Multi-scale Calibration Board Based Accurate Calibration of a Scanning Electron Microscope

Abstract

We present a new calibration method based on multi-scale calibration board for scanning electronic microscopy (SEM) at low magnifications. The implementation of the calibration mainly depends on automatic center-point features detection from the calibration board image instead of corner detection. The center-point feature is the centroid of each squares, which is obtained via morphological processing. We use the calibration model consisting of projective model and static distortion model. It is proved the point detection proposed in this paper is more accurate, efficient and robust, compared with corner point detection in the traditional calibration process. Then, reprojection error is used to evaluate the accuracy of the calibration results. At last, the method is efficient and reliable with accuracy, which is testified by experiment.