EFFICIENCY OF THE SPECTRAL-SPATIAL CLASSIFICATION OF DATA OF HYPERSPECTRAL OBSERVATIONS
S. M. Borzov1, O. I. Potaturkin1,2
1Institute of Automation and Electrometry, Siberian Branch, Russian Academy of Sciences, 630090, Novosibirsk, prosp. Akadmika Koptyuga, 1
2Novosibirsk State University, 630090, Novosibirsk, ul. Pirogova, 2
Keywords: дистанционное зондирование Земли, гиперспектральные изображения, фрагменты, спектральные и пространственные признаки, классификация типов подстилающих поверхностей, remote sensing, hyperspectral images, fragments, spectral and spatial attributes, classification of underlying surface types
Abstract
The efficiency of methods of the spectral-spatial classification of difficult-to-distinguish types of vegetation on the basis of hyperspectral data of remote sensing of the Earth, which take into account local neighborhoods of analyzed image pixels, is experimentally studied. Algorithms that involve spatial pre-processing of the raw data and post-processing of map charts of pixel-by-pixel spectral classification are considered. Results obtained both for a large-scale hyperspectral image and for its test fragment with different methods of learning sample formation are reported. The classification accuracy in all cases is estimated through comparisons of sub-satellite data and map charts of classes formed by using the compared methods. The reasons for the differences in these estimates are discussed.
|
