Quality inspection management system for textile fabric production
By combining quantum dot sensors and artificial intelligence image recognition technology, all-round detection of textile fabrics is achieved, and the misjudgment and misjudgment problems of traditional quality inspection methods are solved, detection accuracy and production efficiency are improved, and product quality is ensured.
Patent Information
- Application Number
- CN202411466925.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-21
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The quality inspection methods for the production of traditional textile fabrics rely on manual inspection, and there are problems of misjudgment and misjudgment. The microstructure and macro appearance inspection of the fabrics is not comprehensive and accurate enough, so it is impossible to effectively evaluate the quality of different types of fabrics.
Combining quantum dot sensor technology and artificial intelligence image recognition technology, we can realize all-round detection of textile fabric quality, detect microstructure through quantum dot sensors and analyze macro appearance through artificial intelligence, establish a correlation model, and automatically adjust production equipment parameters in real-time feedback system to develop VR/AR quality inspection software provides an immersive quality inspection experience.
It improves the accuracy and reliability of quality inspection, ensures product quality meets standards, realizes real-time control and optimization of the production process, improves production efficiency and quality inspection efficiency, and reduces human error.
Smart Images

Figure CN118982541A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of textile fabric production, in particular to a quality inspection management system for textile fabric production. Background Art
[0002] In traditional textile fabric production quality inspection, there are often some limitations, which makes it difficult to effectively ensure the accuracy of quality inspection. For example, previous quality inspection methods may rely mainly on manual inspection, and the subjective factors of the inspectors will have a greater impact on the results, which is prone to misjudgment and omission. Moreover, manual inspection is inefficient and difficult to meet the needs of large-scale production.
[0003] In addition, traditional quality inspection technology may not be comprehensive and in-depth enough when testing the microstructure and macroscopic appearance of fabrics. There may be a lack of effective means and precise analysis for testing the microstructure of fabrics, such as fiber molecular arrangement and crystallinity, and it is impossible to accurately evaluate the quality of fabrics. At the same time, in terms of macroscopic appearance testing, the analysis of characteristics such as color, texture and pattern may not be accurate and detailed enough, making it difficult to detect some subtle flaws and defects.
[0004] Traditional quality inspection technologies may not have targeted testing methods for natural fiber fabrics and synthetic fiber fabrics, and cannot fully consider their respective characteristics and quality requirements, resulting in inaccurate quality assessments of different types of fabrics.
[0005] This has led to a demand for more advanced and accurate quality inspection management systems for textile fabric production to improve the accuracy and reliability of quality inspection. Summary of the invention
[0006] The purpose of the present invention is to make up for the shortcomings of the prior art and provide a quality inspection and management system for textile fabric production. It can significantly improve the accuracy of quality inspection, reduce misjudgments and missed judgments, and ensure that product quality meets standards by integrating quantum dot sensor technology and artificial intelligence image recognition technology, as well as setting specific quality inspection configurations for different fabrics. At the same time, the establishment of a real-time feedback system realizes real-time control and optimization of the production process, can adjust production equipment parameters in time, ensure product quality stability, and improve production efficiency. Through the development of VR / AR quality inspection software, an immersive quality inspection experience is provided, so that quality inspectors can view fabric defects more intuitively. The automatic recognition function of the software can also provide detection standards and processing methods, thereby improving quality inspection efficiency and accuracy.
