University of Twente Student Theses

Login
This website will be unavailable due to maintenance December 1st between 8:00 and 12:00 CET.

Comparison of thermal infrared and multispectral UAV imagery for detecting pine trees (Pinus Brutia) health status in Lefka Ori National Park in West Crete, Greece

Kahsay, Azeb Gidey (2022) Comparison of thermal infrared and multispectral UAV imagery for detecting pine trees (Pinus Brutia) health status in Lefka Ori National Park in West Crete, Greece.

[img] PDF
8MB
Abstract:Natural and anthropogenic stressors such as drought, pests, and diseases exert increasing pressure on the forests' condition. Forest health assessment, mapping, and monitoring are crucial for targeted management interventions and conservation. Direct forest health assessment in the field, despite considers as accurate, is a labour-intensive approach. Remote sensing (RS) is widely used in forest health assessment to create standardized methods that reduce subjectiveness, extrapolate observations in unvisited, inaccessible areas, and reduce labour and costs. Unmanned aerial systems (UAS) have gained popularity in many forest-related management activities and research. Stress in trees causes a change in their physiological process, resulting in a change in the reflectance of multispectral bands (visible; 0.55 - 0.735 µm and near-infrared (NIR) 0.79µm bands) and a temperature rise in the canopy. Thermal infrared (TIR, 7.5 -13.5 μm) remote sensing data can detect such canopy temperature changes. Previous research has confirmed the ability of UAS imagery to detect plants' health status. This study aims to investigate whether UAS-TIR imagery can be used to accurately map the health and infestation status of Pine trees (Pinus brutia) and compare the prediction accuracy with results obtained using multispectral remote sensing (MS) data. The usefulness of UAS-acquired TIR and multispectral data were examined in an open Mediterranean Pine forest in west Crete, Greece. The UAS campaign was conducted between 30 August and 1 September 2021, covering 0.4 km2. During fieldwork, the defoliation as an indicator of the health assessment and discoloration for Marchalina hellenica infestation assessment of individual trees were recorded, and preliminary analysis was done using 105 observation data. Canopy temperature and vegetation indices were computed and further, extracted for the delineated tree crowns, and used to classify trees' health and infestation status; RGB image output was also used to improve the segmentation accuracy. In line with past research in other ecosystems, the results from present study indicate that canopy temperature was able to show the separability between health classes using defoliation as an indicator; however, the difference in discoloration-based infestation class was not significant. Alongside, vegetation indices find it difficult to show a defined relation with defoliation-based health class, although the separability between infestation classes was significantly demonstrated. Among the calculated vegetation indices, SAVI obtained the highest separability in the discoloration-based infestation classes. A weak negative correlation was observed between canopy temperature and vegetation indices. Further investigation is needed to assess the performance of TIR data hyperspectral. Keywords: Forest health, UAS, UAV, Thermal infrared, Canopy temperature, Infestation.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Subject:38 earth sciences, 43 environmental science, 48 agricultural science
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:https://purl.utwente.nl/essays/92086
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page