Semi-automatic Landslide Inventory Mapping with Multiresolution Segmentation Process: A Case Study from Ulus District (Bartin, NW Turkey)

Autor: Harun Sonmez, Murat Ercanoglu, Gulseren Dagdelenler
Rok vydání: 2020
Předmět:
Zdroj: Understanding and Reducing Landslide Disaster Risk ISBN: 9783030602260
DOI: 10.1007/978-3-030-60227-7_8
Popis: Landslide inventory mapping studies have been received special attention from a wide range of specialists. In accordance with this situation, the main purpose of this study was to produce a semi-automatic GIS-based inventory mapping for locating landslides by a multiresolution segmentation process, which is the first phase of Object-Oriented Analyses (OOA). Ulus district of Bartin located in the Western Black Sea Region of Turkey was chosen as the study area. For the multiresolution segmentation process, a total of 1132 objects were automatically extracted using the first 4 bands of the Landsat ETM + satellite image and three thematic maps (slope, curvature and Normalized Difference Vegetation Index) of the study area. Multiresolution segmentation process was performed with Definiens Professional Earth (DPE in Definiens professional 5 UserGuide, Definiens AG, Munchen, Germany, 2006) software. In order to determine semi-automatically landslide locations in the study area, Artificial Neural Networks (ANN) method has been applied to the study area with 8 parameters (brightness, shape index, GLDV Contrast, Length/Width, Maximum Difference, Asymmetry, GLCM Contrast and GLCM Homogeneity) as an input and landslide (1) and no landslide (0) information as an output. This semi-automatically inventory map of the area estimated 78.2% of the existing landslides accurately. The applicability of this method is easy and quick, but when applying this method to the study area, thematic, spectral, shape and textural properties of the landslides must be revealed accurately and in detail.
Databáze: OpenAIRE
načítá se...