A Robust Method for Vanishing Point Estimation in Road Image

This paper presents a method for vanishing point(VP) estimation in the complex road environment images. Through prior information, VP predicated region in the image are computed, which can remove a large number of noise Line segments. Then using Hough Transform the sample Line segments in image plane are projected into the points in parameter space where VP is estimated by a weighted least median square method based on length of Line segments. During the experiment, many complex road environment images were used to test our method. The experimental results show that our method can correctly estimate vanishing point, and is more accuracy and robust than traditional method.