A Comparison of Ridge Detection Methods for DEM Data
Authors:
- Shinji KokaSKShinji Koka
- Koichi AnadaKAKoichi Anada
- Yasunori NakayamaYNYasunori Nakayama
- Kimio SugitaKSKimio Sugita
- Takeo YakuTYTakeo Yaku
- Ryusuke YokoyamaRYRyusuke Yokoyama
- 1 August 2012
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 513-517
Abstract
We deal with ridge detection methods from digital elevation map (DEM) data. As ridge detection methods, the O (N2) -time steepest ascent method and the O (N) -time discrete Lap lace transform (D.L.T.) method are known, where N is the number of cells. However, the D.L.T. method is too blurry to form ridge lines. In this paper, we introduce a 12 neighbor D.L.T. method which is a modification of the 4 neighbor D.L.T. method. And we also introduce another ridge detection method by the classification of local shapes around each cell. We can consider 32 patterns for ridges or valleys. Furthermore, we compare and evaluate their ridge detection methods in a certain area. We note that our two methods provide blurry terrain maps, but it require only O (N) -time for N cells, in comparison with the steepest ascent method.Keywords
This publication has 5 references indexed in Scilit:
- Ridge Detection with the Steepest Ascent MethodProcedia Computer Science, 2011
- Watershed Cuts: Minimum Spanning Forests and the Drop of Water PrincipleIEEE Transactions on Pattern Analysis and Machine Intelligence, 2008
- Geographical Concept Recognition With the Octgrid Method for Learning Geography and Geology2007
- Ridge-valley lines on meshes via implicit surface fittingACM Transactions on Graphics, 2004
- Evaluation of methods for ridge and valley detectionIEEE Transactions on Pattern Analysis and Machine Intelligence, 1999