Due to low environmental impact and inexpensive cost, natural fibres like flax fibre (FF) increasingly gain interest as reinforcement structure for composite materials [1]. The properties of FF reinforced plastics (FFRP) highly depend on the textile composition of the FF-semi finished product, fundamentally on the FF-yarn structure. For load bearing FFRP applications only a homogeneous distribution and uni¬directional orientation of the fibres will result in an optimal utilisation of the FFs’ advantageous strength and stiffness properties. Due to the natural growth of the fibres, irregularities like uneven length and thickness occur. Textile yarn spinning normally requires applying yarn twist to ensure sufficient tensile strength and level compaction for follow-up processes. The resulting influences on the yarn structure must be characterised to take the effects on the FFRP in consideration.

Important structural-textile features of FF yarns are thickness, hairiness, twist angle, twist direction, and number of twists [2]. For some of these characteristics standard physical tests and according test devices exist. Yet, most of these tests are not suitable for inline-testing. This work introduces methodologies of characterizing these characteristics using digital image processing in Python to achieve fast, accurate, repeatable and non-destructive tests. Appropriate methods of image segmentation and feature extraction for the FFs’ natural character are introduced and compared. To determine yarn twist related features, a novel application of Fast Fourier Transformation in combination with magnitude spectra is developed. A comparison of the results to physical tests reveals the accuracy of the introduced methods of digital image processing.

References

[1] K.L.Pickering, M.G. Aruan Efendy, T.M. Lea, Composites Part A: Applied Science and Manufacturing, 2016, 83, 98-112.

[2] L. Mohammed, M. N. M. Ansari, G. Pua, M. Jawaid, M. Saiful Islam, International Journal of Polymer Science, 2015, 2015, 1-15.