[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a quality inspection and management system for textile fabric production, the system comprising an intelligent detection module, the intelligent detection module combines quantum dot sensor technology with artificial intelligence image recognition technology to achieve all-round detection of textile fabric quality; The quantum dot sensor is used to detect the microstructure of textile fabrics, including but not limited to the arrangement of fiber molecules and crystallinity. Assume that the quantum dot sensor detection signal intensity of the fabric sample is , the fiber molecular arrangement parameter is and the crystallinity parameter is , then the detection result function ,in, and The determination method is as follows: collect a large number of different types of standard fabric samples, and detect these samples with quantum dot sensors to obtain the corresponding signal strength , fiber molecular arrangement parameters and crystallinity parameters Then, using the regression analysis algorithm, the actual fabric quality evaluation results are taken as the target value. and Optimize and solve so that the function The error between the output result and the actual quality assessment result is minimized. The specific regression analysis algorithm can use the least square method to determine the optimal parameter value by minimizing the sum of squares of the errors. The artificial intelligence image recognition technology is used to analyze the macroscopic appearance of fabrics, including but not limited to color, texture and pattern. The input fabric image is subjected to feature extraction and classification through a deep learning image recognition algorithm, thereby identifying the macroscopic features of the fabric. In the feature extraction process, a convolutional neural network (CNN) can be used to automatically learn the features of the image, and then classification is performed through a fully connected layer; In the intelligent detection process, the microscopic detection results of the quantum dot sensor are associated with the macroscopic detection results of artificial intelligence image recognition to establish a correlation model ,in, is the artificial intelligence image recognition result parameter, and The method of determining and Similarly, by collecting a large number of standard fabric samples, taking the comprehensive quality assessment results as the target value, and using the regression analysis algorithm for optimization and solution, the output results of the association model can accurately reflect the actual quality of the fabric. In this way, comprehensive and accurate detection of the quality of textile fabrics can be achieved, and the detection results can be fed back to the production equipment in real time to enable real-time control and optimization of the production process.
[0008] Furthermore, the quantum dot sensor has specific configuration parameters for natural fiber fabrics and synthetic fiber fabrics; For natural fiber fabrics, the quantum dot sensor not only detects the impurity content and aging degree of the fiber, but also analyzes the relationship between the natural texture characteristics and quality of the fiber through a specific algorithm. Suppose the texture complexity of the natural fiber fabric is , the impurity content is The aging degree is , then the natural fiber fabric quality evaluation function ,in, , , The determination process is as follows: collect a large number of natural fiber fabric samples, test the texture complexity, impurity content and aging degree of each sample, and at the same time combine professional quality assessors to conduct quality assessment on the samples, use regression analysis algorithm, take the quality assessment results as the target value, and conduct quality assessment on the samples. , , An optimization solution is performed so that the quality evaluation function can accurately reflect the quality of natural fiber fabrics. In the detection of texture complexity, image analysis technology can be used to extract the characteristic parameters of the texture, including but not limited to the directionality and periodicity of the texture.
[0009] Furthermore, for synthetic fiber fabrics, the quantum dot sensor monitors the polymerization degree and molecular weight distribution of the fiber, and combines the artificial intelligence image recognition technology to analyze the glossiness of the synthetic fiber fabric. Assuming that the polymerization degree of the synthetic fiber fabric is , the molecular weight distribution parameter is , the glossiness parameter is , then the synthetic fiber fabric quality evaluation function ,in, , , The method of determining the quality of synthetic fiber fabrics is as follows: a large number of synthetic fiber fabric samples are collected, and the degree of polymerization, molecular weight distribution and glossiness of the samples are tested, and quality evaluation is performed at the same time. Through the regression analysis algorithm, the quality evaluation results are used as the target value. , , Optimization solutions are performed to ensure that the quality evaluation function can accurately reflect the quality of synthetic fiber fabrics. Spectral analysis technology and other related physical and chemical methods can be used in the detection of degree of polymerization and molecular weight distribution.
[0010] Furthermore, a dynamic calibration system for quantum dot sensors is established. The system monitors the performance changes of the sensors in real time. By comparing the differences between the test results of the standard samples and the actual test results, the calibration algorithm is used to adjust the parameters of the sensors. The test results of the standard samples are set to The actual test result is , the difference function ,When the difference exceeds a certain threshold, the calibration algorithm is started to adjust the parameters of the sensor to minimize the difference. The specific calibration algorithm is as follows: According to the working principle and performance characteristics of the sensor, a parameter adjustment model is established. By continuously trying different parameter combinations, the value of the difference function is gradually reduced until the preset accuracy requirements are met. In the parameter adjustment model, it is necessary to consider the sensitivity and resolution factors of the sensor, as well as the impact of environmental factors on sensor performance.
[0011] Furthermore, a special VR / AR quality inspection software is developed, which is seamlessly integrated with the intelligent inspection equipment. Through virtual reality and augmented reality technology, the inspection results are presented to the quality inspectors in an immersive way. The quality inspectors can interact with the virtual scene through gesture control and voice commands, and view the location and type of fabric defects more intuitively. At the same time, the software also has the function of automatically identifying uncertain defect types and providing inspection standards and processing methods. The automatic recognition algorithm in the software is trained based on artificial intelligence image recognition technology and defect samples in the database. It can quickly and accurately identify various defect types and provide reference for quality inspectors according to preset standards and processing methods. In the implementation of virtual reality and augmented reality technology, the hardware equipment of head-mounted display devices and gesture recognition sensors, as well as the corresponding software algorithms, can be used to create an immersive experience.
