Abstract:

In order to realize the identification of fabric defects and the matching location of the defective samples in garment production and to solve the problem of rapid detection, extraction and replacement of defective samples, taking the dyed denim as the research object, the common defect marks were classified according to the characteristics, and based on the topological relationship model of spatial objects, the matrix relationship and topological intersection model between convex and concave samples and block and linear defects were deduced based on 9-intersection matrix model and 4-intersection matrix model. This topology model was applied to the extraction process of defective sample in the production of cowboy clothing, and an algorithm for extracting defective samples was proposed by combining the image information of the defect identification and the sample data from the automatic discharge system. Finally, a single-layered denim cloth with a width of 1.52 m and a length of 8 m was used as the experimental object to verify the proposed method for extracting defective samples. The results show that the method can accurately locate the coordinate information of the defects on the cloth, and successfully extract the information of the samples containing the defects to supply the supplementary discharge of the automatic discharge system, and the accuracy rate is above 95%.

Key words: defective sample, topological relation, defect extraction , itersection model