Pattern images input to The knitting CAD system have a large number of different colors, and it is necessary to reduce the number of colors in the image by merging similar colors through color-separation algorithms. However, current pattern images usually have the problem of image degradation, which seriously affects the accuracy of color-separation algorithms. In addition, the traditional color-separation algorithm needs to rely on the manual setting of clustering parameters, which is very time-consuming and laborious. To solve these problems, In this paper, we propose a density peak clustering color-separation algorithm based on self-organizing mapping (SOM) neural network, which firstly uses enhanced super-resolution generative adversarial network (Real-ESRGAN) blind super-resolution reconstruction network to clarify the degraded image and obtain a high-resolution image with clear boundaries; secondly, we carry out the initial clustering through the SOM neural network to simplify the image information; and then we use an improved density peak clustering (DPC) algorithm to calculate clustering centers under the conditions of conforming to the perception of human eyes, and carry out secondary clustering on the image; finally, carry out image post-processing through the variegated spots merging algorithm based on the connected component analysis. The experimental results show that the algorithm proposed in this study can effectively deal with degraded pattern images, and the clustering effectiveness evaluation indices perform well.

Extracting standardized digital design patterns from real knitted fabric images is critical for textile reverse engineering and digital archiving. Unlike smooth graphics, knitted fabrics exhibit high-frequency textures from yarn loop interlacing, introducing significant grayscale variations within same-color regions. Existing algorithms struggle to distinguish these from pattern edges, causing color quantization and segmentation failures. To suppress yarn texture while preserving edges between color blocks, we propose an adaptive pattern extraction method using Bayesian-optimized bilateral filtering. The primary contribution lies in providing a domain-specific, application-focused integrated framework. Specifically, (1) a knitting-texture-aware multidimensional evaluation parameter is constructed by integrating physical-cause-based texture features (gray-level co-occurrence matrix (GLCM) contrast, homogeneity, and Laplacian variance) with perception-based edge preservation metrics (the Sobel operator and the structural similarity index (SSIM)), enabling accurate discrimination between yarn-level texture noise and pattern-level color block boundaries—a distinction that generic image quality metrics cannot make. (2) Then, this domain-specific objective function is embedded within a Bayesian optimization framework to achieve automatic, zero-shot, per-image parameter adaptation across different knitting processes, without requiring any external training data. K-means color quantization maps in continuous tones to discrete classes, generating standardized patterns meeting knitting requirements. Experiments on 316 samples covering six processes show our method outperforms standard denoising and advanced algorithms like relative total variation (RTV), achieving an average SSIM of 0.83 and PSNR of 26.92 dB, reducing processing time from 15–30 min to 21 s per image, providing efficient automation for knitted Computer-Aided Design (CAD) systems.