A multilevel context-based system for classification of very high spatial resolution images

This paper proposes a novel pixel-based system for the supervised classification of very high geometrical (spatial) resolution images. This system is aimed at obtaining accurate and reliable maps both by preserving the geometrical details in the images and by properly considering the spatial-context information. It is made up of two main blocks: 1) a novel feature-extraction block that, extending and developing some concepts previously presented in the literature, adaptively models the spatial context of each pixel according to a complete hierarchical multilevel representation of the scene and 2) a classifier, based on support vector machines (SVMs), capable of analyzing hyperdimensional feature spaces. The choice of adopting an SVM-based classification architecture is motivated by the potentially large number of parameters derived from the contextual feature-extraction stage. Experimental results and comparisons with a standard technique developed for the analysis of very high spatial resolution images confirm the effectiveness of the proposed system.