Lane Detection and Tracking through Affine Rectification

Inversion of the perspective mapping in acquired images enables an autonomous vehicle to perform vision-based navigation tasks, such as obstacle detection. We present work on affine rectification that restores the parallel property of a traffic lane, using the vanishing point of equally spaced parallel lane boundaries. Experimental results demonstrate the potential of our new algorithm.

• Type: Journal
• Year: 2007
• Where: MVA
• Authors: Qiang He, Chee-Hung Henry Chu
• Keywords: Affine Rectification, Computer Vision, MVA 2007, Parallel Lane Boundaries, Vision-based Navigation Tasks