Democritus University of Thrace (DUTH)
Τεχνικές ανάλυσης βλάστησης μέσω εναέριας και δορυφορικής τηλεπισκόπησης με χρήση τεχνητής νοημοσύνης και δικτύων νέας γενιάς
Abstract
dc:descriptionThis PhD thesis focuses on developing innovative methods for vegetation analysis and early detection of plant stress using remote sensing imagery from Unmanned Aerial Vehicles (UAV), Internet of Things (IoT) sensors, and satellites. The main objective is to support efficient resource management by optimizing vegetation monitoring processes. This thesis integrates remote sensing techniques, computer vision, machine learning, and modern infrastructures like cloud computing and 5G networks. The methodology begins with an extensive literature review in Coverage Path Planning (CPP) for UAVs, aiming at optimizing energy consumption and time-efficient coverage of areas of interest. Subsequently, methods for collecting and processing aerial optical and Thermal Infrared (TIR) images were developed and applied in experimental vineyards. These methods included background noise removal and clustering of thermal values, resulting in high-precision detection of stressed areas, with significant improvements over other spectral indices. A hybrid framework for early water stress detection was proposed in the next phase, combining TIR and optical imagery through data fusion and wavelet denoising. The methodology demonstrated the ability to detect water stress up to 84 hours earlier than conventional soil moisture sensors. Moreover, its deployment on IoT devices and cloud computing environments over 5G networks showcased notable processing speed and efficiency advantages. The research was further extended to explore the capabilities of satellite remote sensing, focusing on nanosatellites and their propulsion systems, which are deemed critical for the operational reliability of Earth observation missions. Subsequently, a novel method was developed for detecting plant stress using satellite imagery from the WorldView-3 satellite. Despite differences in spatial resolution, the method achieved reliable and accurate plant stress detection comparable in quality to UAV imagery. The thesis concludes with a comparative evaluation of the proposed methodologies, emphasizing the advantages of integrating aerial and satellite imaging and the importance of exploiting remote sensing technologies for precision agriculture. This work's key research contribution lies in the development of an integrated and scalable framework for early detection of plant stress, suitable for deployment in real-world operational settings using both aerial and satellite remote sensing.
Degree
thesis:*- Grantor dc:publisher
- Democritus University of Thrace (DUTH)
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Fevgas, Georgios
- Φεύγας, Γεώργιος
Subjects
dc:subject × 14- Φυτική Καταπόνηση
- Τηλεπισκόπηση
- Διαδίκτυο των Πραγμάτων - IoT
- Παρατήρηση της Γης
- Μη επανδρωμένα εναέρια οχήματα
- Vegetation stress
- Unmanned Aerial Vehicle
- Remote sensing
- Internet of Things
- Earth observation
- Περιβαλλοντική Μηχανική
- Επιστήμες Μηχανικού και Τεχνολογία
- Environmental Engineering
- Engineering and Technology
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/59335