An Automated Directed Spectral Search Methodology for Small Target Detection
Date
2014-05
Authors
Grossman, Stanley I.
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Much of the current efforts in remote sensing tackle macro-level problems such as determining the extent of wheat in a field, the general health of vegetation or the extent of mineral deposits in an area. However, for many of the remaining remote sensing challenges being studied currently, such as border protection, drug smuggling, treaty verification, and the war on terror, most targets are very small in nature - a vehicle or even a person. While in typical macro-level problems the objective vegetation is in the scene, for small target detection problems it is not usually known if the desired small target even exists in the scene, never mind finding it in abundance. The ability to find specific small targets, such as vehicles, typifies this problem. Complicating the analyst's life, the growing number of available sensors is generating mountains of imagery outstripping the analysts' ability to visually peruse them.
Description
Keywords
Remote sensing, Analytic sweet spot, Automated target detection, Directed search, Nearest neighbor inflation, Spectral searach