P stage pressing process parameter sets and the P fitness of the P pressing process parameter sets meets the preset fitness difference; if not, update the P stage pressing process parameter sets as adjustment objects, and perform adjustment iterations according to the P parameter adjustment direction sets and the P parameter adjustment step sets until the preset number of iterations is met, thereby obtaining the P fine-tuning pressing process parameter sets.

7. The method for adaptively controlling trouser waist pressing parameters according to claim 2, characterized in that: The preset pressing control index set includes pleat straightness, pleat depth, fabric flatness, dimensional stability, and symmetry of trouser waist and lining.

Images