Driver assistance systems that continuously monitor the surroundings or interior of a vehicle, perform situation assessment, and either warn the driver or apply vehicle control, hold great potential to increase the safety and convenience of driving. This lecture gives an overview of such intelligent vehicles available “today”, “tomorrow” and “next week”. Applications for both the highway and urban scenario are covered. In particular, this lecture discusses modules developed at DaimlerChrysler Research for the recognition of traffic infrastructure (lanes, traffic signs, traffic lights) and other traffic participants (vehicles, pedestrians). Extensive video footage will be shown illustrating various intelligent vehicles in action. The lecture concludes with a discussion of the multitude of factors which determine whether a certain driver assistance function can successfully be brought onto the market.

The ability to perceive humans and their activities by vision is key for a machine to interact intelligently and effortlessly with a human-inhabited environment. Because of many potentially important applications, ``Looking at People'' is currently one of the most active application domains in computer vision. This lecture discusses a number of promising applications and provides an overview of recent developments in this domain. The emphasis lies on discussing various methodologies; they are grouped in 2-D approaches with or without explicit shape models and 3-D approaches. The visual surveillance application and associated image processing techniques are featured prominently.

This lecture describes a multiyear effort at DaimlerChrysler towards developing a video-based pedestrian safety system. It covers the typical stages of system design: from market need analysis, system specification and implementation, to evaluation and user acceptance. It first discusses how accident statistics, existing pedestrian protection concepts, and technical feasibility, influence system specifications. The resulting prototype system consist of three components: pedestrian sensing (perception), risk estimation (reasoning), and warning/control initiation (action). The sensing component uses a stereo camera to detect pedestrians in front of a moving vehicle, and estimate relative trajectories. The lecture describes the various vision modules that have been developed for this purpose, based on depth, shape, motion, and texture cues. Finally, recent results are present from large-scale field tests on pedestrian protection, on a test track and in real urban traffic, as part of the EU project SAVE-U, completed August 2005.

Copyright © 2001-2008 Gavrila. All rights reserved. Disclaimer.