Obtaining of Support Spectral Signals during Hyperspectral Survey

Abstract

Application of hyperspectral survey of earth’s surface is one of the most informative option of data using from remote sensing of the Earth. Advantage of hyperspectral images during remote monitoring of environment is connected with possibility of creation of surface’s qualification cards with isolation of similar plots, where it is possible a presence of any pollutant. In the absence of spectral standard classification of pixels on the images is carrying out “without learning”. Classification “with learning” is supposed use of database of spectral signals that substantially accelerated and simplified further work on type surface’s identification. There were obtained data of spectral signals of different surfaces that were tabulated in stand-alone library. Images obtained during flight tests were investigated with the use of different software packages, designated for topical data processing. There were determined possibilities of pollutions’ different types’ identification on the surface with the use classification “with learning”. There was carrying out test on detection of plots with presence of oil pollution