167
© Springer International Publishing AG 2017
R. Díaz-Delgado et al. (eds.), The Roles of Remote Sensing in Nature Conservation,
DOI 10.1007/978-3-319-64332-8_9
Mapping the Distribution of Understorey
Rhododendron Ponticum Using Low-Tech
Multispectral UAV Derived Imagery
Abigail Sanders
Abstract Invasive species, such as Rhododendron ponticum, are an issue of global
concern as they are a threat to biodiversity, ecosystem provisioning services and
economies and act as a reservoir for emerging pathogen dispersal. R. ponticum is a
Class A prohibited plant under Schedule 9(2) of the UK wildlife and countryside
act, and removal costs in the UK in 2010 were approximately £86 million. This
figure is not inclusive of removal costs on private land, an area where R. ponticum
is spreading rapidly, and which amounts to 74% of the UK woodland. R. ponticum
is a sporulation host which is tolerant to the highly virulent tree-killing root-fungus
pathogen Phytothphera, in particular P. ramorum and P. kernoviae. These were
introduced in 2003 and have now become widespread and are a particular threat to
species such as the English Oak (Quercus robur). Limited access and high costs of
aerial and satellite imagery restrict the progress of research into the spatial distribution patterns of R. ponticum and thus their effective removal strategy. Unmanned
Aerial Vehicles (UAV) are currently used for conservation and agriculture but are as
yet prohibitively expensive. Using commercially available action cams (Sony and
Mobius: open source camera adapted for Infrared detection using a Rosco Red
Filter) and a DJI Phantom 3 professional quadcopter a multispectral image composed of 7 bands (440–800 nm range) was generated for sites containing invasive
rhododendron and other non-target species. Mosaicking efficiency reduced the
integrity of images but was rectified through georeferencing and Polynomial (spline)
transformation. R. ponticum was shown in pairwise comparisons with non-target
species to show greatest separability at 540 (p < 0.05) and least at 550 (p > 0.05) in
multivariate analysis. Intra-species difference between two R. ponticum populations
growing in differing locations was also significant at p = 0.01 in the same pairwise
comparison. Ivy was not significantly different to larch and R. ponticum in all but
the 540 and 750 bandwidths (p > 0.01). Seven band spectral signatures were all
shown to be significantly different at F(1, 42) = 6.795, p < 0.001. This was confirmed during supervised classifications of the composite spectral checkerboard,
A. Sanders (*)
Computing, Engineering and Science Department. Glyntaff, Trefforest Campus, University of
South Wales, South Glamorgan, UK
e-mail: abigailsanders1234@gmail.com
© Springer International Publishing AG 2017
R. Díaz-Delgado et al. (eds.), The Roles of Remote Sensing in Nature Conservation,
DOI 10.1007/978-3-319-64332-8_9
Mapping the Distribution of Understorey
Rhododendron Ponticum Using Low-Tech
Multispectral UAV Derived Imagery
Abigail Sanders
Abstract Invasive species, such as Rhododendron ponticum, are an issue of global
concern as they are a threat to biodiversity, ecosystem provisioning services and
economies and act as a reservoir for emerging pathogen dispersal. R. ponticum is a
Class A prohibited plant under Schedule 9(2) of the UK wildlife and countryside
act, and removal costs in the UK in 2010 were approximately £86 million. This
figure is not inclusive of removal costs on private land, an area where R. ponticum
is spreading rapidly, and which amounts to 74% of the UK woodland. R. ponticum
is a sporulation host which is tolerant to the highly virulent tree-killing root-fungus
pathogen Phytothphera, in particular P. ramorum and P. kernoviae. These were
introduced in 2003 and have now become widespread and are a particular threat to
species such as the English Oak (Quercus robur). Limited access and high costs of
aerial and satellite imagery restrict the progress of research into the spatial distribution patterns of R. ponticum and thus their effective removal strategy. Unmanned
Aerial Vehicles (UAV) are currently used for conservation and agriculture but are as
yet prohibitively expensive. Using commercially available action cams (Sony and
Mobius: open source camera adapted for Infrared detection using a Rosco Red
Filter) and a DJI Phantom 3 professional quadcopter a multispectral image composed of 7 bands (440–800 nm range) was generated for sites containing invasive
rhododendron and other non-target species. Mosaicking efficiency reduced the
integrity of images but was rectified through georeferencing and Polynomial (spline)
transformation. R. ponticum was shown in pairwise comparisons with non-target
species to show greatest separability at 540 (p < 0.05) and least at 550 (p > 0.05) in
multivariate analysis. Intra-species difference between two R. ponticum populations
growing in differing locations was also significant at p = 0.01 in the same pairwise
comparison. Ivy was not significantly different to larch and R. ponticum in all but
the 540 and 750 bandwidths (p > 0.01). Seven band spectral signatures were all
shown to be significantly different at F(1, 42) = 6.795, p < 0.001. This was confirmed during supervised classifications of the composite spectral checkerboard,
A. Sanders (*)
Computing, Engineering and Science Department. Glyntaff, Trefforest Campus, University of
South Wales, South Glamorgan, UK
e-mail: abigailsanders1234@gmail.com