Registering Particle Data Sets Using a Rotation and Translation Invariant Nearest-Neighbor Algorithm.

Abstract

It can be useful to register (or align) two sets of particle data measured from the same physical sample. However, if the two data sets were collected at different translational or rotational offsets, finding the optimal registration can be a challenge. We will present an algorithm that efficiently determines the rotation and translational offset that best registers (in a least-squares sense) the corresponding particles in two or more data sets measured from the same sample. This algorithm can be used to merge two data sets that have been collected on overlapping but otherwise distinct regions on the sample. Alternatively, it can be used to overlay data sets that have been collected on the same sample area to compare replicate data for quality control and measurement efficiency purposes.