COMPUTER AIDED CONTROL OF CUTTING ERROR IN TEXTILE PRODUCTS


Creative Commons License

Çevik K. K., Kocer H. E.

TEKSTIL VE KONFEKSIYON, cilt.27, sa.3, ss.300-308, 2017 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 27 Sayı: 3
  • Basım Tarihi: 2017
  • Dergi Adı: TEKSTIL VE KONFEKSIYON
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, TR DİZİN (ULAKBİM)
  • Sayfa Sayıları: ss.300-308
  • Anahtar Kelimeler: Textile cutting errors, Textile quality control, Image processing, DEFECT DETECTION, ACTIVE CONTOURS, INSPECTION, FILTERS, EDGES
  • Akdeniz Üniversitesi Adresli: Hayır

Özet

At present, the audits about the cutting error of textile products (leather, fabric, etc.) are made by the human by the eye via the template. Making these audits that necessitate accurate measurement by eye both takes so much time and enhance the risk occurrence risk. In this article, the image processing based industrial quality control system that determines the cutting errors of textile products automatically and discriminates between faulty and faultless products is explained. The system minimizes the faults based upon the human auditing and increases the number of pieces that are controlled by the unit of time. The performed system is composed of Panel PC, line scan camera, system of conveyor, basket control unit, image processing software and control user interface. The textile pieces (cuts) to be inspected come into the part by the conveyor where the camera and illumination unit are available, and the image is captured. This captured image is sent to the Panel PC and controlled whether there is a cutting error via image processing software. According to the result of the audit, the basket system at the end of the conveyor (conveyor belt) moves back and forth on wheel rail, and the textile pieces are provided to fall into the required basket. The performed system was tested on the leather pieces that were taken from a company in the leather sector. Totally it was tried by 150 times for 50 pieces of leather in 5 different templates and these pieces felt into the required basket correctly by discriminating for faulty/faultiness ones by 149 times (99,33% success ratio).