Multiresolution Snakes for urban road extraction from Ikonos and Quickbird images

This article addresses the problem of street extraction in high resolution images from a topologically correct graph of the network. The extraction algorithm makes use of specific active contours (snakes) combined with a multiresolution analysis (MRA). The use of the MRA for computing the snake’s energy enables to increase the convergence of the algorithm by minimizing the problem of noise (vehicles, ground markings,. . . ). This phase is composed of two sequential steps: the extraction of street segments and the extraction of street intersections. Indeed, these two objects present too many differences in both topology and shape to be processed in the same way. Results of the street network extraction are presented in order to illustrate the different steps of the method and future prospects are exposed. 1 ROAD NETWORK EXTRACTION 1.1 State of the art Road extraction from remotely sensed images has been the purpose of many works in the image processing field, and because of its complexity, is still a challenging topic. These methods are based on generic tools of image processing, such as linear filtering (Wang and Howarth 1987), mathematical morphology (Destival 1987), Markov fields (Merlet and Zérubia 1996), neural networks (Bhattacharya and Parui 1997), dynamic programming (Gruen and Li 1995), or multiresolution analysis (Baumgartner et al. 1999; Couloigner and Ranchin 2000). Road models are common for all authors, i.e. the radiometry along one road is relatively homogeneous and contrasted compared to its background. Moreover the width of the road and its curvature are supposed to vary slowly, and the road network is supposed to be connex. Promising studies try to take the context of the road into account in order to focus the extraction on the most promising regions (Baumgartner et al. 1999). The recent possibility to have satellite images with a very high spatial resolution (1 meter or less) has reboosted the interest for road extraction (especially for the applications in urban areas). This increased resolution enables a more accurate localization of the road sides as well as its extraction as a surface element. In return, it generates a higher complexity of the image and an increase of geometric noise (vehicles, trees along the road, occlusions, . . . ).