[0012] Furthermore, VR / AR technology is combined with remote collaboration functions to achieve cross-regional quality inspection collaboration. Quality inspectors in different locations can enter the same virtual inspection scene through a network connection to jointly inspect and evaluate fabrics. During the collaboration process, the system uses data transmission and sharing algorithms to ensure that each quality inspector can obtain inspection data and image information in real time and communicate and collaborate effectively. The data transmission and sharing algorithms use encryption technology and efficient data compression algorithms to ensure data security and transmission efficiency. At the same time, the collaboration platform also provides real-time voice and video communication functions to facilitate communication and discussion among quality inspectors. In the data transmission and sharing algorithms, network communication technology based on the TCP / IP protocol and data encryption algorithm AES can be used to ensure the secure transmission of data.
[0013] Furthermore, in the process of establishing the association model between quantum dot sensor data and image recognition data, a multimodal data fusion algorithm is used. This algorithm fuses the micro data of the quantum dot sensor and the macro data of artificial intelligence image recognition to extract more representative features and improve the accuracy and reliability of the association model. The specific algorithm steps include data preprocessing, feature extraction, feature fusion and model training. In the data preprocessing stage, the quantum dot sensor data and image recognition data are normalized to eliminate the dimensional difference of the data. In the feature extraction stage, feature extraction methods suitable for their respective data types are used to extract micro and macro features. In the feature fusion stage, weighted fusion, splicing fusion and other methods are used to fuse the two features. In the model training stage, the fused features are used for training to continuously optimize the parameters of the association model. In the feature extraction and fusion process, convolutional neural network and recurrent neural network technologies in deep learning are used to extract and fuse features.
[0014] Furthermore, a real-time feedback system is developed to feed back the detection results of quantum dot sensors and artificial intelligence image recognition to the production equipment in real time. The feedback system adopts a closed-loop control algorithm to automatically adjust the parameters of the production equipment according to the detection results. Suppose the parameter adjustment amount of the production equipment is , the test result parameters are , then adjust the algorithm function ,in, The method of determining is: according to the production process and equipment characteristics, a mathematical model between the production equipment parameters and the test results is established, and the function is determined by analyzing and optimizing a large amount of production data. The specific form of the test results enables the production equipment to be adjusted in time according to the test results to ensure the stability of product quality. When establishing the mathematical model, it is necessary to consider the working principle and process factors of the production equipment, as well as the impact of the test results on product quality.
[0015] Furthermore, in order to improve the accuracy and efficiency of artificial intelligence image recognition, data enhancement technology is used to expand the training data. Data enhancement methods include random rotation, flipping, cropping and brightness adjustment. By transforming the original image, more training samples are generated to improve the generalization ability of the model. At the same time, the optimized deep learning model architecture and training algorithm are used to accelerate the convergence speed of the model and improve the detection accuracy. The optimization of the deep learning model architecture includes adjusting the number of network layers, the number of neurons and the activation function to adapt to different types of textile fabric image recognition tasks. The optimization of the training algorithm uses adaptive learning rate adjustment and batch normalization technology to improve training efficiency and stability.
[0016] Compared with the existing technology, the quality inspection and management system for textile fabric production has the following beneficial effects: 1. The present invention combines quantum dot sensor technology with artificial intelligence image recognition technology to achieve all-round detection of textile fabric quality. The quantum dot sensor is used to detect the microstructure of the fabric, including the fiber molecular arrangement and crystallinity, and artificial intelligence image recognition technology is used to analyze the macroscopic appearance of the fabric, such as color, texture and pattern. In addition, specific quality inspection configurations are set for natural fiber fabrics and synthetic fiber fabrics, which can more accurately evaluate the quality of different types of fabrics. The comprehensive application of these technologies can improve the accuracy of quality inspection and ensure that product quality meets standards.
