Repository logo
 

Native Vegetation Classification Using Remote Sensing Techniques: A Case Study of Dairy Flat Regrowth Bush by Using the AUT Unmanned Aerial Vehicle

Files

Whole thesis(23.73 MB)

Date

2014

Authors

Supervisor

Breen, Barbara

Item type

Thesis

Degree name

Master of Applied Science

Journal Title

Journal ISSN

Volume Title

Publisher

Auckland University of Technology

Abstract

Traditional field-based methods of habitat mapping to determine and classify vegetation on private land have been proven unsatisfying in terms of coverage, and time and cost-effectiveness. Remote sensing using Unmanned Aerial Vehicles (UAVs) is a new technology which is able to acquire land resources and environmental gradients as well as other spatial information. Although research using UAV techniques has been active since the beginning of the 21st century in New Zealand, it still has tremendous potential value for further and deeper exploration of UAV use. There has been, to date, little academic research based on UAVs’ remote sensing apart from commercial and military use. The aim of this study was to develop effective UAV-based remote sensing methods to classify native New Zealand vegetation on private land using an easily accessible area of regenerating bush. Results of this research provide a systematic method for UAV remote sensing classification. The object-based maximum likelihood supervised classification produced the most accurate classification result of approximately 80% using the true colour imagery mosaic. The results of this thesis suggest that the UAV remote sensing technique is capable of acquiring sufficiently high quality data from private land that can be used to mosaic and produce accurate vegetation classification at a species level.

Description

Keywords

Remote sensing, UAV, Vegetation classification, Supervised classification, Object based classification, Pix4D

Source

DOI

Publisher's version

Rights statement

Permanent link

https://hdl.handle.net/10292/8430

Collections

Masters Theses