Quantitative determination of morphological features of triaxially braided composites by the use of machine vision

The morphologies of the reinforcement strongly influence the material properties of fiber-reinforced composites. The ability to quantify the morphological features of the composites is important for the prediction of material properties and for quality control of the manufacturing processes. This paper presents a method for quantitatively measuring fiber braid angles at the surface of a part by the use of computer-aided image-processing analysis. The materials under investigation are triaxially braided glass/urethane composites. The determination of braid angle at a sample surface is performed non-destructively and is based on the use of a standard fast Fourier transform (FFT) method coupled with a novel directional scanning method developed as part of this work. Calibration tests were performed for the purpose of comparison between computer-based pattern recognition and manual determinations. Good agreement between machine vision results and manual measurements for braid angle was found. The machine vision system is fully automated and features rapid image acquisition and processing with good measurement accuracy. It can be applied to on-line quantitative measurements of geometrical features and other applications that are important for quality control and engineering design.