Automatic Recognition of Woven Fabric Patterns by Stretching & Correcting Technique in Mechanics and Image Processing
In order to overcome the difficulties and low precision for automatic recognition of woven fabric patterns such as tight structures, skew, or nonuniform structures, in this paper, we present a new approach and develop a novel instrument for this purpose. First, the smaple is forced nondestructively deformation by biaxial stretching, it can not only improve clearly the woven fabric structure, but also produce or enhance image features for Pattern Recognition. Second, fabric crossover points are located accurately and their useful information can be extracted by using dynamic windows, and then the crossover points are classified by using a Fuzzy cmeans clustering algorithm by meanings of grey-level and texture features. Finally, we can effectively correct the crossover points identified incorrectly by Template superimposing algorithm and plot automaticlly the pattern drafts of a weave repeat.
Liqing Xie Weidong Yu
Textiles Materials and Technology Lab, Donghua University, Shanghai 201620,R.P.China College of Text Textiles Materials and Technology Lab, Donghua University, Shanghai 201620,R.P.China
国际会议
上海
英文
486-490
2009-05-27（万方平台首次上网日期，不代表论文的发表时间）