A cuff structure glove with RFID traceability function and manufacturing method thereof

Abstract

The application discloses a cuff structure glove with an RFID tracing function and a manufacturing method thereof, relates to the technical field of cuff structure gloves, and aims to solve the technical problem that current RFID tags are prone to damage and falling off on gloves, and high rejection rate is caused by detection lag, comprising a glove body made of rubber material, and a cuff structure is arranged at an opening end of a wrist part of the glove body; and an RFID tag is completely covered and packaged in the inside of the cuff part. The RFID tag is completely covered and packaged in the interlayer cavity of the cuff part through design, physical isolation effect of the tag and external friction and bending area is realized, the annular thickened part of the cuff structure is integrally folded and bonded to the glove body, the cooperative and compatible effect of the tag packaging and the sealing performance of the glove is realized, a multimodal sensing fusion detection system is arranged before vulcanization, a synchronous online mechanism of machine vision detection and radio frequency performance detection is introduced, and products with hidden dangers are prevented from flowing into subsequent links.

Description

Technical Field

[0001] This invention relates to the field of rolled-edge glove technology, and more specifically, to a rolled-edge glove with RFID traceability function and its manufacturing method. Background Technology

[0002] In high-end manufacturing, biomedicine, fine chemicals, and semiconductor industries, where cleanliness and personnel safety are paramount, disposable protective gloves are critical consumables for ensuring process safety and personnel health. With the increasing intelligence and sophistication of industrial quality management systems, there is an urgent need for precise, individualized traceability throughout the entire lifecycle of protective gloves. Integrating radio frequency identification (RFID) technology into protective gloves enables digital management of production, warehousing, requisition, use, and disposal.

[0003] Currently, there are technical solutions in the industry that attempt to attach or encapsulate RFID tags on the wrist, palm, and other parts of gloves. However, most solutions attach the tags to the surface of the glove or in a simple interlayer. With frequent wearing, bending, and friction with work surfaces and equipment, the tags are very easy to fall off, wear out, or be physically damaged, causing the traceability function to fail.

[0004] Furthermore, existing testing procedures typically perform label validity checks after the gloves have undergone vulcanization. However, the vulcanization process causes irreversible cross-linking and curing of the rubber substrate. If label malfunction or physical packaging defects are discovered at this stage, the glove blank cannot be repaired and must be scrapped entirely, resulting in a significant waste of raw materials and production energy, and a substantial increase in production costs. Therefore, we propose a rolled-edge glove with RFID traceability and its manufacturing method. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art, adapt to practical needs, and provide a rolled-edge glove with RFID traceability function and its manufacturing method, so as to solve the technical problems of RFID tags being easily damaged and falling off on gloves and the high scrap rate caused by detection lag.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a rolled-edge glove with RFID traceability function, comprising a glove body made of rubber material, wherein the wrist opening of the glove body is provided with a rolled edge; the inside of the rolled edge is completely covered and encapsulated with an RFID tag.

[0007] Preferably, the rolled edge is formed by folding the edge of the wrist opening end of the glove body outward and bonding it to form an annular thickened part, and the RFID tag is placed inside the cavity formed after the folding.

[0008] Preferably, the RFID tag's chip storage area contains digital quality profile information uniquely bound to it. The profile information includes at least the physical packaging quality data of the rolled edge obtained through online testing of the glove body before vulcanization, the electrical performance data of the RFID tag, and the production batch information.

[0009] A method for manufacturing a rolled-edge glove with RFID traceability includes the following steps:

[0010] S100, the mold is impregnated with adhesive to form a glove substrate. After multiple impregnation and drying cycles, a glove substrate with uniform thickness and stable molding is obtained.

[0011] S200: The glove substrate is rolled at the end, and an RFID tag is inserted into the cavity of the rolled edge, so that the RFID tag is initially attached to the rolled edge structure to form a glove blank with an RFID tag.

[0012] S300: The glove blank is transferred to the inspection station. The physical packaging quality inspection of the rolled edge structure and the electrical performance inspection of the RFID tag are carried out simultaneously through the multimodal sensor fusion inspection system. The two types of inspection results are combined with the glove production information to generate a digital quality file. The file is bound to the unique identification code of the RFID tag and written into the tag storage area. At the same time, it is stored in the database. If the inspection fails, the abnormal rework process is triggered.

[0013] S400. The glove blank that has passed inspection and completed data binding is vulcanized and molded to integrate the RFID tag with the rubber substrate with rolled edges. After vulcanization, it is demolded to obtain the finished glove with rolled edge structure that has RFID traceability function.

