HR: 0830h
AN: B41B-0879    [PDF]
TI: Using Declassification Intelligence Satellite Pictures With Quickbird Imagery to Study Urban Land Cover Dynamics: a Case Study from Kazakhstan
AU: * Ratcliffe, I C
EM: iratclif@calmit.unl.edu
AF: University of Nebraska-Lincoln, School of Natural Resources, 113 Nebraska Hall, Lincoln, NE 68588-0517 United States
AU: Henebry, G M
EM: ghenebry@calmit.unl.edu
AF: University of Nebraska-Lincoln, School of Natural Resources, 113 Nebraska Hall, Lincoln, NE 68588-0517 United States
AB: Recent declassification of high-resolution intelligence satellite pictures from the CORONA, ARGON, LANYARD, GAMBIT, and HEXAGON (KH-1 through KH-9) programs has provided more than 910,000 scenes of the Earth's surface collected between 1960 and 1980. Public access to this historical image record enables investigation into urban land cover dynamics. The spatial resolutions of these photographic images are nominally comparable to current high-resolution commercial satellite imagery-IKONOS, QuickBird. However, the declassified imagery suffers from multiple limitations including variable dynamic ranges, geometric distortions, and sparse imaging system metadata. Can these data be used to characterize urban land cover dynamics in conjunction with current sensors? Within the context of a NASA LCLUC project on Kazakhstan, we acquired QuickBird imagery for the cities of Astana (2002/05/04) and Almaty (2002/09/08). The QuickBird data were converted to spectral radiance and geometrically rectified. CORONA and GAMBIT data of Astana and Almaty were acquired from the USGS archives. Only a single CORONA image was found for Astana (1971/09/17). For Almaty, two GAMBIT images (1964/04/23 and 1966/01/27) and two CORONA images (1968/08/13 and 1971/09/17) were found. The images, as film diapositives, were scanned at 1600 dpi on a flatbed scanner. The CORONA and GAMBIT images were co-registered to the QuickBird images. Image co-registration proved problematic due to the differing sun/target/sensor geometries of the CORONA and GAMBIT imagery and the coarser spatial resolution of the CORONA data. While the declassified imagery nominally has very high spatial resolution (CORONA ~2m; GAMBIT ~1m), it appears fuzzier than the native raster data from QuickBird, due to the analog data source and multiple image transfer steps. In the 1968 and 1971 images for Almaty and the 1971 image of Astana, each city extended over the edge of the film. Subsequent mosaicking resulted in seams that were evident across both images due to uncontrolled radiometric differences and some urban features did not align due to registration errors. Subsets of the imagery were generated for areas that depicted significant land cover change and were compared with subsequent dates. Types of observed urban land cover change that were observed include urban expansion, urban densification, landscape transformation, and afforestation. These results demonstrate the utility of harvesting historical land cover data from the declassified intelligence satellite datasets for the study of urban land cover dynamics.
UR: http://www.calmit.unl.edu/kz/
DE: 1640 Remote sensing
DE: 9320 Asia
SC: Biogeosciences [B]
MN: 2003 Fall Meeting