[0017] 2. The present invention realizes real-time control and optimization. The system establishes a real-time feedback system to feed back the test results to the production equipment in real time. The feedback system adopts a closed-loop control algorithm to automatically adjust the parameters of the production equipment according to the test results to ensure the stability of product quality. In this way, problems in the production process can be discovered in time, and real-time control and optimization can be carried out to improve production efficiency and product quality.
[0018] 3. The present invention provides an immersive quality inspection experience. The developed VR / AR quality inspection software is seamlessly integrated with the intelligent inspection equipment. Through virtual reality and augmented reality technology, the inspection results are presented to quality inspectors in an immersive manner. Quality inspectors can interact with the virtual scene through gesture control and voice commands, and view the defect location and type of the fabric more intuitively. At the same time, the software also has the function of automatically identifying uncertain defect types and providing inspection standards and processing methods, which can provide quality inspectors with a more accurate reference. This immersive quality inspection experience can improve quality inspection efficiency and accuracy and reduce human errors.
[0019] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0021] Figure 1 This is a flow chart of a quality inspection management system for textile fabric production. DETAILED DESCRIPTION
[0022] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0023] Embodiment 1: This embodiment aims to build a comprehensive quality inspection and management system for textile fabric production, which fully integrates intelligent detection modules, quantum dot sensors, artificial intelligence image recognition technology, association models and real-time feedback systems to achieve accurate detection and real-time control of textile fabric quality.
[0024] Quantum dot sensors provide microscopic information for fabric quality assessment by detecting the microstructure of textile fabrics, including parameters such as fiber molecular arrangement and crystallinity. Suppose the quantum dot sensor detection signal intensity of the fabric sample is , the fiber molecular arrangement parameter is , the crystallinity parameter is , then the detection result function , and It is a weight coefficient used to adjust the influence of fiber molecular arrangement and crystallinity parameters on the test results.
[0025] By collecting a large number of different types of standard fabric samples, these samples are tested by quantum dot sensors to obtain the corresponding signal strength: , the fiber molecular arrangement parameter is and the crystallinity parameter is At the same time, combined with the actual quality assessment results of samples by professional quality assessors, the regression analysis algorithm (using the least squares method) is used for optimization and solution. Specifically, the goal of the least squares method is to minimize the sum of squares of errors, that is, ,in Indicates the sample number, For the function in the sample The predicted value on For sample The actual quality assessment results are continuously adjusted and The value of Minimize to determine the optimal and , to ensure that the function The output results can accurately reflect the microscopic quality characteristics of the fabric.
[0026] The deep learning image recognition algorithm is used to extract and classify the input fabric images to analyze the macroscopic appearance of the fabric, including features such as color, texture, and pattern. In the feature extraction process, a convolutional neural network (CNN) is used to automatically learn the features of the image. CNN extracts low-level to high-level features of the image through multi-layer convolution and pooling operations. Then, classification is performed through the fully connected layer, and the extracted features are mapped to different categories to identify the macroscopic features of the fabric.
[0027] Correlate the microscopic detection results of quantum dot sensors with the macroscopic detection results of artificial intelligence image recognition to establish a correlation model , and is the weight coefficient used to balance the impact of microscopic and macroscopic detection results on the output of the correlation model. and The determination method is similar to that of collecting a large number of standard fabric samples and obtaining the microscopic detection results of quantum dot sensors (including , and And the macroscopic detection result parameters of artificial intelligence image recognition At the same time, taking the comprehensive quality evaluation result as the target value, the regression analysis algorithm is used to optimize the solution. and The value of makes the output result of the correlation model the target value, and the regression analysis algorithm is used to optimize the solution. and The value of makes the output result of the association model can accurately reflect the actual quality of the fabric.
[0028] The detection results of quantum dot sensors and artificial intelligence image recognition are fed back to the production equipment in real time so that the parameters of the production equipment can be adjusted in time to ensure the stability of product quality. The feedback system adopts a closed-loop control algorithm to automatically adjust the parameters of the production equipment according to the detection results. Suppose the parameter adjustment amount of the production equipment is , the test result parameters are , then adjust the algorithm function According to the production process and equipment characteristics, a mathematical model between production equipment parameters and test results is established. By analyzing and optimizing a large amount of production data, the function is determined. In order to find the specific form of the function, when establishing the mathematical model, it is necessary to fully consider the working principle of the production equipment, the process flow and the impact of the test results on the product quality. The parameters enable the production equipment to be adjusted accurately and timely according to the test results, thus ensuring the stability of product quality.
