Title :
The application of directional wavelets in multiscale representation of pulp fibre image
Author :
Hou, Bei-Ping ; Wen Zhu
Author_Institution :
Inst. of Ind. Process Control, Zhejiang Univ., Hangzhou, China
Abstract :
This work proposes a novel algorithm to decompose fibre image and extract its edge characteristics. Because the characteristics of the paper fibre such as length and width play an important role in papermaking industry, so how to measure its related characteristics is important to improve the paper quality and production. An online fibre analyzer is designed based on the computer vision theory, and the distribution structure of the online fibres on image presents the multi-directional properties, so the multi-direction and multi-scale algorithm is proposed, and the directional wavelet transform is applied to the analysis of fibre image. Our analysis shows that directional wavelet transform can better reflect the edge information of images, because direction and texture information of fibre image can be better expressed. The experiment proves that the directional characterizations of fibre image can be extracted effectively.
Keywords :
computer vision; edge detection; fibres; image representation; image texture; paper industry; paper pulp; wavelet transforms; computer vision theory; directional wavelet transforms; edge characteristics extraction; fibre distribution structure; fibre image decomposition; image edge information; image extraction; image texture; multidirectional algorithm; multiscale algorithm; online fibre analyzer; paper production; paper quality; papermaking industry; pulp fibre image analysis; Algorithm design and analysis; Computer vision; Image analysis; Image texture analysis; Length measurement; Optical fiber theory; Production; Pulp and paper industry; Wavelet analysis; Wavelet transforms;
Conference_Titel :
Machine Learning and Cybernetics, 2004. Proceedings of 2004 International Conference on
Print_ISBN :
0-7803-8403-2
DOI :
10.1109/ICMLC.2004.1384595
