Title: Comparative analysis on SIFT features in visible and infrared aerial imaging

Authors: Xiaodong Li; Nabil Aouf; Mark Richardson

Addresses: Department of Informatics and Systems Engineering, Cranfield University, Shrivenham, SN6 8LA, UK ' Department of Informatics and Systems Engineering, Cranfield University, Shrivenham, SN6 8LA, UK ' Department of Informatics and Systems Engineering, Cranfield University, Shrivenham, SN6 8LA, UK

Abstract: This paper addresses an in-depth analysis on the features extracted from both visual and infrared images in characterising natural environments for vision-based airborne navigation problems, in particular, the performance of feature extraction algorithms and matching when variant SIFT feature detection and matching techniques were utilised successfully. The application of both feature extraction and matching methods on various characteristic aerial images were investigated based on the performance metrics including processing time, number of matched features and matching rates without/with RANSAC outlier rejection applied, etc., and relevant discussion and conclusion were given.

Keywords: affine SIFT; scale invariant feature transform; ASIFT; speeded-up robust features; SURF; SIFT features; visible aerial imaging; infrared aerial imaging; random sample consensus; RANSAC; feature extraction; vision; airborne navigation; feature matching; aerial images.

DOI: 10.1504/IJAPR.2014.065767

International Journal of Applied Pattern Recognition, 2014 Vol.1 No.3, pp.273 - 297

Received: 08 Apr 2013
Accepted: 13 Aug 2013

Published online: 17 Nov 2014 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article