• DocumentCode
    2897929
  • Title

    High-speed vision challenge for 100% online on-Strip inspection

  • Author

    Wanliang, Chen ; Rong, Xiang Yong ; Jianhua, Ruan

  • Author_Institution
    Leshan Phoenix Semicond. Co. Ltd., Leshan, China
  • fYear
    2010
  • fDate
    Nov. 30 2010-Dec. 2 2010
  • Firstpage
    1
  • Lastpage
    6
  • Abstract
    The paper presents the challenges of effective high-speed detection for screen out incomplete fill (IF) reject via online Vision System at Trim/Form process. In general, most of the assembly factories screen the incomplete fill reject manually or through a vision on Test Handler. The speed of this Unit by Unit inspection is very low. Usually, the effective UPH is less than 30k. Accordingly, a new idea of online Strip by Strip inspection is bring out. By this way, total 36 units (or even more) can be detected every step. And the inspection UPH can reach to more than 500k. To achieve this target, the biggest challenge comes from that large image data/information need to be processed in tens of microseconds. As presented in this paper, based on advanced algorithms, excellent program structure and optimized hardware composition, the system can realize 100% online unit inspection with only 1 camera at a Trim/Form machine. The speed of this vision can freely match with any high-speed running Trim/Form machine, which will gain competitive advantages.
  • Keywords
    assembling; computer vision; electronics industry; inspection; production engineering computing; assembly factory; form process; high-speed detection; high-speed vision; online on-strip inspection; online vision system; screen out incomplete fill reject detection; test handler; trim process; Algorithm design and analysis; Image edge detection; Image resolution; Lighting; Manuals; Pixel;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Electronic Manufacturing Technology Symposium (IEMT), 2010 34th IEEE/CPMT International
  • Conference_Location
    Melaka
  • ISSN
    1089-8190
  • Print_ISBN
    978-1-4244-8825-4
  • Type

    conf

  • DOI
    10.1109/IEMT.2010.5746670
  • Filename
    5746670