Vegetation pattern recognition using hyperspectral air sounding dataстатья
Информация о цитировании статьи получена из
Web of Science,
Scopus
Статья опубликована в журнале из списка Web of Science и/или Scopus
Дата последнего поиска статьи во внешних источниках: 18 июля 2013 г.
Аннотация:For the problem of pattern recognition of natural and man-made objects using remote hyperspectral imaging data, we propose an approach that is based on both the criterion of a minimal Euclidean distance relative to spectra of some reference objects and specific features of wavelength shift of the area of transition from the chlorophyll absorption band to the spectral reflectivity maximum that is characteristic to vegetation. The database of this pattern-recognition method is constructed on the basis of pixel radiance histograms for particular spectral channels. The histogram in the maximum separability wavelength of object classes characteristic of the chosen test area is divided into a certain number of spectral intervals, which are grouped with respect to the above-mentioned shift. Using computational techniques for separating out these spectral groups, we point to new possibilities in the recognition of different vegetation types with the help of high-resolution spatial and spectral air sounding data.