用样条插值算法实现集成块引脚的高精度检测

Improving Precision of IC Chip Measurement Based on Spline Interpolation

灰度图像边缘检测的精度受到 CCD摄像机的采样频率的限制 ,且 CCD为一衍射受限系统 ,灰度图像的边缘变得模糊。因此 ,亚像素精度的算法在高精度的边缘检测中受到重视。在此 ,首先对样条插值法和最小二乘法进行比较和分析 ,得出了用样条插值法进行边缘检测的合理性。将该算法应用于集成块引脚的边缘检测 ,实验表明 ,该算法的重复性很好 ,分辨率可以达到四十分之一个像素 ,能够达到工业检测的标准。

The precision of edge inspection in graylevel image is limited by the sampling resolution of CCD camera, and CCD is a diffractive and confined system, the edge of graylevel image is becoming vague. So subpixel algorithms are at a premium in higher precision measurement. Firstly this paper analysis and compares the algorithms of spline interpolation and the least-square method, it proposes that spline interpolation is more reasonable when applying in the measurement of an edge. Using this algorithm in the edge measurement of IC chip, the results of the experiments show that this algorithm has good repetitions, its resolution can reach one fortieth pixel, and can fulfil the criterion of industry inspection.