A remote sensing technique for the assessment of stable interannual dynamical patterns of vegetation

Описание

Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Год издания: 2011

Идентификатор DOI: 10.1117/12.896748

Ключевые слова: EVI, Forest, Kernel k-means, Minimum noise fraction, MODIS, NDVI, Principal component analysis, Time series, Principal Components, Agriculture, Ecosystems, Hydrology, Remote sensing, Satellite imagery, Space optics, Vegetation

Аннотация: The time series of various parameters of satellite imagery (NDVI/EVI, temperature) during the growing season were considered in this work. This means that satellite images were considered not like a number of single scenes but like temporal sequences. Using time series enables estimating the integral phenological properties of vegeПоказать полностьюtation. The basis of the developed technique is to use one of the methods of transformation of the multidimensional space in order to get the principal components. The technique is based on considering each dimension of the multidimensional space as satellite imagery for a specific date range. The technique automatically identifies spatial patterns of vegetation that are similar by phenology and growing conditions. Subsequent analysis allowed identification of the belonging of derived classes. Thus, the technique of revealing the spatial distribution of different dynamical vegetation patterns based on the phenological characteristics has been developed. The technique is based on a transformation of the multidimensional space of states of vegetation. Based on the developed technique, areas were obtained with similar interannual trends. © 2011 SPIE.

Ссылки на полный текст

Издание

Журнал: Proceedings of SPIE - The International Society for Optical Engineering

Выпуск журнала: Vol. 8174

Персоны

Вхождение в базы данных