IEICE Information and Communication Technology Forum
Fusion of Spectral and Spatial Information for Land Cover Classification
Aleksandra Pizurica, Renbo Luo, Rui Wang, Shaoguang Huang, Jie Li, Hongyan Zhang, Wenzhi Liao,
PDF download (339.2KB)
Hyperspectral imagery contains a wealth of spectral and spatial information that can improve target detection and recognition performance. Existing feature extraction methods cannot fully utilize both the spectral and spatial information. Data fusion by simply stacking different feature sources together does not work well either, as it not takes into account the differences between feature sources. In this paper, we present our recent graph-based approach for fusing spectral information and spatial information of hyperspectral imagery, and we show a case study on how our graph-based fusion method combines multiple features, which can be applied in other applications. Our approach takes into account the properties of different data sources, and makes full advantage of both the spectral and the spatial features through the fusion graph. Experimental results on the classification of fusing real hyperspectral images are very encouraging.