Date of Award

Fall 12-1-2008

Document Type

Thesis

Degree Name

Master of Science in Information Systems (MSIS)

First Advisor

Surendra Sarnikar

Second Advisor

Ronghua Shan

Third Advisor

Stephen Krebsbach

Abstract

Satellite images are used more and more to analyze global tree and forest cover change. It is also extensively used in modeling the climatic change. These are becoming very important tools for the policy makers to discuss treaties and policies for global climatic change. Geographic Information System Center of Excellence (GIScCE) at South Dakota State University has become leader in development of such information (products). It extensive contracts with various world organizations is a testament to their work in this area. Hence it has become imperative that they would be able to analyze satellite images and provide the products in timely manner with few or no discrepancies. GIScCE processes satellite images from various satellites based on the project requirement. During this processing, elimination of artifacts and normalizing the vegetative cover over large areas of land, they use many statistical models. One such statistical model which is heavily used is regression tree. The objective of this project is to develop a regression tree to address their need to processes large datasets. This is currently not supported by commercially available software's. A regression tree model was developed and tested against the currently used commercial software S-Plus. The results from the both the implementations are identical. The speed improvement in setting up the program is improved. The results from S-Plus are normally go through post processing phase which is now integrated into the code. This saves more time for the users to work on other aspects. The program is believed to use less memory than the traditional commercial software. The program is now delivered to GIScCE and is under further testing and pending full deployment. Rapid Development (RAD) software development methodology is used in the development of the project.

Comments

dsu-th-208

Download

Share

COinS