Region-based perceptual grouping for road extraction from high-resolution images

is carried out on region basis. An imagery scene is modeled by the spectrally homogeneous regions restrained in shape and size. Given a set of low level image features for each region, perceptual grouping is performed on the regions to generate a higher level of structure, from which road segments are extracted. In our unique approach, perceptual grouping is based on the similarity of spectral, orientation and proximity of the candidate regions. For this, a crisp classifier is employed for similarity measurement to find the road candidates; the road candidates are further merged into larger regions in terms of their orientation and adjacency. The road segments are extracted from these merged regions on geometric measurement of a candidate road segment. In order to evaluate the performance of the proposed algorithm, experiments are carried out on imagery of real scenes. The results showed the approach has the advantage of distinguishing roads from artifacts caused by parking lots and buildings.