Online Measurement of Sizing Yarn Hairiness Based on Computer Vision

In weaving production, the function of sizing to cover the hairs is of great significance for ensuring the subsequent weaving efficiency. While measuring hairiness in sizing yarns still relies on manual evaluation or offline measurement, this paper proposes a computer-vision-based method for measuring the yarn hairiness on sizing machine, including an instrument and image analysis algorithm. Considering the layered parallel arrangement of yarns on the sizing machine and the undesirable imaging conditions for online measuring, the algorithm is built based on the Hessian multi-scale adaptive threshold and hair length index. The proposed method showed a 94% correlation coefficient with the commercial offline hairiness tester, verifying the effectiveness. The discussion demonstrated the desirable robustness of the proposed method for yarn variety, sizing degree, yarn speed, imaging condition, and algorithm parameters. This study can support for the sizing tasks such as real-time parameter adjustment, online quality monitoring, historical problem tracing.