Dacron fabric defect detection method and system based on visual image

Abstract

The invention discloses a polyester fabric defect detection method and system based on a visual image, and relates to the field of fabric detection.The method comprises the steps that a multi-scale feature map of a polyester fabric is obtained; generating change track characteristics of each key process; extracting integrated features from the multi-scale feature map of the polyester fabric by using a convolutional neural network, and constructing a multi-granularity feature matrix in combination with the change trajectory features of each key process; based on federated learning and a random forest technology, performing importance scoring on each feature in the multi-granularity feature matrix, and performing feature screening according to importance scores to obtain polyester fabric defect identification input features; and performing score prediction on the polyester fabric defect identification input features through a neural network, and identifying a potential polyester fabric defect area. According to the method, the accuracy and robustness of defect detection of the polyester fabric are improved, and the method has the advantages of data privacy protection and multi-source information integration.

Description

Technical Field

[0001] The present invention relates to the field of fabric detection, and in particular to a polyester fabric defect detection method and system based on visual images. Background Art

[0002] Polyester fabric is a synthetic fabric made from polyester fibers, which are made from polyester polymers through a spinning process. Polyester fabric is widely used in clothing, home textiles and industrial textiles due to its excellent wear resistance, wrinkle resistance and good elasticity. Polyester fabric has the characteristics of light texture, easy washing and quick drying, and bright colors, making it an ideal choice for a variety of applications.

[0003] Polyester fabric defect detection refers to the identification and evaluation of defects in polyester fabrics during production or processing through specific methods and techniques. These defects may include fiber breakage, fabric fuzzing, color difference, holes or other unevenness. The purpose of defect detection is to ensure the quality and consistency of fabrics and prevent unqualified products from entering the market, thereby improving the overall quality of products and user satisfaction. Among them, polyester fabric defect detection based on visual images is a method that uses computer vision technology and image processing algorithms to automatically detect defects on polyester fabrics. Fabric images are obtained through high-resolution camera equipment to identify possible defect areas. This method can improve the speed and accuracy of detection, reduce the errors and labor intensity of manual detection, and is widely used in quality control in the textile industry.

[0004] For example, Chinese patent CN111415349B discloses a polyester filament hair detection method based on image processing technology, which includes processing polyester filament images using a dual-image dual-threshold silk stem and hair segmentation method, segmenting the hair, determining the background position from the non-silk stem position and the non-hair position, calculating the length of each hair strand using a view calibration method, performing statistical processing, and completing the detection of polyester filament hair strands.

[0005] However, the above-mentioned polyester filament hair detection method based on image processing technology still has the following shortcomings: lack of extraction of multi-granular features of polyester images, especially the characteristic changes of each key process in the production process of polyester fabrics, which leads to the segmentation algorithm can only rely on the visual features of the current image, and lacks important contextual information about the changes in the fabric state, which is not conducive to the overall effect and accuracy of defect detection.

[0006] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0007] In view of the problems in the related art, the present invention proposes a polyester fabric defect detection method and system based on visual images to overcome the above-mentioned technical problems existing in the existing related technology.

[0008] To this end, the specific technical solution adopted by the present invention is as follows: According to one aspect of the present invention, a polyester fabric defect detection method based on visual images is provided, the method comprising: S1. Perform multi-scale analysis on the standardized polyester fabric image based on a multi-level dilated convolution structure to obtain a multi-scale feature map of the polyester fabric. S2. Obtain the characteristic changes of each key process in the production process of polyester fabrics, and generate the change trajectory characteristics of each key process; S3. Use convolutional neural network to extract integrated features from the multi-scale feature map of polyester fabric, and combine the change trajectory characteristics of each key process to construct a multi-granularity feature matrix; S4. Based on federated learning and random forest technology, each feature in the multi-granularity feature matrix is ​​scored for importance, and features are screened according to the importance scores to obtain the input features for polyester fabric defect recognition; S5, scoring and predicting the polyester fabric defect recognition input features through a neural network to identify potential polyester fabric defect areas; Furthermore, based on the multi-level dilated convolution structure, a multi-scale analysis is performed on the standardized polyester fabric image, and the multi-scale feature map of the polyester fabric is obtained, including: The standardized polyester fabric image is divided into several image blocks, and a multi-level dilated convolutional layer is designed. Each level has a different dilation rate to form a convolutional layer with different receptive fields. The standardized image is sequentially passed through convolutional layers with different receptive fields to extract the feature maps at each level; The feature maps at each level are adjusted to the same size through sampling operations and weighted fusion is performed to obtain multi-scale feature maps of polyester fabrics.

