Abstract: Section Dyed Yarn, characterized by unique colors and textures, is widely utilized in the field of fashion textiles. Fabrics made from section dyed yarn are highly favored for their vibrant and diverse color appearances. To address the randomness and energy consumption issues during the production and sampling process of section dyed yarn, as well as the low fidelity in simulating fabric authenticity, this paper proposes a simulation method for section dyed yarn knitted fabrics based on the B-spline coil model.
Initially, in order to extract the main surface texture information of the yarn from the original images, yarn images are acquired using a scanner. Image processing techniques, including threshold segmentation using the Otsu algorithm and morphological opening and closing operations, are then applied for the preprocessing of the collected section dyed yarn images.
Subsequently, to map the extracted main yarn texture into coil shapes, data points are extracted along the course of the knitting loops from the real fabric image. Using the least squares fitting technique, B-spline coil curves are fitted and constructed. During the mapping process based on these curves, appropriate pixel offsets are needed according to the B-spline normals to achieve the curved effect. As the collected yarn texture is linear, resulting in blank pixels at curved sections of the knitted loop after mapping, nearest-neighbor interpolation is employed to fill these gaps and ensure the integrity of the mapped coils.
Ultimately, a method of arranging coil images horizontally and vertically is used to simulate the fabric. Viewing fabric weaving as a process in which coil images increase from left to right and from bottom to top, starting from the lower-left corner, coils are arranged accordingly. To establish the interlocking relationship between coils, a Boolean matrix is utilized to eliminate pixel overlap positions in intertwined coils based on the simulated fabric organization. Under certain lighting conditions, different positions of the coils exhibit varying brightness effects on the fabric surface. To enhance the realism of the fabric, this study adjusts the grayscale ratios of the coils to create brightness variations. Linear interpolation is employed to achieve a smooth transition between shaded and unshaded areas for a natural gradient effect. Moreover, the proposed simulation method, after obtaining yarn texture and color parameters, enables the rapid display of serialized products by changing yarn colors and arrangement cycles.
The research results demonstrate that the simulated fabric is generally consistent with real fabric in terms of appearance and color arrangement. The simulation method, based on the real section dyed yarn texture, allows for the rapid change of yarn colors and the display of the fabric's color, texture, and pattern effects. This method effectively enhances the authenticity of simulating section dyed yarn fabrics, reducing the difficulty and cost of sample production for businesses.

Key words: Fabric Simulation, Section Dyed Yarn, Least Squares Fitting, B-Spline Curves, Coil Mapping

摘要： 针对段染纱针织物花纹预测困难且现有织物模拟方法真实感有待提升的问题，提出了一种基于真实段染纱的针织物外观模拟方法。首先，使用图像处理技术对获取的段染纱图像进行阈值分割和形态学处理等预处理，从原始图像中提取纱线主体区域。然后，从真实织物中提取线圈点位信息，借助最小二乘拟合构建B样条线圈曲线，并根据曲线将段染纱真实图像映射为线圈图像。最后，在考虑线圈交织覆盖关系以及光照因素后，通过排列线圈实现纬平针针织物及其系列针织面料的外观模拟。研究结果表明：所模拟的线圈模型能够还原真实线圈中的纱线走向，且模拟织物外观花纹效果与实物一致，整体呈现出较好的真实感。

关键词: 织物模拟, 段染纱, 最小二乘拟合, B样条曲线, 线圈映射