Principles of Geographic

Information Systems

Dr. Nabil Moustafa

Raster Data Model

Chapter Four
 Agenda

• Introduction
• Elements of Raster Data Model
• Types of Raster Data
• Raster Data Structure
• Raster Data Compression
• Data Conversion and Integration
Introduction
• A vector data model is ideal for discrete features with well defined locations and
shapes.

• The vector data model does not work well with spatial phenomena that vary
continuously over the space such as precipitation, elevation and soil erosions.

• A better option for representing continuous phenomena is the raster data model,
also called the field based model.

• The raster data model uses a regular grid to cover the space. The value in each
grid cell corresponds to the characteristic of a spatial phenomena at the cell
location, and the changes in the cell value reflect the spatial variation of the
phenomena.
A continuous elevation raster with darker shades for
higher elevations.
 Agenda

• Introduction
• Elements of Raster Data Model
• Types of Raster Data
• Raster Data Structure
• Raster Data Compression
• Data Conversion and Integration
Elements of Raster Data Model
• A raster is also called a grid or an image in GIS.
• A raster represents a continuous surface, but for data
storage and analysis, a raster is divided into rows, columns,
and cells (also called pixels).
• The origins of rows and columns is typically at the upper-left
corner of the raster.
• Rows function as y-coordinates, and columns as x-
coordinates, each cell in the raster is explicitly defined by its
row and column position.
Elements of Raster Data Model
• Raster data represent point with single cells, lines with
sequences of neighboring cells, and polygons with collection
of contiguous cells
• Cell values
– Each cell in a raster carries a value, which represents the
characteristics of a spatial phenomena at the location denoted by it
row and column.
– Depending of the coding of the cell values, a raster can be either
categorical or numeric.
– Integer cell values usually represent categorical data, which may or
may not be ordered. For example a land cover raster may use 1 for
urban and use, 2 for forested land, 3 for water body, and so on.
Representation of point, line, and
area features: raster format on the
left and vector format on the right.
Elements of Raster Data Model
– floating point cell values represent continuous, numeric data. For
example, a precipitation raster may have precipitation values of
20.15, 12.23.
– A floating point raster requires more computer memory than an
integer raster.
• Cell size
– The cell size determine the resolution of a raster. A cell size of 10
meters means that each cell measures 100 square meters.
– A 10 meter raster has a finer (higher) resolution than a 30-meter
raster.
– A large cell size cannot represent the precise location of spatial
features, thus increasing the chance of having mixed features such as
forest, and water in a cell.
DEMs at three resolutions: 30 meters, 10 meters, and 3 meters. The 30-m
and 10-m DEMs are USGS DEMs. The 3-m DEM is a derived product from
LIDAR data.
Elements of Raster Data Model
• Cell Depth
– Refers to no. of bits for storing cell values.
– A higher cell depth means that the cell can store a wide range of
values
– 8 bit raster can store 28 256 possible values
– 16 bit raster can store 216 65,536 possible values
• Raster bands
– A raster ay have a singe band or multiple bands.
– Each cell in a single band raster has only one cell value.
– Each cell in a multiband raster is associated with more than one cell
value.
Elements of Raster Data Model
• Spatial reference
– Raster data must have the spatial reference information so
that they can align spatially with other data sets in a GIS.
– A raster that has been processed to much a projected
coordinate system is often called a georeferenced raster.
UTM coordinates for the extent and the center of a 30-meter cell.
 Agenda

• Introduction
• Elements of Raster Data Model
• Types of Raster Data
• Raster Data Structure
• Raster Data Compression
• Data Conversion and Integration
Types of Raster Data
• Satellite Imagery
• USGS Digital Elevation Models (DEMs)
• Non- USGS DEMs
• Global DEMs
• Digital Orthophotos
• Bi-level Scanned Files
• Digital Raster graphics (DRGs)
• Graphic Files
• GIS Software-Specific raster Data
Types of Raster Data
1. Satellite Imagery: Imagery acquired from
satellites such as Landsat and Spot.
2. Digital Elevation Models (DEMs): The
representation of continuous elevation values
over a topographic surface by a regular array of
z-values referenced to a common datum.
3. Digital Orthophotos Quadrangles (DOQ): The
scanned photographic images that have been
corrected for distortions due to camera tilt,
terrain displacement, and other factors.

17
Types of Raster Data (continued)

4. Bi-Level Scanned Files: The scanned images

containing values of 1 and 0, they are usually
scanned from paper maps.
5. Digital Raster Graphics (DRGs): The scanned
images of USGS topographic maps.
6. Graphic Files: Maps, photographs, and images
can be stored as raster graphic files such as
TIFF (tagged image file format),
GIDF( graphics interchange format), and
JPEG( joint photographic expert group).