[0009] Furthermore, the characteristic changes of each key process in the production process of polyester fabric are obtained, and the change trajectory characteristics of each key process are generated, including: Obtain the key processes and process sequences of polyester fabric production, and record the baseline characteristic values ​​of polyester fabric in the initial state of each key process; Based on the baseline characteristic values ​​of polyester fabric in the initial state of each key process, the characteristic change data after each key process is obtained; according to the characteristic change data after each key process, the change trajectory characteristics of each key process are generated.

[0010] Furthermore, based on the baseline characteristic value of the polyester fabric in the initial state of each key process, the characteristic change data after each key process is obtained; according to the characteristic change data after each key process, the change trajectory characteristics of each key process are generated, including: The characteristic data after each key process is compared with the baseline characteristic value in the initial state to obtain the characteristic change data after each key process; Based on the characteristic change data after each key process, numerical methods are used to quantify the change trajectory characteristics of each key process.

[0011] Furthermore, a convolutional neural network is used to extract integrated features from the multi-scale feature map of polyester fabric, and combined with the change trajectory characteristics of each key process, a multi-granularity feature matrix is ​​constructed, including: The multi-scale feature map of polyester fabric is input into the convolutional neural network, and the convolutional neural network is used to extract different levels and scale features of the multi-scale feature map of polyester fabric; The standardized change trajectory features of each key process are fused with the features in the multi-scale feature map to obtain a multi-granularity feature matrix.

[0012] Furthermore, based on federated learning and random forest technology, each feature in the multi-granularity feature matrix is ​​scored for importance, and feature screening is performed based on the importance comprehensive score. The input features for polyester fabric defect recognition include: A random forest model is trained using a training dataset based on a multi-granularity feature matrix, and the importance score of the features in each multi-granularity feature matrix is ​​output by the random forest model; Using the federated learning framework, the feature importance scores of each polyester fabric production line source are encrypted and aggregated to obtain a comprehensive score of feature importance; Using the preset importance comprehensive scoring threshold and the importance comprehensive score of each feature, each feature in the multi-granularity feature matrix is ​​screened to obtain the input features for polyester fabric defect recognition.

[0013] Furthermore, a training data set based on a multi-granularity feature matrix is ​​used to train the random forest model, and the importance scores of the features in each multi-granularity feature matrix output by the random forest model include: Construct a training data set based on a multi-granularity feature matrix and assign a label to each sample in the training data set; Use the particle swarm algorithm to preliminarily screen the features in the multi-granularity feature matrix to obtain the preliminarily screened features; Use grid search algorithm to optimize the parameters of random forest model; The random forest model is trained and validated using the training dataset, and the tree structure is adjusted through multiple iterations to minimize the classification error. The information gain algorithm in the trained random forest model is used to calculate the importance score of each initially screened feature; the initially screened features are sorted from high to low according to their importance score.

[0014] Furthermore, the particle swarm algorithm is used to preliminarily screen the features in the multi-granularity feature matrix, and the features obtained after preliminary screening include: Initialize a particle swarm, and each particle represents a feature subset in a multi-granularity feature matrix; Calculate the fitness value of each particle, and the fitness value is determined by the classification accuracy of the random forest model; Update the velocity and position of the particle, select the optimal feature subset, and use the optimal feature subset as the initial screening feature.

[0015] Furthermore, the federated learning framework is used to encrypt and aggregate the feature importance scores of each polyester fabric production line source, and the comprehensive feature importance scores are obtained, including: Configure the security protocol of federated learning; each polyester fabric production line transmits the local feature importance score to the central server after encryption; Perform encrypted aggregation operations on the central server to calculate the comprehensive importance score of each feature after preliminary screening; The comprehensive importance score of each initial screening feature is sent to each polyester fabric production line; Among them, when the central server performs the encrypted aggregation operation, a weighted mechanism is used to assign corresponding weights according to the amount of data for each production line, and the calculation formula for the comprehensive importance score of each feature after preliminary screening is: ; in, Y f Indicates f The comprehensive importance score of the features after the initial screening; n i Indicates i The amount of data for a polyester fabric production line, n j Indicates j The amount of data for a polyester fabric production line, N represents the total number of polyester fabric production lines participating in federated learning; x f,i Indicates i Production line f The feature importance scores of the features after preliminary screening.

[0016] Furthermore, the neural network is used to score and predict the input features of polyester fabric defect recognition, and the potential polyester fabric defect areas are identified, including: Pre-training a neural network for polyester fabric defect recognition, and using the neural network for polyester fabric defect recognition to score polyester fabric defect recognition input features; Based on the scores of polyester fabric defect identification input features, potential polyester fabric defect areas are identified.

[0017] According to another aspect of the present invention, a polyester fabric defect detection system based on visual images is also provided, which includes: a multi-scale analysis module, a change trajectory feature acquisition module, a feature integration module, a feature screening module and a defect detection module.