[0029] Implementation steps: Sample collection and testing: Collect a large number of different types of standard fabric samples, covering fabrics with various fiber compositions and tissue structures. Perform quantum dot sensor testing on the samples and record the signal strength. , the fiber molecular arrangement parameter is and the crystallinity parameter is At the same time, artificial intelligence image recognition technology is used to analyze the macroscopic appearance of the sample and obtain the image recognition result parameters ; Quality assessment and data recording: Professional quality assessors conduct quality assessment on samples, obtain actual quality assessment results, and record relevant data in detail; Parameter optimization and model determination: Using regression analysis algorithm (least square method), taking the actual quality assessment results as the target value, , , and Perform optimization and solve, and determine the detection result function through repeated iterative calculations and associated models , so that it can accurately reflect the quality characteristics of the fabric; Real-time monitoring and feedback adjustment: During the production process, the detection results of quantum dot sensors and artificial intelligence image recognition are collected in real time, and the detection result parameters are Substitute into the adjustment algorithm function , automatically calculate the parameter adjustment of production equipment , and adjust the parameters of production equipment in time to achieve real-time control and optimization of the production process.
[0030] Embodiment 2: This embodiment provides specific quality inspection configurations for natural fiber fabrics and synthetic fiber fabrics to meet the quality inspection requirements of different types of fabrics.
[0031] Natural fiber fabric quality inspection: In addition to detecting the impurity content and fiber aging degree in the fiber, the quantum dot sensor also analyzes the relationship between the natural texture characteristics of the fiber and its quality through a specific algorithm.
[0032] Assume that the texture complexity of natural fiber fabric is , the impurity content is The aging degree is , then the natural fiber fabric quality evaluation function .
[0033] , , The determination process is as follows: a large number of natural fiber fabric samples are collected, and each sample is tested for texture complexity, impurity content and aging degree. At the same time, professional quality assessors are used to assess the quality of the samples, and regression analysis algorithms are used to determine the quality of the samples with the quality assessment results as the target value. , , An optimization solution is performed so that the quality evaluation function can accurately reflect the quality of natural fiber fabrics.
[0034] In the detection of texture complexity, image analysis technology is used to extract the characteristic parameters of the texture, including but not limited to the directionality and periodicity of the texture.
[0035] Quality inspection of synthetic fiber fabrics: Quantum dot sensors monitor the degree of polymerization and molecular weight distribution of fibers, and combine artificial intelligence image recognition technology to analyze the gloss of synthetic fiber fabrics.
[0036] Assume the degree of polymerization of synthetic fiber fabric is , the molecular weight distribution parameter is , the glossiness parameter is , then the synthetic fiber fabric quality evaluation function .
[0037] , , Determination method: Collect a large number of synthetic fiber fabric samples, test the degree of polymerization, molecular weight distribution and glossiness of the samples, and conduct quality assessment at the same time. Through regression analysis algorithm, take the quality assessment results as the target value, and , , An optimization solution is performed to ensure that the quality evaluation function can accurately reflect the quality of synthetic fiber fabrics.
[0038] In the detection of polymerization degree and molecular weight distribution, spectral analysis technology and other relevant physical and chemical methods are used.
[0039] Implementation steps: For natural fiber fabrics: Collect a large number of natural fiber fabric samples; The texture complexity of each sample is tested, and the characteristic parameters of the texture, such as the directionality and periodicity of the texture, are extracted using image analysis technology to obtain the texture complexity. ; Detection of impurity content and aging degree ; Professional quality assessors conduct quality assessment on samples and obtain quality assessment results; Using regression analysis algorithm, taking quality assessment results as target values, , , Perform optimization to determine the quality evaluation function of natural fiber fabrics .
[0040] For synthetic fabrics: Collect a large number of synthetic fabric samples; Use spectral analysis technology and other relevant physical and chemical methods to detect the degree of polymerization and molecular weight distribution parameters ; Analyzing gloss parameters using artificial intelligence image recognition technology ; Conduct quality assessment and obtain quality assessment results; Through regression analysis algorithm, taking the quality assessment result as the target value, , , Perform optimization to determine the quality evaluation function of synthetic fiber fabrics .
