Tracking of objects in motion-distorted scanning electron microscope images

Microrobotics is a field of high interest in the recent past. Commonly used imaging modalities in this field include optical microscopes and scanning electron microscopes (SEM). Due to the scanning principle of the SEMs, objects in motion may not be represented in the SEM images the same way as they really are. Movement leads to movement artifacts. This effect strongly limits the possibilities for the use of image based methods for tracking and recognition and it also limits the possible speeds achievable for automated procedures. In this work, a tracking algorithm is proposed which is robust to artifacts and distortions generated by object motion. The algorithm is tested and validated in experiments.