[0018] The multi-scale analysis module is used to perform multi-scale analysis on the standardized polyester fabric image based on the multi-level dilated convolution structure to obtain the multi-scale feature map of the polyester fabric; the change trajectory feature acquisition module is used to obtain the feature changes of each key process in the production process of the polyester fabric and generate the change trajectory features of each key process; the feature integration module is used to extract the integrated features from the multi-scale feature map of the polyester fabric using a convolutional neural network, and combine the change trajectory features of each key process to construct a multi-granularity feature matrix; the feature screening module is used to score the importance of each feature in the multi-granularity feature matrix based on federated learning and random forest technology, and screen the features according to the importance score to obtain the input features for polyester fabric defect recognition; the defect detection module is used to score and predict the input features for polyester fabric defect recognition through a neural network to identify potential polyester fabric defect areas.

[0019] The beneficial effects of the present invention are: (1) Improve detection accuracy and robustness: The present invention uses a multi-level dilated convolution structure to perform multi-scale analysis on polyester fabric images, extract multi-scale feature maps, effectively capture image features of different scales and angles, adapt to defects of various sizes and shapes on polyester fabrics, and greatly improve detection accuracy. At the same time, combined with the change trajectory characteristics of each key process in the production process, the characteristic differences of the fabric caused by process changes during production can be perceived during detection. This integration can help better distinguish normal process changes from actual defects, and improve the overall detection robustness and accuracy.

[0020] (2) Data privacy protection and multi-source information integration: The federated learning framework and random forest technology are used to score feature importance. This ensures that when data is shared between multiple production lines, the privacy of each data can still be protected. At the same time, the combination of random forest and particle swarm algorithm is used to efficiently screen features, ensuring high relevance and low redundancy of input features, thereby improving the efficiency and performance of the model. This integration and secure collaboration of multi-source information provides a powerful tool for optimizing the detection process.

[0021] (3) Accurate identification: In the final defect identification stage, the pre-trained neural network model uses the screened high-quality features for scoring and prediction, which can effectively identify potential defect areas on polyester fabrics and achieve efficient and accurate automated detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] 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 will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 is a flow chart of a polyester fabric defect detection method based on visual images according to an embodiment of the present invention; Figure 2 : is a module block diagram of a polyester fabric defect detection system based on visual images according to an embodiment of the present invention In the figure: 1. Multi-scale analysis module; 2. Change trajectory feature acquisition module; 3. Feature integration module; 4. Feature screening module; 5. Defect detection module. DETAILED DESCRIPTION

[0024] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0025] According to an embodiment of the present invention, a polyester fabric defect detection method and system based on visual images are provided.

[0026] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, according to one embodiment of the present invention, a polyester fabric defect detection method based on visual images is provided, the method comprising: S1. Based on the multi-level dilated convolution structure, a multi-scale analysis is performed on the standardized polyester fabric image to obtain the multi-scale feature map of the polyester fabric.

[0027] In one embodiment, a multi-scale analysis is performed on a standardized polyester fabric image based on a multi-level dilated convolution structure to obtain a multi-scale feature map of the polyester fabric, including: The standardized polyester fabric image is divided into several image blocks, and multi-level dilated convolutional layers are designed. Each level has a different dilation rate to form convolutional layers with different receptive fields.

[0028] The standardized image is passed through convolutional layers with different receptive fields in turn to extract the feature maps at each level.

[0029] The feature maps at each level are adjusted to the same size through sampling operations and weighted fusion is performed to obtain multi-scale feature maps of polyester fabrics.

[0030] It should be noted that the purpose of segmenting the image into several image blocks is to process large-size images more efficiently and provide finer-grained feature extraction. The main function of the dilated convolution is to expand the receptive field by inserting holes between the convolution kernels without increasing the number of parameters. The choice of dilation rate should be adjusted according to the fabric characteristics and the expected size of the defects. The feature maps at each level are adjusted to the same size through sampling operations, which usually involves interpolation techniques (such as bilinear interpolation). Weighted fusion is to integrate information at each level, and the choice of weights should be based on the contribution of features at each level to the final goal (such as defect identification).

[0031] Multi-scale feature maps not only enhance the sensitivity to defects of different sizes and shapes, but also provide rich and cross-scale consistent input features for subsequent neural network models. These feature maps can help improve the generalization ability of the model in subsequent model training.

[0032] S2. Obtain the characteristic changes of each key process in the production process of polyester fabrics, and generate the change trajectory characteristics of each key process.

[0033] In one embodiment, the characteristic changes of each key process in the production process of polyester fabric are obtained, and the change trajectory characteristics of each key process are generated, including: The key processes and process sequences of polyester fabric production were obtained, and the baseline characteristic values ​​of polyester fabric in the initial state of each key process were recorded.

[0034] Based on the baseline characteristic values ​​of polyester fabric in the initial state of each key process, the characteristic change data after each key process is obtained; according to the characteristic change data after each key process, the change trajectory characteristics of each key process are generated.

