Video Processing for Dynamic Evaluation of Yarn Diameter

Yarn diameter is one of the important properties that affect the fabric structure and performance. The laboratory method of measuring the yarn diameter under the microscope is more precise than other methods. However, this method is laborious and localized at certain positions which may not give continuous information about the diameter behavior. This work introduces a computer vision method for continuously capturing the images of a long yarn sample. The analysis of these images was performed using a robust image processing algorithm that has the ability to remove the yarn hairiness and fill-in the yarn body for determining its boundaries and measuring the diameter. The continuity of data produced allows the time-series analysis of the data which reveals a lot of the yarn diameter characteristics including its short term, long term, and periodic variations. The developed algorithm was applied to study the diameter of a slub yarn and the data produced were very comparable to the measured data produced by USTER tester IV.