Raster Data Reading Assignment, Chapters 11 and 12

1
Raster Data(Reading Assignment, Chapters 11 and
12)

• Applications of GIS
• Week 2
• 2004 Fall

2
Elements of the Raster Data Model

• Grid consists of rows, columns and cells.
  Origins of rows and columns is at upper-left
  corner of the grid. Rows is like y and columns x
  coordinates.
• Values of cells either integer or
  floating-point value.
• Resolution 10m, 30m and 90m

3
Types of Raster Data

• Satellite Imagery (landsat7.usgs.gov,
  terra.nasa.gov)
• DEM (digital elevation models)
• 7.5-minute DEM (30m,10m in UTM) corresponds to a
  USGS 124000 scale quadrangle (no-data slivers
  box 7.2), Level 1- vertical accuracy of 15m,
  Level 2 vertical accuracy of 7 m, derived from
  contour lines
• LIDAR (light detection and ranging) to produce
  detailed DEM
• DOQ (Digital Orthophoto quad) from aerial photos.
  Mostly 3.75-minute quarter with 1-meter ground
  resolution
• DRG scanned 7.5-min topo maps with 250 dpi
  2.4 meter resolution, georeferenced to the UTM
• TIFF, JPEG other image files

4
Data Structure

• Three methods for recoding two-dimensional
  regions.
• Chain code (figure 7.5)
• Block code (fig 7.6)
• Regional quad (fig 7.7)

5
Data Compression

• MrSID (Multi-resolution Seamless Image Database)
  recalling the image data at different
  resolutions or scales. Compress a large image
  such as a DOQ with subblocks and eliminates the
  artificial block boundaries during the
  compression process.

6
Georeferening

• Geometric transformation of satellite imagery is
  called georeferening
• Affine transformation
• through rotation, translation, and scaling
• Polynomial equations
• Requiring selection of control points
• After transformed, a world file is created (such
  as tmrect.bil -gt tmrect.biw)
• Polynomial equation (is called warping or rubber
  sheeting)
• A new grid created from polynomial equation.
• Data resampling consists of three methods
  (nearest neighbor, bilinear interpolation, and
  cubic convolution).
• Nearest neighbor is primarily used for
  categorical data. However, three methods can be
  used for elevation data. Smoothness of the
  resampled output increase from nearest neighbor,
  bilinear to cubic convolution)

7
Data Conversion - Rasterization and Vectorization

• Rasterization, three steps
• Assign all cells with 0 in coverage area
• Change values of those cells corresponding to
  points, lines and polygons. (1 for points, line
  value for line and polygons value for a polygon
  boundary)
• Fill the interior of the polygon outline with the
  polygon value

8
Integration of Raster and Vector Data

• DOQs, DRGs and graphics file used for background.
• DEMs can be processed to obtain contour, slope,
  aspect, watersheds, shades, stream networks.
• DEMs can be converted to TIN (triangulated
  irregular network)