Explore the vast range of titles available.

This paper presents a turnkey integrated system that can be operated in real time for real textile manufacturers. Eight types of defects in woven fabric, including stain, broken end, broken weft, hole, nep, double pick, kinky weft and float can be recognized and classified. First, an image is captured by a CMOS industrial camera with a pixel size of 4600 × 600 above the batcher at 20 m/min. After that, the four-stage image processing procedure is applied to detect defects and for classification. Stage 1 is image pre-processing; the filtration of the image noise is carried out by a Gaussian filter. The light source is corrected to reduce the uneven brightness resulting from halo formation. The improved mask dodging algorithm is used to reduce the standard deviation of the corrected original image. Afterwards, the background texture is filtered by an averaging filter, and the mean value is corrected for histogram shifting, so that this system is robust to the texture and color changes of woven fabric. The binary segmentation threshold is determined using the mean value and standard deviation of an image with a normal sample. Stage 2 uses adaptive binarization for separation of the background and defects and to filter the noise. In Stage 3, the morphological processing is used before the defect contour is circled, i.e., four features of each block, including the defect area, the aspect ratio of the defect, the average gray level of the defect and the defect orientation, which are calculated according to the range of contour. The image defect recognition dataset consists of 2246 images. The results show that the detection success rate is 96.44%, and the false alarm rate is 3.21%. In Stage 4, the defect classification is implemented. The support vector machine (SVM) is used for classification, 230 defect images are used as training samples, and 206 are used as test samples. The experimental results show that the overall defect recognition rate is 96.60%, providing that the software and hardware equipment designed in this study can implement defect detection and classification for woven fabric effectively.

People wear clothes for warmth, survival and necessity in modern life, but in the modern era, eco-friendliness, shortened production times, design and intelligence also matter. To determine the relationship between data series and verify the proximity of each data series, a gray relational analysis, or GRA, is applied to textiles, where seamless bonding technology enhances the bond between components. In this study, a polyurethane prepolymer, 2-hydroxyethyl acrylate (2-HEA) as an end-capping agent and n-octyl acrylate (ODA) as a photoinitiator were used to synthesize a dual-curing polyurethane hot-melt adhesive. Taguchi quality engineering and a gray relational analysis were used to discuss the influence of different mole ratios of NCO:OH and the effect of the molar ratio of the addition of octyl decyl acrylate on the mechanical strength. The Fourier transform infrared spectroscopy (FTIR) results showed the termination of the prepolymer’s polymerization reaction and the C=O peak intensity at 1730 cm−1, indicating efficient bonding to the main chain. Advanced Polymer Chromatography (APC) was used to investigate the high-molecular-weight (20,000–30,000) polyurethane polymer bonded with octyl decyl acrylate to achieve a photothermosetting effect. The thermogravimetric analysis (TGA) results showed that the thermal decomposition temperature of the polyurethane hot-melt adhesive also increased, and they showed the highest pyrolysis temperature (349.89 °C) for the polyhydric alcohols. Furthermore, high peel strength (1.68 kg/cm) and shear strength (34.94 kg/cm2) values were detected with the dual-cure photothermosetting polyurethane hot-melt adhesive. The signal-to-noise ratio was also used to generate the gray relational degree. It was observed that the best parameter ratio of NCO:OH was 4:1 with five moles of monomer. The Taguchi quality engineering method was used to find the parameters of single-quality optimization, and then the gray relation calculation was used to obtain the parameter combination of multi-quality optimization for thermosetting the polyurethane hot-melt adhesive. The study aims to meet the requirements of seamless bonding in textile factories and optimize experimental parameter design by setting target values that can effectively increase production speed and reduce processing time and costs as well.

Five-axis CNC tool machines have been widely used in the aerospace and automotive industries for complex and special-purpose mold production as smart machines. Obtaining good surface quality and manufacturing speed for the designed parts is always a challenging issue in a high-speed machining process. Researchers have provided several algorithms to smooth the trajectory profile or servo controller design to get good product quality. For the multi-axis synchronous motion, only a few researchers have discussed the interaction between the machine’s dynamic ability and tool path/trajectory generation. To ensure synchronous motion, the dynamic performance of each axis becomes a constraint for the frequency of the designed trajectory command. According to this requirement, a maximum jerk value in the trajectory generator could be calculated first, and then, combined with the machining method selected for the parts. A novel jerk value decision-making process is proposed for the parts machining process in this study. This new approach to obtain a better finished surface quality and shorten the machining time without adding supplemental equipment or information is demonstrated experimentally on a five-axis CNC machine for a free-form vane that could help machining operators take maximum advantage of the smart production techniques.

