Georeferencing, Feature Extraction and Modeling From Raster and Point Data Information.

Tyler J. Alumbaugh and Peter Bajcsy

Technical Report NCSA-ALG04-0002, April 2004

In this report we present an overview of the GIS processing steps for predictive modeling from raster and point data and we describe the software tools developed for this purpose. The GIS processing steps include data georeferencing, visualization, feature extraction from raster data over locations defined by point data, and predictive modeling of (a) cross-feature dependencies by a correlation matrix and (b) spatial dependencies by a kriging technique. The described work leverages from our previous development of georeferencing capabilities and is driven by agricultural and environmental applications.

This document outlines the theory and implementation of each processing step and serves as a guideline for other scientists trying to work with geo-registered raster and point data.