Skip to main content

Iterative Corner Extraction and Matching for Mosaic Construction

Alkaabi, Sultan, Deravi, Farzin (2005) Iterative Corner Extraction and Matching for Mosaic Construction. In: Proceedings of the Second Canadian Conference on Computer and Robot Vision (CRV’05). . pp. 468-475. (The full text of this publication is not currently available from this repository. You may be able to access a copy if URLs are provided) (KAR id:8899)

The full text of this publication is not currently available from this repository. You may be able to access a copy if URLs are provided.

Abstract

A rapid and automatic iterative corner extraction

and matching for 2D mosaic construction is presented.

This new system progressively estimates the geometric

transformation parameters between two misaligned

images. It combines corner extraction, matching, and

transformation parameters estimation into an iterative

scheme. By aligning the images in successive iterations,

accuracy improves significantly. The accurately

aligned images are used to re-extract new features,

which are subsequently matched to select

correspondences used to estimate a transformation with

n-degrees of freedom. The false correspondences are

suppressed progressively to achieve an accurate

transformation estimate. The system is used to construct

a mosaic from two misaligned images. The

performance of the system is demonstrated

experimentally using various images of differing

complexity.

Item Type: Conference or workshop item (Paper)
Subjects: T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing
Divisions: Divisions > Division of Computing, Engineering and Mathematical Sciences > School of Engineering and Digital Arts
Depositing User: Yiqing Liang
Date Deposited: 15 Aug 2009 18:16 UTC
Last Modified: 16 Nov 2021 09:46 UTC
Resource URI: https://kar.kent.ac.uk/id/eprint/8899 (The current URI for this page, for reference purposes)
Deravi, Farzin: https://orcid.org/0000-0003-0885-437X
  • Depositors only (login required):