This project is funded by the Ministry of Research and Innovation, CNCS – UEFISCDI, within PNCDI III
Project number PN-III-P4-ID-PCE-2016-0222
Landslide inventory maps are the key component of a landslide risk assessment and are required to document the extent of landslide phenomena in a region, to investigate the distribution, types, pattern, recurrence and statistics of slope failures, and to study the evolution of landscapes dominated by mass-wasting processes. Although important research has been carried out so far, no semi-automated and transferable method for generating landslide inventories exists. The main challenge still remains transferability. The overall goal of this research is to develop a coherent and intelligent object-based framework to extract landslides from Digital Elevation Models (DEM) for creating landslide inventory maps. In order to address this important topic, innovative solutions to tackle the current challenges, particularly the transferability of the model, are proposed.
Specifically, the project aims at:
-Developing, testing and applying an automate technique to calibrate the input variables to the characteristics of a given study area;
-Evaluating the potential of calibrating the geographic extent to segment landslides as single objects;
-Coupling neighbourhood indicators, i.e. landscape metrics, with land-surface variables to improve detection and delineation of landslides;
-Developing and demonstrating a semi-automatic approach to extract landslides from DEMs for creating landslide inventory maps.
By proposing a new framework to map landslides from DEMs we provide a direct and significant contribution to the field of geomorphology. Derivatives of the proposed work may also contribute to other fields where understanding and modelling morphological patterns in surfaces needs to be addressed, e.g. geomorphometry, geographic object-based image analysis, and geoinformatics. The approach proposed here shall lead to an operational solution for creating landslide inventories, which has practical applicability in land management.