Frederik Meysel

Situation analysis and atypical event detection with multiple cameras and multi-object tracking

This paper discusses a trajectory based recognition algorithm to expand the common approaches for atypical event detection in multi-object traffic scenes and to obtain area-based types of information (e.g. maps of speed patterns, trajectory curvatures or erratic movements). Different views of the same area by more than one camera sensor are necessary, because of the typical limitations of single camera systems, resulting from occlusions by other cars, trees and traffic signs. Furthermore, distributed cooperative multi-camera system (MCS) enables a significant enlargement of the observation area. The fusion of object data from different cameras is done by a multi-target tracking approach. This approach opens up opportunities to identify and specify traffic objects, their location, speed and other characteristic object information. New and consolidated information of traffic participants is derived from the system.

Testing a collision avoidance display with high-precision navigation

Recent years have seen a rise in sophisticated navigational positioning techniques. Starting from classical GPS, differential GPS, ground-based augmentation, and raw data submission have opened possibilities for high precision lateral positioning for beyond what was thinkable before. This yields new perspectives for technologies like ACAS/TCAS, by enabling last-minute lateral avoidance as a supplement to the established vertical avoidance maneuvers. Working together with Ohio University’s Avionics Department, DLR has developed and tested a set of displays for situational awareness and lateral last-minute avoidance in a collision situation, implementing some state-of-the art ideas in collision avoidance. The displays include the possibility to foresee the hazard zone of a possible intruder and thus avoid that zone early. The displays were integrated into Ohio University’s experimental airplane, and a flight experiment was conducted to make a first evaluation of the applicability. The tests were carried out in fall 2012. We will present the principal architecture of the displays and detail the implementation into the flight carrier. Furthermore, we will give first results of the displays’ performance.

Map-based change detection (MBCD) in urban traffic scenes

In traffic surveillance applications, automatic methods to detect abrupt changes in a traffic scene are highly desirable. A robust method is introduced for camera based assessment of scene changes based on the dynamics derived from the detected objects in the scene. The approach needs a calibrated and oriented camera but no further contextual knowledge and works even with highly fragmented trajectories. It is suitable to detect unexpected scene changes that have not been trained.