[0041] Embodiment three: This embodiment focuses on the development of VR / AR quality inspection software and its integration with other technologies.
[0042] VR / AR quality inspection software: seamlessly integrated with intelligent inspection equipment, using virtual reality and augmented reality technology to present inspection results to quality inspectors in an immersive way.
[0043] Quality inspectors can interact with virtual scenes through gesture control and voice commands to more intuitively view the location and type of fabric defects.
[0044] The software has the function of automatically identifying uncertain defect types and providing detection standards and processing methods. The automatic recognition algorithm is trained based on artificial intelligence image recognition technology and defect samples in the database. It can quickly and accurately identify various defect types and provide reference for quality inspectors based on preset standards and processing methods.
[0045] In the implementation of virtual reality and augmented reality technologies, hardware devices such as head-mounted display devices and gesture recognition sensors, as well as corresponding software algorithms, are used to create an immersive experience.
[0046] Quantum dot sensor dynamic calibration system: monitors the performance changes of the sensor in real time, and adjusts the sensor parameters using the calibration algorithm by comparing the test results of the standard sample with the actual test results.
[0047] Assume that the test result of the standard sample is The actual test result is , the difference function .
[0048] When the difference exceeds a certain threshold, the calibration algorithm is activated to adjust the sensor parameters to minimize the difference.
[0049] The specific calibration algorithm is as follows: According to the working principle and performance characteristics of the sensor, a parameter adjustment model is established. By continuously trying different parameter combinations, the value of the difference function is gradually reduced until the preset accuracy requirements are met. In the parameter adjustment model, it is necessary to consider the sensor's sensitivity, resolution factors, and the impact of environmental factors on sensor performance.
[0050] Multimodal data fusion algorithm: In the process of establishing the association model between quantum dot sensor data and image recognition data, a multimodal data fusion algorithm is used to fuse the micro data of quantum dot sensors and the macro data of artificial intelligence image recognition, extract more representative features, and improve the accuracy and reliability of the association model.
[0051] The specific algorithm steps include data preprocessing, feature extraction, feature fusion and model training; In the data preprocessing stage, the quantum dot sensor data and image recognition data are normalized to eliminate the dimensional differences of the data; In the feature extraction stage, feature extraction methods suitable for respective data types are used to extract micro and macro features; In the feature fusion stage, weighted fusion, splicing fusion and other methods are used to fuse the two features; In the model training phase, the fused features are used for training to continuously optimize the parameters of the associated model; In the process of feature extraction and fusion, convolutional neural network and recurrent neural network techniques in deep learning are used to extract and fuse features.
[0052] Implementation steps: Develop VR / AR quality inspection software and integrate it with intelligent testing equipment; Collect a large number of fabric defect samples and build a defect database; Use artificial intelligence image recognition technology and defect database to train automatic recognition algorithms; Configure hardware devices such as head-mounted display devices and gesture recognition sensors; Write corresponding software algorithms to achieve immersive experience of virtual reality and augmented reality; Establish a dynamic calibration system for quantum dot sensors to monitor changes in sensor performance in real time; Compare the standard sample test results with the actual test results and calculate the difference function ; When the difference exceeds a threshold, a calibration algorithm is initiated to adjust the sensor parameters according to a parameter adjustment model; When establishing the association model, a multimodal data fusion algorithm is used to perform data preprocessing, feature extraction, feature fusion and model training; Continuously optimize the parameters of the correlation model to improve the accuracy and reliability of the quality inspection management system.
[0053] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.
