Advanced mathematical analysis to optimize image processing for pore size assessment of knitted fabrics

Abstract

Abstract Pore size is one of the significant factors influencing fabric performance, such as comfort and protection. The reliable measurement of pore size is fundamental to studying fabric structure and performance. Image processing is the popular method to characterize pore size because of its convenience and efficiency. However, there is no standard procedure to retrieve pores from fabric images. Furthermore, many uncertainties exist when deciding on critical thresholds, making the process subjective and less comparable. The study proposed a standardized workflow to determine the pore size and the distribution of knitted fabrics using image processing. The major contributions include discussing the impact of structuring element size on fabric structure, suggesting an objective way to select the optimal structuring element size, and proposing an approach to isolate noise pores. As a result, the measurement of pore size using image processing is more accurate and reliable, reflecting the true pore size distribution of fabrics, establishing a solid foundation for producing qualified fabrics. Moreover, the proposed method includes specific steps for image processing and objective criteria for selecting thresholds, enabling the generation of computer algorithms that automatically process fabric images on a large scale, making it more efficient and saving labor, time, and money.