Detection and extraction of road networks from high resolution satellite images

This article addresses the problem of road extraction from new high resolution satellite images. The proposed algorithm is divided in two sequential modules : a topologically correct graph of the road network is first extracted, and roads are then extracted as surface elements. The graph of the network is extracted by a following algorithm which minimizes a cost function. The extraction algorithm makes use of specific active contours (snakes) combined with a multiresolution analysis (MRA) for minimizing the problem of geometric noise. This reconstruction phase is composed of two steps : the extraction of road segments and the extraction of road intersections. Results of the road 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 ([1]), mathematical morphology ([2]), Markov fields ([3]), neural networks ([4]), dynamic programming ([5]), or multiresolution analysis ([6] ; [7]). 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 ([6] ; [8]). The recent possibility to have satellite images with a high spatial resolution (1 meter or less) has re-boosted the interest for road extraction (especially for the applications in This work was supported by a CNRS/DGA grant of the french Ministry of Defence. The authors would like to thank the firm G.I.M. (Geographic Information Management) for the IKONOS image. 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, . . .). 2. THE PROPOSED APPROACH 2.1. Description A method has been developed in order to extract and characterize the road network from high resolution images. Inputs of the algorithm, besides the high resolution satellite image, are models of roads (using roads properties defined by [7]) and properties of road network (such as connexity). Our algorithm is composed of two sequential modules (fig.1). Fig. 1. The methodology including topology management and road reconstruction Firstly, a topologically correct graph of the road network is extracted. This step aims at giving correct spatial connections between roads as well as an approximation of their location. The next step is the actual road reconstruction. Due to the high resolution of the images, a surface reconstruction has to be performed. This step uses the previous step ha l-0 03 37 18 5, v er si on 1 6 N ov 2 00 8 Author manuscript, published in "IEEE International Conference on Image Processing, Barcelona : Espagne (2003)"