Identification Of Fabric Surface Color Texture Features And Its Application

This paper engaged to utilize computer vision or image processing technologyinto textile field, deeply researched the appearance of fabric, devoted to colordetection and texture identification. Color detection and texture recognition are twoindependently main line in this paper, a fabric color analysis and management systemdesign scheme and a new fabric pattern method were proposed based on thosetheories. How to accurately simulate human vision by computer vision and how toapply computer vision to auxiliary control textile designation or production andinspect the quality of textile is the core content of this research.In the measurement and analysis of fabric surface color feature, firstly analyzethe accuracy and applicability of several different color systems in color description,aim to seeking out the optimal color system which is close to human’s eye, the resultshows that the CIE Lab color system, developed by the International IlluminationCommission(CIE), is the best. In the result of that the color features directly relate tofabric color measurement, chromatic aberration contrast, the quality of fabric dyeing,dyeing rate and the calculation proportion of each dye dyed in variety of dyes,Corresponding mathematical models of each system will be specific introduced inchapter2(the theoretical model of measurement and evaluation of the surface of thefabric color features). In order to contrast the color difference between two differentfabrics, the group has compared the characteristics and application scope of severalcommon chromatic aberration formulas.The theoretical model is the basis of establishment of the fabric color analysisand management system, which is consist of computer standard light box, GiabitEthernet camera and relevant data line, specifically derives from Visual C++,OpenCV (open Computer Vision), Access, SQL programming environment, dividesinto color characterization description, color difference comparative analysis, fabricsoiled/discolored analysis, fabric dyeing dye uptake rate measurement and fabricdyeing dyes dosage calculation five subsystems, simultaneously detailed descriptionsof the establishment technological process of those subsystems, the relationshipbetween them, the application method of each subsystem are particular introduced. Inorder to verify and optimize the system, the measurement results were specificanalyzed in matlab environment, an effective method was promoted to solvequantization parameters simultaneously.For identification fabric surface texture features introduced segmentation feature,firstly an algorithm to segment general fabric images was introduced, then anotheralgorithm to segment texture fabric image be proposed, next a conjunction of the twoclustering algorithm by logic operation can used to segment complex fabric images.On the basis, extract the active contour information of the segmented ones to detectthe fabric pattern aliased or not, extract fabric pattern to solve the problem of compleximages can’t be woven though CAM/CAD system.