Cardiff University | Prifysgol Caerdydd ORCA
Online Research @ Cardiff 
WelshClear Cookie - decide language by browser settings

Ship wake detection in X-band SAR images using sparse GMC regularization

Karakuş, Oktay and Achim, Alin 2019. Ship wake detection in X-band SAR images using sparse GMC regularization. Presented at: 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Brighton, England, 12-17 May 2019. ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, pp. 2182-2186. 10.1109/ICASSP.2019.8683489

Full text not available from this repository.

Abstract

Ship wakes have crucial importance in the analysis of SAR images of the sea surface due to the information they carry about vessels. Since ship wakes mostly appear as lines in SAR images, line detection methods have been widely used for their identification. In the literature, common practice for detecting ship wakes is to use Hough and Radon transforms in which bright (dark) lines appear as peaks (troughs) points. In this paper, the ship wake detection problem is addressed as a Radon transform based inverse problem with a sparse non-convex generalized minimax concave (GMC) regularization. Despite being a non-convex regularizer, the GMC penalty enforces the cost function to be convex. The solution to this convex cost function optimisation is obtained in a Bayesian formulation and the lines are recovered as maximum a posteriori (MAP) point estimates with a sparse GMC based prior. The detection procedure consists of a restricted area search in the Radon domain and the validation of candidate wakes. The performance of the proposed method is demonstrated in TerraSAR-X images of five different ships and with a total of 19 visible ship wakes. The results show a successful detection performance of up to 84% for the utilised images.

Item Type: Conference or Workshop Item (Paper)
Date Type: Published Online
Status: Published
Schools: Computer Science & Informatics
Publisher: IEEE
ISBN: 9781479981311
Last Modified: 15 Nov 2021 17:00
URI: https://orca.cardiff.ac.uk/id/eprint/145187

Citation Data

Cited 8 times in Scopus. View in Scopus. Powered By Scopus® Data

Actions (repository staff only)

Edit Item Edit Item