To cite this paper use one of the standards below:
Invasive insect pests such as Gonipterus sp. n. 2 pose a significant threat to the Eucalyptus industry. This economically important leaf-feeding beetle can completely defoliate Eucalyptus trees and lead to shoot and tip death. Early detection of this pest is important for early intervention to prevent pest outbreaks. Conventional insect pest monitoring methods are time consuming and spatially restrictive. The use of remote sensing techniques such as imagery obtained with unmanned aerial vehicle systems presents the opportunity for rapid assessment, at centimeter spatial resolution, to detect damage levels of Gonipterus sp. n. 2. The aim was to develop a remote sensing damage detection model to use as a monitoring tool. A Micasense red-edge P sensor mounted on an unmanned aerial vehicle was used to take images of E. dunnii trees with varying levels of Gonipterus sp. n. 2 infestation. A random sampling approach was used to assess defoliated trees based on visual assessments of six severity classes. Multispectral images were analysed by extracting the mean spectral reflectance of a 0.5m buffer around the central tree crown. Model training and validation were implemented using support vector machines classification in Python, and hyperparameter tuning using grid search. Three classification modelling scenarios were investigated based on spectral reflectance bands, indices, and their combination. Recursive feature elimination was used to select the essential variables. The scenario using both spectral reflectance and vegetation indices had the highest classification accuracy of 0.92%, while spectral bands achieved 0.84%, and vegetation indices had a poor overall accuracy of 0.21%. The recursive feature elimination selected Red-edge717, near-infrared 842, and Enhanced Vegetation Index as the essential variables in classifying defoliation levels. This study provides a baseline for developing an early detection system using UAV multispectral data for Gonipterus sp. n. 2.
With nearly 200,000 papers published, Galoá empowers scholars to share and discover cutting-edge research through our streamlined and accessible academic publishing platform.
Learn more about our products:
This proceedings is identified by a DOI , for use in citations or bibliographic references. Attention: this is not a DOI for the paper and as such cannot be used in Lattes to identify a particular work.
Check the link "How to cite" in the paper's page, to see how to properly cite the paper