Facial reconstruction

Search LJMU Research Online

Browse Repository | Browse E-Theses

EXPLORING STATISTICAL CLASSIFICATION, GIS ANALYSIS AND MAPPING OF HERBIVORE BEHAVIOURS USING ACCELEROMETERS AND HIGH ACCURACY GPS

Hartley, A (2015) EXPLORING STATISTICAL CLASSIFICATION, GIS ANALYSIS AND MAPPING OF HERBIVORE BEHAVIOURS USING ACCELEROMETERS AND HIGH ACCURACY GPS. Masters thesis, Liverpool John Moores University.

[img] Text
158032_2015hartleymphil.pdf - Published Version

Download (2MB)

Abstract

Electronic sensors equipped with accelerometers have the potential to remotely monitor and record herbivore behaviours. In the UK, sheep are a significant consumer in both managed pasture and upland ecosystems. The ability to automatically collect behavioural data could help inform research into ecosystem functioning and animal welfare. This study evaluated the placement of accelerometers and the ability of data generated to automatically classify four behaviours in sheep; grazing, standing (non grazing), lying head up and lying head down. An application of this method was used to analyse and map data in a GIS and investigate if sheep show a preference for areas with higher fructan levels in grass. Three sheep were fitted with two accelerometers each. One attached to a head halter and one centrally located across the withers by means of a dog harness. Training data were collected and discriminant function analysis was used to develop a model that could predict future unobserved behaviours. Correct classification rates of 95.2%, 91.0% and 91.8% were achieved for each sheep. In the fructan study, although no preference was detected, the study did demonstrate that data from accelerometers can be used to generate behavioural distribution maps. The use of accelerometers is a suitable method for classifying a range of behaviours in sheep.

Item Type: Thesis (Masters)
Uncontrolled Keywords: GIS, GPS. Accelerometers, Statistical classification, Sheep, Fructan
Subjects: H Social Sciences > HA Statistics
Q Science > QL Zoology
Divisions: Natural Sciences & Psychology (closed 31 Aug 19)
Date Deposited: 24 Oct 2016 15:25
Last Modified: 03 Sep 2021 23:26
DOI or ID number: 10.24377/LJMU.t.00004397
Supervisors: Sneddon, Jenny, Brown, Richard and Small, Richard
URI: https://researchonline.ljmu.ac.uk/id/eprint/4397
View Item View Item