Self Organizing Neural Network: A Method for Feature Analysis of Lands
- Faculty of Environment, University of Tehran, Iran.
Published in The First NZAAR International Event Series on Natural and Built Environment, Cities, Sustainability and Advanced Engineering, Putrajaya, Malaysia, 26 June 2016, pages 116 to 119. New Zealand Academy of Applied Research. ISSN 2463-5979 (online), 2463-5960 (print).
Abstract
The exploration and categorization of Shuttle Radar Topography Mission (SRTM) data for classification and extraction of topographic features is a challenging task. This paper presents a semi-automatic approach using artificial neural networks -Self Organizing Map (SOM) and SRTM data for analysis and identification of land features in Lut Desert. Lut Desert or Dasht-e Lut is among the hottest, driest desert regions in the world. This area is characterized by homogeneous repetition of wind-eroded landforms, some of the world’s highest mega dunes and salt flats in southeastern Iran. In this hyper area region the extremely unidirectional winds have fluted the bedrock outcrops and old lake beds to aerodynamic forms, called yardangs. A SOM with a low average quantization error was used for further analysis. Feature space analysis, morphometric signatures, three-dimensional inspection and auxiliary data facilitated the assignment of semantic meaning to the output classes in terms of geomorphometric features. Results are provided in a geographic information system as thematic maps of landform entities based on form and slope. The results demonstrate that a SOM is an efficient tool for analyzing and recognition of desert lands features.
Keywords: Self Organizing Map; Radar data; Neural Network; Desert
Rights
Copyright © New Zealand Academy of Applied Research Ltd 2016. All rights reserved. Reproduced in the GDI Academy archive from the original proceedings without change to the authors’ text.