Hyperspectral NIR imaging of plant material

Loading...
Thumbnail Image
Other Title
Authors
Holmes, Wayne
Look, Morgan
Lai, Anthony
Sidhu, Deepinder
Author ORCID Profiles (clickable)
Degree
Grantor
Date
2020-12-07
Supervisors
Type
Conference Contribution - Oral Presentation
Ngā Upoko Tukutuku (Māori subject headings)
Keyword
New Zealand
weed detection and identification
weeds
pastures
spectral reflectance
field spectroscopy
proximal imaging
spectral imaging
Citation
Holmes, W., Look, M., Lai, A., & Sidhu, D. (2020, December). Hyperspectral NIR imaging of Plant Material. Paper presented at the Unitec Research Symposium, Mount Albert Campus, Unitec.
Abstract
This study looks at the classification of plant species and components using Hyperspectral cameras in the Near Infrared region of the spectrum as part of a move towards precision agriculture. The NIR region of the electromagnetic spectrum lies just below the visible spectrum. Its longer wavelength has several advantages over visible light such as the ability to penetrate significantly below the surface of a material and along with absorption peaks for many chemical groups present in this region. In this work proximal spectral reflectance images were used of common New Zealand pasture weeds in order to determine the inter- and intra- species proximal spectral reflectance variations. It examined the ability and extent of accuracy when using hyperspectral cameras to uniquely identify three common species of weeds that grow in pastures based on their reflectance spectra alone. The use of these cameras showed that considerable measurement noise in the spectral data was present. This noise was due to using uncontrolled lighting i.e. solar illumination in field applications and the effect of scattered light on shading in the image. It was shown that a significant reduction of noise can be achieved by careful experimental design prior to acquiring the images. Despite the noise the study was successful in identifying weed species based purely on the reflectance spectra. This work also showed the ability of hyperspectral near-infrared imaging to identify the plant components such as flowers, stems and leaves on individual plants
Publisher
Link to ePress publication
DOI
Copyright holder
Authors
Copyright notice
All rights reserved
Copyright license
Available online at
This item appears in: