Analysing the influence of TNT on Southern African trees grass and shrubs using in-situ hyperspectral data

Landmines pose a significant risk to the health and livelihood of millions of people in war-torn countries. From a humanitarian point of view, these hidden dangers must be detected and removed. Several detection methods exist, including ground-penetrating radar, sniffer dogs and rats. It is a tedious process and can take months to clear only a single minefield. This study investigates whether TNT from leaking landmines can be detected using hyperspectral remote sensing of plant foliage, with the aim of being an area reduction aid. Increasing concentrations of TNT was administered to a study sample of five species of trees, five species of grasses and four species of shrubs, and leaf-clip readings were taken at regular intervals with a field spectrometer. Statistical correlation testing of seven plant health indices (red-edge position, first derivative reflectance, normalised difference water index, moisture stress index, water band index, photochemical reflectance index and nitrogen index) was done on the results of the readings. TNT has a mixed effect on the health of the tested plants, with some species displaying adverse effects of TNT on their health, while others proved to be healthier or more resilient against the effects. Results also varied in magnitude. Even in a single species, differing concentrations TNT lead to varying results. The various indices delivered varying results, with some indices delivering inconclusive results. Positive results were yielded from the REP analysis, indicating this as a possible index to use in landmine detection.

Grantor dc:publisher

University of Pretoria

Year dc:date.issued

2017

Advisor dc:contributor.advisor

• Schmitz, Peter

• UCTD

Statement dc:rights

• © 2018 University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria.

Language dc:language.iso

en

Dc Identifier Other

A2018

OAI identifier oai:identifier

oai:repository.up.ac.za:2263/65887