[0014] Preferably, the multimodal sensing fusion detection system includes a collaborative sensing information acquisition module, a multi-source heterogeneous data processing module, and an intelligent judgment and execution module;

[0015] The sensor information acquisition module is used to acquire, non-contactly, the morphological raw data of the target area of ​​the glove blank and the electromagnetic response raw data of the RFID tag when the detection station is triggered.

[0016] The multi-source heterogeneous data processing module is used to perform machine vision-based feature extraction and pattern recognition analysis on the morphological raw data to generate a first quality vector about the integrity of the packaging, and simultaneously demodulate and parse the electromagnetic response raw data to generate a second quality vector about the integrity of the label function.

[0017] The intelligent judgment and execution module is connected to the multi-source heterogeneous data processing module, the production information database, and the production line control network, respectively. It is used to perform correlation analysis and comprehensive evaluation on the first quality vector, the second quality vector, and the real-time acquired production environment vector based on a preset fusion decision algorithm, and generate a unique digital quality file and a corresponding production line control instruction set. The control instruction set includes at least a release instruction for the transfer of qualified products or a sorting trigger instruction for the disposal of defective products.

[0018] Preferably, the sensing information acquisition module includes an optical sensing unit and a radio frequency sensing unit;

[0019] The optical sensing unit includes a high-resolution industrial camera and a structured light source. The structured light source is configured to illuminate the rolled edge area of ​​the glove blank at a specific angle and light intensity. The industrial camera is configured to acquire images of the rolled edge area under the illumination of the structured light source to enhance surface texture and three-dimensional morphological features.

[0020] The radio frequency sensing unit is adapted to the RFID tag communication protocol and is used to collect the raw electromagnetic response data of the RFID tag.

[0021] Preferably, the multi-source heterogeneous data processing module includes an image processing subunit and a radio frequency signal processing subunit;

[0022] The image processing subunit is configured to: preprocess and extract features from the morphological raw data, and match and compare it with a pre-stored defect feature template library to output a first quality vector containing defect type, defect location and confidence score;

[0023] The preprocessing includes applying a Gaussian filtering algorithm to the original image. Denoising is performed to obtain a valid image. Its formula is:

[0024] ;

[0025] in, The original morphological image data of the curled edge region. The effective image data after Gaussian filtering and denoising. For the filter coefficients, the Defines the number of pixels involved in the filtering calculation as the neighborhood range of the filtering window. , The pixel within the filter window relative to the core pixel coordinate offset, It is a Gaussian function;

[0026] The feature extraction performed by the image processing subunit includes extracting features from valid images. Extracting defect feature parameters of the curled edge Then, normalization is performed to obtain normalized feature quantization values. Its formula is:

[0027] ;

[0028] in, For the first Normalized quantized values ​​of each defect feature For the first The original measurement values ​​of each defect feature, For the first The minimum value of a defect feature in the defect feature template library For the first The maximum value of each defect feature in the defect feature template library;

[0029] The defect template matching performed by the image processing subunit includes matching the normalized feature vector. , and pre-stored standard defect feature vectors Perform cosine similarity matching to calculate the confidence score for defect determination. The formula is as follows:

[0030] ;

[0031] in, This is a confidence score for feature matching; the closer the value is to 1, the higher the defect matching degree. The extracted defect feature vector, For standard defect feature vectors, For feature vectors The first in One element, Standard feature vector The first in One element, , They are the feature vectors , The modulus length;

[0032] The first mass vector regarding the integrity of the rolled edge packaging is generated using the following formula:

[0033] ;

[0034] in, This is the first quality vector, used to quantify and describe the physical packaging quality of the rolled edge. Defect types are coded, with unique numerical codes assigned to defects such as curled edges, wrinkles, bubbles, and exposed / offset labels. These are the pixel coordinates of the defect within the curled edge region, used for precise defect location. The confidence score for defect matching characterizes the reliability of defect determination.

[0035] Preferably, the radio frequency signal processing subunit is configured to: decode the raw electromagnetic response data, obtain the tag ID, analyze the signal strength index, bit error rate and response time parameter, and output the second quality vector containing the tag ID and various radio frequency performance parameters;

[0036] Among them, the radio frequency signal processing subunit has electrical performance parameters Normalization is performed to obtain the performance quantification value. Its formula is:

[0037] ;

[0038] in, For the first Normalized quantized values ​​of electrical performance parameters For the first The original measured values ​​of each electrical performance parameter, For the first The minimum acceptable threshold value for each electrical performance parameter. For the first The maximum acceptable threshold value for each electrical performance parameter;

[0039] The second mass vector is generated using the following formula:

[0040] ;

[0041] in, This is the second mass vector. A unique identifier for an RFID tag. This is the normalized quantized value of the signal strength. This is the normalized quantized value of the bit error rate. This is the normalized quantized value for the response time.

