Authors:
Sergiu Nedevschi
1
;
Radu Danescu
1
;
Dan Frentiu
1
;
Tiberiu Marita
1
;
Florin Oniga
1
;
Ciprian Pocol
1
;
Rolf Schmidt
2
and
Thorsten Graf
2
Affiliations:
1
T.U. of Cluj-Napoca, Romania
;
2
Volkswagen AG, Germany
Keyword(s):
Stereovision, non-flat road, road-obstacle separation, 3D points grouping, object detection, object tracking.
Related
Ontology
Subjects/Areas/Topics:
Informatics in Control, Automation and Robotics
;
Intelligent Transportation Technologies and Systems
;
Robotics and Automation
;
Vision, Recognition and Reconstruction
Abstract:
This paper presents a high accuracy stereovision system for obstacle detection and vehicle environment perception in various driving scenarios. The system detects obstacles of all types, even at high distance, outputting them as a list of cuboids having a position in 3D coordinates, size, speed and orientation. For increasing the robustness of the obstacle detection the non-planar road model is considered. The stereovision approach was considered to solve the road-obstacle separation problem. The vertical profile of the road is obtained by fitting a first order clothoid curve on the stereo detected 3D road surface points. The obtained vertical profile is used for a better road-obstacle separation process. By consequence the grouping of the 3D points above the road in relevant objects is enhanced, and the accuracy of their positioning in the driving environment is increased.