DocumentCode :
2014624
Title :
Application of the Haar wavelet to mura detection for polarizer
Author :
Chi-Cheng Cheng ; Hsu-Ming Jao
Author_Institution :
Dept. of Mech. & Electro-Mech. Eng., Nat. Sun Yat-Sen Univ., Kaohsiung, Taiwan
fYear :
2013
fDate :
25-28 Feb. 2013
Firstpage :
1080
Lastpage :
1085
Abstract :
“Mura” is a local lightness variation without a clear contour on a uniform surface image, which imparts an unpleasant sensation to human vision. There exist three common types of mura including line mura, band mura, and area mura. In this study, the Haar wavelet transform method is applied to detect muras on the polarizer. Because of its capability of multi-resolution analysis for digital images, different sized muras at different locations can be segmented. Since the Haar wavelet transform cannot extract all high frequency components from an image at one time, the original image is decomposed into a main image and a sub-image at the beginning. Then the wavelet transform technique is implemented to extract all high frequency components from these two images. Experiments were extensively conducted on different frequencies and sizes of muras. Experimental results show that the presented approach is able to detect all line muras, as well as most band and area muras.
Keywords :
Haar transforms; image resolution; image segmentation; optical polarisers; vision; wavelet transforms; Haar wavelet transform; area mura; band mura; digital images; high frequency components; human vision; image decomposition; line mura; local lightness variation; multiresolution analysis; mura detection; polarizer; uniform surface image; Fabrication; Films; Filtering; Inspection; Wavelet transforms; Haar transform; mura detection; polarizer; wavelet transform;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Industrial Technology (ICIT), 2013 IEEE International Conference on
Conference_Location :
Cape Town
Print_ISBN :
978-1-4673-4567-5
Electronic_ISBN :
978-1-4673-4568-2
Type :
conf
DOI :
10.1109/ICIT.2013.6505822
Filename :
6505822
Link To Document :
بازگشت