Bubble Defects Inspection on LED Sealing Glue Images

• Main Authors: Yi-Feng Lin, 林宜鋒
• Other Authors: Chuan-Yu Chang
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/97676775986568405068

碩士 === 國立雲林科技大學 === 資訊工程系碩士班 === 100 === Bubble defect inspection is an important step in light-emitting diode (LED) packaging processes. It is difficult to detect bubble defects because bubbles are transparent and have irregular shape in LED sealing glue images. Traditionally, experienced engineers visually check defect regions and exclude defective elements. However, the inspection processes are costly and time-consuming. This paper proposes an automatic bubble defect inspection method for LED sealing glue images. The system automatic locates the sealing glue region. A decision tree is adopted to outline regions of interest (ROI), and many textural and geometric features are extracted from the ROIs. A support-vector-machine (SVM) is used to classify the ROIs as acceptable or defective. Experimental results demonstrate effectiveness of the proposed approach.