DocumentCode :
2554963
Title :
Automatic Detection of Defects in Solar Modules: Image Processing in Detecting
Author :
Nian, Bei ; Fu, Zhizhong ; Wang, Li ; Cao, Xiaoxuan
Author_Institution :
Pattern Recognition & Intell. Syst., Univ. of Shanghai for Sci. & Technol., Shanghai, China
fYear :
2010
fDate :
23-25 Sept. 2010
Firstpage :
1
Lastpage :
4
Abstract :
Image acquisition devices which can get infrared image of solar modules is designed by using the principles of the semiconductor´s electroluminescence, and image processing is applied to the detection system which can detect the defects automatically including black pieces, fragmentation, broken grid, crack and so on. At first the defects of the infrared image are classified and then the defects´ types and locations are marked out after filtering, single-chip division, gray-scale transformation, binary, feature description and extraction, finally the results are feeded back to the database. This method increases the defects´ types (such as invisible crack) which the manual testing is difficult to identify, it also can eliminate human errors which manual testing may produce possibly and can reduce labor costs, defects´ rates, further it can improve the detection´s efficiency and productivity of production line.
Keywords :
cracks; fault diagnosis; image processing; solar cells; automatic detection; black pieces; broken grid; crack; defects; detection system; feature description; feature extraction; fragmentation; gray-scale transformation; image acquisition devices; image processing; infrared image; semiconductor electroluminescence; single-chip division; solar modules; Cameras; Computers; Conferences; Feature extraction; Image recognition; Photovoltaic cells;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Wireless Communications Networking and Mobile Computing (WiCOM), 2010 6th International Conference on
Conference_Location :
Chengdu
Print_ISBN :
978-1-4244-3708-5
Electronic_ISBN :
978-1-4244-3709-2
Type :
conf
DOI :
10.1109/WICOM.2010.5600703
Filename :
5600703
Link To Document :
بازگشت