CHARACTERIZING FIBER CRIMP BY IMAGE-ANALYSIS - DEFINITIONS, ALGORITHMS, AND TECHNIQUES

Crimp is one of the most important aspects of fiber geometry. Image analysis provides a means for quantifying crimp morphology in an efficient and reliable manner. This study presents a protocol for the automated analysis of two-dimensional crimp data. Preprocessing operations consist of histogram equalization, boundary definition with the use of image morphology, and boundary information storage by chain codes. Descriptors for characterizing and quantifying crimps are defined and algorithms for calculating these descriptors are developed. Preliminary results of a trial analysis indicate that these techniques may prove useful in the general application of crimp characterization and classification.