Intelligent management system and method for sanitary towel production line

Abstract

The invention relates to the field of sanitary napkin production, and particularly discloses an intelligent management system and method for a sanitary napkin production line, and the method comprises the steps: firstly obtaining a sanitary napkin gluing equipment gluing production video collected by a camera, sanitary napkin production order information collected by a database, and a sanitary napkin gluing equipment glue head coordinate parameter collected by a sensor; and then feature extraction and correlation analysis are performed on the three parts by using a deep learning technology, and finally, a glue head motion parameter at the next moment is obtained through a decoder, so that a gluing strategy is automatically adjusted according to requirements of different orders, various sanitary towel designs are adapted, and the flexibility and the production efficiency of a production line are improved.

Description

Technical Field

[0001] The present application relates to the field of sanitary napkin production, and more specifically, to an intelligent management system and method for a sanitary napkin production line. Background Art

[0002] Sanitary napkins are absorbent care products made of multiple layers of absorbent material. The core layer is designed to quickly absorb and lock in liquids, while the outer layer prevents leakage and ensures the user's comfort and dryness. Sanitary napkins usually have adhesive on the back to fix them to underwear and prevent them from sliding during use.

[0003] At present, in the production of sanitary napkins, the gluing template of the gluing equipment is preset and fixed, but there are many types of existing sanitary napkins, such as various wing-type or oversized back-swing sanitary napkins, or sanitary napkins without wings, etc., and the positions where gluing is required are different. This requires that the gluing template be replaced according to the type of sanitary napkin when gluing the sanitary napkin. This requires manual replacement of the gluing template during the production process, resulting in high labor consumption and low work efficiency, and high limitations during processing.

[0004] Therefore, an intelligent management system and method for a sanitary napkin production line is desired. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides an intelligent management system and method for a sanitary napkin production line, which first obtains the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the coordinate parameters of the glue head of the sanitary napkin gluing equipment collected by the sensor, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally obtains the glue head motion parameters at the next moment through a decoder, so as to automatically adjust the gluing strategy according to the requirements of different orders, adapt to various sanitary napkin designs, and improve the flexibility and production efficiency of the production line.

[0006] According to one aspect of the present application, there is provided an intelligent management system for a sanitary napkin production line, comprising:

[0007] The sanitary napkin production data acquisition module is used to obtain the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the coordinate parameters of the gluing head of the sanitary napkin gluing equipment collected by the sensor;

[0008] A sanitary napkin production data extraction module, used to extract optimized sanitary napkin gluing multimodal association feature vectors and sanitary napkin gluing equipment glue head coordinate association feature vectors from the sanitary napkin gluing equipment gluing production video collected by the camera, the sanitary napkin production order information collected by the database, and the sanitary napkin gluing equipment glue head coordinate parameters collected by the sensor;

[0009] The glue head motion parameter decoding module is used to obtain the glue head motion parameters at the next moment based on the optimized sanitary napkin gluing multi-modal association feature vector and the sanitary napkin gluing equipment glue head coordinate association feature vector.

[0010] According to another aspect of the present application, there is provided an intelligent management method for a sanitary napkin production line, comprising:

[0011] Obtain the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the gluing head coordinate parameters of the sanitary napkin gluing equipment collected by the sensor;

[0012] Extracting optimized sanitary napkin gluing multimodal association feature vectors and sanitary napkin gluing equipment glue head coordinate association feature vectors from the sanitary napkin gluing equipment gluing production video collected by the camera, the sanitary napkin production order information collected by the database, and the sanitary napkin gluing equipment glue head coordinate parameters collected by the sensor;

[0013] Based on the optimized multi-modal association feature vector of sanitary napkin gluing and the glue head coordinate association feature vector of the sanitary napkin gluing equipment, the glue head motion parameters at the next moment are obtained.

[0014] Compared with the prior art, the present application provides an intelligent management system and method for a sanitary napkin production line, which first obtains the gluing production video of the sanitary napkin gluing equipment collected by a camera, the sanitary napkin production order information collected by a database, and the coordinate parameters of the glue head of the sanitary napkin gluing equipment collected by a sensor, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally uses a decoder to obtain the glue head motion parameters at the next moment, thereby automatically adjusting the gluing strategy according to the requirements of different orders, adapting to various sanitary napkin designs, and improving the flexibility and production efficiency of the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 This is a block diagram of an intelligent management system for a sanitary napkin production line according to an embodiment of the present application.

[0017] Figure 2 This is a block diagram of a sanitary napkin production data extraction module in an intelligent management system for a sanitary napkin production line according to an embodiment of the present application.

[0018] Figure 3 This is a block diagram of a sanitary napkin appearance feature extraction unit in an intelligent management system for a sanitary napkin production line according to an embodiment of the present application.

[0019] Figure 4 It is a block diagram of a pad motion parameter decoding module in an intelligent management system of a sanitary napkin production line according to an embodiment of the present application.

[0020] Figure 5 This is a flow chart of an intelligent management method for a sanitary napkin production line according to an embodiment of the present application.

[0021] Figure 6 is a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0023] Figure 1 FIG. 1 is a block diagram of an intelligent management system for a sanitary napkin production line according to an embodiment of the present application. Figure 1 As shown, the intelligent management system 100 of the sanitary napkin production line according to the embodiment of the present application includes: a sanitary napkin production data acquisition module 110, which is used to obtain the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database and the glue head coordinate parameters of the sanitary napkin gluing equipment collected by the sensor; a sanitary napkin production data extraction module 120, which is used to extract the optimized sanitary napkin gluing multimodal association feature vector and the sanitary napkin gluing equipment glue head coordinate association feature vector from the sanitary napkin gluing production video collected by the camera, the sanitary napkin production order information collected by the database and the sanitary napkin gluing equipment glue head coordinate association feature vector collected by the sensor; a glue head motion parameter decoding module 130, which is used to obtain the glue head motion parameters of the next moment based on the optimized sanitary napkin gluing multimodal association feature vector and the sanitary napkin gluing equipment glue head coordinate association feature vector.

