Automatic detection of the layout of color yarns with logical analysis

In this paper, a novel method based on logical analysis is proposed to recognize the layout of color yarns of yarndyed fabric from the color pattern. The mathematic expressions of the color pattern and the layout of color yarns are introduced first, and then the fitness of the layout of color yarns to the color pattern is defined by the relation between them. The principle of the proposed recognition method begins with the assumption of the color of the first warp yarn in the color pattern. The whole algorithm procedure is then analyzed in detail with an actual sample color pattern. Experiments on some color patterns recognized from actual yarn-dyed fabrics, some color patterns simulated manually and some color patterns including error color information of floats prove that the method proposed in this paper is effective for detecting the layout of color yarns from the color pattern of yarn-dyed fabric and it has fault-tolerance ability in some degree. The research in the paper can be used to construct the whole recognition system of the parameters of yarn-dyed fabric.     

Automatic recognition of woven fabric patterns based on pattern database

In traditional method, the woven fabric pattern can not be easily recognized because of the error judgments of float type. To solve this problem, a novel method based on the pattern database was proposed to detect the woven fabric pattern in this study. FCM algorithm was used to classify the floats in the fabric image into two categories, warp floats and weft floats. The classification result then was then compared with the standard patterns in the pattern database, and the similar coefficients between them were calculated. The pattern which has the largest similar coefficient with the classification result was considered as the correct woven pattern, and its diagram and name were output in the system. The test for actual fabric images indicated that the method proposed can recognize the fabric patterns effectively, and the diagram and name of the fabric pattern can be recognized simultaneously. The experiments can also prove that the algorithm which is based on pattern database has good fault tolerant ability, and it can recognize the fabric patterns with a range of error judgments.     

Self-organizing map network for automatically recognizing color texture fabric nature

The method of recognizing color texture brought forth in the present study is to employ unsupervised learning network to automatically recognize the fabric type and the main texture types. Firstly, the color scanner is adopted to extract fabric image which is afterwards saved as the digital image. Secondly,CIE-Lab color model is taken to obtain the feature value and wavelet transform is utilized to display the texture of the fabric image. Thirdly, co-occurrence matrix is employed to figure out the feature values of the texture structure such as angular second moment, entropy, homogeneity, contrast. Finally, self-organizing map (SOM) network is used as the classifier. The experiment result shows that the study can automatically and accurately classify the fabric types (including shuttle-woven fabric, jersey fabric and non-woven fabric) and main texture type of the fabric (such as plain weave, twill weave, satin weave, single jersey, double jersey and non-woven fabric).     

Simulation of the weave structural design of synthetic woven fabrics

This study surveys the basic procedure of data base system of the fabric structural design which can be linked with existing pattern design and garment design CAD systems. For this purpose, the theoretical and empirical equations related to the fabric structural design are analyzed and discussed with various fabric specimens. The fabric structural parameters such as weave density coefficient, cover factor and yarn density coefficient of various kinds of fabrics are calculated using the empirical equations. These calculated fabric structural parameters of many kinds of polyester and nylon fabrics are compared and discussed with weave pattern, and materials such as polyester and nylon. Furthermore the difference between fabric structural parameters calculated by empirical equations are analyzed with polyester and nylon fabrics as a basic study for data base system of the fabric structural design. Finally, the weave density coefficients of polyester and nylon fabrics were analysed and discussed with shrinkages of dyeing and finishing processes, and also surveyed according to the weaving company and weave structural parameters such as weave pattern and denier.     

Automatic inspection of double-system-mélange yarn-dyed fabric density with color-gradient image

According to the color yarns in the fabric, the yarn-dyed fabrics are divided into two categories: single-systemmélange color fabrics and double-system-mélange color fabrics. The method for inspecting the density of double-systemmélange color fabrics is discussed in this study. By analyzing the pattern and color characters of double-system-mélange color fabrics, color-gradient image is proposed to detect the density. The gray-projection method and correlation coefficient method are selected to locate the wefts and warps. With the help of Fourier low-pass filter, the positions of yarns in double-system-mélange color fabric are found, and then the density can be obtained by counting the yarns in a unit length automatically. The experiment proved that the method proposed can detect double-system-mélange color fabric density successfully.     