18
Types of Raster Data (continued)
7. GIS Software-Specific Raster Data: GIS packages use
raster data that are imported from DEMs, satellite
images, scanned images, graphic files, and ASCII files
or are converted from vector data. The GIS software
packages store raster data in different formats, for
example, ArcGIS stores raster data in the ESRI Grid
format.

19
 Agenda

• Introduction
• Elements of Raster Data Model
• Types of Raster Data
• Raster Data Structure
• Raster Data Compression
• Data Conversion and Integration
Raster Coding
➢ In the data entry process, maps can be digitized or
scanned at a selected cell size and each cell assigned a
code or a value.
➢ The methods for assigning cell codes are:
1. Presence/Absence
2. Cell Center
3. Dominant Area
4. Percent Coverage (advanced method)

21
 Presence/Absence

➢The most basic method is to record a feature if some of it

occurs in the cell space.

➢The only practical way of coding point and line features,

because they don’t take up much area of a cell.

22
Cell Center
➢ This method involves reading only the center of the cell
and assigning the code accordingly.

➢ When there are multiple features in a cell area, the one in

the center wins the code.

➢ This is no a good scheme for points or lines, because

they don’t necessarily pass through the exact center.

23
Dominant Area
➢ A common method is to assign the cell code to the
feature with the largest (dominant) share of the cell.

➢ This is suitable primarily for polygons, although line

features could be assigned according to which one has
the most linear distance in a cell.

24
25
Raster Data Structure
• Raster data structure refers to the method for storing raster
data.
• Three common methods are examined here:
– Cell by cell encoding
– Run-length encoding
– Quad tree
• Cell by cell encoding
– Provides the simplest raster data structure.
– A raster is stored as a matrix. And its cell values are written into a file
by row and column.
– this method is an ideal choice if the cell values of a raster change
continuously.
The cell-by-cell data structure
records each cell value by row
and column.
Raster Data Structure
• Run-Length Encoding
– Raster data with many repetitive cell values can be more efficiently
stored using run-length encoding (RLE).
– RLE records the cell values by row and by group.
– A group refers to a series of adjacent cells with the same cell value.
– Many GIS software use RLE in addition to cell by cell encoding for
storing raster data such as GRASS, IDRISI, and ArcGIS.
• Quad Tree
– Quad tree uses recursive decomposition to divide a raster into a
hierarchy of quadrants.
Raster Data Structure
– recursive decomposition refer to a process of continuous subdivision
until every quadrant in a quad tree contain only one cell value.
– The quad tree contains nodes and branches (subdivision).
– A node represents a quadrant , and it can be a nonleaf node or leaf
node.
– A nonleaf node is a branch point , meaning that the quadrant is
subject to division.
– A leaf node represent a quadrant that has one cell and it is an end
point which then be coded with its value.
– Quad tree is efficient way for storing area data, especially if the data
contain few categories, it is also efficient for data processing.
The regional quad tree
method divides a raster into a
hierarchy of quadrants. The
division stops when a
quadrant is made of cells of
the same value (gray or
white). A quadrant that
cannot be subdivided is
called a leaf node. In the
diagram, the quadrants are
indexed spatially: 0 for NW, 1
for SW, 2 for SE, and 3 for
NE. Using the spatial
indexing method and the
hierarchical quad tree
structure, the gray cells can
be coded as 02, 032, and so
on. See text for more
explanation.
Raster Data Structure
– Header file
• To import a raster data from a DEM or satellite image, a GIS
package requites info about the raster such as cell size, no of
bands, no of rows,….
• Satellite image
 Agenda

• Introduction
• Elements of Raster Data Model
• Types of Raster Data
• Raster Data Structure
• Raster Data Compression
• Data Conversion and Integration
Raster Data Compression
• Raster data require considerable memory space. The
memory requirement becomes higher for high resolution
DEMs and Satellite images.
• Data compression refers to the reduction of data volume. A
variety of techniques are available for data compression,
they can be lossless or lossy.
• Lossless compression preserves the cell or pixel values and
allows the origin image or raster to be precisely
reconstructed.
Raster Data Compression
• Lossy compression: cannot reconstruct fully the original
image but can achieve higher compression ratios than a
lossless compression.
• Wavlet transformation
 Agenda

• Introduction
• Elements of Raster Data Model
• Types of Raster Data
• Raster Data Structure
• Raster Data Compression
• Data Conversion and Integration
Data Conversion and Integration
• Rasterization converts vector data to raster data.
• Vectorization converts raster data to vector data,
vectorization involves three basic elements
– Line thining
– Line extraction
– Topological reconstruction
• Integrating of raster and vector data
Figure 4.12
On the left is an example of conversion from vector to raster
data, or rasterization. On the right is an example of conversion
from raster to vector data, or vectorization.