Title :
Spatio-temporal wavelet statistics of SAR backscatter for the characterization of forest degradation in Cameroon
Author :
De Grandi, Elsa Carla ; Mitchard, Edward ; Woodhouse, Iain
Author_Institution :
Sch. of Geosci., Univ. of Edinburgh, Edinburgh, UK
Abstract :
Spatio-temporal analysis of ENVISAT SAR backscatter statistics was undertaken to characterise different landcover types including grassland, forest/agriculture mosaic, degraded forest and intact forest in Eastern Cameroon, Central Africa. The spatial statistics analysis was based on texture measures including wavelet variance and spectra. While, temporal analysis called into play multi-temporal features (swing and linear regression parameters) of pixel trajectories along image acquisitions for tracking landcover changes between 2003 and 2009.
Keywords :
geophysical image processing; image segmentation; image texture; land cover; radar imaging; spatiotemporal phenomena; statistical analysis; synthetic aperture radar; vegetation mapping; wavelet transforms; AD 2003 to 2009; ENVISAT SAR backscatter statistics; SAR backscatter; agriculture mosaic; central Africa; degraded forest; eastern Cameroon; forest degradation characterization; forest mosaic; grassland; image acquisitions; landcover changes; landcover types; multitemporal features; spatiotemporal wavelet statistics; temporal analysis; texture measures; wavelet spectra; wavelet variance; Backscatter; Correlation; Degradation; Discrete wavelet transforms; Extraterrestrial measurements; Trajectory; Wavelet analysis; ENVISAT ASAR; degraded forest; multi-temporal features; spatio-temporal statistics; wavelet variance;
Conference_Titel :
Geoscience and Remote Sensing Symposium (IGARSS), 2014 IEEE International
Conference_Location :
Quebec City, QC
DOI :
10.1109/IGARSS.2014.6946935