Automatic color separating system for printed fabric using the self-organizing map network approach

We can only use color numbers, color values and design to describe the color pattern of printed fabrics, which is different from woven fabrics with yarn disposition and texture as pattern determinants. Since most printed fabrics contain many different patterns nowadays, we need more than words and simple methods to describe the color patterns. The complication in pattern identification has made the analysis and comparison difficult and will have to be conducted manually. The automatic computer color separating system for printed fabrics proposed in this paper uses unsupervised learning network to automatically separate printed colors. The system first uses color scanner to pick the image of the printed fabrics and stores it as digital image. Then, it uses wavelet transformation to minify the fabric image to reduce the calculation load of color separation and also reserve the printing structure and color distribution of the original image. It also uses LAB color model to acquire characteristic value of the colors and the Self-Organizing Map Network (SOMN) to conduct color separation. According to our experimental results, this system can rapidly and automatically complete color separation and identify repeating patterns for printed fabrics’ images.     

基于深度学习芒果图像在线识别与计数方法研究

为实现果树产量的智能评估,本研究对自然环境下的圣心芒果树图像进行果实识别研究,提出基于深度学习算法的芒果图像在线识别计数方法。首先,采用Faster R-CNN深度学习模型构建芒果图像识别算法;接着基于微信小程序与网页平台开发芒果图像上传模块,实现随时随地上传芒果图像至服务器;然后采用基于TCP协议的服务器客户端通信模式,并结合基于MATLAB平台的Faster R-CNN程序集,构建上传图像的在线分析模块,实现线上芒果图像的实时识别与计数;最终,芒果图像的识别与计数结果通过微信小程序和Web页面程序反馈给用户,内容包括单张图片和1个果园区域内所有图片的识别与计数结果,并实现青色芒果和红色芒果的分类统计。应用本研究构建的在线深度学习识别计数方法,在自然环境下采集125幅芒果图像进行测试试验。结果表明：芒果图像识别算法的计数识别准确率达到82.3%,其中漏检率与误检率分别为11.7%和8.6%,平均计数误差与计数误差率分别为4.2和7.9%;芒果图像在线识别计数方法能有效实现果树图像的采集、上传、识别与计数、分类统计和结果反馈,对整个果园区域内结果数量进行统计与分析,为果园的智慧管理提供科学决策依据。     

Image analysis of the luster of fabrics with modified cross-section fibers

We have investigated the luster of modified cross-sectional fiber fabrics as one of the essential quality estimates for clothing development. We have confirmed an objective evaluation method, and have determined the experimental luster characteristics of modified cross-section fibers. The cross-section of the fibers in a fabric affects the appearance of a textile. We used the image analysis method to investigate the luster to determine the critical factors influencing the appearance of modified cross-section fiber fabrics. For similarly structured textiles in a component fabric, clear differences were observed in the fabric weave, density, percentage, and total area of blobs, which is image region. Color played a decisive role in the luster of the textiles, and luster was not significantly influenced by the modified cross-section fabric weave. In addition, the degree of luster did not increase in the order plain to twill to satin for modified cross-sectional fiber fabrics. All the split-type microfibers exhibited higher numerical luster values (percentage of pixels, and number and total area of blobs) than sea-island microfibers did. The degree of luster of the modified cross-sectional fiber fabrics was not high at specular reflection angles.     

A general geometrical model for predicting color mixing of woven fabrics from colored warp and filling yarns

Despite the advances in woven fabrics, CAD systems, and weaving technologies, the process of weave/color selection for each area of a Jacquard pattern still requires the intervention of the CAD system operator and/or designer, who works from color gamut. Relying on the designer subjective assessment, multiple weaving trials may be needed to produce a fabric that matches the target artwork or sample. In this paper, a general geometric model is provided to predict the color contribution of warp and filling yarns of a given woven fabric in terms of warp and pick densities, warp and filling yarns sizes, weave, size of the color repeat of warp and filling yarns, and the number of yarns of different colors. Such geometrical modeling, combined with sound existing color mixing equations, paves the road for the automation of the process of weaves and color selection and thus dramatically reduces the production cycle.     

