Research On The Construction And Application Of The Color Representation Model Of Colored Spun Fabric

As an important attribute of the appearance of an object and the first element that affects vision,color is an important part of textile and apparel design and development,as well as production and sales.Colored spun fabrics are spun by two or more pre-colored fibers blending of different colors,which has a hazy three-dimensional feeling and a unique color style.With the advancement of digital imaging systems,how to use the digital image processing technology to measure and represent color of fabrics is the current mainstream research direction.In this paper,aiming at the randomness and non-uniformity of coloration of colored spun fabrics,combing with digital image processing technology and pattern recognition to construct a color representation model based on colored spun fabrics,and use it to analyze the factors affecting the color of colored spun fabrics.The main research contents are as follows:(1)Aiming at the problem that the overall coloration of dyed fibers of colored spun fabrics is difficult to describe accurately,a color representation model based on mixed color space has been established.This model is based on the HSV and Lab color spaces,fusing the components with the same attributes.On this basis,the color moment and LBP-gray co-occurrence matrix image method is used to extract the color and texture features of the fabric.The experimental results show that for colored spinning knitted fabrics and woven fabrics,the established color representation model can accurately characterize the tiny changes of mass ratio parameters and types of fibers.(2)Aiming at the problem of the limited ability of LBP-gray co-occurrence matrix image method to represent the texture of colored spun fabrics,a texture feature extraction algorithm based on Three Structure Descriptor is established.This algorithm is based on HSV,Lab and YUV color spaces,extracting Three Structure texture features and global color features from the brightness and color channels of each color space,and performs feature-level fusion.Meanwhile,the Pearson correlation coefficient method is used to quantitatively analyze the effectiveness of the color representation model.The experimental results show that the established color representation model can accurately characterize the tiny changes of ratio of dyed fibers of fabrics.(3)Research on the color difference and the general law of coloration caused by changes in the fiber morphology and weaving process parameters of colored spun fabrics.Based on the Three Structure feature extraction model,the uniform mode is introduced and a lowdimensional Uniform Three Structure Descriptor is established.At the same time,the " intra class distance " is used to construct the criterion of coloration difference.The results show that the established color representation model has ideal color description ability and can clarify the law of color changes between fabrics with different fabric structures.The research in this paper has important significance for constructing a scientific color representation model of colored spun fabrics.It can also provide a strong guarantee for the digital management and analysis of colored textiles,and lay a solid foundation for the digital transformation and upgrading of the entire industrial chain of colored textiles industry.