DocumentCode :
430773
Title :
Nondestructive exported mangosteen grading by inspection of external features
Author :
Rujapha, P. ; Kosin, C.
Author_Institution :
Dept. of Comput. Eng., King Mongkut´s Univ. of Technol. Thonburi, Bangkok, Thailand
Volume :
1
fYear :
2004
fDate :
26-29 Oct. 2004
Firstpage :
465
Abstract :
In mangosteen grading for quality warranty, we have to classify mangosteen into (1) extra class, (2) class I, and (3) class II. However some factors such as color texture and flaw that indicate the class are ambiguous. This paper proposes a method of non-destructive exported mangosteen grading by inspection of external features. In this method, we employ hue feature to separate the calyx of the mangosteen, value feature to detect flaw, and red and green color to determine color texture. To evaluate the performance of proposed method, the experiments with twenty harvested mangosteen have been performed, and the results reveal approximately 95% accuracy, compared with an expert.
Keywords :
agricultural products; edge detection; feature extraction; flaw detection; food products; image classification; image colour analysis; image texture; inspection; nondestructive testing; ambiguous factors; calyx separation; color texture; external features inspection; flaw detection; green color; harvested mangosteen; hue feature; mangosteen classification; nondestructive exported mangosteen grading; quality warranty; red color; value feature; Color; Computer vision; Digital images; Electronic mail; Image edge detection; Inspection; Performance evaluation; Shape measurement; Size measurement; Visualization;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Communications and Information Technology, 2004. ISCIT 2004. IEEE International Symposium on
Print_ISBN :
0-7803-8593-4
Type :
conf
DOI :
10.1109/ISCIT.2004.1412889
Filename :
1412889
Link To Document :
بازگشت