[0042] Preferably, the intelligent judgment and execution module includes a data fusion engine and a decision controller;

[0043] The data fusion engine is used to receive the first quality vector and the second quality vector, and to retrieve the production environment vector corresponding to the current glove blank from the production information database in real time. The production environment vector includes mold number, rubber batch, number of impregnations, edge rolling process parameters and operator identification.

[0044] The data fusion engine further aligns and correlates the first quality vector, the second quality vector, and the production environment vector in time and space, and performs fusion analysis through a predefined rule engine or machine learning model to generate the digital quality profile.

[0045] The decision controller determines whether the current glove blank is qualified based on the fusion analysis results output by the data fusion engine, and generates the corresponding production line control instruction set.

[0046] Preferably, the data structure of the digital quality archive includes at least: an associated primary key, a visual inspection result set, a radio frequency inspection result set, a production process parameter set, a timestamp, and a comprehensive quality judgment mark; wherein, the associated primary key is the unique identifier of the RFID tag.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] 1. This invention designs a system that completely encapsulates the RFID tag within the cavity of the rolled edge, achieving physical isolation between the tag and external friction and bending areas. The annular thickening design of the rolled edge provides a natural protective barrier for the tag. Furthermore, the rolled edge structure serves as an integrally folded and glued annular thickening part of the glove body, achieving a synergistic compatibility between tag encapsulation and glove sealing performance. A multimodal sensor fusion detection system is set up before vulcanization, introducing a synchronous online mechanism for machine vision inspection and radio frequency performance inspection. This prevents potentially defective products from entering subsequent stages, solving the current problems of RFID tags being easily damaged and falling off on gloves, and the high scrap rate caused by delayed detection.

[0049] 2. This invention also utilizes a Gaussian filtering algorithm to preprocess the acquired rolled edge images, effectively eliminating noise interference. Through feature extraction and normalization, it quantifies physical defect characteristics. Furthermore, a cosine similarity matching algorithm is employed to calculate defect confidence scores, accurately identifying physical packaging defects such as rolled edges, wrinkles, bubbles, and exposed labels. Simultaneously, the system analyzes the electromagnetic response signals of RFID tags, evaluating their signal strength, bit error rate, and other electrical performance through parameter normalization. This allows for the timely detection and rejection of defective products before vulcanization and curing, fundamentally preventing potentially hazardous products from entering subsequent stages and significantly improving the long-term reliability and traceability of the finished gloves.

[0050] 3. This invention also utilizes a data fusion engine within the intelligent judgment and execution module to perform weighted fusion analysis on a first quality vector representing physical quality, a second quality vector representing electrical performance, and a production environment vector containing mold, batch, and process parameters. This process, based on a preset rule engine or machine learning model, generates a comprehensive judgment result reflecting the quality status of each individual glove. Finally, the system automatically generates a structured digital quality archive using the unique identifier of the RFID tag as the primary key. This archive not only contains the final judgment result but also integrates the original result sets of visual and RFID inspections, production process parameter sets, and timestamps, achieving real-time, accurate, and deep binding between production data and physical entities. This establishes a unique and complete digital twin for each glove, providing a solid data foundation for full lifecycle traceability. Attached Figure Description

[0051] Figure 1 This is a schematic cross-sectional view of the rolled-edge glove structure of the present invention;

[0052] Figure 2 This is a flowchart of the overall manufacturing method of the present invention;

[0053] Figure 3 This is a detailed flowchart of the multimodal sensing fusion detection and data binding of the present invention.

[0054] The labels in the diagram are as follows: 1. Glove body; 2. Wrist opening; 3. Rolled edge structure; 4. Annular thickened part; 5. Interlayer cavity; 6. RFID tag. Detailed Implementation

[0055] Example 1

[0056] like Figure 1 As shown, this embodiment provides a rolled-edge glove with RFID traceability function, including a glove body 1 made of rubber material, and a rolled-edge portion 3 provided at the wrist opening end 2 of the glove body 1.

[0057] The inside of the rolled edge 3 is completely encapsulated with an RFID tag 6.

[0058] Furthermore, the rolled edge 3 is formed by folding the edge of the wrist opening 2 of the glove body 1 outward and bonding it to form an annular thickened part 4, and the RFID tag 6 is placed inside the sandwich cavity 5 formed after the folding.

