The process of converting (usually analog)
information into a (usually digital)
representation
that can be processed easily using a computer
Some Examples:
Scanning images using a a camera
Creating 3D-models using a laser scanner
Sampling and quantizing analog audio
Our Interest
The Source:
A map or globe
The Result:
A vector representation (i.e., a description of
the geographic features in the map) typically consisting of
points and line segments
Common Geographic Features
Points:
At a low resolution, cities and towns
At a high resolution, intersections
Line Segments:
Streets and street segments
Rivers
Communications lines
Polygons:
Town/city boundaries
County/state/regional boundaries
Country boundaries
Approaches
Manual:
Draw a grid on the map to be digitized and interpolate
within cells
Computer Assisted:
Use a pointing device (e.g., a mouse or puck) with
a known resolution and collect "clicks"
Automated:
Use an algorithm to perform a raster-to-vector
conversion
Drive/walk the geographic features with appropriate sensors
(e.g., GPS)
A Common Process
The Steps:
Give each point an identifier and then collect the
coordinates of all points (e.g., intersections when digitizing
a street map)
Give each line segment an identifier and define it using
the identifiers of its end points (e.g., define street segments
in terms of the intersections at each end)
Give each polygon an identifier and define it using
the identifiers of its line segments (e.g., define a county
in terms of the street segments on its boundary) using a
consistent winding rule
Add shape points to line segments as needed
The Advantages:
Reduces errors/inconsistencies (e.g., the coordinates of the
same point need not be measured more than once)
Provides topological (loosely, the relationships between
the features) information in addition to geographic
information