Raster data model

1. 1
AGS 722
Raster Data Model
(Chang’s Chapter 7)
Elements of the Raster Data Model
ƒRaster model divides the area into grid cells
or pixel.
ƒEach grid cell is filled with the measured
attribute values.
ƒIt can represent points, lines and area (Figure
7.1).
ƒResolution depends on real world area
represented by each grid cell.
AGS 722

2. 2
AGS 722
Raster Data Model
ƒThe larger the area represented, the lower
the resolution of data.
ƒCells are identified by their positions in the
grid.
ƒRaster data is geo-referenced by:
•Real world coordinates of the reference
point
•Cell size in real world distance
•Use the upper-left or lower-left corner of
grid as the reference point.
AGS 722
IDRISI Metadata

3. 3
AGS 722
Raster Data Model
ƒStorage requirement is high.
ƒEx: If the area is 100 km x 100 km and cell
size is 10 m. It needs 10,000 rows x 10,000
columns or 100,000,000 pixels.
ƒIf one byte is used per pixel, it requires 100
MB storage
AGS 722
Types of Raster Data
1.Satellite Imagery
ƒRemotely sensed satellite data are
recorded in raster format.
ƒSpatial resolution varies:
• 30 m. for Landsat 4 and 5 (use the
Thematic Mapper scanner), and
Landsat 7 (use Enhanced Thematic
Mapper-Plus, ETM+ scanner).

4. 4
AGS 722
• 20 m. for SPOT images (Multi-spectral
sensor), and 10 m. for SPOT
Panchromatic sensor).
• 4 m. and 1 m. for IKONOS Multi-spectral
and Panchromatic images respectively.
ƒThe pixel value in a satellite image represents
light energy reflected or emitted from the
Earth’s surface.
AGS 722
ƒThe measurement of light energy is based on
electromagnetic spectrum.
ƒPanchromatic images are comprised of a
single spectral band.
ƒMulti-spectral images have multiple bands.
– Landsat TM has 7 band.
ƒLand use, land cover and hydrography can
be classified from image processing system.
ƒSatellite images can be diaplayed in black
and white or in color.

5. 5
AGS 722
Landsat TM Bands
AGS 722
Composite Color Images

6. 6
AGS 722
2. Digital Elevation Models (DEM)
ƒDEM consists of an array of uniformly spaced
elevation data.
ƒDEM are produced from:
– a stereoplotter and aerial photograph with
overlapping areas.
– Satellite imagery such as SPOT stereo
model using special software.
AGS 722
3. Digital Orthophotos
ƒPrepared from aerial photograph or other
remotely sensed data.
ƒDisplacement caused by camera tilt and
terrain relief has been removed.
ƒThey are geo-referenced and can be
registered with topographic and other maps.

7. 7
AGS 722
Digital Orthophoto
AGS 722
4. Binary Scanned Files
ƒScanned image containing values of 1 and 0.
ƒMaps to be digitized are typically scanned at
300 or 400 dpi (dots per inch).
5. Graphic Files
ƒMaps, photographs and images can be stored
as digital graphic files.
– e.g. TIFF (Tagged Image File Format), GIF
(Graphic Interchangeable Format), JPEG
(Joint Photographic Exports Group), etc.
– GeoTIFF is a geo-referenced version of
TIFF format.

8. 8
AGS 722
Raster Data Structure
ƒRefers to storage of raster data so that
it can be processed by the computer.
Cell-by Cell Encoding
ƒA raster model is stored as a matrix.
ƒIts cell values are written into a file by
row and column. (Figure 7.2)
ƒIdeal to store the cell values that
change continuously, e.g.,DEM.
AGS 722

9. 9
AGS 722
ƒFor multi-spectral satellite image, each
cell has more than one value, data are
stored in either of the following formats.
–The band interleaved by line (.bil):
this method stores the 1
st
value of
every row sequentially, followed by
the second value of every row, and so
on in one image.
AGS 722
Multi-band Satellite Data Structure
.bsq
.bil
.bip
Figure 7.x

10. 10
AGS 722
The Band Sequential (.bsq) method:
stores values of each band sequentially
in one image.
The Band Interleave by Pixel (.bip):each
row of an image is stored sequentially,
row 1 all bands, row 2 all bands, and so
on.
(See Figure 7.x)
AGS 722
Multi-band Satellite Data Structure
.bsq
.bil
.bip
Figure 7.x

11. 11
AGS 722
Run-length Encoding
ƒRecords the cells by row and by group
ƒEach group includes a cell value and the
number of cells with that value.
ƒIf all cells in a row contain the same value,
only one group is recorded, hence save
computer memory.
ƒSee Figure 7.3.
AGS 722

12. 12
AGS 722
Chain Code Method
ƒRepresent the boundary of a region by using
a series of cardinal directions and cells.
– Ex: N1 means moving north by 1 cell,
S4 means moving south by 4 cells.
ƒSeeFigure 7.4
AGS 722

13. 13
AGS 722
Block Code Method
ƒUses square blocks to represent the region.
– A unit square represents 1 cell.
– 4-square block represents 2 x 2 cells
– 9-square block represents 3 x 3 cells, and
so on.
ƒEach square block is coded only with the
location of a cell (lower left of the block), and
the side length of the block.
ƒSee Figure 7.5
AGS 722

14. 14
AGS 722
Quad Tree Method
ƒUses recursive decomposition to divide a grid
into a hierarchy of quadrants. (Figure 7.6).
ƒA quadrant having cells with the same value
will not be sub-divided, and it is stored as a
leaf node.
ƒLeaf nodes are coded with the value
homogeneous quadrant.
ƒA quadrant having different cell values will be
subdivided until a quadrant at the finer level
contains only one value.
AGS 722

15. 15
AGS 722
ƒThis method is efficient for storing and
processing data.
ƒDifferent raster GIS software use different
method of storing data.
– IDRISI and GRASS use either cell-by-cell
or run length encoding method.
– SPANS uses a quad-tree data structure.
AGS 722
Data Compression
ƒRefers to the reduction of raster data
volumes.
ƒRun length encoding method may reach 10:1
compression ratio.
ƒTIFF and GIF files use lossless compression
which allows the original image to be
precisely reconstructed.

16. 16
AGS 722
Data Compression
ƒJPEG files use lossy compressionwhich can
achieve high compression ratios but can not
reconstruct the original image fully.
ƒMrSid (Multi-resolution Seamless Image
Database) has capability of recalling image
data at different resolution or scales and also
can compress a large image.
AGS 722
Projection of Raster Data
ƒProjected raster data are based on rows and columns but the rows and columns are measured in real-world coordinates.
–Ex:
• Rows: 463, Columns: 318, Cell size: 30
m
• UTM coordinates at the lower left corner:
499995, 5177175
• UTM coordinates at the upper right
corner: 509535, 5191065

17. 17
AGS 722
• The cell in Row 1 and Column 1 at the
upper left corner has UTM coordinates of
499995, 5191035.
Data Conversion
ƒConversion of vector to raster data is called
rasterization.
ƒConversion of raster to vector data is called
vectorization. (Figure 7.8)
ƒBoth require use of computer algorithms which
most GIS software have.
AGS 722

18. 18
AGS 722
Integration of Raster and Vector Data
ƒCan take place in data display, data
processing, data conversion, or data analysis.
ƒDEM are input data to extract topographic
features such as contour, drainage network,
watersheds, etc.
ƒMost GIS packages allow simultaneous
display of raster and vector data.
ƒData conversion must be performed first if the
analysis of both raster and vector data is
required.