Claims
1. A quality inspection management system for textile fabric production, characterized in that: The system includes an intelligent detection module, which combines quantum dot sensor technology with artificial intelligence image recognition technology to achieve all-round detection of textile fabric quality; The quantum dot sensor is used to detect the microstructure of textile fabrics, including but not limited to the arrangement of fiber molecules and crystallinity. Assume that the quantum dot sensor detection signal intensity of the fabric sample is , the fiber molecular arrangement parameter is and the crystallinity parameter is , then the detection result function ,in, and The determination method is as follows: collect a large number of different types of standard fabric samples, and detect these samples with quantum dot sensors to obtain the corresponding signal strength , fiber molecular arrangement parameters and crystallinity parameters Then, using the regression analysis algorithm, the actual fabric quality evaluation results are taken as the target value. and Optimize and solve so that the function The error between the output result and the actual quality assessment result is minimized. The specific regression analysis algorithm can use the least square method to determine the optimal parameter value by minimizing the sum of squares of the errors. The artificial intelligence image recognition technology is used to analyze the macroscopic appearance of fabrics, including but not limited to color, texture and pattern. The input fabric image is subjected to feature extraction and classification through a deep learning image recognition algorithm, thereby identifying the macroscopic features of the fabric. In the feature extraction process, a convolutional neural network (CNN) can be used to automatically learn the features of the image, and then classification is performed through a fully connected layer; In the intelligent detection process, the microscopic detection results of the quantum dot sensor are associated with the macroscopic detection results of artificial intelligence image recognition to establish a correlation model ,in, is the artificial intelligence image recognition result parameter, and The method of determining and Similarly, by collecting a large number of standard fabric samples, taking the comprehensive quality assessment results as the target value, and using the regression analysis algorithm for optimization and solution, the output results of the association model can accurately reflect the actual quality of the fabric. In this way, comprehensive and accurate detection of the quality of textile fabrics can be achieved, and the detection results can be fed back to the production equipment in real time to enable real-time control and optimization of the production process.
2. A quality inspection management system for textile fabric production according to claim 1, characterized in that: The quantum dot sensor has specific configuration parameters for natural fiber fabrics and synthetic fiber fabrics; For natural fiber fabrics, the quantum dot sensor not only detects the impurity content and aging degree of the fiber, but also analyzes the relationship between the natural texture characteristics and quality of the fiber through a specific algorithm. Suppose the texture complexity of the natural fiber fabric is , the impurity content is The aging degree is , then the natural fiber fabric quality evaluation function ,in, , , The determination process is as follows: collect a large number of natural fiber fabric samples, test the texture complexity, impurity content and aging degree of each sample, and at the same time combine professional quality assessors to conduct quality assessment on the samples, use regression analysis algorithm, take the quality assessment results as the target value, and conduct quality assessment on the samples. , , An optimization solution is performed so that the quality evaluation function can accurately reflect the quality of natural fiber fabrics. In the detection of texture complexity, image analysis technology can be used to extract the characteristic parameters of the texture, including but not limited to the directionality and periodicity of the texture.
3. The quality inspection management system for textile fabric production according to claim 2 is characterized in that: For synthetic fiber fabrics, the quantum dot sensor monitors the polymerization degree and molecular weight distribution of the fiber, and combines artificial intelligence image recognition technology to analyze the glossiness of the synthetic fiber fabric. Assuming that the polymerization degree of the synthetic fiber fabric is , the molecular weight distribution parameter is , the glossiness parameter is , then the synthetic fiber fabric quality evaluation function ,in, , , The method of determining the quality of synthetic fiber fabrics is as follows: a large number of synthetic fiber fabric samples are collected, and the degree of polymerization, molecular weight distribution and glossiness of the samples are tested, and quality evaluation is performed at the same time. Through the regression analysis algorithm, the quality evaluation results are used as the target value. , , Optimization solutions are performed to ensure that the quality evaluation function can accurately reflect the quality of synthetic fiber fabrics. Spectral analysis technology and other related physical and chemical methods can be used in the detection of degree of polymerization and molecular weight distribution.
4. The quality inspection management system for textile fabric production according to claim 1, characterized in that: A dynamic calibration system for quantum dot sensors is established. The system monitors the performance changes of the sensors in real time. By comparing the differences between the test results of the standard samples and the actual test results, the calibration algorithm is used to adjust the parameters of the sensors. The test results of the standard samples are set to The actual test result is , the difference function ,When the difference exceeds a certain threshold, the calibration algorithm is started to adjust the parameters of the sensor to minimize the difference. The specific calibration algorithm is as follows: According to the working principle and performance characteristics of the sensor, a parameter adjustment model is established. By continuously trying different parameter combinations, the value of the difference function is gradually reduced until the preset accuracy requirements are met. In the parameter adjustment model, it is necessary to consider the sensitivity and resolution factors of the sensor, as well as the impact of environmental factors on sensor performance.
