Hough变换在表面污渍检测中的应用

Using Hough Transform in Surface Defect Detection of Chip

针对具体芯片图像的形状特征,提出了基于改进Hough变换的几何配准算法。该算法利用边缘点的梯度方向信息,以及通过圆上任意一点并沿着该点法线方向的直线必定经过该圆的圆心这一几何性质,将三维Hough变换空间简化为二维空间,较好地解决了传统广义Hough变换运算时间长、存储空间大的问题。实验表明,可以有效地对两幅图像进行配准,从而检测出芯片表面的污渍。

According to the shape of the specific chip image, a new geometric image registration algorithm based on improved Hough transform is presented. By using the gradient direction information of edges and the specific geometric property of circle, the three-dimensional Hough transform space is reduced to a two-dimensional one, thus the problems such as long computational time and large storage in common general Hough transform can be solved. Experiments show that the algorithm is effective to match the two images for surface defect detection.