Ma, Guanglin:
ISBN 9783899599831

Real-time Vision Based Pedestrian Detection. Beiträge zur Informationstechnologie Bd. 3 # Pb., 150 S., 69 Abb., 5 Tab.

SCHLAGWORTE:
Vision based pedestrian detection
image processing
active safety

This thesis intensively focuses on the development of a robust real-time pedestrian candidate detection algorithm which can be integrated into an on-vehicle real-time vision based pedestrian detection unit. A pedestrian candidate detection algorithm must detect pedestrians as a subset of all obstacles in front of the vehicle. In this thesis, the detection step is performed by an obstacle detection algorithm which satisfies the requirements of pedestrian detection (i.e. pedestrians can be considered as a subset belonging to an obstacles set). For a alternative warning system application, the detection step would also have to yield information about possible "targets" that are relevant for such applications.

	^