Title :
Automatic extraction of component inspection regions from Printed Circuit Board
Author :
Jun-Oh Kim ; Youn-Ae Lee ; Tae-Hyoung Park
Author_Institution :
Mirae Co., Cheonan, South Korea
Abstract :
The inspection machine in PCB(Printed Circuit Board) assembly line checks for assembly errors by examining the images inside the component inspection region. The component inspection region consists of a component package region and a soldering region. The regions need to be extracted automatically so the auto-teaching system of the inspection machine can work. An image segmentation method was proposed to extract the component inspection regions automatically from the images of the PCB. The acquired image was transformed to HSI color model, and then segmented by several regions by a clustering method. A modified K-means algorithm was developed to increase the accuracy of extraction. The heuristics generating the initial clusters and merging the final clusters were newly proposed. The vertical and horizontal projection was also developed to distinguish the region of the component package and the region of soldering. The experimental results are presented to highlight the utility of the proposed method.
Keywords :
image colour analysis; image segmentation; inspection; microassembling; pattern clustering; printed circuit manufacture; production engineering computing; HSI color model; PCB assembly line checks; auto-teaching system; automatic extraction; component inspection regions; horizontal projection; image segmentation method; modified K-means algorithm; printed circuit board assembly line checks; soldering region; vertical projection; Clustering algorithms; Clustering methods; Image color analysis; Image segmentation; Inspection; Soldering; Assembly Inspection; Auto-Teaching System; Clustering; Image Segmentation; Printed Circuit Board (PCB);
Conference_Titel :
System Integration (SII), 2012 IEEE/SICE International Symposium on
Conference_Location :
Fukuoka
Print_ISBN :
978-1-4673-1496-1
DOI :
10.1109/SII.2012.6427276