[0059] Furthermore, the chip storage area of ​​the RFID tag 6 contains digital quality file information uniquely bound to it. The file information includes at least the physical packaging quality data of the rolled edge 3 obtained by online testing of the glove body 1 before vulcanization molding, the electrical performance data of the RFID tag 6, and the production batch information.

[0060] Example 2

[0061] like Figures 1 to 2 As shown, this embodiment provides a method for manufacturing a rolled-edge glove with RFID traceability function, including the following steps:

[0062] S100. Immerse the glove mold in a corrosion-resistant rubber adhesive solution for impregnation. After removal, send it to a drying device for drying. Repeat the above impregnation and drying cycle multiple times until the glove substrate reaches the preset thickness and density requirements, resulting in a glove substrate with uniform thickness and stable molding. This step, through gradient adhesive adhesion and drying, provides a structurally stable substrate foundation for subsequent edge-rolling and labeling, avoiding label insertion misalignment due to uneven substrate thickness.

[0063] S200: At the end of the glove substrate, perform edge curling. Simultaneously, precisely insert the RFID tag 6 into the interlayer cavity 5 of the curled edge 3, allowing the RFID tag 6 to initially adhere to the curled edge structure, forming a glove blank with the RFID tag 6. This step performs edge curling and RFID insertion simultaneously, utilizing the flexibility of the rubber substrate during edge curling to achieve a tight fit between the tag and the curled edge 3, avoiding damage to the substrate from secondary operations.

[0064] S300: The glove blank is automatically transferred to the inspection station via the production line's linked conveyor mechanism, triggering the multimodal sensor fusion inspection system to start inspection. The system simultaneously performs physical packaging quality inspection of the rolled edge structure and electrical performance inspection of the RFID tag 6. It integrates the results of the two types of inspections with glove production information to generate a digital quality file, binds the file to the unique identification code of the RFID tag 6 and writes it into the tag storage area, and simultaneously stores it in the database.

[0065] If the inspection fails, the system automatically triggers the abnormal rework process: it issues a graded alarm signal, transfers the unqualified glove blank to a dedicated rework area, and records information such as the type of non-conformity, inspection station, and inspection time; the operator performs targeted treatments such as re-implanting labels and repairing rolled edges according to the abnormal information, and after rework, it is transferred back to the inspection station for a second inspection. Only after passing the inspection can it proceed to the next process.

[0066] This step, through the design of "pre-vulcanization inspection", avoids the overall scrapping of defective products due to vulcanization and curing, and greatly reduces the waste of production raw materials and energy consumption;

[0067] S400. The qualified and data-bound glove blank is sent into the vulcanizing furnace and vulcanized using a high-temperature and high-pressure integrated hot vulcanization process. During the vulcanization process, the rubber substrate of the rolled edge 3 undergoes a cross-linking and curing reaction, enabling the RFID tag 6 to achieve a tight molecular-level bond and gapless integrated sealing with the rubber substrate of the rolled edge 3.

[0068] After vulcanization, the finished gloves are obtained by demolding. After demolding, the finished product traceability verification is required: the RFID tag 6 of the finished gloves is read by the dedicated RFID authorized reader and writer device, the corresponding digital quality file is retrieved from the cloud database, and the tag reading status, file information integrity and consistency between local and cloud data are verified. Only after the verification is qualified can the gloves be shipped out of the factory.

[0069] Example 3

[0070] This embodiment provides a multimodal sensor fusion detection system for manufacturing a rolled-edge glove with RFID traceability function. The multimodal sensor fusion detection system includes a sensor information acquisition module that works collaboratively, a multi-source heterogeneous data processing module, and an intelligent judgment and execution module.

[0071] The sensor information acquisition module is used to acquire the morphological raw data of the target area of ​​the glove blank and the electromagnetic response raw data of the RFID tag 6 in a non-contact manner when the detection station is triggered, so as to avoid physical damage to the uncured rubber substrate; the sensor information acquisition module includes an optical sensing unit and a radio frequency sensing unit.

[0072] The optical sensing unit includes a high-resolution industrial camera and a structured light source. The structured light source is configured to illuminate the rolled edge region of the glove blank at a specific angle and intensity, enhancing the surface texture and three-dimensional morphological features of the rolled edge. The industrial camera is configured to acquire images of the rolled edge region under the illumination of the structured light source, obtaining raw morphological data. (Two-dimensional image pixel matrix) provides basic data for subsequent physical packaging quality inspection;

[0073] The radio frequency sensing unit is a radio frequency communication component adapted to the communication protocol of the RFID tag 6. It can establish a contactless radio frequency connection with the RFID tag 6 within the rolled-edge structure and collect the raw electromagnetic response data of the RFID tag 6. (Time-domain radio frequency signal sequence), used for subsequent electrical performance testing.

