Robust Phase-Correlation Based Registration of Airborne Videos Using Motion Estimation

Abstract

This paper presents an algorithm for near-real time registration of airborne video sequences with reference images from a different sensor type. Phase-correlation using Fourier-Melin Invariant (FMI) descriptors allow to retrieve the rigid transformation parameters in a fast and non-iterative way. The robustness to multi-sources images is obtained by an enhanced image representation based on the gradient norm and by the extrapolation of registration parameters by motion estimation between frames. A phase-correlation score, indicator of the registration quality, is introduced to regulate between frame-toreference image registration and extrapolation from previous frames only. Our Robust Phase-Correlation based registration algorithm using Motion Estimation (RPCME) is compared with a Mutual Information (MI) algorithm for the registration of two different panchromatic airborne videos with Geoeye reference images. The RPCME algorithm registered most of the frames accurately, retrieving much better orientation than MI. It shows robustness and good accuracy to multisource images with the advantage of being a direct (non-iterative) method.

Illumination-invariant image matching for autonomous UAV localisation based on optical sensing

“…Phase Correlation has been applied for airborne video motion estimation by De Morsier et al (2013), who generated gradient image pairs to cope with illumination variant problem. However, this approach mainly focuses on global illumination variation, caused by the change in solar illumination intensity.…”