Surface-mount device light emitting diode (SMD-LED) is characterized by small size, wide viewing angle and light weight. It becomes the main package type of LED gradually. The traditional visual inspection is likely to cause misrecognition due to personal subjectivity and different defect recognition standards. Therefore, this study develops an automatic SMD-LED defect detection system, which is characterized by non-contact inspection, defect recognition standardization and upgrading product quality. It detects the common and important defects in LED package components, including missing component, no chip, wire shift and foreign material. In this study the gray scale characteristic of histogram is used as the rapid sieving analysis indicator of missing component defect, and then the component and solder joint are positioned by using fast normalized cross-correlation, and the maximum correlation coefficient value is used as judgment indicator of no chip defect. In order to overcome the difficult identification as the weld line is subject to light rays, the improved Michelson-like contrast (MLC) enhancement is proposed, and the segmentation threshold is selected by entropy information to segment the weld line successfully. Furthermore, in order to overcome the effect of the tolerance of component size and internal electrode and unfixed weld line position resulted from lead frame process on foreign material detection result, the multiscale adaptive Fourier analysis (MAFA) is proposed in the concept of texture anomaly detection for foreign material defect detection. The result proves that the proposed method can segment the defect effectively and correctly compared with the phase-only transform (PHOT) and multiscale phase-only transform (MPHOT), and it can be used in other fields of texture anomaly detection. The overall recognition rate of this system is 98.25%, contributing to the large market demand and high component quality of LED industry.

Silicone resin containing phenyl group was synthesized by hydrolysis–condensation reaction using tetraethoxysilane (TEOS), chlorotrimethylsilane (TMCS) and phenyltriethoxysilane (PhTES). 1 H-NMR and 29 Si-NMR spectroscopy characterizations also confirmed that silicone resin was successfully obtained based on the combination of TEOS, PhTES and TMCS in a crosslinked network structure. Silicone resin has been thoroughly characterized using gel permeation chromatography, thermogravimetric analysis, ultraviolet–visible spectroscopy, contact angle and softening point measurements. A steady increase in the molecular weight, hydrophobic, softening point and maximum degradation temperature of the silicone resin has been observed with the increasing weight percentage of the PhTES crosslink, but transmittance properties decreased. It has been shown that the non-polar component of the higher contact angle of silicone resin can be increased up to 113° by 3 % PhTES.Silicone resin applied to silicone pressure-sensitive adhesives (SPSA) to increase the hydrophobic indicated a decrease in surface energy, leading to improved wettability for polytetrafluoroethylene. The peel strength of SPSA can be increased up to 8 N/2.5 cm by content PhTES 2 % of silicone resin and increased 54 % without PhTES. Graphical Abstract

This study discusses the optimal design of an automatic inspection system for processing light-emitting diode (LED) chips. Based on support vector machine (SVM) with optimal theory, the classifications of micro-defects in light area and electrode area on the chip surface, and develop a robust classification module will be analyzed. In order to design the SVM-based defect classification system effectively, the multiple quality characteristics parameter design. The Taguchi method is used to improve the classifier design, and meanwhile, PCA is used for analysis of multiple quality characteristics on influence of characteristics on multi-class intelligent classifier, to regularly select effective features, and reduce classification data. Aim to reduce the classification data and dimensions, and with features containing higher score of principal component as decision tree support vector machine classification module training basis, the optimal multi-class support vector machine model was established for subdivision of micro-defects of electrode area and light area. The comparison of traditional binary structure support vector machine and neural network classifier was conducted. The overall recognition rate of the inspection system herein was more than 96%, and the classification speed for 500 micro-defects was only 3 s. It is clear that we have effectively established an inspection process, which is highly effective even under disturbance. The process can realize the subdivision of micro-defects, and with quick classification, high accuracy, and high stability. It is applicable to precise LED detection and can be used for accurate inspection of LED of mass production effectively to replace visual inspection, economizing on labor cost.

Embroidered textile is a highly valued artwork. Three-dimensional patterns can be created by variable stitches and the material characteristics of embroidery thread. At present, computer plate-making is widely used in embroidery operation, allowing factories to mass-produce efficiently, but in the quality control stage, manpower is needed for visual inspections. The embroidered textile is classified after quality control. The conforming products are classified into three different categories: non-defective products, defective products but improvable after reworking and re-inspection, and disposable products when defective products cannot be improved. However, there is no specific standard for embroidered textile defects. In order to increase efficiency of human resources, elucidate the defects on embroidery, and achieve automated defect classification, we proposed methods on the recognition and classification of embroidered textile defects. After interviewing producers of embroidered textiles and manufacturers of computer embroidery machines, we identified four types of defect in embroidery textile patterns: foundation yarn floating knit, stitch missing, joint defect, and misregistration for defect recognition. The back-propagation neural network (BPNN) and characteristic procedure classification were used for defect classification. The characteristic procedure classification yielded more accurate results (accuracy rate: training: 100%, testing: 100%), but the process requires more time and cost for training. The BPNN requires the least amount of time, but the recognition rate is slightly lower (accuracy rate: training: 100%, testing: 98.80%) when compared to the characteristic procedure classification. The embroidered textile pattern recognition and classification methods proposed by this study are expected to provide an automated inspection procedure for the embroidery textile industry.

