| Sign In to gain access to subscriptions and/or personal tools. |
Automatic Identification of Ramie and Cotton Fibers Using Characteristics in Longitudinal View, Part II: Fiber Stripes AnalysisCollege of Textiles, Donghua University, Shanghai, 200051, China, nbwrw{at}163.com
Shanghai Entry-Exit Inspection and Quarantine Bureau, Shanghai, 200135, China
College of Textiles, Donghua University, Shanghai, 200051, China
Department of Human Ecology, The University of Texas at Austin, 219 Gearing Hall, Austin, TX 78712, USA Based on the analysis of stripes on fiber surfaces, a new method for cotton and ramie fiber identification is introduced in this paper. The stripes of a fiber surface were extracted by segmentation, edge detection, and thinning, and then they were orthogonally projected along the curving skeleton of the fiber. In addition, six characteristic parameters for identification were obtained, and based on the method of maximum probability, equations for identification were established on the six probability distribution curves of the characteristic parameters. Finally, weight coefficients of the equations were obtained from self-adapting identification tests. The experiments showed that the overall tolerance for false identification of cotton or ramie fiber was under 7%.
Key Words: characteristic parameters • equations for recognition • orthogonal projection • probability distribution curves • self-adapting identification test • stripes on fiber surfaces
This version was published on November
1, 2009 Textile Research Journal, Vol. 79, No. 17,
1547-1556 (2009) |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
