Study on detection of boats using satellite imagery for use on unmanned aerial vehicles

Sie sind hier: Homepage > Suche

Study on detection of boats using satellite imagery for use on unmanned aerial vehicles

da Silva, B. R. F. / Nogueira, M. B. / Alsina, P. J. / de Albuquerque, G. L. A. / Dantas, J. B. D. / de Medeiros, A. A. D. / Santiago, G. Santos

Abstract

This work presents the development of a system that performs the detection of boats from aerial images acquired by satellites. This will serve as a knowledge base for the implementation of a system to be implemented in a fleet of unmanned aerial vehicle. The main purpose of the system is to detect ships that are in the risk zone of rocket trajectory launched from Barreira do Inferno Launch Center — CLBI. In previous work the authors proposed an algorithm to perform boat detection, however the results showed a high incidence of false positives. In order to improve those results, we propose the use of the Histogram of Oriented Gradients descriptor in candidate images followed by machine learning teachniques such as Suport Vector Machine and K Nearest Neighbours to classify them. To validate the system some experimental results are shown using satellite images, since the aerial vehicle is under constuctions and there is no image database yet.

Zugriff

Zugriff prüfen

Verfügbarkeit in meiner Bibliothek prüfen

Bestellung bei Subito €

Exportieren, teilen und zitieren

Dokumentinformationen

Titel :

Study on detection of boats using satellite imagery for use on unmanned aerial vehicles

Beteiligte:

da Silva, B. R. F. (Autor:in) / Nogueira, M. B. (Autor:in) / Alsina, P. J. (Autor:in) / de Albuquerque, G. L. A. (Autor:in) / Dantas, J. B. D. (Autor:in) / de Medeiros, A. A. D. (Autor:in) / Santiago, G. Santos (Autor:in)

Erschienen in:

2017 Latin American Robotics Symposium (LARS) and 2017 Brazilian Symposium on Robotics (SBR) ; 1-5

Erscheinungsdatum :

2017-11-01

Format / Umfang :

474595 byte

ISBN :

978-1-5386-0956-9

DOI :

https://doi.org/10.1109/SBR-LARS-R.2017.8215300

Medientyp :

Aufsatz (Konferenz)