Towards an electronic orientation table: using features extracted from the image to register Digital Elevation Model

Looking at a countryside, human has no difficulty to identify salient information such as skyline in case of benign weather. At the same time, during the last decade, smartphones are more and more abundant in our daily life with always new efficient proposed services but augmented reality systems suffer from a lack of performance to offer a well adapted tool in this context. The aim of this paper is then to propose a new approach coupling image processing and Digital Elevation Model (DEM) exploitation to remedy that shortcoming. The proposed method first discriminates skyline and non-skyline pixels and then introduces a realtime matching between previously extracted map and DEM. In order to evaluate objectively each step and our finalized tool, a new tagged dataset with ground-truth is created and will benefit the entire scientific community.