Efficient algorithms for obstacle detection using range data
Computer Vision, Graphics, and Image Processing
Dynamic integration of height maps into a 3D world representation from range image sequences
International Journal of Computer Vision - Special issue on machine vision research at Osaka University
Automatic feature point extraction and tracking in image sequences for arbitrary camera motion
International Journal of Computer Vision - Special issue: image understanding research at the University of Maryland
Model-Based Object Tracking in Traffic Scenes
ECCV '92 Proceedings of the Second European Conference on Computer Vision
Knowledge-based anytime computation
IJCAI'95 Proceedings of the 14th international joint conference on Artificial intelligence - Volume 1
| Hi-index | 0.00 |
Most road-accidents are caused by a driver not paying attention to the traffic situation. This problem can be solved by assistance systems which warn the driver of potential dangers. Recent progress in sensor technology makes range sensors very attractive for automotive applications, but up to now little research has been done into the interpretation of range image sequences in outdoor environments. In this article an overall approach to the analysis of range image sequences for use in future distance warning systems is presented. A method for obstacle detection is described using radial slope evaluation and 3D connected component analysis. For obstacle tracking a new approach based on an anytime algorithm is proposed. This algorithm has an interpretation of the input data, possibly sub-optimal, ready whenever it is stopped and improves the interpretation with additional computing time. The different approaches proposed in this paper have been tested on a data set of 11 range image sequences consisting of 595 frames in total. Experimental results reveal that a reliable interpretation can be obtained in real time.