Development of improved image processing algorithms for an automated inspection system using line scan cameras

Line scan camera를 이용한 검사 시스템에서의 새로운 영상 처리 알고리즘

  • 장동식 (고려대학교 산업공학과) ;
  • 이만희 (고려대학교 산업공학과) ;
  • 부창완 (큐빅테크(주) CAD/CAM 연구소)
  • Published : 1997.08.01

Abstract

A real-time inspection system is developed using line scan cameras. Several improved algorithms are proposed for real-time detection of defects in this automated inspection system. The major improved algorithms include the preprocessing, the threshold decision, and the clustering algorithms. The preprocessing algorithms are for exact binarization and the threshold decision algorithm is for fast detection of defects in 1-D binary images. The clustering algorithm is also developed for fast classifying of the defects. The system is applied to PCBs(Printed Circuit Boards) inspection. The typical defects in PCBs are pits, dent, wrinkle, scratch, and black spots. The results show that most defects are detected and classified successfully.

Keywords

References

  1. Sampling Inspection and Quality Control G. Barrie
  2. Solid State Camera Product Don W. Lake
  3. 미세 결점 인식을 위한 Line Scan 시스템 개발에 관한 연구 최인걸
  4. 컴퓨터 비젼 입문(제 3장) 최형일
  5. IEEE, T-PAMI v.13 Gray level thresholding in badly illuminated images J. R. Parker
  6. Handbook of Pattern Recognition and Computer Vision Cluster analysis and related issues Richard C. Dubes
  7. Pattern Recognition v.22 no.4 An efficient sequential clustering Method P. Trahanias;E. Skordalakist
  8. 교환렌즈 기술 백과 일신 사진서적 편집실
  9. 패턴인식의 원리(제 3장) 이성환