Question 4
How can you represent spatial relationships in a data
structure?
Every GIS stores information using some data structure, most
likely based on one of these models:
- Spaghetti
Imagery
Shapefiles
Topological data structure
Triangular Irregular Network (TIN)
The topological structure can be implemented using a variety
of techniques :
- lists of links around objects (chains around polygons, chains
around nodes...)
- Winged Edge (and variants)
What are the spatial relationships a user might demand?
- Topological queries: polygon overlay, point in polygon, neighboring
objects
- Distance based queries: buffers, setbacks, neighborhood filters
- Transformations between data models...
In response to this question, pick at least
one of the options below:
- Sketch out the processing required to convert spaghetti into
a topological structure. What is different between this process
and the conversion between shapefiles and topological structure?
- In image processing, what is required to convert an image
into a vector representation. Is this process different for an
image of lines versus an image that classifies the pixels into
classes (objects)?
- Given a particular polygon in a shape file, how would you
figure out the identity of neighboring polygons? What processing
would make this faster?
- TINs are topological data structures, but they are also surface
structures. What are some limitations in converting arbitrary
vector objects into a TIN framework?
Your answers shouldn't be much longer than one page.
Index from here: Back to 465
Main Page | Questions / Assignments
| Glossary from Exploring
GIS | Schedule | Bibliography
| Course Policies | Resources
Version of 24 January 2001