[0035] In one embodiment, based on the baseline characteristic value of polyester fabric in the initial state of each key process, characteristic change data after each key process is obtained; according to the characteristic change data after each key process, the change trajectory characteristics of each key process are generated, including: The characteristic data after each key process is compared with the baseline characteristic value in the initial state to obtain the characteristic change data after each key process.

[0036] Based on the characteristic change data after each key process, numerical methods (such as difference calculation and percentage change) are used to quantify the change trajectory characteristics of each key process. For example, color is represented by ΔE, texture change is calculated by image analysis software to calculate the change in texture complexity, thickness is measured in microns, and strength is measured by strength change value obtained by mechanical testing.

[0037] Among them, the baseline characteristic value can clearly reflect the characteristics of the changes in polyester fabrics in each key process. When measuring the baseline characteristic value, it should be ensured that accurate and standardized methods are used (such as color measurement under standard light sources, standard mechanical testing equipment). The frequency of collecting characteristic change data during the process should be determined according to the process characteristics and the dynamics of fabric changes. ΔE for color change is based on the color difference formula of the International Illumination Commission standard and should be adjusted according to the actual light source and observation conditions. The calculation of texture complexity uses a grayscale co-occurrence matrix or other image analysis techniques. The elasticity and compressibility of the fabric should be considered when measuring thickness, and mechanical testing should be carried out under conditions where the ambient temperature and humidity can be controlled.

[0038] The change trajectory features are stored in a structured form to facilitate subsequent analysis and processing. Data analysis using the generated change trajectory features can identify problems or trends in the production process. This not only helps to adjust production parameters immediately, but also prevents potential production problems through trend prediction.

[0039] S3. Use convolutional neural network to extract integrated features from the multi-scale feature map of polyester fabric, and combine the change trajectory characteristics of each key process to construct a multi-granularity feature matrix.

[0040] In one embodiment, a convolutional neural network is used to extract integrated features from the multi-scale feature map of polyester fabric, and combined with the change trajectory features of each key process, a multi-granularity feature matrix is ​​constructed including: The multi-scale feature map of polyester fabric is input into the convolutional neural network, and the convolutional neural network is used to extract different levels and scale features of the multi-scale feature map of polyester fabric.

[0041] The standardized change trajectory features of each key process are fused with the features in the multi-scale feature map to obtain a multi-granularity feature matrix.

[0042] It should be noted that the multi-scale feature map is input into the convolutional neural network to allow the convolutional neural network to extract features at different scales. This multi-scale nature helps detect polyester fabric defects of different sizes and shapes, allowing the convolutional neural network to not only identify local details but also understand broader contextual information.

[0043] The fusion of change trajectory features with multi-scale features usually adopts feature connection, weighted average or other more complex fusion methods, such as attention mechanism. The purpose of constructing a multi-granularity feature matrix is ​​to unify features of different levels and types in a structured representation. This matrix not only enhances the expressive power of features, but also facilitates subsequent machine learning algorithms to process and analyze. It usually contains feature vectors output from convolutional neural networks and feature vectors generated by each process change trajectory, forming a comprehensive feature set.

[0044] S4. Based on federated learning and random forest technology, each feature in the multi-granularity feature matrix is ​​scored for importance, and features are screened according to the importance scores to obtain the input features for polyester fabric defect identification.

[0045] In one embodiment, based on federated learning and random forest technology, each feature in the multi-granularity feature matrix is ​​scored for importance, and feature screening is performed based on the importance comprehensive score, and the polyester fabric defect recognition input features obtained include: The random forest model is trained using a training dataset based on a multi-granularity feature matrix, and the importance scores of the features in each multi-granularity feature matrix are output by the random forest model.

[0046] Using the federated learning framework, the feature importance scores of each polyester fabric production line source are encrypted and aggregated to obtain a comprehensive score of feature importance.

[0047] Using the preset importance comprehensive scoring threshold and the importance comprehensive score of each feature, each feature in the multi-granularity feature matrix is ​​screened to obtain the input features for polyester fabric defect recognition.

[0048] In one embodiment, a training data set based on a multi-granularity feature matrix is ​​used to train a random forest model, and the importance score of the features in each multi-granularity feature matrix output by the random forest model includes: Construct a training dataset based on a multi-granularity feature matrix and assign a label to each sample in the training dataset.

[0049] The particle swarm algorithm is used to preliminarily screen the features in the multi-granularity feature matrix to obtain the preliminarily screened features.

[0050] The grid search algorithm is used to optimize the parameters of the random forest model, including the number of trees, the maximum number of features, the maximum tree depth, and the minimum number of samples.

[0051] The training dataset is used to train and validate the random forest model, and the structure of the tree is adjusted through multiple iterations to minimize the classification error.

