Dissertation
Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device

Research on Identification System of Cashmere and Wool Fiber

Author MengFeiFei
Tutor ShangShuYuan
School Beijing Institute of Clothing Technology
Course Mechanical and Electronic Engineering
Keywords cashmere fiber identification fuzzy image processing feature extraction wavelet transform pattern recognition
CLC TS101.921
Type Master's thesis
Year 2012
Downloads 21
Quotes 0
Download Dissertation

Cashmere fiber is a kind of rare and expensive animal fiber, which has high performance. Its production is expensive and high-grade. Thus the content of cashmere is not only affecting the performance of manufactured goods, and also closely related with the cost and the pricing of the production. Cashmere fiber processing is based on digital image processing techniques. After image processing the characteristic parameters of the fiber are extracted by some extraction method. At last, research the method of the classify recognition for cashmere fiber. The main work is using a large number of cashmere fiber images to do digital image processing, extract and measure characteristic parameters, and do data analysis, etc. The main contributions are given as following:A method of image restoration for fuzzy image is researched. This method is used to restore a lot of the fuzzy cashmere image, which called blind deconvolution algorithm. Cashmere fiber processing is using digital image processing techniques. By image processing, the gray and binary images can be presented. The methods of extraction for characteristic parameters are taken to study. For characteristic parameters with many kinds, such as the diameter, height, thickness, density, area, perimeter, ratio of diameter and height, the scales square factor, inside square factor, relative perimeter and relative area. All these parameters are extracted and saved in the database. The parameter information in the database can be picked up. The method of BP neural network is presented to make the analysis and research the kind of the fiber from the database data. Results of the simulation have made the error analysis. The cashmere fiber identification system is designed based the former analysis.

Related Dissertations
More Dissertations