Estimating the Distance to Free Area Boundary on a Train Platform from an Image
Yuna NISHIBE, Tadayoshi SHIOYAMA, Haiyuan Wu, Qian CHEN, Tetsuo SHIMADA
Train platform is one of the most dangerous place for the blind. This paper presents a method for extracting free area (the region without obstacles) in train platform and estimating the distance to the boundary of free area from an image. First, the input image is segmented by using Gabor filter as well as greyscale information and edge information. Then the image region containing the place where the blind stands is determined as the free area. Next, the boundary of free area is extracted and the vertical lines are extracted, which are then used to determine the distance from the blind to the boundary of free area based on 3D vector analysis.