[0052] The information gain algorithm in the trained random forest model is used to calculate the importance score of each initially screened feature; the initially screened features are sorted from high to low according to their importance score.

[0053] In one embodiment, a particle swarm algorithm is used to preliminarily screen the features in the multi-granularity feature matrix, and the features obtained after the preliminary screening include: Initialize the particle swarm, and each particle represents a feature subset in the multi-granularity feature matrix.

[0054] The fitness value of each particle is calculated, and the fitness value is determined by the classification accuracy of the random forest model.

[0055] Update the velocity and position of the particle, select the optimal feature subset, and use the optimal feature subset as the initial screening feature.

[0056] Specifically, the first step in constructing a training dataset is to extract samples from the multi-granular feature matrix and assign labels to each sample. The labels can be binary (e.g., defective or not) or multi-category (different types of defects).

[0057] The particle swarm algorithm is an optimization algorithm based on swarm intelligence. It simulates information sharing between individuals in a group and dynamically adjusts the position and speed of particles to find the feature combination with the best classification performance. The features after preliminary screening can reduce the data dimension and improve the efficiency of model training. Grid search is a hyperparameter optimization method that defines a grid of candidate parameters (such as the number of trees, the maximum number of features, the maximum tree depth, the minimum number of samples, etc.) and evaluates the model performance of each parameter combination one by one. The optimization goal is usually to improve the accuracy or F1 score of the model. The results of grid search directly affect the performance and stability of the random forest model.

[0058] Information gain in random forests is used to measure the impact of a feature on the classification result. Information gain is usually based on indicators such as entropy and Gini impurity, reflecting the contribution of the feature in reducing the uncertainty of the data set. By calculating the split contribution of each feature at each tree node, the features can be ranked to identify the most discriminative features.

[0059] In one embodiment, the federated learning framework is used to encrypt and aggregate the feature importance scores of each polyester fabric production line source to obtain a comprehensive feature importance score including: Configure security protocols for federated learning, such as homomorphic encryption or secure multi-party computing (MPC), to ensure that data is always encrypted during transmission and aggregation; each polyester fabric production line transmits the local feature importance score to the central server after encryption.

[0060] An encrypted aggregation operation is performed on the central server to calculate the comprehensive importance score of each feature after preliminary screening.

[0061] The combined importance score of each initial screening feature is sent to each polyester fabric production line.

[0062] Among them, when the central server performs the encrypted aggregation operation, a weighted mechanism is used to assign corresponding weights according to the amount of data for each production line, and the calculation formula for the comprehensive importance score of each feature after preliminary screening is: ; in, Y f Indicates f The comprehensive importance score of the features after the initial screening; n i Indicates i The amount of data for a polyester fabric production line, n j Indicates j The amount of data for a polyester fabric production line, N represents the total number of polyester fabric production lines participating in federated learning; x f,i Indicates i Production line f The feature importance scores of the features after preliminary screening.

[0063] Specifically, homomorphic encryption is the direct addition and multiplication of encrypted data. The result is still correct after decryption, and aggregation can be performed without decrypting the data. Secure multi-party computation ensures that multiple participants complete computing tasks without disclosing any private data.

[0064] The purpose of the weighting mechanism is to adjust the score according to the amount of data of each production line, ensuring that the production line with a large amount of data has a greater impact on the final result. The calculated comprehensive importance score is distributed back to each participating production line in an encrypted state. This ensures that each production line can obtain information from a global perspective without leaking local data, helping to optimize local feature selection and model training.

[0065] S5. Use a neural network to score and predict the input features of polyester fabric defect recognition and identify potential polyester fabric defect areas.

[0066] In one embodiment, a neural network is used to score and predict polyester fabric defect recognition input features, and potential polyester fabric defect areas are identified, including: A neural network for polyester fabric defect recognition is pre-trained and the input features for polyester fabric defect recognition are scored using the neural network for polyester fabric defect recognition.

[0067] Based on the scores of polyester fabric defect identification input features, potential polyester fabric defect areas are identified.

[0068] Specifically, the pre-trained neural network uses a large amount of labeled data (including defective and non-defective samples) to build a model that can identify defects in polyester fabrics. The pre-training steps include: (1) Collect and label a large number of polyester fabric image datasets to ensure that the dataset contains various types of defects.

[0069] (2) Choose a suitable neural network structure, such as convolutional neural network (CNN), because CNN performs well in image recognition tasks.

[0070] (3) Use the labeled dataset to train the model and adjust the network weights to minimize the training error. In the process, data augmentation techniques (such as flipping, rotation, and scaling) may be used to increase data diversity and improve the generalization ability of the model.

[0071] (4) Perform model validation and testing to ensure that the model performs well on unseen data.