[0024] In the intelligent management system 100 of the sanitary napkin production line, the sanitary napkin production data acquisition module 110 is used to obtain the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the glue head coordinate parameters of the sanitary napkin gluing equipment collected by the sensor. It should be understood that the sanitary napkin is a care product with an absorption function, which is composed of multiple layers of hygroscopic materials. The core layer is designed to quickly absorb and lock the liquid, while the outer layer prevents leakage to ensure that the user remains dry and comfortable. The back is usually provided with an adhesive to fix the sanitary napkin on the underwear to prevent it from sliding. In the current production process, the gluing template of the gluing equipment is fixed. However, due to the wide variety of sanitary napkins, such as wing-type, large back swing type or wingless design, the gluing position is different, so the template must be replaced according to the type of different sanitary napkins. This requires manual template replacement during the production process, resulting in large human resource consumption, low work efficiency, and certain limitations in the processing process. Therefore, in the technical solution of the present application, by acquiring the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the gluing head coordinate parameters of the sanitary napkin gluing equipment collected by the sensor, and combining with deep learning technology, the gluing head motion parameters at the next moment are predicted to automatically adjust the gluing strategy according to the specific needs of different orders. This adjustment enables the production line to adapt to various sanitary napkin designs, thereby improving the flexibility and efficiency of the production line.

[0025] Specifically, obtaining the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the glue head coordinate parameters of the sanitary napkin gluing equipment collected by the sensor is crucial to the realization of an intelligent management system. The gluing production video captured by the camera provides real-time visual data of the gluing process, which can help identify the status of the sanitary napkin on the production line, detect possible defects, and monitor the gluing quality. The production order information provides the specific requirements of the production task, including production quantity, specifications, and priority, which are key to adjusting production parameters and optimizing production scheduling. The glue head coordinate parameters obtained by the sensor provide precise location information for the gluing equipment to ensure the accuracy and consistency of the gluing process. By comprehensively analyzing these diverse data sources, the system can more accurately control the production process, adjust the movement and operation of the gluing equipment in a timely manner, optimize production efficiency, and ensure product quality.

[0026] In the above-mentioned intelligent management system 100 of the sanitary napkin production line, the sanitary napkin production data extraction module 120 is used to extract the optimized sanitary napkin gluing multimodal correlation feature vector and sanitary napkin gluing equipment glue head coordinate correlation feature vector from the sanitary napkin gluing equipment gluing production video collected by the camera, the sanitary napkin production order information collected by the database, and the sanitary napkin gluing equipment glue head coordinate parameters collected by the sensor. In this way, it is helpful to accurately reflect the production status of sanitary napkins, reveal the correlation between different data sources, and provide a basis for optimizing the gluing process, predicting equipment behavior, and making real-time adjustments, thereby improving production efficiency, ensuring product quality, and achieving a higher level of automation and intelligence in the production process.

[0027] Figure 2 FIG. 1 is a block diagram of a sanitary napkin production data extraction module in an intelligent management system for a sanitary napkin production line according to an embodiment of the present application. Figure 2 As shown, in a specific embodiment of the present application, the sanitary napkin production data extraction module 120 includes: a sanitary napkin appearance feature extraction unit 121, which is used to extract features from the sanitary napkin gluing equipment gluing production video collected by the camera to obtain the appearance monitoring and tracking feature vector of the sanitary napkin to be glued; a sanitary napkin production order feature extraction unit 122, which is used to extract features from the sanitary napkin production order information collected by the database to obtain the sanitary napkin production order information associated feature vector; a sanitary napkin multimodal feature construction unit 123, which is used to construct the appearance monitoring and tracking feature vector of the sanitary napkin to be glued and the sanitary napkin production order information associated feature vector to obtain the optimized sanitary napkin gluing multimodal associated feature vector; a glue head coordinate parameter feature extraction unit 124, which is used to extract features from the sanitary napkin gluing equipment glue head coordinate parameters collected by the sensor to obtain the sanitary napkin gluing equipment glue head coordinate associated feature vector.

[0028] It should be understood that the camera video provides real-time visual data on the production line, which can reflect the changes in the position, posture and appearance of the sanitary napkin during the gluing process. Through feature extraction, these video data are converted into appearance monitoring and tracking feature vectors, and the appearance information of the sanitary napkin can be standardized and digitized for subsequent analysis and comparison. Among them, the feature vector includes key visual information such as the edge contour, surface texture, and color characteristics of the sanitary napkin, which helps to monitor and identify the difference between the actual state and the expected state of the sanitary napkin during the gluing process. Through these feature vectors, the production progress of the sanitary napkin can be tracked in real time, possible defects or inconsistencies can be detected, and the production process can be adjusted to ensure the uniformity and quality of the gluing.

[0029] Furthermore, feature extraction is performed on sanitary napkin production order information collected from the database, mainly to improve the accuracy and efficiency of production scheduling and process management. Production order information usually includes key data such as order number, production quantity, product specifications, delivery date, priority, etc., which are crucial to the planning and execution of the production process. However, the original order information often exists in the form of text or tables, and its direct use may be difficult to effectively integrate with real-time data in the production process. Through feature extraction, the production order information is converted into associated feature vectors, which can convert complex order data into a unified and quantifiable feature set, which can facilitate comprehensive analysis with other data sources (such as production videos, equipment sensor data, etc.).

[0030] Furthermore, by combining visual features with production order information, factors that may lead to product defects can be identified more accurately, which helps to monitor the production process of sanitary napkins in all aspects and ensure the consistency and stability of the product in each production link.

[0031] In particular, feature extraction is performed on the coordinate parameters of the gluing head of the sanitary napkin gluing equipment collected by the sensor to obtain the associated feature vector of the gluing head coordinates of the sanitary napkin gluing equipment, mainly to achieve precise control and optimization of the gluing process. The sensor provides real-time position data about the gluing head of the gluing equipment, which includes information such as the coordinates, motion trajectory, and position change of the gluing head. The original coordinate data may be very complex and large, and directly using this data for analysis and processing may not be efficient enough, so it is necessary to convert it into a concise and meaningful feature vector through feature extraction.

[0032] Figure 3 FIG. 1 is a block diagram of a sanitary napkin appearance feature extraction unit in an intelligent management system for a sanitary napkin production line according to an embodiment of the present application. Figure 3 As shown, in a specific embodiment of the present application, the sanitary napkin appearance feature extraction unit 121 includes: a gluing production key frame extraction subunit 1211, which is used to extract the gluing production video of the sanitary napkin gluing equipment collected by the camera at a predetermined sampling frequency to obtain a plurality of appearance monitoring key frames of the sanitary napkin to be glued; an appearance monitoring image pixel enhancement subunit 1212, which is used to respectively pass the plurality of appearance monitoring key frames of the sanitary napkin to be glued through a sanitary napkin appearance monitoring pixel enhancement convolutional neural network as a generator to obtain a plurality of appearance monitoring image pixel enhancement key frames of the sanitary napkin to be glued; a sanitary napkin appearance monitoring feature encoding subunit 1213, which is used to feature encode the plurality of appearance monitoring image pixel enhancement key frames of the sanitary napkin to be glued to obtain the appearance monitoring tracking feature vector of the sanitary napkin to be glued.

