Inshore ship detection in multispectral satellite images

Author

Besbinar, Beril ; Gurbuz, Yeti Ziya ; Alatan, A. Aydin

Author_Institution

ODTU Elektrik ve Elektron. Muhendisligi Bolumu, Ankara, Turkey

fYear

2015

fDate

16-19 May 2015

Firstpage

2029

Lastpage

2032

Abstract

In this paper, a novel method is proposed for the detection of sea and inshore ships in satellite images which contain red, green, blue and near-infrared bands. For sea detection, an initial mask is obtained by thresholding the water index calculated using green and near infrared bands. At the second stage, models of sea and land learned via the initial mask are used in graph-cut method and a sea-land mask with high accuracy is obtained. Linear extension of the line segments on the sea-land mask boundary is aimed for the detection of inshore ships.

Keywords

geophysical image processing; graph theory; image colour analysis; image segmentation; infrared imaging; masks; oceanographic techniques; ships; blue band; graph-cut method; green band; image segmentation; inshore ship detection; multispectral satellite imaging; near-infrared band; red band; sea detection; sea-land mask; water index thresholding calculation; Ground penetrating radar; Image segmentation; Integrated optics; Marine vehicles; Optical imaging; Optical sensors; Satellites; graph cut; line segment merging; multispectral satellite images; remote sensing; sea detection; ship detection;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing and Communications Applications Conference (SIU), 2015 23th

Conference_Location

Malatya

Type

conf

DOI

10.1109/SIU.2015.7130264

Filename

7130264