Monitoring diseases by using Hyperspectral and Thermal techniques at two different spatial scales: A feasibility study in alpine regions

Abstract

<p> </p><p>Since some decades, cultural and traditional alpine farming have been changed enormously, mainly due to anthropogenic activities and economic factors: some pastures were abandoned, some others changed from farming to touristic areas (eg. ski resorts) and some others have been dramatically intensified by changing to monoculture. In consequence, these activities allow practices of deforestation, the massive use of fertilizers and pesticides, the excessive use of machinery, grading, drainage among others. Additionally, human activity have impacted on weather by resulting on low rainfall or drought extended periods. The combination of these factors result that vegetation and species are more vulnerable to the infestation of pests and diseases.</p><p>On this feasibility study, we propose the identification, mapping and classification of individual trees affected by fungal species (alternaria) in apple orchards located in South Tyrol, Italy, based on hyperspectral and thermal imagery. We have conducted terrestrial and UAV-based surveys to identify (un)healthy indivuals (trees). High spatial resolution scale consisted on terrestrial monitoring approaches based on the identification of trees and leaves, the collection of leave spectral signature based on a dedicated spectroradiometer (300 to 2000 nm) and spectral imagery of individuals. Medium spatial resolution consisted on UAV-based spectral data collection and interpretation. 30 hyperspectral bands (400 to 900 nm) in the VIR range and thermal imagery (14 µm) in combination with leave-spectral bands allowed the identification and mapping of un-healthy individuals for further treatment.</p>