Representation of Terrain

1
Representation of Terrain

• Jaehyung Yu

2
Topics to be covered by Jay

• Central role of terrain representation
• Digital Terrain Modeling vs. DEM
• Techniques for creating DEM
• Interpolation methods for DEM generation
• DEM quality measures
• DEM interpretation

3
Central role of terrain representation

• Cardinal examples for many activities in GIS and
  environmental modeling
• A prominent role in the development of methods
  for conceptualization and visualization of 3-D
  data
• Understanding of the nature of the earth surface
  and atmospheric processes in subjective and
  analytical terms

4
Digital Terrain Modeling vs.
Digital Elevation Model

• Digital Elevation Model
• an ASCII or binary file that contains only
  spatial elevation data in a regular gridded
  pattern in raster format
• Digital Terrain Model
• Allows the possibility of including landscape
  attributes other than topography as a means of
  improving the digital representation of a section
  of terrain
• Representation of Terrain
• DEM generation
  interpretation

5
Elevation Source

• Surface specific point elevation data high and
  low points, saddle points, and points on streams
  and ridges make up the skeleton of terrain
• Contour and stream-line data
• Remotely sensed elevation data

6
Techniques for creating DEM

• Digitizing contour lines
• Aerial photogrammetry
• Stereoscopy using optical satellite imagery
• Laser scanning
• Radar techniques

7
Digitizing contour lines

• derived from existing topographic maps
• various interpolation methods
• weighted moving averages, bicubic splines, and
  finite elements
• still the main source for creating DEMs

Raster
DEM through interpolation
Vector with elevation tag
8
Aerial photogrammetry

• bundle block adjustment
• aerial triangulation
• GPS reduces number of ground control points

9
Optical satellite data

• stereo SPOT Pan images
• two processing approaches
• geometric model of a CCD line scanner
• automatic matching techniques

10
Optical satellite data (SPOT)
11
Optical satellite data

• Modular Optoelectronic Multispectral Scanner
  (MOMS)
• stereo acquisition is along-track direction
• flown on the MIR station

12
Optical satellite data (ASTER)
Begin Acquisition 3N
Complete Acquisition 3B
0 sec
9
55
64
6.7 km/sec
27.6o
705 km orbit
3N
3B
Ground
0 km
60
370
430
Stereo Scene
13
Optical satellite data (ASTER)
14
Laser scanning (LIDAR)

• Avoids the problems of aerial triangulation and
  orthorectification
• Accuracies within 213 cm

15
Laser scanning (LIDAR)
16
Radar techniques

• Radargrammetry
• same area from two different angles
• used for long time with airborne systems
• some potential for RADARSAT data
• usually requires ground control points

17
Radar techniques

• Radarclinometry
• also referred to as shape-from-shading
• inversion of the radiometric incidence angle
  correction enables an estimate of local terrain
  slope
• filtering of data improves quality

18
Radar techniques
(left) Height map derived using stereo SAR
technique only (right) Improved map created by
augmenting original stereo result with
shape-from-shading technique.
19
Radar techniques

• SAR interferometry
• repeat-pass geometry (satellite)
• covering large areas in short time frame
• computationally intensive
• variety of data available

20
Interpolation methods for DEM generation

• Point Elevation
• - Voronoi (Thiessen) polygon
• - TIN
• - Inverse distance weighting
• - Kriging
• - Spline
• Contour data
• - TIN
• - Topogrid (local adaptive gridding)

21
Voronoi Polygon

• A polygon whose interior consists of all points
  in the plane which are closer to a particular
  lattice point than to any other.

22
TIN based interpolation

• Triangulation locally fitting polynomials
• Based on elevation spot points, and contours
• TIN based linear interpolation
• Z a bx cy

23
Inverse distance weighting

• assumes that the unknown value of a point is
  influenced more by nearby control points than
  those farther away
• The degree of weight is expressed by the inverse
  of the distance between points raised to a power

24
Kriging

• The rate at which the variance between points
  changes over space
• Uses a weighted linear combinations of a number
  of neighboring sample values to model the spatial
  variation within a local area bounded by the
  input sample points

25
Spline

• Thin-plate splines create a surface, which passes
  through control points and has the least possible
  change in slope at all points
• Regularized Splines
• The approximation of regularized splines has the
  following form
• where a represents the trend function, and the
  basis function R(r) is

26
TOPOGRID

• Iterative finite difference interpolation
  procedure
• Uses a nested multi-resolution computation
  structure, starting with an initial coarse grid
  and successively halving the grid spacing until
  the final specified grid resolution is obtained

27
DEM quality measures

• lineage (metadata)
• data history, processing, assumptions etc.
• positional accuracy
• root mean square (measure of overall accuracy)
• standard deviation (measure of precision)
• attribute accuracy
• data completeness
• error of omission (measurable)
• logical consistency
• structural integrity of the data
• fidelity of relationships
• semantic accuracy
• number of features, relationships or attributes
  which agree with the selected model

28

• temporal information
• type of update
• validity period of the data set
• accuracy of derivatives of height
• slope gradient
• maximum rate of change of altitude
• slope aspect
• compass direction of maximum
• profile convexity
• rate of change of gradient
• plan convexity
• rate of change of aspect
• Ground control points
• 20 interior and 8 edge points for determination
  of RMS
• points well distributed
• points representative for the terrain

29
DEM interpretation

• Scale
• - Matching the resolution of grid DEM to
  source data
• - Spectral and fractal analyses of scale
• Terrain parameters
• Features
• - Catchments and sub catchments
• - Contour flow line networks
• - Multi scale feature model

30
Thank you!!! Any Question?
Jay