![]() |
ИСТИНА |
Войти в систему Регистрация |
ИПМех РАН |
||
Starting from the first precision satellite systems in the 1970s-1980s, remote sensing Earth sciences have passed through a wide range of regional studies from visual decoding images from photo and scanning instruments to updated automation techniques of multi-spectral (6-7 spectral bands of visible and near infrared wavelengths) and hyper-spectral (hundreds of spectral bands) imagery processing. Pattern recognition of regional objects in a particular subject area is a major discipline that encompasses algorithmic and programmatic tools of the relevant remote sensing data processing. A specialized sector of remote sensing Earth sciences research was created in Museum of Earth Sciences late in the 1980-s on the basis of physical-geographic regions and natural zoning and soil formation departments. Three volumes of remote sensing Earth sciences were published Moscow University Publ. called as “Geophysical basis” (1992), “Information and mathematical basis” (1998), “Dialogue between nature and society. Sustainable development” (2000). Imaging spectroscopy with hundreds of spectral channels as a part of remote sensing Earth sciences has evolved later towards wider understanding the information content of the air-borne and space-borne images of high spatial resolution. Being oriented on the science and technology disciplines of data processing, the related techniques are based on the available chronicles of nature, particular protected regions, forest inventory, forest typology, geo-botanical and other ground-based observations. Validation of the remote sensing information products of data processing is an important part of the techniques elaborated. Some results of the air-borne campaigns with a domestic imaging spectrometer and ground-based campaign on a test forested area are presented using the listed categories of observations. These results published in high rating foreign journals deal with retrieval of forest stand attributes using hyper-spectral air-borne remote sensing data of forest recognition for different species and ages on the test area.