In the last few years, thanks to high resolution DTMs derived by LiDAR technology, new methodologies for features extraction based on landform curvature have been developed. In a small dolomitic basin, the Rio Col Duro, a series of elaborations were carried out in order to test the effectiveness of these new methods for channel network recognition respect to traditional methods such as Area threshold, Slope-Area threshold, and Strahler order. Initially the traditional methods have been used at different DTM grid cell size. The results suggested that 1 m DTM should be considered as optimal for channel network extraction. Nevertheless compared to traditional methods landform curvature is proposed as a useful approach for channels extraction. It recognizes incision and convergence typical of channel morphology. The advantage to investigate micro-topography of an area, especially in the Alps, is crucial to understand the range of real morphological forms. Threshold values for channel network extraction calculated as multiples (1, 2 and 3 times) of curvature standard deviation have been used. Different kernels of curvature calculation have been applied to limit the “noise” and refine the extraction. Curvature, related only to geometric surface, is more suitable to describe real convergence in the area, even if it is not able to trace a continuous path in the drainage network. However, this method is useful only when using a high resolution DTM and an intermediate (not too small, not too large) kernel for curvature calculation. The traditional methods are still considered a valid approach for channel network extraction, but they identify channels also within large and uniform hillslopes where actual channelized forms are not present. Curvature is therefore proposed as a new challenge for such kind of analysis.

Metodologie per l'estrazione del reticolo idrografico da topografia ad alta risoluzione

Costa, Andrea
2010/2011

Abstract

In the last few years, thanks to high resolution DTMs derived by LiDAR technology, new methodologies for features extraction based on landform curvature have been developed. In a small dolomitic basin, the Rio Col Duro, a series of elaborations were carried out in order to test the effectiveness of these new methods for channel network recognition respect to traditional methods such as Area threshold, Slope-Area threshold, and Strahler order. Initially the traditional methods have been used at different DTM grid cell size. The results suggested that 1 m DTM should be considered as optimal for channel network extraction. Nevertheless compared to traditional methods landform curvature is proposed as a useful approach for channels extraction. It recognizes incision and convergence typical of channel morphology. The advantage to investigate micro-topography of an area, especially in the Alps, is crucial to understand the range of real morphological forms. Threshold values for channel network extraction calculated as multiples (1, 2 and 3 times) of curvature standard deviation have been used. Different kernels of curvature calculation have been applied to limit the “noise” and refine the extraction. Curvature, related only to geometric surface, is more suitable to describe real convergence in the area, even if it is not able to trace a continuous path in the drainage network. However, this method is useful only when using a high resolution DTM and an intermediate (not too small, not too large) kernel for curvature calculation. The traditional methods are still considered a valid approach for channel network extraction, but they identify channels also within large and uniform hillslopes where actual channelized forms are not present. Curvature is therefore proposed as a new challenge for such kind of analysis.
2010-03-19
162
estrazione reticolo idrografico, curvatura, lidar
File in questo prodotto:
File Dimensione Formato  
Andrea_Costa_TesiSpecialisticaSFA.pdf

accesso riservato

Dimensione 18.87 MB
Formato Adobe PDF
18.87 MB Adobe PDF

The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12608/14722