DEM products Elevation 0- ~10,000 (earth) 16 bit (signed) Shaded relief 0-255 Slope 0-90 (if in degrees) Aspect 0-360 16 bit (flat = ?) Incidence 0-90 requires orbital metadata Visual: Anaglyphs and 3D perspectives
DEM layers a. Elevation DN = (metres, 16 bit): represented as grayscale or pseudocolour tints; Raster DEMs are stored as elevations (metres) or 32 bit (after interpolation) some DEM tiles in Canada may be in feet (how would the values compare ?) conversion = .3048
c. Slope Calculated in degrees (0-90) or % (0 -> infinity) slope is rise/run = vertical change over the horizontal distance 8 bit results should be adequate for most purposes
4. Aspect The direction in which a slope is facing, measured by azimuth (0-360) This raises three question for analysis: a north facing slope has both extreme values, 0 and 360 flat slopes have no value (they are given an arbitrary numerical value, e.g. 555) 0-360 requires 16 bit data; some software (PCI) converts to 8 bit by dividing by 2
ASTER image: Svalbard, Norway (15 metre resolution)
Evolution of DEM creation 1950s Generation of contours from stereo photos 1980s Mass points from stereo photos 1990s Automated generation of masspoints 2000s Direct generation of grids from stereo-imagery (and expansion of LiDAR)
SPOT 5 DEM (DSM) : 2002 30m High Resolution Stereoscopy (HRS) instrument (the high-res multispectral sensor is HRG) http://www.spot.com/web/SICORP/1809-spot-3d.php
Directed high res sensors- Ikonos 1999 2-5m http://www.satimagingcorp.com/galleryimages/ikonos-high-resolution-dem-eritrea.jpg
Photogrammetry: Tatras, Slovakia 2m
(Airborne) LiDAR 10cm – 10m (Mt.St. Helens) http://www.nasa.gov/centers/goddard/images/content/67409main_dh2.gif
LiDAR Forests for the World, 2009 – DTM and/or DSM
City 1 metre DEM (from 1m contours)
TRIM DEM
DEM quality and errors Scale – Match with needs and imagery Type of terrain - pixel homogeneity Horizontal and vertical accuracy DSM or DTM ? (most satellite imagery create DSMs) Interpolation / resampling method – from points or lines? Null values and spikes Errors are more easily spotted from a hillshaded model
DEM creation by interpolation Inverse distance weighted - simple Nearest neighbour – honours raw values Spline – minimizes curvature -> smooth surface Kriging – uses spatial correlation of points
DEM creation by interpolation Inverse Distance weighted - simple Nearest neighbour – honours raw values Spline – minimizes curvature -> smooth surface Kriging – uses spatial correlation of points (employing semi-variogram of distance v difference)
DEM Interpolation methods
DEM differencing Differences in DEMs (heights at same x,y) result from: Interpolation method Generation mode e.g. DTM v DSM Resolution (scale) Resampling and other processing Real change e.g. landslides, glacier melt, lakes ..
1888 First Canadian Topographic series NTDB 1:50,000 data (1982) 1:40,000 (Eldon sheet), Interval 200’ Interval 40 metres a. Glacier retreat b. New melt lakes north of Vermillion Lake c. Lack of gulleys on north side of Bourgeau Lake
Eldon, 1888
ETM+ 1999
TRIM DEM: PG (vertical accuracy 10 metres)
TRIM DEM – masspoints ~70m spacing captured onscreen from stereo-photography ‘soft copy’ (prior to fully automated image matching)
BC provincial DEM 1985 (1981-1989) – Homathko Icefield - mapplace BC provincial DEM 1985 (1981-1989) – Homathko Icefield - mapplace.ca or lrdw.ca
Assiniboine Spillway, MB http://www.gov.mb.ca/stem/mrd/geo/demsm/metadata.html
SRTM comparison with topographic DEM http://gisweb.ciat.cgiar.org/sig/download/laboratory_gis/srtm_vs_topomap.pdf
SRTM – ASTER comparison Sample area in Indonesia
DEMs : Lab next week: Correction operations require metadata, in this case: Sun azimuth ~ 151.59 Sun angle ~ 45.95