Thesis
Selection of triangle sets for the analysis of randomly spaced data
Honours, Murdoch University
1980
Abstract
A triangulation scheme is proposed to provide a foundation in the direct analysis of randomly spaced data. It presents a framework for producing the maximum possible information about spatial differences in the data set, such as the extraction of a wind field from a set of barometric pressures recorded at stations that report irregularly or are mobile.
Five theorems regarding triangular nets are proven. They allow calculation of the number of non-overlaping triangles, as well as generalizations that prove that any net of triangles can be converted into any other net through the adjustment of diagonals of quadrilaterals a feature of the algorithm.
The triangulation algorithm is given and explained, followed by a FORTRAN program listing, including supporting graphical programs which plot the data, the triangular net and illustrate the concept of the selection process. A definition of optimality is included which is based on triangle shape; this definition is used to derive an algorithm which seeks the most optimal triangular net.
The results indicate that a powerful tool is available for data analysis, a tool limited only by data accuracy which is independent of the inconsistencies of manual, subjective analysis.
Details
- Title
- Selection of triangle sets for the analysis of randomly spaced data
- Authors/Creators
- E.R. Magnus
- Contributors
- Peter E. Kloeden (Supervisor)William Scott (Supervisor)
- Awarding Institution
- Murdoch University; Honours
- Identifiers
- 991005544502307891
- Murdoch Affiliation
- School of Mathematical and Physical Sciences
- Language
- English
- Resource Type
- Thesis
Metrics
239 File views/ downloads
77 Record Views