Assessing the relationship among fabric constructional parameters, fractional reflectances and cover factors of polyester fabrics by experimental and mathematical methods

This paper focuses on the assessment of the relation among constructional properties, fractional reflectances and cover factors of fabrics woven from polyester yarns. A novel equation for the calculation of the relation between fractional reflectance and fabric cover factor was proposed and the usage of the equation was assessed by reflectance measurements. 48 polyester fabrics having different constructional parameters were used and the fabrics differed from each other by their cover factors. The warp yarn type and count, warp density and warp yarn twist were the same but weft yarn count, weft yarn fiber count and weft density were different for the fabrics in the experimental sub-groups. The reflectance measurements were conducted on the pretreated but undyed fabric samples as well as on the individual yarn systems of the same fabrics. Fabrics with the same cover factors exhibited different fractional reflectances. Reflectances were found to be dependent on the cover factor as well as on yarn fiber fineness, yarn count, yarn density and fabric weave. The changes in crimp of the yarns according to different construction parameters also governed the changes in fractional reflectances of fabric surfaces. The proposed equation was tested according to different fabric construction parameters and it was concluded that fiber fineness and weave pattern were among the most important parameters which govern the total light reflectances from the fabric surfaces, although they are not incorporated in the calculation of the fabric cover factors. The proposed equation was used to explain the effects of these components on the reflectance behavior of the fabric surfaces and on fabric cover.     

The effect of fabric balance and fabric cover on surface roughness of polyester fabrics

The effects of fabric balance and fabric cover on surface roughness values of textured polyester woven fabrics with different constructional parameters were investigated. The warp yarn properties (type, count and warp density) were kept constant while the effect of variation in weft yarn density and weave pattern were studied. Measurements were conducted on pre-treated white fabric samples and the results assessed in relation to their constructional properties. A general overview of the results showed that surface roughness values of polyester fabrics affected by fabric balance and fabric cover and the effects were related to fabric thickness, yarn densities, yarn crimp, positioning of yarns in fabric structure. A change in weave pattern from sateen to plain increased the fabric balance and fabric cover, but decreased the surface roughness. Similarly, an increase in weft density increased the fabric balance and fabric cover, but decreased surface roughness. In order to produce fabrics with smooth surface properties yarn density should be increased, yarn float lengths decreased, cover of fabrics increased and fabric balance improved.     

Crimp analysis of worsted fabrics in the terms of fabric extension behaviour

Inside a woven fabric structure, warp and weft yarns acquire crimp as a result of yarns interlacing according to the weave pattern. Since warp and weft yarns are oriented in two perpendicular directions, applying tensile load in one direction causes extension in the load side and fabric contraction in the opposite direction. This process was investigated in this study by using an image processing procedure and it was found that fabric’s extension is in coincidence with yarn’s de-crimping process in the same direction. After the de-crimping stage, yarns in the load direction will be extended and at the same time crimp in the other direction will be increased, until jamming phenomenon happens in the fabric structure. The crimp interchange between warp and weft yarns follows a three-order polynomial function with a turning point in which the yarns in the load direction have no crimp.     

Assessment of color strength and color difference values of polyester fabrics containing continuous weft yarns after abrasion

This paper assesses the color difference and color strength values (K/S) obtained for eight disperse-dyed polyester fabric samples with different fabric construction parameters (weft yarn type, weft yarn count, weft density and fabric weave) after four sets of abrasion cycles. Warp yarn type and count, warp density, and warp yarn twist are the same for all fabrics. Fabric samples are dyed in a commercial red disperse dye (C.I. Disperse Red 74:1) and four different abrasion cycles (2500, 5000, 7500, 10000) are used. TheK/S values of the abraided fabrics and color difference values between the control fabric (dyed but not abraided) and abraded fabrics are calculated. The main differences in theK/S and color difference values are observed between 0–2500 abrasion cycles. The high tenacity of the polyester fibers and continuous polyester yarns causes some fuzz but no pilling formation on the fabric surface that lead to increasedK/S values and color differences. Fiber dullness, yarn thickness, yarn density and fabric weave are concluded to have different effects on the appearance after abrasion.     

Electromagnetic interference shielding effectiveness of SS/PET hybrid yarn incorporated woven fabrics

A detailed study of electromagnetic shielding effectiveness (EMSE) of woven fabrics made of polyester and stainless steel/polyester blended conductive yarn was presented in this research work. Fabrics with different structures were analyzed and their shielding behavior was reported under different frequencies. Shielding efficiency of fabric was analyzed by vector network analyzer in the frequency range of 300 kHz to 1.5 GHz using coaxial transmission line holder. The effects of different fabric parameters such as weft density, proportion of conductive weft yarn, proportion of stainless steel content, grid openness, weave pattern and number of fabric layers on EMSE of fabrics were studied. The EMSE of fabric was found to be increased with increase in proportion of conductive yarn in the weft way. With increase in overall stainless-steel content in the fabric, the EMSE of fabric was increased. As such weave is considered, it did not have significant effect on EMSE of fabrics. But fabric with lower openness and aperture ratio showed better conducting network, hence better shielding. With increase in number of layers of fabric and ply yarns, EMSE of fabric was increased.     

