Building height retrieval from WorldView-2 multi-angular images

The paper presents a new algorithm for the estimation of the heights of buildings acquired by multiangular multispec-tral/panchromatic acquisitions. The proposed method aims at exploiting the complete multiangular spatial and spectral information provided by the WorldView-2 system with controlled computational complexity of the overall automated system. It adopts a hierarchical classification scheme to detect the tops of the buildings from multangular pansharpened multispectral images of the considered urban scenario. Given five angular acquisitions, the (up to) four displacements of a rooftop with respect to the most nadiral view provide an accurate estimate of the height of the building. The experimental results show that the proposed method provides reliable and accurate building height estimates when compared to the true heights provided by publicly available building databases.