Fast image drift compensation in scanning electron microscope using image registration

Scanning Electron Microscope (SEM) image acquisition is mostly affected by the time varying motion of pixel positions in the consecutive images, a phenomenon called drift. In order to perform accurate measurements using SEM, it is necessary to compensate this drift in advance. Most of the existing drift compensation methods were developed using the image correlation technique. In this paper, we present an image registration-based drift compensation method, where the correction on the distorted image is performed by computing the homography, using the keypoint correspondences between the images. Four keypoint detection algorithms have been used for this work. The obtained experimental results demonstrate the method's performance and efficiency in comparison with the correlation technique.