A Study on Surface Defect Inspection for Salar cells

• Main Authors: Ruo-mian Weng, 翁若綿
• Other Authors: Jyh-bin Suen
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/55327352131660327585

碩士 === 義守大學 === 工業工程與管理學系碩士班 === 97 === In this paper, we propose a surface defect inspection method by using the machine vision and image processing techniques for poly-silicon solar cells. The objective is to perform quality assurance with high inspection rate and reduce manufacturing cost. The defects to be detected include edge cracks, finger interruptions, and stains. The solar cell images are grabbed by a scanner, then we use Fourier transformation (FT) and morphology to orientate solar cells image. The edge crack defect can be segmented by projection. For the printed lines, we use gray-scale morphology to remove vertical busbars, and find finger interruptions by image subtraction and thresholding. The stain defects are detected by gray-scale morphology and image subtraction. We also compare the effect of using FT on stain defect detection rate. Experimental results indicate the proposed orientation and defect detection methods perform well. The detection rate of stain defect is higher when we utilize Fourier transformation.