IC真实缺陷的边界提取和缺陷尺寸与形状的表征

为了对ＩＣ制造中真实多余物缺陷进行分类与识别，ＩＣ多余物缺陷的特征提取是非常重要的一步，文中首先提出一种基于数学形态学的ＩＣ真实多余物缺陷边界的检测方法，其次对边界进行了链码描述，最后对边界所表示的多余物缺陷进行了尺寸测量与形状分析，在预处理阶段，利用彩色ＨＳＶ模型分割原ＩＣ图像，然后用图形学开运算消除背景噪音，对开后的结果图像进行形态膨胀及形态腐蚀运算，消除多余物缺陷中的小洞噪音以获得从复杂背景

Boundary Extraction Based on Morphology and Size-Shape Measurement for the IC Real Defect

Extracting the extra materia defect features from an IC real image is very important to the classification and recognition of the defect in the IC real manufacturing process. In this paper, a boundary detecting method based on mathematical morphology for IC real extra material defects is presented at first. Secondly, the boundary chain code of the extra material defect is generated from the IC boundary image. Finally, several size and shape features are detected so that the IC extra material defect can be identified. In preprocessing phase, the IC image is segmented using HSV color model and the background noises are removed from the IC noisy image by the morphological operating. This step is followed by morphological eroding and dilating for filling holes in the segmented IC image obtain the isolation and extraction of IC extra material defects from a complex scene. Once isolated, the extra material defect boundary can be examined by using morphology operators. The group size features (those are area, perimeter, object core, width and length) and the shape feature (those are called rectangularity and circularity measures) are detected from the boundary image. Experimental result shows that proposed methods make the shape analyzing of IC extra material defects simple and detectable.