KOPRI Repository

Object-based detecting and counting penguins in unmanned aerial vehicle images

Metadata Downloads
Title
Object-based detecting and counting penguins in unmanned aerial vehicle images
Authors
Hyun, Chang-Uk
Kim, Hyun-cheol
Hong, Soon Gyu
Kim, Jeong-Hoon
Subject
UAV; Object-based image analysis; Antarctica
Issue Date
2016
Citation
Hyun, Chang-Uk., et al. 2016. Object-based detecting and counting penguins in unmanned aerial vehicle images. International Symposium on Remote Sensing 2016, JeJu. 2016.04.20-22.
Abstract
This study presents an application of unmanned aerial vehicle (UAV) images to detect and count penguins in Barton Peninsula, King George Island, Antarctica. The area around Nar?bski Point located on the southeast coast of the Barton Peninsula was designated as Antarctic Specially Protected Area No. 171 (ASPA 171), and Chinstrap and Gentoo penguins inhabit in the area. It is important to fulfill periodic wildlife monitoring in the ASPA 171, because variations of its population reflect climate and environmental changes and disturbance from human activities. In this study individual penguin was detected and counted in a part of the Barton Peninsula using images obtained from UAV and object-based image analysis technique. About 1,500 images acquired in January 2014, and about 360 images were selected and mosaicked as an image of about 3 cm spatial resolution. A subset of the mosaicked image including representative rookeries was segmented based on gradient of pixel values, and spectral and spatial attributes were assigned to each segment. The segments indicating individual penguins were effectively detected from adjacent ground segments with high contrasts in spectral and spatial attributes. It is expected that the very-high-resolution UAV image analysis approach can be applied to other wildlife habitats in Antarctica, and the results can be used to establish effective periodic wildlife monitoring plans.
Conference Name
International Symposium on Remote Sensing 2016
Conference Place
JeJu
Conference Date
2016.04.20-22.
Files in This Item
General Conditions
      ROMEO Green
    Can archive pre-print and post-print or publisher's version/PDF
      ROMEO Blue
    Can archive post-print (ie final draft post-refereeing) or publisher's version/PDF
      ROMEO Yellow
    Can archive pre-print (ie pre-refereeing)
      ROMEO White
    Archiving not formally supported

    qrcode

    Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

    Browse