Robust multi-sensor image registration using pixel migration

Y. Keller, A. Averbuch

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Scopus citations

Abstract

This paper presents an energy minimization approach to registration of significantly dissimilar images, acquired by sensors of different modalities. The proposed algorithm introduces a robust matching criterion by aligning the locations of gradient maxima. The alignment is formulated as a parametric variational optimization problem which is solved iteratively by considering the intensities of a single image. The locations of the maxima of the second image's gradient are used as initialization. This implicit matching criterion is achieved while utilizing the full spatial information, without the need to use invariant image representations. We were able to robustly estimate affine and projective global motions using 'coarse to fine' processing, even when the images are characterized by complex space varying intensity transformations. These cause current state-ofthe- Art algorithms to fail. Finally, we present the registration of real images, which were taken by multi-sensor and multi-modality using affine and projective motion models.

Original languageEnglish
Title of host publication2002 IEEE Sensor Array and Multichannel Signal Processing Workshop Proceedings, SAME 2002
PublisherIEEE Computer Society
Pages100-104
Number of pages5
ISBN (Electronic)0780375513
DOIs
StatePublished - 2002
EventIEEE Sensor Array and Multichannel Signal Processing Workshop, SAME 2002 - Rosslyn, United States
Duration: 4 Aug 20026 Aug 2002

Publication series

NameProceedings of the IEEE Sensor Array and Multichannel Signal Processing Workshop
Volume2002-January
ISSN (Electronic)2151-870X

Conference

ConferenceIEEE Sensor Array and Multichannel Signal Processing Workshop, SAME 2002
Country/TerritoryUnited States
CityRosslyn
Period4/08/026/08/02

Fingerprint

Dive into the research topics of 'Robust multi-sensor image registration using pixel migration'. Together they form a unique fingerprint.

Cite this