Road elevation map estimation based on affine transformation and stereo matching

This paper presents a road elevation map estimation method based on affine transformation. A geometric model of reference plane and stereo rig is proposed. The model showed that the height of 3-D points relative to the reference plane can be described by the proposed disparity relative to the reference plane. The reference plane is implicitly established by performing an affine transformation on target image. A local based stereo matching is applied to estimate subpixel disparity map. The elevation map can be easily estimated from subpixel disparity map using proposed elevation computation model. The experimental results on real-word road scenes show that the proposed method can provide accurate road elevation estimation.