[0033] It should be understood that the video generated by the camera contains a large amount of continuous frame data. Although this data can provide detailed information about the production process, a lot of information in the continuous frame images may be redundant. Therefore, analyzing all frames may not only be computationally burdensome, but also difficult to extract useful key information from them. By setting a predetermined sampling frequency and extracting multiple key frames from the video, the amount of data can be significantly reduced while retaining key production process information. Each key frame represents an important moment in the production process, which is usually at a specific stage or critical operation moment in the production process. This method can effectively capture the appearance characteristics of sanitary napkins at different time points, such as the uniformity of glue coating, position accuracy, etc., so as to achieve accurate monitoring and analysis of production quality. Among them, the setting of the sampling frequency needs to comprehensively consider the actual situation of the production line and the monitoring objectives. If the sampling frequency is too low, it may not be able to accurately capture subtle changes in the production process, while if the sampling frequency is too high, it may lead to excessive computing and storage burdens. By setting the sampling frequency reasonably, it is possible to ensure that the selection of key frames is representative while controlling the complexity of data processing.

[0034] Furthermore, during the production process, although the camera can capture the appearance information of the sanitary napkin, the original image may be blurred, noisy, or have lost details due to interference from lighting conditions, camera resolution, or production environment. These problems may affect the accurate detection of glue uniformity and other production defects. Therefore, image enhancement technology is needed to improve the clarity and detail of the image. Using pixel-enhanced convolutional neural networks as generators for image enhancement, convolutional neural networks can learn the mapping relationship from low-quality images to high-quality images. By training the network, the resolution and details of the image can be improved, so that the visual effect and amount of information of the image are significantly improved.

[0035] Furthermore, feature encoding is performed on multiple pixel-enhanced key frames of the appearance monitoring images of the sanitary napkins to be glued to obtain the appearance monitoring tracking feature vector of the sanitary napkins to be glued, in order to convert high-quality image information into concise and useful feature representations for effective analysis and tracking. Through feature encoding, these complex visual information are compressed into high-dimensional feature vectors, making the data easier to process and analyze. This vectorized representation can effectively reduce data redundancy, improve computational efficiency, and retain information that is valuable for analysis. The feature encoding process usually involves mapping the pixel values ​​of the image to a higher-level feature space, and extracting abstract features of the image through a convolutional neural network or other deep learning models.

[0036] In a specific embodiment of the present application, the sanitary napkin appearance monitoring feature encoding subunit 1213 includes: passing the multiple pixel enhanced key frames of the sanitary napkin appearance monitoring images to be glued through a sanitary napkin appearance monitoring feature encoder based on a void convolutional neural network to obtain an appearance monitoring tracking feature map of the sanitary napkin to be glued; downsampling the appearance monitoring tracking feature map of the sanitary napkin to be glued to obtain the appearance monitoring tracking feature vector of the sanitary napkin to be glued.

[0037] It should be understood that the application of dilated convolutional neural networks in feature encoding can effectively extract and represent important information in images while retaining a wider range of contextual information, which is critical for high-resolution production monitoring. Dilated convolution is a technique that expands the receptive field by introducing dilation in the convolution operation, which means that it can capture a wider range of image contextual information without increasing the amount of computation. Traditional convolution operations may not be able to fully capture all relevant features in an image due to the limitation of the receptive field, especially when there are large ranges of changes or details in the image. Dilated convolution can capture image features more widely while maintaining the spatial resolution of the image by inserting holes in the convolution kernel. This is particularly effective for dealing with complex situations in production, such as subtle inconsistencies in the glue coating area. Specifically, each layer of the first convolutional neural network model in the feature encoder for monitoring the appearance of sanitary napkins to be glued based on the atrous convolutional neural network performs atrous convolution processing based on the first convolutional kernel, pooling processing along the channel dimension and nonlinear activation processing on the input data in the forward pass of the layer, so that the first convolutional neural network model outputs a first feature matrix from the last layer; each layer of the sanitary napkin production order convolutional neural network model as a feature encoder in the feature encoder for monitoring the appearance of sanitary napkins to be glued based on the atrous convolutional neural network performs atrous convolution processing based on the second convolutional kernel, pooling processing along the channel dimension and nonlinear activation processing on the input data in the forward pass of the layer, so that the second feature matrix is ​​output from the last layer of the sanitary napkin production order convolutional neural network model as a feature encoder; the first feature matrix and the second feature matrix are fused to obtain the appearance monitoring tracking feature map of the sanitary napkin to be glued.

[0038] Furthermore, feature maps are highly detailed feature representations extracted from images by deep learning models, usually containing rich spatial information and a large amount of pixel data. However, such high-resolution feature maps may incur large computational overhead in practical applications, especially when real-time monitoring and processing are required. Downsampling is the process of combining the data of multiple adjacent pixels in a feature map into a single representative value. This can be achieved through a variety of methods, such as max pooling, average pooling, or other more complex sampling techniques. Through downsampling, the size of the feature map can be effectively compressed, converting it from high-dimensional data to a more concise feature vector. This process not only reduces the computational and storage requirements, but also retains the core feature information in the image, such as key spatial structure and texture features. The feature vector can therefore represent the main information of the image more compactly, making subsequent analysis and decision-making more efficient.

[0039] In a specific embodiment of the present application, the sanitary napkin production order feature extraction unit 122 includes: passing the sanitary napkin production order information collected from the database through a sanitary napkin production order information context encoder including an embedding layer to obtain multiple sanitary napkin production order information feature vectors; arranging the multiple sanitary napkin production order information feature vectors in two dimensions to obtain a sanitary napkin production order associated feature matrix; passing the sanitary napkin production order associated feature matrix through a sanitary napkin production order convolutional neural network as a feature encoder to obtain the sanitary napkin production order information associated feature vector.

