The classification of galaxies is morphological. Some of the physical characteristics of the galaxies are strongly related to their appearances including the types of stars that the galaxy contains; activity in the nucleus; and some information about the age. Although the categorization can be done manually the abundance of data being received by the modern telescopes necessitates automation of the process.
Shapes of the galaxies range from very simple (e.g. elliptical galaxies) to highly complex (e.g. irregular galaxies). If one could obtain a measure of complexity for such shapes, different ranges of magnitude of the measure can point to different classes. Fractal Dimension can be such a measure.
Fractal Dimension is a quantity related to the complexity of a given shape. The range of such dimensions for two dimensional shapes is from one to two. Simple shapes like lines, circles and triangles have Fractal Dimensions of one while complicated shapes have Fractal Dimensions close to two.
There are several methods to calculate Fractal Dimensions. Box-counting is one of such and is considered ideal for complex shapes. The method involves constructing a box around a given image, consecutively dividing the box into smaller size (s) boxes, and counting the number(n) of boxes covering the shapes for each step. A log-log plot of n verses 1/s provides data for a straight line. The slop of such a line gives the Fractal Dimension of the shape.
Our goal is to develope a method for calculating the fractal dimensions of galaxy images based on their isophotes. The method will be applied to a set of galaxy images whose morphologies are known and whose qualities are comparable to that of those that will be produced by the large surveys. We may also include an application of the method to N-body simulations with a view to classify merging galaxies. #
Thesis Committee Members
Advisors:
George Rhee,
Stephen Lepp
Lon Spight
Donna Weistrop
Stanley Smith