Research On Defect Detection And Location System Of Garment Piece

Ordinary garments are mostly made of stitched pieces of clothes,and the quality of cut pieces greatly influences the quality of clothes.If defects can be detected before sewing the clothes,repairing or removing the defective pieces can reduce unnecessary sewing work and improve product quality.The traditional fabric defect detection is done by manual,and the detection process is subject to subjective factors,and the efficiency is deficiency.With the development of computer vision to the textile industry,automatic detection system for fabric defects gradually replace the manual detection,and become an important means to ensure the quality of the fabric.At present,some domestic advanced textile and garment enterprises have begun to explore the automatic garment technology,robot and machine vision and other technologies to achieve less artificial or even no automatic garment.Therefore,it is of great significance to study the automatic sewing of garment pieces,and machine vision-based garment defect detection and location method is an important link.In this paper,the defect detection and location method of the garment pieces are studied on the basis of summarizing the existing literature and research at home and abroad.The main contents are as follows:(1)To analyze the existing methods of segmentation and the characteristics of garment cutting,an improved region growing segmentation method is proposed.(2)According to the actual industrial application,the article proposed the method of the front and back pieces of the garment piece to the horizontal strip,analyzed the piece texture,determined the minimum texture period to determine the piece matching method,and discussed the influence for alignment by choosing different matching template.(3)Combining with the existing results of deep learning,classifying the defective and defect-free tissue samples,it is proved that Lenet-5 and R-CNN structure has a good effect in defect detection.(4)Develope garment defect detection and positioning system application software.Finally,summarize the content of the full text,analyze the shortcomings of the article.