[0040] It should be understood that production order information usually includes data of various dimensions, such as order number, production time, product specifications, production batches, demand, etc. In its original form, this information may be difficult to directly use for analysis and processing due to its high dimensionality and diversity. Embedding Layer is an effective way to process this type of data. The embedding layer can convert this information into a more semantically meaningful feature representation by mapping discrete and sparse data (such as order number, product specifications, etc.) into a continuous, low-dimensional vector space. This mapping not only retains the structural information of the original data, but also captures the potential associations and contextual information between the data by learning to map similar inputs to similar vector space positions. The sanitary napkin production order information context encoder further processes these embedded features and deeply analyzes and encodes the information through a neural network model. These encoders usually include convolutional layers, recurrent layers, or self-attention mechanisms to capture and integrate complex patterns and relationships in the production order information. The sanitary napkin production order information feature vector obtained through this processing flow can effectively summarize and represent the key information in the production order. Specifically, the sanitary napkin production order information collected from the database is segmented to obtain a production order word sequence; the embedding layer of the sanitary napkin production order information context encoder including the embedding layer is used to map each production order word in the production order word sequence into a word embedding vector to obtain a sequence of production order word embedding vectors; the converter-based Bert model of the sanitary napkin production order information context encoder including the embedding layer is used to perform global context semantic encoding on the sequence of production order word embedding vectors to obtain multiple sanitary napkin production order information feature vectors.

[0041] Furthermore, feature vectors are obtained by embedding and encoding production order information, and they effectively summarize the key information and contextual features of each order. However, these feature vectors may be difficult to use directly for comparison and analysis if they are not properly organized and arranged. Arranging the feature vectors in two dimensions, that is, organizing them in a two-dimensional space in the form of a matrix, can effectively show the relationship and data structure between orders. This two-dimensional arrangement process usually includes arranging the feature vectors according to some rules (such as the time sequence of the order, product type or other relevant factors) to form a feature matrix. This matrix not only retains the semantic information of the original features, but also can show the similarities and differences between different orders through the structure of the matrix.

[0042] Furthermore, the sanitary napkin production order associated feature matrix is ​​formed by two-dimensionally arranging the feature vectors of the order information, which integrates various information of the order into a matrix. However, this matrix may contain a large number of dimensions and complex relationships, making direct analysis difficult. Convolutional neural networks (CNNs), especially CNNs for image processing, have powerful feature extraction and pattern recognition capabilities. Applying convolutional neural networks to order-associated feature matrices can effectively extract key associated features from them. Specifically, each layer of the sanitary napkin production order convolutional neural network used as a feature encoder performs convolution processing, mean pooling processing based on local feature matrices, and nonlinear activation processing on the input data in the forward pass of the layer to output the sanitary napkin production order information associated feature vector by the last layer of the sanitary napkin production order convolutional neural network used as a feature encoder, wherein the input of the sanitary napkin production order convolutional neural network used as a feature encoder is the sanitary napkin production order associated feature matrix.

[0043] In a specific embodiment of the present application, the sanitary napkin multimodal feature construction unit 123 includes: fusing the appearance monitoring and tracking feature vector of the sanitary napkin to be glued and the sanitary napkin production order information associated feature vector to obtain a sanitary napkin gluing multimodal associated feature vector; based on the sanitary napkin gluing equipment glue head coordinate associated feature vector, performing fine-grained distribution difference correction on the sanitary napkin gluing multimodal associated feature vector based on deep-level shift information to obtain an optimized sanitary napkin gluing multimodal associated feature vector.

[0044] It should be understood that the fusion of the appearance monitoring and tracking feature vector of the glued sanitary napkin and the associated feature vector of the sanitary napkin production order information is to achieve more comprehensive and accurate production process control and optimization. Among them, the appearance monitoring and tracking feature vector provides information about the actual status and quality of the sanitary napkin on the production line, including visual data such as appearance, glue coating uniformity and position. The associated feature vector of the production order information contains the specific requirements of the order, such as production quantity, specifications and priority. These feature vectors describe the key elements of the production process from different perspectives, but using any type of data alone may not fully capture the complex situation of the production process. By fusing these two feature vectors, the needs of the production order and the actual production status of the sanitary napkin can be comprehensively considered. This fusion can not only combine the detailed requirements of the production order with the status data in the actual production process, but also reveal the matching degree and difference between the order requirements and the actual production.

[0045] In particular, in the technical solution of the present application, considering that the multimodal correlation feature vector of sanitary napkin gluing integrates visual monitoring features and production order information features, its feature dimensions involve complex relationships between image pixels, order data and their combination. These features usually have high dimensions and complex distributions, and the correlation between data is not simply linear. The feature vector associated with the glue head coordinates of the sanitary napkin gluing equipment mainly comes from sensor data, reflecting the physical position and motion state of the equipment. The feature dimension is low, and the device coordinates are usually represented in a more direct way. The differences in the sources, processing and representation methods of these two types of features lead to fine-grained inconsistencies in their distribution in the feature space. This inconsistency is reflected in the different local structures of the feature manifold. The fine-grained inconsistency will cause problems when the features are fused into the feature vector for decoding the glue head motion parameters, resulting in abnormal stacking, which is manifested in that the decoded values ​​may be unnaturally distributed or aggregated on the boundary of the feature manifold, thereby affecting the accurate prediction and adjustment of the glue head motion parameters. Therefore, in the technical solution of the present application, based on the gluing head coordinate association feature vector of the sanitary napkin gluing equipment, the fine-grained distribution difference correction based on deep-level shift information is performed on the multimodal association feature vector of the sanitary napkin gluing to obtain an optimized multimodal association feature vector of the sanitary napkin gluing.

[0046] Among them, based on the sanitary napkin gluing equipment glue head coordinate associated feature vector, the sanitary napkin gluing multimodal associated feature vector is corrected for fine-grained distribution differences based on deep-level shift information to obtain an optimized sanitary napkin gluing multimodal associated feature vector, including: calculating the positional difference between the sanitary napkin gluing multimodal associated feature vector and the sanitary napkin gluing equipment glue head coordinate associated feature vector to obtain a sanitary napkin production shift positional difference representation vector; after the sanitary napkin gluing multimodal associated feature vector and the sanitary napkin gluing equipment glue head coordinate associated feature vector are cascaded into a sanitary napkin production joint feature vector, the sanitary napkin production joint feature vector is input into a one-dimensional convolutional layer to obtain a sanitary napkin production shift convolutional coding representation vector; using the first sanitary napkin production linear embedding weight matrix to perform sanitary napkin production linear embedding modulation coding on the sanitary napkin gluing multimodal associated feature vector to obtain a sanitary napkin gluing multimodal associated feature vector linear modulation vector. After the measurement, the positional difference between the linear modulation vector of the multimodal association feature vector of the sanitary napkin gluing and the glue head coordinate association feature vector of the sanitary napkin gluing equipment is calculated to obtain the sanitary napkin production shift modulation difference representation vector; the sanitary napkin production shift positional difference representation vector is used as the value vector, the sanitary napkin production shift convolutional coding representation vector and the sanitary napkin production shift modulation difference representation vector are used as the query vector and the key vector, and the sanitary napkin production shift positional difference representation vector, the sanitary napkin production shift convolutional coding representation vector and the sanitary napkin production shift modulation difference representation vector are input into the shift corrector based on the simulated converter structure to obtain the sanitary napkin production multi-scale hierarchical fusion shift modulation weight vector; the positional point multiplication between the sanitary napkin production multi-scale hierarchical fusion shift modulation weight vector and the sanitary napkin gluing multimodal association feature vector is calculated to obtain the optimized sanitary napkin gluing multimodal association feature vector.