[0074] The multi-source heterogeneous data processing module is used to perform machine vision-based feature extraction and pattern recognition analysis on raw morphological data to generate a first quality vector regarding encapsulation integrity. Simultaneously, the raw electromagnetic response data is demodulated and parameter parsed to generate a second mass vector relating to the tag's functional integrity. ;

[0075] The multi-source heterogeneous data processing module includes an image processing subunit and a radio frequency signal processing subunit;

[0076] The image processing subunit is configured to: preprocess the raw morphological data, extract features, and match and compare it with a pre-stored defect feature template library, and output a first quality vector containing defect type, defect location and confidence score;

[0077] Specifically:

[0078] Image preprocessing (raw data normalization): Gaussian filtering algorithm is used to normalize the raw image. Denoising processing is performed to eliminate imaging noise and background interference, resulting in an effective image of the curled edge region. The formula is as follows:

[0079] ;

[0080] in, This is the morphological raw image data (two-dimensional pixel matrix) of the curled edge region. The effective image data after Gaussian filtering and denoising. These are the filter coefficients, used to adjust the noise reduction intensity and adapt to the imaging characteristics of different rubber substrates. Defines the number of pixels involved in the filtering calculation as the neighborhood range of the filtering window. , The pixel within the filter window relative to the core pixel coordinate offset, This is a Gaussian function used to calculate the weight of each pixel within the filtering window;

[0081] Feature extraction (defect parameter quantization): from valid images Defect feature parameters (such as defect area, contour roughness, label offset distance, etc.) of the rolled edge 3 are extracted, and all feature parameters are normalized to obtain dimensionless feature quantization values. The formula is as follows:

[0082] ;

[0083] in, For the first The normalized quantized value of each defect feature, with a value range of [value range missing]. , For the first The original measurement values ​​of each defect feature, For the first The minimum value of a defect feature in the defect feature template library For the first The maximum value of each defect feature in the defect feature template library;

[0084] Defect template matching (confidence score): This involves matching the normalized feature vectors... Compared with standard defect feature vectors in a pre-stored defect feature template library Perform cosine similarity matching to calculate the confidence score for defect determination. The formula is as follows:

[0085] ;

[0086] in, The confidence score for feature matching, with a value range of [value range missing]. The closer the value is to 1, the higher the defect matching degree. The extracted defect feature vector contains One normalized feature parameter, The standard defect feature vector is retrieved from the defect feature template library. For feature vectors The first in One element, Standard feature vector The first in One element, , They are the feature vectors , The modulus length;

[0087] First quality vector construction: Encoding defect types Defect location coordinates Confidence score Perform structured integration to generate a first quality vector regarding the integrity of the rolled edge packaging. The formula is as follows:

[0088] ;

[0089] in, This is the first quality vector, used to quantify and describe the physical packaging quality of the rolled edge 3. Defect types are coded, with unique numerical codes assigned to defects such as curled edges, wrinkles, bubbles, and exposed / offset labels. These are the pixel coordinates of the defect within the curled edge region, used for precise defect location. The confidence score for defect matching characterizes the reliability of defect determination.

[0090] The radio frequency signal processing subunit is configured to: decode the raw electromagnetic response data, obtain the tag ID, analyze the signal strength index, bit error rate, and response time parameters, and output a second quality vector containing the tag ID and various radio frequency performance parameters. ;

[0091] Specifically:

[0092] RF signal preprocessing and parameter analysis: Preprocessing the raw time-domain RF signal Demodulation, filtering, and analog-to-digital conversion were performed to extract the core electrical performance parameters of RFID tag 6: signal strength. Bit error rate Response time All parameters are then normalized to obtain dimensionless performance quantification values. The formula is as follows:

[0093] ;

[0094] in, For the first The normalized quantized values ​​of the electrical performance parameters are within the range of [value range missing]. , For the first The original measured values ​​of each electrical performance parameter (such as the original signal strength value and the original bit error rate value). For the first The minimum acceptable threshold value for each electrical performance parameter. For the first The maximum acceptable threshold value for each electrical performance parameter;

[0095] Second quality vector construction: RFID tag 6 unique identifier codes Normalized RF performance parameter set Perform structured integration to generate a second mass vector. The formula is as follows:

[0096] ;

[0097] in, This is the second quality vector, used to quantify and describe the electrical performance of RFID tag 6. This serves as the unique identifier for the RFID tag, acting as the core key for subsequent data binding and traceability. This is the normalized quantized value of the signal strength. This is the normalized quantized value of the bit error rate. This is the normalized quantized value for the response time.

