2011 | OriginalPaper | Buchkapitel
Classifying Wetland Vegetation Type from MODIS NDVI Time Series Using Fourier Analysis
verfasst von : Xiaodong Na, Shuying Zang
Erschienen in: Applied Informatics and Communication
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
On the basis of multi-temporal MODIS (Moderate Resolution Imaging Spectroradiometer) data, this paper used a kind of supervised classification method to extract wetlands distribution of Sanjiang Plain. The NDVI (normalized difference vegetation index) time series data were got from 16-day MODIS data after processing by Harmonic Analysis of Time Series (HANTS). The shape of NDVI time series, which is diagnostic for certain vegetation phenology, was our primary classifier. A similarity measure based directly on the components of the Discrete Fourier Transform which introduced by J.P. Evans was used to determine a pixels class membership. Based on the difference between vegetation phenology, seven kinds of vegetation had been classified: swamp, meadow swamp, bottomland, paddy field, dry land, bush and woodland. The spatial distribution of natural wetland (swamp, meadow swamp, open water) and manpower wetland (paddy field) were extracted. Validation shows that total classification accuracy is 79.67%, Kappa coefficient is 0.7525. Results indicated that this Fourier component similarity measure produced an objective, computationally inexpensive and rapid method of wetlands classification.