[0047] The following optimization formula is used to represent that the fine-grained distribution difference correction based on the deep-level shift information of the multi-modal correlation feature vector of the gluing of the sanitary napkin is performed based on the gluing head coordinate correlation feature vector of the sanitary napkin gluing equipment to obtain the optimized multi-modal correlation feature vector of the gluing of the sanitary napkin;

[0048] Wherein, the optimization formula is:

[0049]

[0050] V 2d =Conv 1D ([V 1 ; V 2 ])

[0051]

[0052] V 1 ′=V 1⊙V d

[0053] Among them, V 1d The vector representing the displacement of sanitary napkin production is expressed by position difference, V 1 represents the multi-modal correlation feature vector of sanitary napkin gluing, V 2 represents the associated feature vector of the glue head coordinates of the sanitary napkin gluing equipment, V 2d represents the sanitary napkin production shift convolutional coding representation vector, V 3d represents the shift modulation difference representation vector of sanitary napkin production, W 1 represents the linear embedding weight matrix of the first sanitary napkin production line, b 1 represents the modulation bias vector for sanitary napkin production, V d represents the multi-scale hierarchical fusion shift modulation weight vector of sanitary napkin production, V 3d T represents the transposed vector of the sanitary napkin production shift modulation differential representation vector, d represents the length of the sanitary napkin production shift convolutional coding representation vector, Indicates position difference, Conv 1D represents a one-dimensional convolutional layer, [V 1 ; V 2 ] represents the concatenation of the multi-modal correlation feature vector of the sanitary napkin gluing and the glue head coordinate correlation feature vector of the sanitary napkin gluing equipment, Relu represents the linear rectification function, represents matrix multiplication, Softmax represents the normalized exponential function, ⊙ represents point multiplication by position, V 1 ' represents the optimized multi-modal correlation feature vector of sanitary napkin gluing.

[0054] In the technical solution of the present application, in the process of fusing the multimodal correlation feature vector of gluing the sanitary napkin and the glue head coordinate correlation feature vector of the gluing device for the sanitary napkin, the fine-grained inconsistency of the feature distribution between the multimodal correlation feature vector of gluing the sanitary napkin and the glue head coordinate correlation feature vector of the gluing device for the sanitary napkin is taken into consideration, which will cause the fused glue head motion parameter decoding feature vector to have abnormal stacking on the boundary of the feature manifold. Based on this, in the technical solution of the present application, before fusing the multimodal correlation feature vector of gluing the sanitary napkin and the glue head coordinate correlation feature vector of the gluing device for the sanitary napkin to obtain the glue head motion parameter decoding feature vector, the multimodal correlation feature vector of gluing the sanitary napkin is corrected for fine-grained distribution differences based on deep-level shift information based on the glue head coordinate correlation feature vector of the gluing device for the sanitary napkin.

[0055] In the technical solution of the present application, a fine-grained distribution difference correction method based on deep-level shift information is proposed to address the problem of fine-grained inconsistency in feature distribution that occurs in the process of fusing the multimodal correlation feature vectors of sanitary napkin gluing and the correlation feature vectors of the glue head coordinates of the sanitary napkin gluing equipment. The method aims to reduce the abnormal stacking phenomenon of the glue head motion parameter decoding feature vectors on the boundary of the feature manifold, so as to improve the robustness and classification performance of the model.

[0056] Specifically, the method first calculates the element-by-element difference between the multimodal correlation feature vector of gluing on sanitary napkins and the feature vector associated with the glue head coordinate of the gluing equipment for sanitary napkins to obtain the element-by-element difference representation vector of the production displacement of sanitary napkins, thereby capturing the difference between the two feature vectors. Then, the multimodal correlation feature vector of gluing on sanitary napkins and the feature vector associated with the glue head coordinate of the gluing equipment for sanitary napkins are concatenated into a sanitary napkin production joint feature vector and input into a one-dimensional convolutional layer to generate a convolutional coding representation vector of the production displacement of sanitary napkins. That is, the deep-level displacement implicit information between the multimodal correlation feature vector of gluing on sanitary napkins and the feature vector associated with the glue head coordinate of the gluing equipment for sanitary napkins is captured by one-dimensional convolutional coding. At the same time, the first sanitary napkin production linear embedding weight matrix is ​​used to perform sanitary napkin production linear embedding modulation coding on the sanitary napkin gluing multimodal correlation feature vector into a sanitary napkin gluing multimodal correlation feature vector linear modulation vector, and the element-by-element difference between the sanitary napkin gluing multimodal correlation feature vector linear modulation vector and the sanitary napkin gluing equipment glue head coordinate correlation feature vector is calculated to obtain the sanitary napkin production shift modulation differential representation vector. The sanitary napkin production shift modulation differential representation vector is used to represent the position expression of the displacement information of the sanitary napkin gluing multimodal correlation feature vector and the sanitary napkin gluing equipment glue head coordinate correlation feature vector in the linear space. That is, in the technical solution of the present application, the sanitary napkin production shift position difference, the sanitary napkin production shift convolution coding representation vector and the sanitary napkin production shift modulation differential representation vector respectively represent the explicit expression of the fine-grained distribution difference displacement information between the sanitary napkin gluing multimodal correlation feature vector and the sanitary napkin gluing equipment glue head coordinate correlation feature vector in different feature spaces.

[0057] Next, the shifted element-by-element differential representation vector of sanitary napkin production is used as the value vector, the shifted convolutional coding representation vector of sanitary napkin production and the shifted modulation differential representation vector of sanitary napkin production are used as the query vector and the key vector respectively, and these vectors are input into the shift corrector based on the simulated converter structure to generate the multi-scale hierarchical fusion shift modulation weight vector of sanitary napkin production. In this way, the importance of features of different scales is evaluated by the attention mechanism, that is, the attention mechanism is used to simulate the hierarchical precipitation expression of the fine-grained difference feature distribution between the multi-modal association feature vector of the sanitary napkin gluing and the glue head coordinate association feature vector of the sanitary napkin gluing equipment in different feature spaces. Finally, the element-by-element product of the multi-scale hierarchical fusion shift modulation weight vector of sanitary napkin production and the multi-modal association feature vector of sanitary napkin gluing is calculated to obtain the optimized multi-modal association feature vector of sanitary napkin gluing, that is, the weight is assigned to each element in the multi-modal association feature vector of sanitary napkin gluing through the weighted operation, which reduces the influence of local disturbances and improves the quality of the feature vector.