[0098] The intelligent judgment and execution module is connected to the multi-source heterogeneous data processing module, the production information database, and the production line control network, respectively. It is used to determine the first quality vector based on a preset fusion decision algorithm. Second mass vector The system performs correlation analysis and comprehensive evaluation on real-time acquired production environment vectors to generate a unique digital quality file and a corresponding production line control instruction set. The control instruction set includes at least release instructions for qualified product transfer or sorting trigger instructions for defective product disposal.

[0099] The intelligent judgment and execution module includes a data fusion engine and a decision controller;

[0100] The data fusion engine is used to receive the first quality vector. Second mass vector It also retrieves the production environment vector corresponding to the current glove blank from the production information database in real time. Production environment vector This includes mold number, rubber batch, number of impregnations, edge curling process parameters, and operator identification;

[0101] The data fusion engine further integrates the first quality vector. Second mass vector Production Environment Vector Spatiotemporal alignment and correlation are performed, and fusion analysis is conducted through a predefined rule engine or machine learning model to generate a digital quality archive. The data structure of the digital quality archive includes at least: a correlation primary key, a visual inspection result set, an RFID inspection result set, a production process parameter set, a timestamp, and a comprehensive quality judgment mark; wherein, the correlation primary key is the unique identifier of the RFID tag 6.

[0102] The decision controller determines whether the current glove blank is qualified based on the fusion analysis results output by the data fusion engine, and generates the corresponding production line control instruction set.

[0103] Specifically:

[0104] Production Environment Vector Retrieval and Construction: Calculating the Production Environment Vector The formula is:

[0105] ;

[0106] in, This is a production environment vector used to quantitatively describe the entire production process information of the glove preform. A structured code for mold numbering. Structured coding for rubber batches This is a normalized quantized value for the number of immersions. Structured coding for edge curling process parameters Structured coding for operator identification;

[0107] Multi-source vector weighted fusion: Considering the different weights of physical packaging quality, tag electrical performance, and production environment on the overall glove quality, a weighted fusion algorithm is used to... , ,and The mixture is then fused to generate a comprehensive quality vector. The formula is as follows:

[0108] ;

[0109] in, The overall quality vector characterizes the overall quality level of the glove preform. The first mass vector The fusion weight, The second mass vector The fusion weight, Production environment vector The fusion weights and weight constraints are as follows: It can be dynamically adjusted according to actual quality control requirements (e.g., increasing the sealing performance when stringent requirements are met). );

[0110] Digital quality archive generation:

[0111] Using RFID tags with unique identification codes For related primary keys , to integrate the quality vector Original results set of visual inspection RF detection raw result set Production process parameter set Detection timestamp and comprehensive quality assessment criteria Perform structured associations to generate digital quality archives. The formula is as follows:

[0112] ;

[0113] in, This is a digital quality archive containing comprehensive traceability data of the glove blank. As the primary key, it is associated with the RFID tag's unique identifier. Maintain consistency This is the raw result set for visual inspection, containing preprocessed image data, defect feature vectors, etc. This is the raw result set of radio frequency testing, including raw electromagnetic response data and raw values ​​of electrical performance parameters. This is a set of production process parameters, containing complete parameters for processes such as impregnation, drying, and edge rolling. This serves as a timestamp to record the specific time of the detection. It serves as a comprehensive quality assessment criterion, used to characterize whether the glove blank is qualified.

[0114] Dual data binding and storage: The core index information of the digital quality archive is encrypted and associated with the unique identification code of the RFID tag 6, and written to the preset encrypted storage area of ​​the RFID tag 6 through the radio frequency reading and writing module to realize the local physical binding of the test data and the tag; at the same time, the complete digital quality archive and the tag's unique identification code association data are encrypted and uploaded to the local database and the cloud database to realize the cloud logical binding of data, and the local and cloud data maintain real-time synchronization and consistency verification.

[0115] Based on the comprehensive quality vector The fusion results are used to intelligently determine the pass / fail status of the glove blank through a threshold determination algorithm, as shown in the following formula:

[0116] ;

[0117] in, As a comprehensive quality assessment criterion, This indicates that the test is successful. This indicates that the product is unqualified.

