|
ИСТИНА |
Войти в систему Регистрация |
ИСТИНА ИНХС РАН |
||
RETRIEVAL OF BIOLOGICAL PRODUCTIVITY PARAMETERS OF FOREST ECOSYSTEMS USING OPTICAL REMOTE SENSING DATAстатья
- Авторы: Kozoderov V.V., Dmitriev E.V., Sokolov A.A.
- Журнал: Climate&Nature
- Номер: 1(3)
- Год издания: 2016
- Первая страница: 2
- Последняя страница: 19
- Аннотация: Typical understanding in studying fluxes of elements for different reservoirs is seen as a necessity to combine remote sensing data and ground-based observations for various ecosystems. This paper deals with the biological productivity estimates retrieved from optical airborne data of high spectral and spatial resolution for the forest patterns of different species and ages within a test area. The main attention is paid to compare the available forest inventory maps of this area and the results of the pattern recognition using hyper-spectral airborne imagery processing and projective cover parameters for the near to nadir view angles. We construct an improved classifier to enhance a computational efficiency of the procedures employed while minimizing the posteriori energy category of the spectral radiances registered for selected forest classes by the imaging spectrometer on the test area and reducing redundancy of the spectral channels by their optimization without diminishing accuracy of the pattern recognition. The next step of the imagery processing is to retrieve the phyto-mass amount and the total biomass amount for the recognized forest classes based on our original approach of the outgoing radiances formation for each image pixel using the end-members characterization instead of the common-used ‚vegetation indices‛ concept. The end-members are described for a particular forest class by pixels relating to the sunlit tops, completely shaded background and partly illuminated by the Sun and partly shaded. The final procedure is in constructing the models of the forest growth in the form of Net Primary Productivity (NPP) and similar other parameters retrieved from the information products. The listed stages of the forest class recognition, the biomass retrieval and NPP estimates are designed for enhanced parameterization of forested environments in climate models.
- Добавил в систему: Козодеров Владимир Васильевич