University of Patras
Μεθοδολογίες αποτίμησης της επιδεκτικότητας προς κατολίσθηση μέσω της ανάπτυξης πολυδιαχειριστικών συστημάτων
Abstract
dc:descriptionThe aim of this thesis is to introduce the concept of Information System Multi Administration (ISMA) in landslide research. These expandable Multi-Administration Systems implement web technologies to provide information that is directly related to topological (geospatial) data, and can be accessed by simultaneous users."Landslide Administration System for the region of Western Greece", is a prototype ISMA, designed and implemented to record landslides occurences of the Region. Operators can make use of simple mobile phones or laptops to input data directly on the spot and receive measurements. These entries are inserted in forms via a multi-user Online Platform - http://landslide.engeolab.gr - and are updated to offer feedback about the validity of the input. Functional testing has been performed to modify the system several times and improve performance; however, the development process is still in progress towards including spatial information from the whole region of Western Greece. By 2020, more than 1,000 landslide entries had been recorded, 657 of which in the regional units of Achaia and Ilia.Machine Learning was used for landslide susceptibility assessment and prediction in the landslides recorded in the System until 31/12/2019 in Achaia and Ilia. 30% of the System's landslides mapped in the aforementioned regions with four predisposing factors (lithology, altitude, slope, land use) were assessed only after their susceptibility was first calibrated. Data processing was performed with Machine Learning algorithms using WEKA software; the neural network model applied to landslide susceptibility analysis predicted approximately 92% of the data. Landslide Susceptibility Mapping for the regional units of Achaia and Ilia was obtained using an integrated artificial neural network model and its validation. The multi-administration "Landslide Administration System for the region of Western Greece" provides a dynamic content from a Database with detailed recorded landslide events, that have resulted from all available sources. This continually expanding system is freely accessible, providing information on Landslide predisposing factors, with simple and easily applicable analysis that leads to prediction models (susceptibility, risk, etc.).
Degree
thesis:*- Grantor dc:publisher
- University of Patras
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kordouli, Maria
- Κορδούλη, Μαρία
Subjects
dc:subject × 15- Επιδεκτικότητα κατολισθήσεων
- Συστήματα πληροφοριών πολλαπλής διαχείρισης
- Τεχνητή νοημοσύνη
- Μηχανική μάθηση
- Landslide susceptibility
- Susceptibility model
- Landslide simulation
- Neural networks
- Landslide analysis in geographic information systems
- Engineering geology
- Machine learning techniques in landslide susceptibility mapping
- Φυσικές Επιστήμες
- Γεωεπιστήμες και Επιστήμες Περιβάλλοντος
- Natural Sciences
- Earth and Related Environmental Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/48331