[0118] If it is determined to be unqualified ( The system uses an anomaly handling module to locate the core non-conforming factors and generate targeted rework plan codes. The formula is as follows;

[0119] ;

[0120] in, The rework scheme is coded to correspond to different types of non-conformities (such as packaging defects, electrical performance defects, and process parameter deviations). These are the weighted magnitudes of each vector, and the vector corresponding to the minimum value is the source of the core non-compliance factor.

[0121] The production line control instruction set (release instruction or sorting trigger instruction) generated by the decision controller is directly connected to the production line control network, realizing the automatic transfer of qualified products to the vulcanization process and the automatic diversion of unqualified products to the rework area without manual intervention.

[0122] The embodiments disclosed in this invention are preferred embodiments, but are not limited thereto. Those skilled in the art can easily understand the spirit of this invention based on the above embodiments and make different extensions and variations, but as long as they do not depart from the spirit of this invention, they are all within the protection scope of this invention.

Claims

1. A rolled-edge glove with RFID traceability function, comprising a glove body (1) made of rubber material, characterized in that, The wrist opening end (2) of the glove body (1) is provided with a rolled edge (3). The inside of the rolled edge (3) is completely encapsulated with an RFID tag (6).

2. The rolled-edge glove with RFID traceability function according to claim 1, characterized in that, The rolled edge (3) is formed by folding the edge of the wrist opening end (2) of the glove body (1) outward and bonding it to form an annular thickened part (4), and the RFID tag (6) is placed inside the sandwich cavity (5) formed after the folding.

3. A rolled-edge glove with RFID traceability function according to claim 1, characterized in that, The RFID tag (6) contains digital quality archive information uniquely bound to it in its chip storage area. The archive information includes at least the physical packaging quality data of the rolled edge (3) of the glove body (1) obtained by online detection before vulcanization, the electrical performance data of the RFID tag (6), and the production batch information.

4. A method for manufacturing a rolled-edge glove with RFID traceability function, applicable to the rolled-edge glove with RFID traceability function as described in any one of claims 1-3, characterized in that, Includes the following steps: S100, the mold is impregnated with adhesive to form a glove substrate. After multiple impregnation and drying cycles, a glove substrate with uniform thickness and stable molding is obtained. S200. The glove substrate is rolled at the end, and the RFID tag (6) is inserted into the cavity (5) of the rolled edge (3) so that the RFID tag (6) is initially attached to the rolled edge (3) to form a glove blank with the RFID tag (6). S300. The glove blank is transferred to the inspection station. The physical packaging quality inspection of the rolled edge (3) and the electrical performance inspection of the RFID tag (6) are carried out simultaneously through the multimodal sensor fusion inspection system. The two types of inspection results are combined with the production information of the glove body (1) before vulcanization to generate a digital quality file. The file is bound to the unique identification code of the RFID tag (6) and written into the tag storage area. At the same time, it is stored in the database. If the inspection is unqualified, the abnormal rework process is triggered. S400. The glove blank that has passed the inspection and completed the data binding is vulcanized and molded so that the RFID tag (6) and the rubber substrate with rolled edge (3) are integrated and sealed. After vulcanization, the glove with rolled edge structure with RFID traceability function is obtained by demolding.

5. A method for manufacturing a rolled-edge glove with RFID traceability function according to claim 4, characterized in that, The multimodal sensing fusion detection system includes a collaborative sensing information acquisition module, a multi-source heterogeneous data processing module, and an intelligent judgment and execution module; The sensor information acquisition module is used to acquire, without contact, the morphological raw data of the target area of ​​the glove blank and the electromagnetic response raw data of the RFID tag (6) when the detection station is triggered. The multi-source heterogeneous data processing module is used to perform machine vision-based feature extraction and pattern recognition analysis on the morphological raw data to generate a first quality vector about the integrity of the packaging, and simultaneously demodulate and parse the electromagnetic response raw data to generate a second quality vector about the integrity of the label function. The intelligent judgment and execution module is connected to the multi-source heterogeneous data processing module, the production information database, and the production line control network, respectively. It is used to perform correlation analysis and comprehensive evaluation on the first quality vector, the second quality vector, and the real-time acquired production environment vector based on a preset fusion decision algorithm, and generate a unique digital quality file and a corresponding production line control instruction set. The control instruction set includes at least a release instruction for the transfer of qualified products or a sorting trigger instruction for the disposal of defective products.

