A new approach for a topographic feature-based characterization of digital elevation data

  • Authors:
  • Eric Saux;Rémy Thibaud;Ki-Joune Li;Min-Hwan Kim

  • Affiliations:
  • Naval Academy Research Institute, Brest Armées, France;Naval Academy Research Institute, Brest Armées, France;Pusan National University, Pusan, South Korea;Pusan National University, Pusan, South Korea

  • Venue:
  • Proceedings of the 12th annual ACM international workshop on Geographic information systems
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Triangular Irregular Network (TIN) and Regular Square Grid (RSG) are widely used for representing 2.5 dimensional spatial data. However, these models are not defined from the topographic properties of the terrain (i.e., ridge lines, valley lines, saddle points, etc.). This paper introduces a three-step feature-based approach for topographic properties extraction on scattered elevation data modeled by a TIN. Firstly, a segmentation process extracts homogeneous morphological areas bounded by critical lines and points. Secondly, these lines and points are displaced using a deformable process in order to derive the terrain feature points, lines and areas. Thirdly, a classification process labels any topographic feature. This three-step approach relies on the definition of an adapted model of representation (SPIN) and data structure (DCFL2). The proposed approach is validated on a real case study (Seolak mountain in South Korea). Consistent results with the morphology of terrain are displayed.