[0072] After the neural network has gone through the pre-training phase, the pre-trained neural network processes each input feature and calculates the score or probability of each feature through forward propagation. The score represents the possibility that the area is a defective area. A threshold is set, and areas with scores exceeding the threshold are marked as potential defective areas.

[0073] like Figure 2 As shown, according to another embodiment of the present invention, a polyester fabric defect detection system based on visual images is also provided, which includes: a multi-scale analysis module 1, a change trajectory feature acquisition module 2, a feature integration module 3, a feature screening module 4 and a defect detection module 5.

[0074] The multi-scale analysis module 1 is used to perform multi-scale analysis on the standardized polyester fabric image based on the multi-level dilated convolution structure to obtain the multi-scale feature map of the polyester fabric; the change trajectory feature acquisition module 2 is used to obtain the feature changes of each key process in the production process of the polyester fabric and generate the change trajectory features of each key process; the feature integration module 3 is used to extract the integrated features from the multi-scale feature map of the polyester fabric using a convolutional neural network, and construct a multi-granular feature matrix in combination with the change trajectory features of each key process; the feature screening module 4 is used to score the importance of each feature in the multi-granular feature matrix based on federated learning and random forest technology, and screen the features according to the importance scores to obtain the polyester fabric defect recognition input features; the defect detection module 5 is used to score and predict the polyester fabric defect recognition input features through a neural network to identify potential polyester fabric defect areas.

[0075] In order to facilitate understanding of the above technical solutions of the present invention, the working principle of the present invention in the actual process is described in detail below.

[0076] 1. Perform multi-scale feature extraction on standardized polyester fabric images.

[0077] The input polyester fabric image size is 224x224 pixels. The image is normalized before entering the network so that each pixel value is between [0, 1]. The normalized image is split into several 56x56 pixel image blocks.

[0078] A multi-level dilated convolutional layer is designed, and the dilation rates of the convolutional layers are 1, 2, and 3. Each convolutional layer is applied to the image block to obtain feature maps with different receptive fields.

[0079] The size of the feature map extracted after each convolution layer is 56x56, which is unified to 224x224 through sampling operation. The feature maps extracted with different expansion rates are weighted fused to obtain the final multi-scale feature map. For example: The feature map F1 extracted by the first convolution layer (dilation rate = 1) has a size of 224x224.

[0080] The feature map F2 extracted by the second convolution layer (dilation rate = 2) has a size of 224x224.

[0081] The feature map F3 extracted by the third convolution layer (dilation rate = 3) has a size of 224x224.

[0082] The fused feature map F combined , expressed as a weighted sum: , where w1, w2, and w3 are the preset weights of F1, F2, and F3 respectively.

[0083] 2. Obtaining the characteristic changes of key processes of polyester fabrics (1) Obtaining key processes The production process of polyester fabric is divided into four key steps: weaving, dyeing, coating and final shaping.

[0084] (2) Record baseline characteristic values For each key process, record the baseline characteristics of its initial state. For example, at the beginning of the weaving process, the baseline characteristic values ​​of polyester fabric are: [0.8, 1.2, 0.9, 1.1] (such as fabric density, fabric stretch and other parameters).

[0085] (3) Obtaining feature changes after each process After the coloring process, the feature changes to: [0.1, -0.05, 0.03, 0.05].

[0086] After coating treatment, the characteristic changes are: [-0.02, 0.1, -0.03, 0.07].

[0087] After the finalization process, the feature changes are: [0.05, 0.03, 0.1, 0.08].

[0088] (4) Generate change trajectory features Based on the feature change data after each process, the change trajectory features of each process [0.9, 1.15, 0.93, 1.15] are generated.

[0089] 3. Use convolutional neural networks to extract and integrate features and construct a multi-granularity feature matrix (1) Input convolutional neural network The multi-scale feature map is input into the convolutional neural network to further extract multi-level features. After the convolutional neural network is trained, the output feature vector obtained is: cnn =[0.12, 0.34, 0.56, 0.78, 0.91].

[0090] (2) Characteristics of fusion change trajectory The multi-granularity feature matrix is: [0.12, 0.34, 0.56, 0.78, 0.91, 0.9, 1.15, 0.93, 1.15].

[0091] 4. Feature screening based on federated learning and random forest (1) Feature Importance Scoring Use the random forest model to train the multi-granularity feature matrix to obtain the importance score of each feature, such as the scores of F1, F2, and F3: F1(0.12): Importance score = 0.85.

[0092] F2 (0.34): Importance score = 0.92.

[0093] F3 (0.56): importance score = 0.75.

[0094] (2) Federated Learning Encryption Aggregation Each polyester fabric production line (assuming there are 2 production lines) sends the local feature importance score to the central server through encryption. The central server uses the federated learning framework for encrypted aggregation to obtain a comprehensive score of feature importance.

