Binocular Vision-Based Yarn Orientation Measurement of Biaxial Weft-Knitted Composites

Abstract

The mechanical properties of fiber-reinforced composites are highly dependent on the local fiber orientation. In this study, a low-cost yarn orientation reconstruction approach for the composite components’ surface was built, utilizing binocular structured light detection technology to accomplish the effective fiber orientation detection of composite surfaces. It enables the quick acquisition of samples of the revolving body shape without blind spots with an electric turntable. Four collecting operations may completely cover the sample surface, the trajectory recognition coverage rate reached 80%, and the manual verification of the yarn space deviation showed good agreement with the automated technique. The results demonstrated that the developed system based on the proposed method can achieve the automatic recognition of yarn paths of views with different angles, which mostly satisfied quality control criteria in actual manufacturing processes.

Keywords

• binocular vision
• textile composite
• preform
• yarn orientation
• non-destructive testing
• image processing