Ionian University
Ανάπτυξη υλικού χαμηλού κόστους και λογισμικού στοχαστικής επεξεργασίας για γεωσήματα και περιβαλλοντολογικά δεδομένα
Abstract
dc:descriptionThis doctoral dissertation addresses the design, construction, deployment and systematic evaluation of low-cost Internet of Things (IoT) systems for monitoring natural and biological hazards in the Ionian Islands Region of Greece. The study area lies among the highest seismic-hazard regions in Europe owing to its position on the Hellenic Subduction Zone and the Kefalonia Transform Fault, while simultaneously sustaining annual losses of 30—35% of olive production from Bactrocera oleae infestations and recurrent earthquake- and rainfall-triggered landslides. The common requirement across these challenges is dense spatial monitoring at island scale, economically sustainable for peripheral communities. The central research hypothesis is that the weaknesses of a single low-cost node — low signal-to-noise ratio, limited dynamic range, susceptibility to local noise — are compensated by the density of the network in which it is embedded, through statistical redundancy, spatial averaging and cross-validation, rendering the aggregate system reliable and comparable to commercial reference-grade equipment. Under this hypothesis, four subsystems were developed on a common technological platform (Raspberry Pi 3B+, Python, 4G/LTE, energy autonomy): (a) a low-cost seismograph based on a GS-11D geophone and a 24-bit ADS1256 digitizer, (b) a Deep Neural Network (DNN) algorithm for seismic event detection under urban noise, (c) a landslide monitoring station with MEMS accelerometers and geotechnical sensors, and (d) a Digital Olive Fly Trap with Computer Vision for automated Bactrocera oleae counting. The systems were installed on four of the seven islands of the Region of Ionian Islands — Lefkada, Kefalonia, Corfu, and Zakynthos — where they remained in continuous operation for seven years (2017–2024). The seismograph (~€300) was characterized through transfer-function H(s) derivation and deconvolution, achieving a mean local magnitude estimation error of |ΔML| = 0.087 for epicentral distances up to 15 km against a reference station of the Hellenic Unified Seismological Network (HUSN). The DNN algorithm achieves F1 = 88% and recall = 100% in distinguishing seismic events from anthropogenic noise in urban environments. The landslide network (~€800 per station) constitutes, to the best of our knowledge, the first documented low-cost landslide monitoring network in the Ionian Islands. The Digital Olive Fly Trap (~€250 per unit) reaches classification accuracy > 85% under field conditions, contributing to compliance with EU Directive 2009/128/EC on Integrated Pest Management. The per-node cost (~€300—800) corresponds to 1—16% of the commercial equivalent.The integration of the four subsystems into a common multi-domain platform realises the specifications of a Multi-Hazard Early Warning System (MHEWS) as defined by the WMO and the Sendai Framework 2015—2030. According to the literature review, this work represents the first multi-domain application of this scale in an island environment worldwide, establishing the viability of unified low-cost IoT platforms as an alternative for peripheral high-risk communities with limited resources.
Degree
thesis:*- Grantor dc:publisher
- Ionian University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Vlachos, Ioannis
- Βλάχος, Ιωάννης
Subjects
dc:subject × 26- Διαδίκτυο των πραγμάτων
- Σεισμογράφος χαμηλού κόστους
- Συνάρτηση μεταφοράς
- Βαθιά νευρωνικά δίκτυα
- Κατολισθήσεις
- Ψηφιακή δακοπαγίδα
- Δάκος της ελιάς
- Υπολογιστική όραση
- Σύστημα έγκαιρης προειδοποίησης πολλαπλών κινδύνων
- Ιόνια Νησιά
- Internet of things
- Low-cost seismograph
- Transfer function
- Deep neural networks
- Landslides
- Digital olive fly trap
- Bactrocera oleae
- Computer vision
- Multi-hazard early warning system
- Ionian Islands
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Διεπιστημονικές εφαρμογές επιστήμης ηλεκτρονικών υπολογιστών
- Computer and Information Sciences
- Natural Sciences
- Computer Science Interdisciplinary Applications
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/61945