[0095] (3) Feature screening Set the comprehensive threshold of importance score to 0.80, and filter out features with importance scores higher than this threshold. For example: F1: Comprehensive importance score = 0.85, retained.

[0096] F2: Comprehensive importance score = 0.92, retained.

[0097] F3: comprehensive importance score = 0.75, discarded.

[0098] 5. Defect Identification and Area Prediction (1) Neural network rating prediction The filtered features are input into the neural network for prediction. After the neural network scores, the defect prediction score of each area is obtained. For example: Position (x=45, y=120) score = 0.95 (indicates that there is a defect at this position) Position (x=120, y=150) score = 0.78 (indicates that the position is normal) (2) Identification of potential defect areas Identify potential polyester fabric defect areas by setting a threshold (for example: 0.85): The position (x=45, y=120) is identified as the defect area.

[0099] The position (x=120, y=150) is considered the normal region.

[0100] In summary, the present invention uses a multi-level dilated convolution structure to perform multi-scale analysis on polyester fabric images, extract multi-scale feature maps, effectively capture image features of different scales and angles, adapt to defects of various sizes and shapes on polyester fabrics, and greatly improve the accuracy of detection. At the same time, combined with the change trajectory characteristics of each key process in the production process, the feature differences caused by process changes in the fabric during production can be perceived during detection. This integration can help better distinguish normal process changes from actual defects, and improve the overall detection robustness and accuracy. The federated learning framework and random forest technology are used to score feature importance. This ensures that when data is shared between multiple production lines, the privacy of each data can still be protected. At the same time, the combination of random forest and particle swarm algorithm is used to efficiently screen features, ensure high correlation and low redundancy of input features, and improve the efficiency and performance of the model. This integration and secure collaboration of multi-source information provides a powerful tool for optimizing the detection process. In the final defect recognition stage, the pre-trained neural network model uses the high-quality features after screening for scoring prediction, which can efficiently identify potential defect areas on polyester fabrics and achieve efficient and accurate automated detection.

[0101] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A polyester fabric defect detection method based on visual images, characterized in that: The method includes: S1. Perform multi-scale analysis on the standardized polyester fabric image based on a multi-level dilated convolution structure to obtain a multi-scale feature map of the polyester fabric. S2. Obtain the characteristic changes of each key process in the production process of polyester fabrics, and generate the change trajectory characteristics of each key process; S3. Use convolutional neural network to extract integrated features from the multi-scale feature map of polyester fabric, and combine the change trajectory characteristics of each key process to construct a multi-granularity feature matrix; S4. Based on federated learning and random forest technology, each feature in the multi-granularity feature matrix is ​​scored for importance, and features are screened according to the importance scores to obtain the input features for polyester fabric defect recognition; S5, scoring and predicting the polyester fabric defect recognition input features through a neural network to identify potential polyester fabric defect areas; The multi-scale analysis of the standardized polyester fabric image based on the multi-level dilated convolution structure to obtain the multi-scale feature map of the polyester fabric includes: The standardized polyester fabric image is divided into several image blocks, and a multi-level dilated convolutional layer is designed. Each level has a different dilation rate to form a convolutional layer with different receptive fields. The standardized image is sequentially passed through convolutional layers with different receptive fields to extract the feature maps at each level; The feature maps at each level are adjusted to the same size through sampling operations and weighted fusion is performed to obtain multi-scale feature maps of polyester fabrics.

2. The polyester fabric defect detection method based on visual images according to claim 1 is characterized in that: The method of obtaining characteristic changes of each key process in the production process of polyester fabric and generating the change trajectory characteristics of each key process includes: Obtain the key processes and process sequences of polyester fabric production, and record the baseline characteristic values ​​of polyester fabric in the initial state of each key process; Based on the baseline characteristic values ​​of polyester fabric in the initial state of each key process, the characteristic change data after each key process is obtained; according to the characteristic change data after each key process, the change trajectory characteristics of each key process are generated.

3. The polyester fabric defect detection method based on visual images according to claim 2 is characterized in that: The characteristic change data after each key process is obtained based on the baseline characteristic value of the polyester fabric in the initial state of each key process; According to the characteristic change data after each key process, the change trajectory characteristics of each key process are generated, including: The characteristic data after each key process is compared with the baseline characteristic value in the initial state to obtain the characteristic change data after each key process; Based on the characteristic change data after each key process, numerical methods are used to quantify the change trajectory characteristics of each key process.

4. The polyester fabric defect detection method based on visual images according to claim 1 is characterized in that: The method of extracting integrated features from the multi-scale feature map of polyester fabric using a convolutional neural network and combining the change trajectory features of each key process to construct a multi-granularity feature matrix includes: The multi-scale feature map of polyester fabric is input into the convolutional neural network, and the convolutional neural network is used to extract different levels and scale features of the multi-scale feature map of polyester fabric; The standardized change trajectory features of each key process are fused with the features in the multi-scale feature map to obtain a multi-granularity feature matrix.

