Vanishing point estimation for challenging road images

Abstract

In this paper, we present an efficient vanishing point detection method for challenging road images. This detection process is based on the geometrical features of the roads. The slope distribution of the line segments is analyzed to reduce the spurious lines. A distance-based weighting scheme is also utilized to eliminate the voting noise in the voting stage. The proposed algorithm has been tested on a natural data set from Defense Advanced Research Projects Agency (DARPA). Experimental results with both quantitative and qualitative analyses are provided, which demonstrate the superiority of the proposed method over some state-of-the-art methods.