Evaluating fibrous architecture of nonwovens with computer-assisted microscopy

A new computerized imaging, circular polarized light microscopy technique i s developed to measure the orientation of fibers in images of nonwovens. Th e system consists of a modified Olympus BX50 polarized microscope, a Sony A VC-D7 video camera, and a Silicon Graphics Indy computer. There are two met hods of image segmentation: Gaussian distribution based minimum error thres holding and Poisson distribution based minimum error thresholding. Segments of fiber midlines are traced with vertical and horizontal scanlines or, al ternatively, whole midlines are identified recursively with the Hilditch te chnique. The fiber midlines produced with this technique are traversed by a midline traversal algorithm, and the orientation distribution is obtained by least squares line fitting. The accuracy of the technique is evaluated a gainst synthetic images composed of straight lines and sinuous curves. Kupi er's statistic is used to evaluate the consistency of the fiber orientation calculations. A statistical analysis of the results shows that the propose d Hilditch connectivity, which preserves skeletonization, produces the most accurate results.