基於影像處理技術的表面缺陷檢測系統

Abstract

我們建立了一個用於晶片表面缺陷的檢測系統.在這個報告中所使用到的 影像分析技術包括:影像切割(Segmentation),二值化(Binary Thresholding),圖樣匹配(Template Matching)及投影法(Projection) .這個系統能夠有效辨認三種主要表面缺陷:邊緣區域的刮傷(Scratch) ,作用區上的污點(Spot)及鋁墊區域上的異常覆蓋物(Coating).實驗結果 顯示在邊緣區域的刮傷辨識率是87%,在作用區域上的污點辨識率是 94%,在鋁墊上的異常覆蓋物辨識率是99.66%. A surface defect inspection system is built up to inspect chip surface defect of silicon wafers. In this report, the techniques of the image analysis include segmentation, binary thresholding, template matching, and projection. The major defects in three regions of chip surface are: scratch in edge region, spots in active region, and dark coating in pad region. They are effectively detected by the designed defect detection modules in the proposed system. Experimental results show that the recognition rate of the proposed detection method for scratch defect in edge region is 87%, for spot defect in active region is 94%, and for coating defect in pad region is 99.66%.