Automatic Identification of Ramie and Cotton Fibers Using Characteristics in Longitudinal View, Part I: Locating Capture of Fiber Images

Abstract

This paper introduces a new fiber imaging system that can automatically capture a series of microscopic fiber images in the longitudinal view to form a panoramic image of long-fiber segments for reliable identifications of blended fibers. The panoramic image shows thumbnails of long fibers without missing or duplicating any segments on the sample slide. Firstly, the sample slide is scanned quickly in the X- and Y-axes to capture sequential images of fibers and to register the locations of each image. Secondly, a panoramic image is formed by stitching the sub-images together according to their positions. Thirdly, the panoramic image is analyzed to calculate the skeletons of long-fiber segments, which provide simple representations of fiber shapes and locations, and to register the locations of individual segments for the second scan of the slide. Finally, the triaxial motorized stage transports the slide to the location of each registered segment, readjusts the focus, and captures high quality images for the formal analysis. The entire capturing process is fully automated, and the panoramic view permits efficient image capturing for the reliable identification of ramie and cotton mixed fibers.