5. The polyester fabric defect detection method based on visual images according to claim 1 is characterized in that: Based on the federated learning and random forest technology, each feature in the multi-granularity feature matrix is ​​scored comprehensively in terms of importance, and features are screened according to the comprehensive importance scores, and the polyester fabric defect recognition input features obtained include: A random forest model is trained using a training dataset based on a multi-granularity feature matrix, and the importance score of the features in each multi-granularity feature matrix is ​​output by the random forest model; Using the federated learning framework, the feature importance scores of each polyester fabric production line source are encrypted and aggregated to obtain a comprehensive score of feature importance; Using the preset importance comprehensive scoring threshold and the importance comprehensive score of each feature, each feature in the multi-granularity feature matrix is ​​screened to obtain the input features for polyester fabric defect recognition.

6. The polyester fabric defect detection method based on visual images according to claim 5 is characterized in that: The random forest model is trained using a training data set based on a multi-granularity feature matrix, and the importance score of the features in each multi-granularity feature matrix output by the random forest model includes: Construct a training data set based on a multi-granularity feature matrix and assign a label to each sample in the training data set; Use the particle swarm algorithm to preliminarily screen the features in the multi-granularity feature matrix to obtain the preliminarily screened features; Use grid search algorithm to optimize the parameters of random forest model; The random forest model is trained and validated using the training dataset, and the tree structure is adjusted through multiple iterations to minimize the classification error. The information gain algorithm in the trained random forest model is used to calculate the importance score of each initially screened feature; the initially screened features are sorted from high to low according to their importance score.

7. The polyester fabric defect detection method based on visual images according to claim 6 is characterized in that: The particle swarm algorithm is used to preliminarily screen the features in the multi-granularity feature matrix, and the features obtained after the preliminary screening include: Initialize a particle swarm, and each particle represents a feature subset in a multi-granularity feature matrix; Calculate the fitness value of each particle, and the fitness value is determined by the classification accuracy of the random forest model; Update the velocity and position of the particle, select the optimal feature subset, and use the optimal feature subset as the initial screening feature.

8. The polyester fabric defect detection method based on visual images according to claim 6 is characterized in that: The federated learning framework is used to encrypt and aggregate the feature importance scores of each polyester fabric production line source to obtain a comprehensive score of feature importance, including: Configure the security protocol of federated learning; each polyester fabric production line transmits the local feature importance score to the central server after encryption; Perform encrypted aggregation operations on the central server to calculate the comprehensive importance score of each feature after preliminary screening; The comprehensive importance score of each initial screening feature is sent to each polyester fabric production line; Among them, when the central server performs the encrypted aggregation operation, a weighted mechanism is used to assign corresponding weights according to the amount of data for each production line, and the calculation formula for the comprehensive importance score of each feature after preliminary screening is: ； in, Y f Indicates f The comprehensive importance score of the features after the initial screening; n i Indicates i The amount of data for a polyester fabric production line, n j Indicates j The amount of data for a polyester fabric production line, N represents the total number of polyester fabric production lines participating in federated learning; x f,i Indicates i Production line f The feature importance scores of the features after preliminary screening.

9. The polyester fabric defect detection method based on visual images according to claim 1 is characterized in that: The scoring and prediction of the polyester fabric defect recognition input features by the neural network to identify potential polyester fabric defect areas includes: Pre-training a neural network for polyester fabric defect recognition, and using the neural network for polyester fabric defect recognition to score polyester fabric defect recognition input features; Based on the scores of polyester fabric defect identification input features, potential polyester fabric defect areas are identified.

10. A polyester fabric defect detection system based on visual images, used to implement the polyester fabric defect detection method based on visual images according to any one of claims 1 to 9, characterized in that: The system includes: a multi-scale analysis module, a change trajectory feature acquisition module, a feature integration module, a feature screening module and a defect detection module; The multi-scale analysis module is used to perform multi-scale analysis on the standardized polyester fabric image based on a multi-level dilated convolution structure to obtain a multi-scale feature map of the polyester fabric; The change trajectory feature acquisition module is used to acquire the feature changes of each key process in the production process of polyester fabrics and generate the change trajectory features of each key process; The feature integration module is used to extract integrated features from the multi-scale feature map of polyester fabric using a convolutional neural network, and to construct a multi-granularity feature matrix in combination with the change trajectory features of each key process; The feature screening module is used to score the importance of each feature in the multi-granularity feature matrix based on federated learning and random forest technology, and screen the features according to the importance scores to obtain input features for polyester fabric defect recognition; The defect detection module is used to score and predict the polyester fabric defect recognition input features through a neural network to identify potential polyester fabric defect areas.

Images