Multispectral imagery from UAVs to classify and monitor vegetation change in semi-natural grasslands

Type Conference Proceeding (Non-Journal item)
Original languageEnglish
Title of host publicationSustainable meat and milk production from grasslands
Subtitle of host publication27th General Meeting of the European Grassland Federation
EditorsB. Horan, D. Hennessey, M. O'Donovan, E. Kennedy, B. McCarthy, J. A. Finn, B. O'Brien
Place of PublicationFermoy
PublisherTeagasc
ISBN (Print)9781841706436
Publication statusPublished - 2018
Event27th EGF General Meeting: Sustainable meat and milk production from grasslands - Cork, Ireland
Duration: 17 Jun 201821 Jun 2018
Conference number: 27

Publication series

NameGrassland Science in Europe
Volume23

Conference

Conference27th EGF General Meeting
CountryIreland
CityCork
Period17 Jun 201821 Jun 2018
View graph of relations
Citation formats

Abstract

The use of multispectral imagery achieved from fixed-wing unmanned aerial vehicles (UAVs) has the capability to deliver high resolution imagery that can be used to monitor vegetative composition and change in semi-natural grasslands. Using a multispectral camera mounted on a fixed wing UAV with autonomous flight capability, this study is classifying and monitoring vegetation at three sites, each >70 ha, in mid-Wales, across a three-year period. Acquired imagery is stitched and georeferenced orthomosaics produced. Vegetation classification is achieved through a supervised pixel-based classification using the random forest classifier, accessed from the open-source python interface software package RSGISLib. The relatively low cost of this method has the potential to deliver a wide range of environmental and agricultural services on differing types of grasslands and wider vegetative communities; whether it be increasing the accuracy and speed of habitat surveys in categorising vegetation (compared with e.g. Joint Nature Conservation Committee Phase 1 surveying), understanding how different livestock grazing species utilise different areas and patches of vegetation, or investigating how management techniques such as grazing, cutting or burning alter the structure of a vegetative community.

Keywords

  • remote sensing, classification, vegetation composition, surveying, imagery