04 Aug Defect detection on thread spools using vision and artificial intelligence.
Playing a leading role in the textile sector of the Girona region, Antex aims to provide a very high quality yarn. During manufacturing and handling, certain defects may appear on the yarn spools. The large daily production of Antex makes the manual inspection of the spools a difficult task, and a single faulty case can lead to the rejection of part of a batch by the customer, resulting in important economic losses.
Consequently, Opsis developed a machine vision system capable of automatically detecting these defective spools. By means of different 3D cameras, their corresponding lasers, a programmable incremental encoder and Opsis’ OVLib libraries, quality control is carried out. The spools passing through the system are scanned and the resulting images allow the identification and measurement of the different types of defect that may be present.
The variety of possible defects is wide, some of which are superficial and easier to detect. Other ones are more complex and even almost imperceptible to the naked eye. That is the reason why both traditional machine vision techniques and machine learning methods were used.
The installation of the machine vision system marks a milestone in the inspection of yarn spools in terms of:
- Efficiency and speed
- Precision and repeatability