Real-time image registration

Abstract

This paper summarizes the developed real-time algorithm for registering subsequent video frames and experiments performed on an embedded processor. The features extracted from subsequent frames are matched using a RANSAC technique and frames are registered with an affine transformation model. The techniques used in the literature are improved to work in real-time and a system that can register 320×240 resolution images at approximately 40 frames per second is developed. Experiments are performed on a BeagleBoard-XM single board computer that contains ARM and DSP processor cores.