The traditional global template matching is time consuming, has low accuracy, and cannot be adapted to rotation and scale change. The template matching technique proposed in this study improves the time, accuracy and robustness for printed circuit boards (PCB). In order to shorten the image positioning time, the image preprocessing is implemented on PCB image and the image blocks are labeled to obtain the tagged image, and the feature vector is extracted and the marking point region image is selected. The feature vector with rotation change and scale change robustness is extracted from the tagged image after labeling in the PCB image by using artificial neural network, combined with image moments for training. The marking point region image in the PCB image is selected. The scale value of the marking point region image is estimated by parametric template vector matching. The deflection angle of marking point region image is calculated by Hough transform. The obtained scale value and deflection angle value are used for fast template matching to determine the marking point positioning. The three-dimensional (3D) parabolic curve fitting is implemented in marking point positioning and adjacent pixel position to reach the sub-pixel level accuracy. The experiment showed that the proposed template matching technique for the PCB image with or without noise or angle rotation, the average position accuracy error of each translated image is lower than 7 \\[\\upmu \\]m, and the error standard deviation is lower than 5 \\[\\upmu \\]m. The rotation angle error average and standard deviation of angular error of Hough transform are lower than 0.2\\[^{\\circ }\\], more accurate than orientation code (OC) method. The scale value estimation, relative error average and error standard deviation are lower than 0.004 and 0.006 for the image with or without noise. The average complete positioning time of PCB image at resolution of \\[2500\\times 2500\\] is only 0.55 s, which is better than the 3.97 s of traditional global template matching. The results prove that the template matching technique of this study not only has sub-pixel level high accuracy and short computing time, but also has the robustness of rotation change and scale change interference. It can implement rapid, efficient and accurate positioning.

Nano-graphene materials have improved many thermal properties based on polymer systems. The additive polymers’ thermal insulation cannot be significantly increased for use as a reinforcement in multifunctional thermally insulating polymer foam. Herein, we present the development of far-infrared emissivity and antistatic properties using multifunctional nano-graphene polyester fibers. Nano-graphene far-infrared thermal insulation polyester was synthesized with 2% nano-graphene and dispersant polypropylene wax-maleic anhydride (PP wax-MA) using the Taguchi method combined with grey relational analysis (GRA) to improve the thermal properties and the performance of the polymer composite. The thermogravimetric analysis (TGA) shows that the pyrolysis temperature of spinning-grade polyester was increased when the nano-graphene powder was added to the polyester. The differential scanning calorimeter (DSC) analysis confirmed the modification of polyester by nano-graphene, showing the effect of the nucleating agent, which ultimately improved the performance of the polyester. The physical properties of the optimized polyester fibers were improved with a yarn count of 76.5 d, tensile strength of 3.3 g/d, and an elongation at break increased from 23.5% to 26.7% compared with unmodified polymer yarn. These far-infrared emission rates increased from 78% to 83%, whereas the far-infrared temperature increased from 4.0 °C to 22 °C, and the surface resistance increased to 108 Ω. The performance of the optimized modified polyester yarn is far better than single-polypropylene-grafted maleic anhydride yarn. The performance of optimized modified polyester yarn, further confirmed using grey correlation analysis (GRA), can improve the yarns’ mechanical properties and far-infrared functions. Our findings provide an alternative route for developing nano-graphene polyester fabrics suitable for the fabric industry.

Melt spinning machines must be set up according to the process parameters that result in the best end product quality. In this study, artificial intelligence algorithms were employed to create a system that detects abnormal processing parameters and suggests strategies to improve quality. Polypropylene (PP) was selected as the experimental material, and the quality achieved by adjusting the melt spinning machine’s processing parameter settings was used as the basis for judgement. The processing parameters included screw temperature, gear pump temperature, die head temperature, screw speed, gear pump speed, and take-up speed as the six control factors. The four quality characteristics included fineness, breaking strength, elongation at break, and elastic energy modulus. In the first part of our study, we applied fast deep-learning characteristic grid calculations on a 440-item historical data set to train a deep learning neural network and determine methods for multi-quality optimization. In the second part, with the best processing parameters as a benchmark, and given abnormal quality data derived from processing parameter settings deviating from these optimal values, several machine learning and deep learning methods were compared in their ability to find the settings responsible for the abnormal data, which was randomly split into a 210-item training data set and a 210-item verification data set. The random forest method proved to be the best at identifying responsible parameter settings, with accuracy rates of single and double identification classifications together of 100%, for single factor classification of 98.3%, and for double factor classification of 96.0%, thereby confirming that the diagnostic method proposed in this study can effectively predict product abnormality and find the parameter settings responsible for product abnormality.