Defective products are a major contributor toward a decline in profits in textile industries. Hence, there are compelling needs for an automated inspection system to identify and locate defects on the fabric surface. Although much effort has been made by researchers worldwide, there are still challenges with computation and accuracy in the location of defects. In this paper, we propose a hybrid semi-supervised method for fabric defect detection based on variational autoencoder (VAE) and Gaussian mixture model (GMM). The VAE model is trained for feature extraction and image reconstruction while the GMM is used to perform density estimation. By synthesizing the detection results from both image content and latent space, the method can construct defect region boundaries more accurately, which are useful in fabric quality evaluation. The proposed method is validated on AITEX and DAGM 2007 public database. Results demonstrate that the method is qualified for automated detection and outperforms other selected methods in terms of overall performance.

Aiming at the problems of low detection accuracy and high leakage rate in traditional detection algorithms, an improved YOLOv8 algorithm is proposed for automatic detection of fabric defects. A swin transformer block was added to the C2f module in the backbone network, which can transfer information between multiple attention layers in parallel to capture fabric defect information and improve the detection accuracy of small-sized defects. To enhance the model’s performance in detecting defects of various sizes, a bidirectional feature pyramid network (BiFPN) was incorporated into the neck. This allows for the assignment of different weights to defect features in different layers. A convolution block attention module (CBAM) was added to the feature fusion layer, enabling the model to automatically increase the weight of essential features and suppress nonessential features during training to solve the problem of leakage detection of small-sized defects due to occlusion and background confusion. The Wise-IoU (WIoU) loss function replaces the conventional loss function, addressing sample imbalance and directing the model to prioritize average-quality samples. This modification contributes to an overall improvement in the model’s performance. The results of the experiment proved that on the self-constructed fabric defect dataset, the algorithm in this paper achieved an accuracy of 97.7%, recall of 95.1%, and mAP of 96.8%, which are 4.4%, 9.4%, and 5.1% higher than those of the YOLOv8 algorithm, respectively. On the AliCloud Tianchi dataset, the algorithm achieves 52.3%, 49.2%, and 49.8% in terms of accuracy, recall, and mAP, respectively, which is an improvement of 4.4% in terms of accuracy, 2.8% in terms of recall, and 2.7% in terms of mAP compared with the baseline algorithm. The improved YOLOv8 algorithm has a high detection accuracy, low leakage rate, and a detection speed of 107.5 FPS, which aligns with the real-time defect detection speed in the industry.

Aiming at the existing fabric defect detection algorithms with their suboptimal detection accuracy, high model complexity, difficult deployment to the edge of the device, and insufficient ability to meet the problem of real-time detection of fabric defects, we propose a lightweight fabric defect detection model (LWFDD-YOLO) based on the improved YOLOv8n. First, a generalized efficient layer attention network with selective kernel attention (GELAN_SKA) is proposed to replace the C2f module, and selective kernel attention is added to the module to adjust the weights of the convolution kernel according to the different features of different scales, reduce the use of computational resources, thus improving the model’s detection performance and efficiency. Second, a cascaded group attention (CGA) mechanism is added to provide a different input segmentation for each head to enhance the feature diversity of the input attention heads and improve the model’s computational efficiency. An ultra-lightweight dynamic up-sampling operator (Dy_sample) is introduced, which employs a point-sampling-based approach to reduce the consumption of computational resources and improve model performance. Finally, the YOLOv8l framework is utilized to construct a complex model teacher network and the features learned from the teacher network are transferred to the lightweight network proposed in this paper, thus further improving the model performance of the algorithm. The experimental results show that on the self-built fabric defect dataset, the accuracy, recall, and mean average precision ( mAP ) of our algorithm reach 89.4%, 85.2%, and 87.9%, which are 8.9%, 7.2%, and 4.5% higher than the original model, respectively, the number of parameters of the model decreases by 23.4%, the amount of GFLOPs decreases by 25.6%, and the file size is only 9.2 MB. The detection speed can reach 163.4 FPS on GPU. On the AliCloud Tian Chi dataset, the accuracy, recall, and mAP were also improved by 7.4%, 2%, and 4.4%, respectively. The LWFDD-YOLO algorithm proposed in this study can realize real-time detection of fabric defects with relatively obvious improvement in accuracy, and the model requires less memory and is easier to deploy to edge devices, so can be used as a reference for real-time detection of fabric defects.

In the electrospinning process, the Taylor cone, as the jet source, directly affects the jet movement and the quality of the fiber membrane. Therefore, to understand the formation mechanism of the Taylor cone intuitively, a multiphysics coupling model that comprehensively considers the gravitational field, electrostatic field, and fluid field is established, and numerical simulations are conducted in this study. First, we construct a level-set function and analyze the force of the droplet. The gravity, surface tension, and electric field force are coupled to the incompressible Navier–Stokes equation as volume forces, and the nonconservation of the droplet area is solved by approximating the Dirac function with a smooth function. Subsequently, the deformation of the electrospun polyacrylonitrile (PAN) Taylor cone under different process parameters is simulated. Finally, data obtained from the numerical simulation and the average diameter of the electrospun PAN fiber membrane are analyzed via gray relational analysis. The results show that the volume force is the key factor affecting the average diameter of the fiber membrane (the correlation is 0.934). This article provides an effective reference and basis for the analysis and control of the electrospinning process.

This paper provides a new method for predicting the diameter of electrospun nanofibers. Based on the grey system theory, the effects of polyacrylonitrile mass fraction, voltage, flow rate, and receiving distance on fiber diameter were studied. The GM(1,1) (grey model) model and DNGM(1,1) (The DNGM (1,1) model is based on the whitening differential equation using parameters to Directly estimate the approximate Non-homogeneous sequence Grey prediction Model) model were established to predict fiber diameter by a single-factor change, and the results showed high prediction accuracy. The multi-variable grey model MGM(1,n) (MGM(1,n) is a Multivariate Grey prediction Model) was used for prediction of fiber diameter when multiple factors change simultaneously. The results showed that the average modeling fitting error is 8.62%. The background value coefficients of the MGM(1,n) model were optimized, the average modeling fitting error was reduced to 1.01%, and the average prediction error was reduced to 1.33%. Combining the fractional optimization with the background-value coefficient optimization, the optimal background-value coefficient $\alpha$ and the order r were selected. The results showed that the average modeling fitting error is 0.85%, and the average prediction error is 0.38%. The results demonstrate that the grey system theory can effectively predict the diameter of polyacrylonitrile electrospinning fibers with high prediction accuracy. This theory can increase the control of nanofiber diameters in production.