Pedestrian Protection Pedestrian Detection Vulnerable Road Users PROTECTOR SAVEU


	Main long-term research theme at Daimler Research is the visual pedestrian detection from a moving vehicle.

	1. Motivation


	Pedestrians are arguably the most vulnerable traffic participants.

	Killed	Injured	Total
Passenger Cars	22.502	995.026	1.017.528
Pedestrians	6.049	155.151	161.200
Mopeds	2.421	141.870	144.291
Bicycles	2.385	139.442	141.827
Motor Cycles	3.821	124.023	127.844
Other	4.559	121.816	126.375
Total	41.737	1.677.328	1.719.065


	Road traffic accidents 1997. Total figures for EU countries Accident Source: UN-ECE)


	More than 150.000 pedestrians are injured yearly in the EU.

		More than 6000 are killed.




	Children are especially at risk.




		Article on the Daimler pedestrian system (abstract, full text)





	Daimler Pedestrian Detection Benchmark

The past few years have seen increased awareness of the plight of vulnerable road users at the national and EU level. In 2003, the EU passed Phase 1 of Directive 2003/102/E on pedestrian protection, focussing on passive safety, i.e. meaning to reduce injury levels upon impact, by specifying various maximum impact criteria (e.g. head, leg). More recently, June 2008, the EU Parliament approved the Phase 2 draft legislation, which specifies a combination of passive and active safety measures. In particular, Phase 2 requires new passenger cars to be fitted with Brake Assist Systems (BAS) as early as 2009. Pedestrian protection is meanwhile also a major theme for consumer rating groups like Euro NCAP.

Passive pedestrian safety measures involve vehicle structures (e.g. bonnet, bumper) that expand during collision in order to minimize impact of the pedestrian leg or head hitting the vehicle.

mercedes pedestrian protection

© Daimler

For example, Mercedes-Benz introduced the active bonnet as standard for the new E-Class (2009). The system includes three impact sensors in the front section as well as special bonnet hinges pretensioned by powerful springs. Upon impact with a pedestrian, the rear section of the bonnet is pushed upwards by 50 millimetres in a fraction of a second, thus enlarging the deformation zone. The system is reversible and can be reset manually by the driver.


	2. Aim & Challenges

	Although important, passive pedestrian safety measures are constrained by the laws of physics in terms of ability to reduce collision energy and thus injury level. Moreover, passive measures cannot account for injuries sustained in the secondary impact of the pedestrian hitting the road.

The aim is to develop active (video-based) driver assistance systems which detect dangerous situations involving pedestrians ahead of time, allowing the possibility to warn the driver or to automatically control the vehicle (e.g. braking). Such systems are particularly valuable when the driver is distracted or visibility is poor.




		Article on the Daimler pedestrian system (abstract, full text)





	Daimler Pedestrian Detection Benchmark

Yet vision-based pedestrian detection is a difficult problem for a number of reasons. The objects of interest appear in highly cluttered backgrounds and have a wide range of appearances, due to body size and pose, clothing and outdoor lighting conditions. They stand typically far away from the camera, and thus appear rather small in the image, at low resolution. A major complication is that because of the moving vehicle, one does not have the luxury to use simple background subtraction methods (such as those used in surveillance applications) to obtain a foreground region containing the human. Finally, there are hard real-time requirements and stringent performance criteria.


	3. Our Stereo Vision System

See right for a diagram of the Daimler stereo vision pedestrian system. The detection component consists of a cascade of module, each utilizing different visual criteria to successively focus on relevant image regions, carefully balancing robustness and efficiency considerations. The tracking component aggregates per-frame detections to trajectories by a tracking module. Finally, the risk assessment and warning/control component evaluates the probability of collision; if the latter exceeds a threshold an acoustic driver warning is given or automatic vehicle braking is applied.

Pedestrians are detected in a range 5-25m and up to 4m lateral, on each side of vehicle. It does so at processing rates of 7-15 Hz, allowing vehicle speeds up to 50 km/h.


	4. Demos

	The prototype system was integrated in a Mercedes-Benz E-Class limousine.


	Below two videoclips with runs on the test track with “Hans”, our fearless pedestrian dummy. The pedestrian system first warns the driver acoustically, then soon thereafter, brakes the vehicle automatically.

Below video clip shows pedestrian detection in urban traffic. The middle window shows the rectangular regions of interest as a result of stereo preprocessing. The left window shows the final detection superimposed in red on the image. The right window shows a bird’s view of the situation in front of the vehicle (Right scale in meters. Red dots indicate current pedestrian locations, yellow dots indicate other obstacles).




	8,7 Mb (mpg)


			© Daimler

	5. Benchmarking

	Pedestrian detection has meanwhile attracted an extensive amount of interest from the computer vision community. Many techniques have been proposed in terms of features, models and general architectures (see recent survey on pedestrian detection). The picture is increasingly blurred on the experimental side. Reported performances differ by up to several orders of magnitude (e.g. within the same study). This stems from the different types of image data used (degree of background change), the limited size of the test datasets, and the different (often, not fully specified) evaluation criteria such as localization tolerance, coverage area, etc. In order to increase visibility by providing a common point of reference from experimental perspective, we published a number of pedestrian benchmarks: data sets and associated performance metrics.



	6. Projects

Research on vision-based pedestrian protection was previously conducted in the EU projects PROTECTOR (2000-2003), SAVE-U (2002-2005) and WATCH-OVER (2005-2008). It now continues in the German AKTIV-SFR (2006-2010) project. More details ...

PROTECTOR web site


	7. The Road Ahead

Recently, several night-time pedestrian detection systems hit the market. For example, Mercedes-Benz introduced the Night View Assist Plus in its latest E- and S-Class (2009). Two special headlights illuminate the scene in the Near-Infra Red (NIR) range. A camera designed to pick up precisely this type of light records the observed images and sends them to the display. Detected pedestrians are highlighted in this greyscale video stream.

There has been remarkable progress on pedestrian detection over the past few years. Yet more research is needed before such systems can perform active vehicle control reliably.

Another step closer towards Daimler’s long-term Vision of Accident-Free Driving


	[Home Gavrila] [Resume] [Research] [Univ. Amsterdam] [People] [Publications] [Media Coverage] [Search]


	Copyright © 2001-2009 Gavrila. All rights reserved. Disclaimer.