Computer vision for textured yarn interlace (Nip) measurements at high speeds

This paper deals specifically with interlace measurements of intermingled, false-twist textured yarns. A system has been built which is capable of analysing yarns at slow-speed/high-resolution, and also at high-speed/low-resolution. An outline description of the system is given here, along with descriptions of techniques used in the hardware and software which improve the system performance. To monitor interlace, a signal processing algorithm is developed and explained which enables reliable interlace detection, and is effective against diametric noise. Several tests are carried out, which prove the ability of the system to match the results from careful manual inspection of the yarn. The final test is to run the yarn at a high speed, and to compare the results with the slow-speed analysis. The results show that at high speeds, the system is reliable.