Simple method of deriving a surface information form series of historical, panchromatic, high resolution satellite imagery (CORONA program)

摘要

From many years, satellite obtaining remote sensing data are used for monitoring and research a surface of Earth. Since 1972 when ERTS-1 satellite (Landsat 1) has started a service on orbit, multispectral satellite data has been available for scientists and general public. With typical spatial resolution of 5 - 25 meters we can use this data to analyze a natural and anthropogenic changes on our planet in small and medium scale. Initial data from early 70's have geometric resolution about 70 meters (MSS scanner) which is insufficient for many analyses. After 1995 we can easily obtain a declassified high-resolution military surveillance satellite imagery from CORONA satellites. Data collected from period 1961-1980 in analog form (photographic film) have typically spatial resolution in range 1-10 meters, but in wide, panchromatic channel. This article proposes a simple method for retrieving valuable information from series of panchromatic imagery collected over the same area in different seasons. This method utilizes a different properties of land cover types, specially from different vegetation canopies to retrieve a superficial, synthetic spectral channels that can be then analyzed and interpreted. This paper analyzes a method, verifies a results and proposes an interpretation methodology to use with multitemporal CORONA imagery. Concluding, we can obtain a valuable high resolution historical data in multichannel form covering 1972-1980 period. This method can be used for long-time analyses of land cover and land use changes in big scales (up to 1:10 000).