[0058] In summary, in the method, the position difference of the sanitary napkin production shift, the convolutional coding representation vector of the sanitary napkin production shift and the modulation difference representation vector of the sanitary napkin production shift respectively represent the explicit expression of the fine-grained distribution difference shift information between the multi-modal association feature vector of the sanitary napkin gluing and the glue head coordinate association feature vector of the sanitary napkin gluing equipment in different feature spaces. The generation of the multi-scale hierarchical fusion shift modulation weight vector of the sanitary napkin production uses the attention mechanism to evaluate the importance of features of different scales and provide weight information for feature shift correction. Finally, the generation of the optimized multi-modal association feature vector of the sanitary napkin gluing reduces the impact of disturbances, improves the robustness of the model to local disturbances in the input data, enhances the model's adaptability to noisy or variant data sets, improves the fusion quality of the feature vector, and improves the overall classification performance and training efficiency of the model.

[0059] In a specific embodiment of the present application, the glue head coordinate parameter feature extraction unit 124 includes: passing the sanitary napkin gluing equipment glue head coordinate parameters collected by the sensor through the sanitary napkin gluing equipment glue head coordinate semantic converter to obtain multiple sanitary napkin gluing equipment glue head coordinate feature vectors; splicing the multiple sanitary napkin gluing equipment glue head coordinate feature vectors into the sanitary napkin gluing equipment glue head coordinate associated feature vector.

[0060] It should be understood that the glue head coordinate parameters collected by the sensor are usually the actual position data of the equipment during the production process. These data are provided in the form of raw numerical values. Although they can describe the location information of the equipment, they may lack sufficient context and semantic association when used directly for analysis and decision-making. The semantic converter of the glue head coordinates of sanitary napkin gluing equipment processes the original coordinate data and maps these data into a more structured and meaningful feature space. The key to this process is to convert the coordinate data from its original quantitative form into a feature vector that can better reflect the operating status and production conditions of the equipment. Among them, the semantic converter usually includes a series of data processing and feature extraction techniques, such as embedding layers, feature encoding or nonlinear mapping, which can convert coordinate parameters into more abstract and expressive feature representations. The semantic converter can identify the potential patterns and relationships in the coordinate data through learning algorithms. By semantically converting these coordinate data, key information related to equipment performance, gluing quality or production consistency can be extracted. Specifically, the sanitary napkin gluing equipment glue head coordinate parameters collected by the sensor are segmented to obtain a glue head coordinate word sequence; the embedding layer of the sanitary napkin gluing equipment glue head coordinate semantic converter is used to map each glue head coordinate word in the glue head coordinate word sequence into a word embedding vector to obtain a sequence of glue head coordinate word embedding vectors; the converter-based Bert model of the sanitary napkin gluing equipment glue head coordinate semantic converter is used to perform global context semantic encoding on the sequence of glue head coordinate word embedding vectors to obtain multiple sanitary napkin gluing equipment glue head coordinate feature vectors.

[0061] Furthermore, the purpose of splicing multiple glue head coordinate feature vectors of sanitary napkin gluing equipment into a glue head coordinate associated feature vector of sanitary napkin gluing equipment is to integrate and synthesize the feature information of each glue head position, thereby forming a high-dimensional feature vector that can fully represent the equipment status and operating conditions. Among them, each glue head coordinate feature vector of sanitary napkin gluing equipment usually represents the specific position information and related features of a single glue head, which may include the current position, motion trajectory, operating status, etc. of the glue head. Although these feature vectors can describe the details of the glue head at the individual level, if the performance and operating status of the entire gluing equipment are to be fully analyzed, a single feature vector may not be sufficient to provide a sufficient global perspective. By splicing multiple glue head coordinate feature vectors, the position information of different glue heads can be integrated into a unified feature vector. This splicing operation arranges the feature data of each glue head in order or rule, thereby forming a high-dimensional comprehensive feature representation. This integrated feature vector can more comprehensively reflect the status of the entire equipment, including the relative position and coordination between each glue head and the consistency of the overall operation.

[0062] In the intelligent management system 100 of the sanitary napkin production line, the glue head motion parameter decoding module 130 is used to obtain the glue head motion parameters at the next moment based on the optimized multi-modal correlation feature vector of the sanitary napkin gluing and the glue head coordinate correlation feature vector of the sanitary napkin gluing equipment. It should be understood that the optimized multi-modal correlation feature vector of the sanitary napkin gluing contains comprehensive information extracted from multiple data sources (such as images, sensor data, production logs, etc.), which can describe various states and conditions in the gluing process. The glue head coordinate correlation feature vector of the sanitary napkin gluing equipment focuses on the specific position and motion trajectory of the equipment glue head, reflecting the position change and movement of the glue head in actual operation. By combining these two feature vectors, a more comprehensive feature description can be formed to predict the glue head motion parameters at the next moment. Through this prediction, the production system can adjust the position and motion strategy of the glue head in real time to cope with various production needs and environmental changes, and ensure the continuity and stability of the gluing process.

[0063] Figure 4 FIG. 1 is a block diagram of a pad motion parameter decoding module in an intelligent management system for a sanitary napkin production line according to an embodiment of the present application. Figure 4 As shown, in a specific embodiment of the present application, the glue head motion parameter decoding module 130 includes: a gluing equipment gluing feature association unit 131, used to associate the optimized sanitary napkin gluing multimodal association feature vector and the sanitary napkin gluing equipment glue head coordinate association feature vector to obtain a glue head motion parameter decoding feature vector; a gluing equipment glue head motion parameter decoding unit 132, used to pass the glue head motion parameter decoding feature vector through a decoder to obtain a decoding value, and the decoding value is used to represent the glue head motion parameter at the next moment.

[0064] It should be understood that by associating these two feature vectors, a glue head motion parameter decoding feature vector containing comprehensive information can be obtained. This decoding feature vector integrates the multimodal data of the glue coating process and the detailed information of the glue head motion, so as to comprehensively reflect the operating status of the equipment and various changes in the glue coating process.