A novel method based on loop shape for simulating knitted fabric using mass spring model

Fabric simulation is result of combining various methods that have been dramatically evolved during the decade. However, there still exist some limitations. One of the limitations in fabric simulation is lack of using fabric properties such as material, weave structure, density and so on in mass spring modeling. In knitted fabrics, this issue is more important due to their different fabric structures. In this paper, a new mesh based on loop shape for simulating 1×1 rib fabric is proposed which is called Loop mesh. By using the Loop and common meshes, 3D model of drape behavior in 18 types of knitted fabric are simulated. Results of simulation are compared with 3D shape of actual drape behavior in fabric samples which are achieved by depth camera. Results show that the Loop mesh is able to predict the drape behavior of knitted fabric with error value of 5 percent as compared with the real result. It can be found that the Loop mesh produced a closest drape shape to the actual fabric drape than other mesh models.     

基于图像和改进U-net模型的小麦赤霉病穗识别

为了快速、有效地监测小麦赤霉病的发生情况,利用数码相机对人工接种赤霉病菌的小麦田进行RGB图像获取,在图像预处理基础上,利用深度学习网络U-net来对人工标记好的发病麦穗图像进行训练。通过深度学习数据集的构建与测试,建立了基于RGB图像和改进U-net网络模型的小麦赤霉病识别与监测模型,并对模型识别结果进行了验证。结果表明,U-net可以很好地提取图像波段信息,但对于比较复杂的麦穗图像,在使用Keras方法进行图像语义分割时,需要对U-net网络结构进行改进,即在下采样部分加入Dropout层。与人工标记结果相比,模型识别结果的一致性较好,具有较高的监测精度。该模型平均精度为0.969 4,损失函数值为0.075 9,平均交并比MIoU为0.799。上述结果说明改进的U-net模型可以很好地识别和监测小麦图像中的发病麦穗,并在发病麦穗的分割上具有很好的效果     

Surface pretreatment of polyester fabric for ink jet printing with radio frequency O2 plasma

Plasma treatment of textiles is becoming more and more popular as a surface modification technique. It not only changes the outermost layer of a material without interfering with the bulk properties but also offers the advantage of greater chemical flexibility to obtain multifunctional textiles. Inkjet printing is becoming increasingly important and popular for the printing of textiles. When polyester fabric is inkjet printed with pigment-based inks, the printed patterns have poor color yield and easily bleeding. As a result, the fabric requires pretreatment prior to the stage of ink-jet printing. In the present study the polyester fabric was printed with magenta pigment ink after radio frequency O₂ plasma surface-treatment. At such condition, polyester fabric could obtain the effects of features with enhanced color yield and excellent pattern sharpness. The results showed that there were better patterns when the polyester fabric was treated for 9 min at a working pressure of 40 Pa and a working power 80 W. SEM images indicated that radio frequency oxygen plasma induced modifications to the surface of polyester fabric with more micro pits. Water absorption time measurement showed that the hydrophilicity of polyester fabric was remarkably improved after treatment. Anti-bleeding performance of the fabric was improved greatly, too. Therefore, radio frequency O₂ plasma treatment with the ink-jet printing technique could improve the final printed properties of polyester fabric.     

基于霍特林变换的稻米彩色图像粒形检测算法

利用稻米分割后轮廓灰度图与背景灰度图的灰度均值之差和灰度方差之差进行米粒图像分割效果定量评价,对7个彩色通道的稻米图像进行分割评判,选取I1(红色、绿色、蓝色通道的平均值)通道进行稻米图像分割。提取分割后标注的单粒米粒边界的二维坐标向量,对坐标向量进行霍特林变换,通过计算变换后米粒最小外接矩阵来表征稻米粒形,简化了现有的稻米粒形检测算法。检测稻米粒型时,算法在MATLAB7.5.0环境下运行。该算法所得米粒长宽比与人工检测结果的平均相对误差为1.65%,每幅图像平均耗时0.323s;而最小外接矩形算法的平均相对误差为2.24%,每幅图像平均耗时2·837s。