6. A method for manufacturing a rolled-edge glove with RFID traceability function according to claim 5, characterized in that, The sensor information acquisition module includes an optical sensing unit and a radio frequency sensing unit; The optical sensing unit includes a high-resolution industrial camera and a structured light source. The structured light source is configured to illuminate the rolled edge area of ​​the glove blank at a specific angle and light intensity. The industrial camera is configured to acquire images of the rolled edge area under the illumination of the structured light source to enhance surface texture and three-dimensional morphological features. The radio frequency sensing unit is adapted to the communication protocol of the RFID tag (6) and is used to collect the raw electromagnetic response data of the RFID tag (6).

7. A method for manufacturing a rolled-edge glove with RFID traceability function according to claim 5, characterized in that, The multi-source heterogeneous data processing module includes an image processing subunit and a radio frequency signal processing subunit; The image processing subunit is configured to: preprocess and extract features from the morphological raw data, and match and compare it with a pre-stored defect feature template library to output a first quality vector containing defect type, defect location and confidence score; The preprocessing includes applying a Gaussian filtering algorithm to the original image. Denoising is performed to obtain a valid image. Its formula is: ； in, The original morphological image data of the curled edge region. The effective image data after Gaussian filtering and denoising. For the filter coefficients, the Defines the number of pixels involved in the filtering calculation as the neighborhood range of the filtering window. , The pixel within the filter window relative to the core pixel coordinate offset, It is a Gaussian function; The feature extraction performed by the image processing subunit includes extracting features from valid images. Extract the defect feature parameters of the curled edge (3). Then, normalization is performed to obtain normalized feature quantization values. Its formula is: ； in, For the first The normalized quantized value of each defect feature. For the first The original measurement values ​​of each defect feature, For the first The minimum value of a defect feature in the defect feature template library For the first The maximum value of each defect feature in the defect feature template library; The defect template matching performed by the image processing subunit includes matching the normalized feature vector. , and pre-stored standard defect feature vectors Perform cosine similarity matching to calculate the confidence score for defect determination. The formula is as follows: ； in, This is a confidence score for feature matching; the closer the value is to 1, the higher the defect matching degree. The extracted defect feature vector, For standard defect feature vectors, For feature vectors The first in One element, Standard feature vector The first in One element, , respectively, feature vectors , The modulus length; The first mass vector regarding the integrity of the rolled edge packaging is generated using the following formula: ； in, This is the first quality vector, used to quantify the physical packaging quality of the rolled edge (3). Defect type coding: A unique numerical code is assigned to defects such as curled edges, wrinkles, bubbles, and exposed / offset labels. These are the pixel coordinates of the defect within the curled edge region, used for precise defect location. The confidence score for defect matching characterizes the reliability of defect determination.

8. A method for manufacturing a rolled-edge glove with RFID traceability function according to claim 7, characterized in that, The radio frequency signal processing subunit is configured to: decode the raw electromagnetic response data, obtain the tag ID, analyze the signal strength index, bit error rate and response time parameters, and output the second quality vector containing the tag ID and various radio frequency performance parameters; Among them, the radio frequency signal processing subunit has electrical performance parameters Normalization is performed to obtain the performance quantification value. Its formula is: ； in, For the first Normalized quantized values ​​of electrical performance parameters For the first The original measured values ​​of each electrical performance parameter, For the first The minimum acceptable threshold value for each electrical performance parameter. For the first The maximum acceptable threshold value for each electrical performance parameter; The second mass vector is generated using the following formula: ； in, This is the second mass vector. A unique identifier for an RFID tag. This is the normalized quantized value of the signal strength. This is the normalized quantized value of the bit error rate. This is the normalized quantized value for the response time.

9. A method for manufacturing a rolled-edge glove with RFID traceability function according to claim 5, characterized in that, The intelligent judgment and execution module includes a data fusion engine and a decision controller; The data fusion engine is used to receive the first quality vector and the second quality vector, and to retrieve the production environment vector corresponding to the current glove blank from the production information database in real time. The production environment vector includes mold number, rubber batch, number of impregnations, edge rolling process parameters and operator identification. The data fusion engine further aligns and correlates the first quality vector, the second quality vector, and the production environment vector in time and space, and performs fusion analysis through a predefined rule engine or machine learning model to generate the digital quality profile. The decision controller determines whether the current glove blank is qualified based on the fusion analysis results output by the data fusion engine, and generates the corresponding production line control instruction set.

10. A method for manufacturing a rolled-edge glove with RFID traceability function according to claim 9, characterized in that, The data structure of the digital quality archive includes at least: an associated primary key, a visual inspection result set, a radio frequency inspection result set, a production process parameter set, a timestamp, and a comprehensive quality judgment mark; wherein, the associated primary key is the unique identifier of the RFID tag (6).

Images