This research was conducted to determine the effectiveness of using commercial satellite data to detect covert nuclear weapons programs. Seven-band Landsat Thematic Mapper data covering the Pahute Mesa (an area within the United States Nevada Nuclear Testing Site), acquired on October 16, 1985, were analyzed to determine if underground nuclear test sites were spectrally distinguishable from the surrounding area. The analysis consisted of four steps: (1) analyzing the raw data, (2) manipulating the raw data through contrast stretching, filter application, matrix algebra, and principal components analyses, (3) identifying parameters that affect classification of underground nuclear tests and (4) selectively limiting parameters. The results of limiting parameters showed that a supervised classification of a signature created with a five-original-pixel seed of one representative, known test site provided an accurate classification of most known test sites. To further eliminate erroneous classification of roads and other areas of similar reflectance, these areas were seeded to create a second signature. This signature, whose spectral responses were different, was then used in a simultaneous classification. This classification further eliminated erroneous classification of non-test site areas, demonstrating that commercial satellite digital data can aid in the detection of covert nuclear weapons programs, in this case, underground nuclear testing. An application of the classification scheme used is proposed to confront a scenario in which a country seeks additional verification of another party's proposed violation of test ban treaties. / Department of Physics and Astronomy
| Identifer | oai:union.ndltd.org:BSU/oai:cardinalscholar.bsu.edu:handle/184783 |
| Date | January 1993 |
| Creators | Lance, Jay Logan |
| Contributors | Ball State University. Dept. of Physics and Astronomy., Howes, Ruth H. |
| Source Sets | Ball State University |
| Detected Language | English |
| Format | iv, 63 leaves : ill. (some col.) ; 28 cm. |
| Source | Virtual Press |
Page generated in 0.002 seconds