5. The quality inspection management system for textile fabric production according to claim 1, characterized in that: Develop specialized VR / AR quality inspection software that is seamlessly integrated with intelligent inspection equipment. Through virtual reality and augmented reality technology, the inspection results are presented to quality inspectors in an immersive manner. Quality inspectors can interact with virtual scenes through gesture control and voice commands to more intuitively view the location and type of fabric defects. At the same time, the software also has the function of automatically identifying uncertain defect types and providing inspection standards and processing methods. The automatic recognition algorithm in the software is trained based on artificial intelligence image recognition technology and defect samples in the database. It can quickly and accurately identify various defect types and provide reference for quality inspectors based on preset standards and processing methods. In the implementation of virtual reality and augmented reality technology, the hardware equipment of head-mounted display devices and gesture recognition sensors, as well as the corresponding software algorithms, can be used to create an immersive experience.
6. The quality inspection management system for textile fabric production according to claim 5, characterized in that: Combining VR / AR technology with remote collaboration functions can realize cross-regional quality inspection collaboration. Quality inspectors in different locations can enter the same virtual inspection scene through network connection to jointly inspect and evaluate fabrics. During the collaboration process, the system uses data transmission and sharing algorithms to ensure that each quality inspector can obtain inspection data and image information in real time and communicate and collaborate effectively. The data transmission and sharing algorithms use encryption technology and efficient data compression algorithms to ensure data security and transmission efficiency. At the same time, the collaboration platform also provides real-time voice and video communication functions to facilitate communication and discussion among quality inspectors. In the data transmission and sharing algorithms, network communication technology based on the TCP / IP protocol and data encryption algorithm AES can be used to ensure the secure transmission of data.
7. The quality inspection and management system for textile fabric production according to claim 1, characterized in that: In the process of establishing the association model between quantum dot sensor data and image recognition data, a multimodal data fusion algorithm is used. This algorithm fuses the micro data of quantum dot sensors and the macro data of artificial intelligence image recognition to extract more representative features and improve the accuracy and reliability of the association model. The specific algorithm steps include data preprocessing, feature extraction, feature fusion and model training. In the data preprocessing stage, the quantum dot sensor data and image recognition data are normalized to eliminate the dimensional difference of the data. In the feature extraction stage, feature extraction methods suitable for their respective data types are used to extract micro and macro features. In the feature fusion stage, weighted fusion, splicing fusion and other methods are used to fuse the two features. In the model training stage, the fused features are used for training to continuously optimize the parameters of the association model. In the feature extraction and fusion process, convolutional neural network and recurrent neural network technologies in deep learning are used to extract and fuse features.
8. The quality inspection and management system for textile fabric production according to claim 1, characterized in that: Develop a real-time feedback system to feed back the detection results of quantum dot sensors and artificial intelligence image recognition to the production equipment in real time. The feedback system uses a closed-loop control algorithm to automatically adjust the parameters of the production equipment according to the detection results. Suppose the parameter adjustment amount of the production equipment is , the test result parameters are , then adjust the algorithm function ,in, The method of determining is: according to the production process and equipment characteristics, a mathematical model between the production equipment parameters and the test results is established, and the function is determined by analyzing and optimizing a large amount of production data. The specific form of the test results enables the production equipment to be adjusted in time according to the test results to ensure the stability of product quality. When establishing the mathematical model, it is necessary to consider the working principle and process factors of the production equipment, as well as the impact of the test results on product quality.
9. The quality inspection management system for textile fabric production according to claim 1, characterized in that: In order to improve the accuracy and efficiency of artificial intelligence image recognition, data enhancement technology is used to expand the training data. Data enhancement methods include random rotation, flipping, cropping and brightness adjustment. By transforming the original image, more training samples are generated to improve the generalization ability of the model. At the same time, the optimized deep learning model architecture and training algorithm are used to accelerate the convergence speed of the model and improve the detection accuracy. The optimization of the deep learning model architecture includes adjusting the number of network layers, the number of neurons and the activation function to adapt to different types of textile fabric image recognition tasks. The optimization of the training algorithm uses adaptive learning rate adjustment and batch normalization technology to improve training efficiency and stability.
Citation Information
Cited By
Method and system for monitoring production quality of home textile fabric based on industrial data
CN119358850A
Sofa fabric defect intelligent detection system and method based on AI image recognition
CN120318157A
Intelligent textile fabric quality detection method and system
CN120668878A
Automatic identification system for fabric weave structure
CN120913154A