[0065] Furthermore, the glue head motion parameter decoding feature vector is a high-dimensional representation obtained by associating the multimodal features of sanitary napkin gluing and the glue head coordinate features of the gluing equipment. It contains multi-dimensional information in the gluing process, such as equipment status, environmental factors, glue head position, etc. Although this feature vector contains rich information, its form is still abstract and high-dimensional, and it is not intuitive and operable to be directly applied to actual operations. The role of the decoder is to convert this complex feature vector into a specific decoding value, that is, the glue head motion parameter at the next moment. The decoder is usually a trained model or algorithm that can generate a predicted output value based on the input feature vector. This process involves mapping high-dimensional features to specific and actual motion parameter values, such as the position, speed, direction, etc. of the glue head. Using these decoded values, the production system can adjust the position and motion strategy of the glue head in real time to ensure the accuracy and consistency of the gluing process.

[0066] In summary, the embodiment of the present application first obtains the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the coordinate parameters of the glue head of the sanitary napkin gluing equipment collected by the sensor, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally uses the decoder to obtain the glue head motion parameters at the next moment, so as to automatically adjust the gluing strategy according to the requirements of different orders, adapt to various sanitary napkin designs, and improve the flexibility and production efficiency of the production line.

[0067] As described above, the intelligent management system 100 of the sanitary napkin production line according to the embodiment of the present application can be implemented in various terminal devices. In one example, the intelligent management system 100 of the sanitary napkin production line can be integrated into the terminal device as a software module and / or a hardware module. For example, the intelligent management system 100 of the sanitary napkin production line can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the intelligent management system 100 of the sanitary napkin production line can also be one of the many hardware modules of the terminal device.

[0068] Alternatively, in another example, the intelligent management system 100 of the sanitary napkin production line and the terminal device may also be separate devices, and the intelligent management system 100 of the sanitary napkin production line may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0069] Figure 5 FIG. 1 is a flow chart of an intelligent management method for a sanitary napkin production line according to an embodiment of the present application. Figure 5As shown, according to the intelligent management method of the sanitary napkin production line of the embodiment of the present application, it includes: S110, obtaining the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database and the glue head coordinate parameters of the sanitary napkin gluing equipment collected by the sensor; S120, extracting the optimized sanitary napkin gluing multimodal association feature vector and the sanitary napkin gluing equipment glue head coordinate association feature vector from the sanitary napkin gluing production video collected by the camera, the sanitary napkin production order information collected by the database and the sanitary napkin gluing equipment glue head coordinate association feature vector; S130, based on the optimized sanitary napkin gluing multimodal association feature vector and the sanitary napkin gluing equipment glue head coordinate association feature vector, obtaining the glue head motion parameters at the next moment.

[0070] Here, those skilled in the art will appreciate that the specific operations of each step in the intelligent management method for the sanitary napkin production line have been described in detail above. Figures 1 to 4 The intelligent management system of the sanitary napkin production line has been introduced in detail, and therefore, its repeated description will be omitted.

[0071] Below, reference Figure 6 To describe an electronic device according to an embodiment of the present application.

[0072] See also Figure 6 , Figure 6 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0073] The processor 11 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solution provided in the embodiment of the present application; the memory 12 can be implemented in the form of a read-only memory (Read Only Memory, ROM), a static storage device, a dynamic storage device or a random access memory (Random Access Memory, RAM), etc.

[0074] The memory 12 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 12, and the processor 11 calls and executes the intelligent management method of the sanitary napkin production line of the embodiment of this application;

[0075] Input / output interface 13, used to implement information input and output;

[0076] The communication interface 14 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0077] A bus 15 that transmits information between the various components of the device (e.g., the processor 11, the memory 12, the input / output interface 13, and the communication interface 14);

[0078] The processor 11 , the memory 12 , the input / output interface 13 and the communication interface 14 are connected to each other in communication within the device via a bus 15 .

[0079] An embodiment of the present application also provides a storage medium, which is a computer-readable storage medium. The storage medium stores a computer program, and when the computer program is executed by a processor, the intelligent management method of the sanitary napkin production line is implemented.

[0080] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0081] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0082] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0083] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0084] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0085] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0086] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0087] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0088] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0089] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method of each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs. The preferred embodiment of the embodiment of the present application is described above with reference to the accompanying drawings, and the scope of rights of the embodiment of the present application is not limited thereto. Any modifications, equivalent substitutions and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of rights of the embodiments of the present application.

Claims

1. An intelligent management system for a sanitary napkin production line, characterized in that: include: The sanitary napkin production data acquisition module is used to obtain the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the coordinate parameters of the gluing head of the sanitary napkin gluing equipment collected by the sensor; A sanitary napkin production data extraction module, used to extract optimized sanitary napkin gluing multimodal association feature vectors and sanitary napkin gluing equipment glue head coordinate association feature vectors from the sanitary napkin gluing equipment gluing production video collected by the camera, the sanitary napkin production order information collected by the database, and the sanitary napkin gluing equipment glue head coordinate parameters collected by the sensor; The glue head motion parameter decoding module is used to obtain the glue head motion parameters at the next moment based on the optimized sanitary napkin gluing multi-modal association feature vector and the sanitary napkin gluing equipment glue head coordinate association feature vector.

2. The intelligent management system for a sanitary napkin production line according to claim 1, characterized in that: The sanitary napkin production data extraction module comprises: A sanitary napkin appearance feature extraction unit, used for extracting features from the gluing production video of the sanitary napkin gluing equipment collected by the camera to obtain a monitoring and tracking feature vector for the appearance of the sanitary napkin to be glued; A sanitary napkin production order feature extraction unit, used to extract features from the sanitary napkin production order information collected from the database to obtain a sanitary napkin production order information associated feature vector; A sanitary napkin multimodal feature construction unit, used for constructing the appearance monitoring and tracking feature vector of the sanitary napkin to be glued and the sanitary napkin production order information association feature vector to obtain the optimized sanitary napkin gluing multimodal association feature vector; The glue head coordinate parameter feature extraction unit is used to extract the feature of the glue head coordinate parameters of the sanitary napkin gluing equipment collected by the sensor to obtain the glue head coordinate associated feature vector of the sanitary napkin gluing equipment.

