DETECTION OF SMALL OBSTACLES AT LONG RANGE USING MULTIBASELINE STEREO

This paper describes a trinocular stereo technique that is designed for long range small obstacle detection on highways. It first describes traditional stereo techniques and how they are inappropriate in highway environments. The proposed technique, which uses the apparent shape of surfaces in an image, is then described. In order to allow for the rejection of false positives that are introduced by image noise, a simple confidence measure is applied. The system is found to be capable of detecting objects as small as 14cm at ranges in excess of 100m.

[1]  Takeo Kanade,et al.  A stereo machine for video-rate dense depth mapping and its new applications , 1996, Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[2]  Charles E. Thorpe,et al.  A specialized multibaseline stereo technique for obstacle detection , 1998, Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231).

[3]  G. Salgian,et al.  Electronically directed "focal" stereo , 1995, Proceedings of IEEE International Conference on Computer Vision.

[4]  F. Ashcroft,et al.  VIII. References , 1955 .

[5]  Martial Hebert,et al.  Weakly-calibrated stereo perception for rover navigation , 1995, Proceedings of IEEE International Conference on Computer Vision.

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