Past Research
Automated Pattern Recognition in Satellite Imagery
NCI SBIR Phase I and II Completed research 2007-2011
Principal Investigator: Pierre Goovaerts, PhD
Summary
This Phase II SBIR project will complete development of software tools for exposure assessment using high spatial resolution, hyperspectral (HSRH) imagery that quantifies the environment at unprecedented spectral and spatial resolution. This quantum leap in resolution has enormous potential for improving our ability to document, monitor and model environmental exposures. But this potential has not been fully realized, in part because existing software primarily uses the spectral information and underutilizes the rich spatial information contained in high resolution imagery. This Phase II project will develop software for exposure assessment that takes advantage of both the spatial and spectral information in HSRH data.
Three key milestones were accomplished during the Phase I project:
- The convening of a workshop of expert panelists in August, 2000 for the purpose of identifying optimal spatial methods and functionality to include in the software.
- The publication in March 2002 of a special issue of the Journal of Geographical Systems with articles by members of the expert panel. The special issue evaluated different spatial modeling approaches for conducting environmental exposure assessment with HSRH data and detailed obstacles and opportunities in this emerging field.
- The development and testing of prototype software, ImageSeer that incorporates two of the spatial methods deemed to have the greatest potential for enhancing exposure assessment with HSRH data.
Phase II of the project will:
- Build, test and implement a complete software package based on results of the prototype.
- Conduct applied studies in vector-borne disease risk assessment to demonstrate use of the novel methods
- Convene a conference on high spatial resolution hyperspectral imagery in exposure assessment to further disseminate knowledge of the new methods and techniques.