9522943 Xu Fast and objective determination of the basic properties of fibers like: length, diameter, maturity, medullation, color and yield is a way to improve the quality of finished textiles and as a way of providing feedback in development of natural fibers. This project will use image processing technology and newly proposed algorithms to solve this technical problem. Considering preliminary research in this domain, new image processing methods of segmentation are needed to measure fiber cross-sections. New quantitative characterizations that are important to fibers need to be derived from the enhanced image, such as crimp or the waviness of fibers, trash content and the hollowness of natural fibers. Cluster analysis will be investigated for classification into different fiber grades. The long term impact of this research will be a measurement tool that can be used to improve fiber development, textile grading, development of automated textile processing, and ultimately world trade. ***
