Road Region Extraction with Longitudinal Slope

Extraction of road region is a core technology for an autonomous vehicle which achieves safer transport. We propose extraction and modeling method for the road region in front of vehicle, using three-dimensional shape that is analyzed by stereovision. In order to model even if the extracted region has longitudinal slope, a continuous planar model is used. This method uses some structural features of road region. Therefore, it is usable even if the road has no traffic lanes. In this paper, we describe the planar approximation based on principal component analysis. Additionally, a technique of road region candidate selection using a coefficient of determination, and a modeling procedure based on characteristic of image sensor have been proposed. Finally, effectiveness of the method is demonstrated on experiments with various real environment scenes.

[1]  Jitendra Malik,et al.  An integrated stereo-based approach to automatic vehicle guidance , 1995, Proceedings of IEEE International Conference on Computer Vision.

[2]  Junichi Maruyama,et al.  Robust estimation of planar regions for visual navigation using sequential stereo images , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[3]  N. Suganuma,et al.  An Obstacle Extraction Method Using Virtual Disparity Image , 2007, 2007 IEEE Intelligent Vehicles Symposium.

[4]  Kei Senda,et al.  Vehicle Front Environment Recognition Using Stereo Vision , 2005 .

[5]  Sergiu Nedevschi,et al.  Stereovision Approach for Obstacle Detection on Non-Planar Roads , 2004, ICINCO.

[6]  H. Hirschmüller.  Accurate and Efficient Stereo Processing by Semi-Global Matching and Mutual Information , 2005, CVPR.

[7]  Sergiu Nedevschi,et al.  Road Surface and Obstacle Detection Based on Elevation Maps from Dense Stereo , 2007, 2007 IEEE Intelligent Transportation Systems Conference.