3. The intelligent management system for a sanitary napkin production line according to claim 2, characterized in that: The sanitary napkin appearance feature extraction unit comprises: A gluing production key frame extraction subunit, used for extracting the gluing production video of the sanitary napkin gluing equipment collected by the camera at a predetermined sampling frequency to obtain a plurality of appearance monitoring key frames of the sanitary napkins to be glued; An appearance monitoring image pixel enhancement subunit is used to pass the multiple appearance monitoring key frames of the sanitary napkins to be glued through the appearance monitoring pixel enhancement convolutional neural network of the sanitary napkins to be glued as a generator to obtain multiple appearance monitoring image pixel enhancement key frames of the sanitary napkins to be glued; The sanitary napkin appearance monitoring feature encoding subunit is used to perform feature encoding on the plurality of sanitary napkin appearance monitoring image pixel enhancement key frames to obtain the sanitary napkin appearance monitoring tracking feature vector to be glued.

4. The intelligent management system for a sanitary napkin production line according to claim 3, characterized in that: The sanitary napkin appearance monitoring feature coding subunit comprises: The plurality of pixel enhanced key frames of the appearance monitoring images of the sanitary napkins to be glued are passed through a feature encoder for appearance monitoring of the sanitary napkins to be glued based on a dilated convolutional neural network to obtain a tracking feature map for appearance monitoring of the sanitary napkins to be glued; The appearance monitoring and tracking feature map of the sanitary napkin to be glued is downsampled to obtain the appearance monitoring and tracking feature vector of the sanitary napkin to be glued.

5. The intelligent management system for a sanitary napkin production line according to claim 4, characterized in that: The sanitary napkin production order feature extraction unit comprises: Passing the sanitary napkin production order information collected from the database through a sanitary napkin production order information context encoder including an embedding layer to obtain a plurality of sanitary napkin production order information feature vectors; Arranging the plurality of sanitary napkin production order information feature vectors in two dimensions to obtain a sanitary napkin production order associated feature matrix; The sanitary napkin production order related feature matrix is ​​passed through a sanitary napkin production order convolutional neural network as a feature encoder to obtain the sanitary napkin production order information related feature vector.

6. The intelligent management system for a sanitary napkin production line according to claim 5, characterized in that: The sanitary napkin multimodal feature construction unit comprises: The appearance monitoring and tracking feature vector of the sanitary napkin to be glued is merged with the associated feature vector of the sanitary napkin production order information to obtain a multi-modal associated feature vector of the sanitary napkin gluing; Based on the gluing head coordinate association feature vector of the sanitary napkin gluing equipment, the sanitary napkin gluing multimodal association feature vector is corrected for fine-grained distribution differences based on deep-level shift information to obtain an optimized sanitary napkin gluing multimodal association feature vector.

7. The intelligent management system for a sanitary napkin production line according to claim 6, characterized in that: Based on the glue head coordinate correlation feature vector of the sanitary napkin gluing device, the sanitary napkin gluing multimodal correlation feature vector is corrected for fine-grained distribution differences based on deep-level shift information to obtain an optimized sanitary napkin gluing multimodal correlation feature vector, including: Calculating the position difference between the multi-modal correlation feature vector of the gluing of the sanitary napkin and the gluing head coordinate correlation feature vector of the sanitary napkin gluing equipment to obtain the position difference representation vector of the sanitary napkin production displacement; After concatenating the sanitary napkin gluing multimodal correlation feature vector and the sanitary napkin gluing equipment glue head coordinate correlation feature vector into a sanitary napkin production joint feature vector, the sanitary napkin production joint feature vector is input into a one-dimensional convolutional layer to obtain a sanitary napkin production shift convolutional coding representation vector; After performing linear embedding modulation encoding of the sanitary napkin gluing multimodal correlation feature vector into a linear modulation vector of the sanitary napkin gluing multimodal correlation feature vector using the first sanitary napkin production line embedding weight matrix, the position difference between the linear modulation vector of the sanitary napkin gluing multimodal correlation feature vector and the sanitary napkin gluing equipment glue head coordinate correlation feature vector is calculated to obtain a sanitary napkin production shift modulation difference representation vector; Taking the sanitary napkin production shift differential representation vector according to position as the value vector, the sanitary napkin production shift convolutional coding representation vector and the sanitary napkin production shift modulation differential representation vector as the query vector and key vector, the sanitary napkin production shift differential representation vector according to position, the sanitary napkin production shift convolutional coding representation vector and the sanitary napkin production shift modulation differential representation vector are input into a shift corrector based on an imitation transformer structure to obtain a sanitary napkin production multi-scale hierarchical fusion shift modulation weight vector; The optimized sanitary napkin gluing multimodal association feature vector is obtained by calculating the position point multiplication between the sanitary napkin production multi-scale hierarchical fusion shift modulation weight vector and the sanitary napkin gluing multimodal association feature vector.

8. The intelligent management system for a sanitary napkin production line according to claim 7, characterized in that: The glue head coordinate parameter feature extraction unit comprises: The coordinate parameters of the glue head of the sanitary napkin gluing device collected by the sensor are passed through a glue head coordinate semantic converter of the sanitary napkin gluing device to obtain a plurality of glue head coordinate feature vectors of the sanitary napkin gluing device; The plurality of sanitary napkin gluing equipment glue head coordinate feature vectors are spliced ​​into the sanitary napkin gluing equipment glue head coordinate associated feature vector.

9. The intelligent management system for a sanitary napkin production line according to claim 8, characterized in that: The rubber head motion parameter decoding module includes: A gluing equipment gluing feature association unit, used to associate the optimized sanitary napkin gluing multimodal association feature vector with the sanitary napkin gluing equipment glue head coordinate association feature vector to obtain a glue head motion parameter decoding feature vector; The glue head motion parameter decoding unit of the glue application equipment is used for passing the glue head motion parameter decoding feature vector through a decoder to obtain a decoding value, and the decoding value is used for representing the glue head motion parameter at the next moment.

10. An intelligent management method for a sanitary napkin production line, characterized in that: include: Obtain the gluing production video of the sanitary napkin gluing equipment collected by the camera, the sanitary napkin production order information collected by the database, and the gluing head coordinate parameters of the sanitary napkin gluing equipment collected by the sensor; Extracting optimized sanitary napkin gluing multimodal association feature vectors and sanitary napkin gluing equipment glue head coordinate association feature vectors from the sanitary napkin gluing equipment gluing production video collected by the camera, the sanitary napkin production order information collected by the database, and the sanitary napkin gluing equipment glue head coordinate parameters collected by the sensor; Based on the optimized multi-modal correlation feature vector of sanitary napkin gluing and the glue head coordinate correlation feature vector of the sanitary napkin gluing equipment, the glue head motion parameters at the next moment are obtained.

Images