Extraction of simple road crossing

Abstract

Road crossings of roads are important components of the road network. However, they generally are not explicit in the existing research on extraction of road network. This is due to the wide variety of road crossings may contain. Thus this paper presents a methodology for the extraction of simples road crossings considering the stages of pre-processing that performs a smoothing of the region of crossing, followed by binarization and skeletonization of the region of road crossing. Soon after the methodology proposes a model skeleton to perform a filtering of the spurious structures. Based on the model filtered performs the extraction of the edges of road crossings. The methodology was evaluated for four images with respect to the criterion of completeness and proved to be quite effective in content page 83.5% completeness of the average images.