Locating emergent trees in a tropical rainforest using data from an Unmanned Aerial Vehicle (UAV)

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaperJournal articleResearchpeer-review

  • Cici Alexander
  • Amanda H. Korstjens, Bournemouth Univ, Bournemouth University, Dept Life & Environm Sci
  • ,
  • Emma Hankinson, Bournemouth Univ, Bournemouth University, Dept Life & Environm Sci
  • ,
  • Graham Usher, PanEco Fdn, Sumatran Orangutan Conservat Programme
  • ,
  • Nathan Harrison, Bournemouth Univ, Bournemouth University, Dept Life & Environm Sci
  • ,
  • Matthew G. Nowak, Southern Illinois Univ, Southern Illinois University System, Southern Illinois University, Dept Anthropol
  • ,
  • Abdullah Abdullah, Syiah Kuala Univ, Universitas Syiah Kuala, Dept Biol
  • ,
  • Serge A. Wich, Univ Amsterdam, University of Amsterdam, Inst Biodivers & Ecosyst Dynam
  • ,
  • Ross A. Hill, Bournemouth Univ, Bournemouth University, Dept Life & Environm Sci

Emergent trees, which are taller than surrounding trees with exposed crowns, provide crucial services to several rainforest species especially to endangered primates such as gibbons and siamangs (Hylobatidae). Hylobatids show a preference for emergent trees as sleeping sites and for vocal displays, however, they are under threat from both habitat modifications and the impacts of climate change. Traditional plot-based ground surveys have limitations in detecting and mapping emergent trees across a landscape, especially in dense tropical forests. In this study, a method is developed to detect emergent trees in a tropical rainforest in Sumatra, Indonesia, using a photogrammetric point cloud derived from RGB images collected using an Unmanned Aerial Vehicle (UAV). If a treetop, identified as a local maximum in a Digital Surface Model generated from the point cloud, was higher than the surrounding treetops (Trees_EM), and its crown was exposed above its neighbours (Trees_SL; assessed using slope and circularity measures), it was identified as an emergent tree, which might therefore be selected preferentially as a sleeping tree by hylobatids. A total of 54 out of 63 trees were classified as emergent by the developed algorithm and in the field. The algorithm is based on relative height rather than canopy height (due to a lack of terrain data in photogrammetric point clouds in a rainforest environment), which makes it equally applicable to photogrammetric and airborne laser scanning point cloud data.

Original languageEnglish
JournalInternational Journal of Applied Earth Observation and Geoinformation
Pages (from-to)86-90
Number of pages5
Publication statusPublished - Oct 2018

    Research areas

  • Habitat mapping, Drones, Point cloud, Sleeping trees, Conservation, Rainforest, Sumatra, SLEEPING SITES, CANOPY, SELECTION, GIBBONS, SYSTEM, MODEL

See relations at Aarhus University Citationformats

ID: 135339473