In this paper, a novel approach to color and pattern analysis is proposed for printed fabric. An unsupervised analysis method is developed using a fuzzy C-means (FCM) clustering algorithm and a specific cluster-validity criterion (SC criterion). First, the printed fabric is captured by a color scanner and converted into full color digital files, then the mean filter is used to smooth the color of the image. The search for good cluster numbers is made by the SC criterion, and the corresponding color clusters are obtained based on the FCM clustering algorithm. Finally, the color and pattern features of the printed fabric are calculated. The experimental results show that this approach is very suitable for analyzing the colors and patterns of printed fabrics.

We introduce a real-time machine vision system we developed with the aim of detecting defects in functional textile fabrics with good precision at relatively fast detection speeds to assist in textile industry quality control. The system consists of image acquisition hardware and image processing software. The software we developed uses data preprocessing techniques to break down raw images to smaller suitable sizes. Filtering is employed to denoise and enhance some features. To generalize and multiply the data to create robustness, we use data augmentation, which is followed by labeling where the defects in the images are labeled and tagged. Lastly, we utilize YOLOv4 for localization where the system is trained with weights of a pretrained model. Our software is deployed with the hardware that we designed to implement the detection system. The designed system shows strong performance in defect detection with precision of [Formula: see text], and recall and [Formula: see text] scores of [Formula: see text] and [Formula: see text], respectively. The detection speed is relatively fast at [Formula: see text] fps with a prediction speed of [Formula: see text] ms. Our system can automatically locate functional textile fabric defects with high confidence in real time.

We need more than words and simple methods to describe the various different color patterns found on printed fabrics nowadays. The complexity in pattern identification has made the analysis and comparison difficult and will have to be managed manually. The automatic computer color separating system for printed fabrics, proposed in this paper, integrates a genetic algorithm (GA) and a self-organizing map network (SOMN) to automatically separate printed colors, so as to eliminate the time-consuming manual color segmentation and registration currently done in the industry. The system first uses a color scanner to record RGB color images of the printed fabrics and uses median filter processing to reduce color changes due to uneven light reflections arising from the fabric surface weaving texture. Then RGB color space is transformed to HSI color space so that color analysis can match human color sense and use customary procedures. Next, the GA is employed to search for color distributions that are the same as the original image of printed fabrics. The area of each sub-image is 9.06% of the original image, not only reducing color segmentation operation time, but completely reserving the print structure and color distribution of the original image. Afterwards, color characteristic values are obtained in HSI color space. Finally, SOMN is adopted for the color segmentation operation. According to our experimental results, this system can rapidly and automatically complete color separation and identify repeating patterns in images from printed fabrics.

This series of study proposed the automatic machine embroidery image color analysis system as an extension of the previous proposed machine embroidery color separation system and repetitive pattern search system. This paper integrated the machine embroidery image color analysis system to achieve system automation. The system is divided into three stages: (1) acquire the embroidery fabric image, filter the acquired image noise, and smooth the embroidery fabric texture before using the transform to reduce image information amount and improve subsequent calculation speed; (2) use the genetic algorithm for the repetitive pattern search, and restore the identified repetitive pattern images to the original size by image pyramids, then use frequency domain template matching to determine the locations of the repetitive patterns to verify repetitive pattern accuracy; (3) apply the Gustafson–Kessel cluster algorithm and cluster validity partition index SC to obtain the machine embroidery image’s number of colors and corresponding areas, then employ half-toning technology to determine color types for chromatography. The integrated system can accurately identify the repetitive patterns for chromatography to realize the automatic drafting of embroidery fabric. It can be further integrated with automatic manufacturing equipment for machine embroidery automation.

No abstract