Title :
Welding inspection with image processing
Author :
Murosaki, Takashi ; Naoki, Takaha ; Takeda, Syuji ; Hayashiguchi, Akihiro ; Sugiura, Fumiaki
Author_Institution :
Machinery&Tools Dept., DENSO Corp., Aichi, Japan
Abstract :
We studied a welding inspection for alternator´s rectifier with image processing. The point is how to apply the SVM method to the area and contrast of the welding extracted by the difference of good work-pieces. Usually, the judgment boundary of Mahalanobis Distance is assumed to be a midway line from the good one´s distribution and the defective one´s distribution. But, the boundary has a problem that misjudged defective one to the good one. To solve this, we applied SVM (Support Vector Machine) with the process capability index.
Keywords :
alternators; arc welding; automatic optical inspection; image processing; learning (artificial intelligence); rectifiers; support vector machines; SVM method; alternator rectifier welding inspection; image processing; learning machine; mahalanobis distance; support vector machine; tungsten inert gas welding; Alternators; Flowcharts; Image processing; Inspection; Kernel; Rectifiers; Supervised learning; Support vector machine classification; Support vector machines; Welding; CP; Image Processing; SVM; Welding Inspection;
Conference_Titel :
ICCAS-SICE, 2009
Conference_Location :
Fukuoka
Print_ISBN :
978-4-907764-34-0
Electronic_ISBN :
978-4-907764-33-3
